chapters of thesis - politieacademie › kennisenonderzoek › k... · nederlandse titel...

Effectiveness of guideline-based care

of workers with mental health problems

David Rebergen

The study presented in this thesis was conducted at the EMGO+ Institute for Health and Care Research, Department of Public and Occupational Health of the VU University Medical Center. The EMGO+ Institute participates in the Netherlands School of Primary Care Research (CaRe) which was re-acknowledged in 2005 by the Royal Netherlands Academy of Arts and Sciences (KNAW). This study was financially supported by the Dutch Ministry of Internal Affairs and Royal Relations (BZK), the Health Insurance Agency of the Dutch Police (PolitieZorgPolis, formerly known as DGVP), the occupational health service Achmea Vitale (formerly known as Commit), and the EMGO+ Institute, VU University Medical Center. Financial support for the printing of this thesis has been kindly provided by the EMGO+ Institute, the VU University, VTS Politie Nederland, the Netherlands Institute of Mental Health and Addiction (Trimbos Institute), Achmea Vitale, Lifeguard BV, and Ausems en Kerkvliet, arbeidsmedische adviseurs. English title Effectiveness of guideline-based care of workers with mental

health problems Nederlandse titel Effectiviteit van op richtlijn gebaseerde begeleiding in de

psychische bedrijfsgezondheidszorg ISBN 9789086593293 Cover design Twan Bultstra, colabdesign.nl Lay-out Jeroen van Wijngaarden, colabdesign.nl Printed by Gildeprint Drukkerijen Enschede © 2009, David Rebergen, The Netherlands No part of this thesis may be reproduced or transmitted in any form or by any means, electronically or mechanically, including photocopying, recording or any information storage and retrieval system, without prior permission of the holder of the copyright.

VRIJE UNIVERSITEIT

Effectiveness of guideline-based care

of workers with mental health problems

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam,

op gezag van de rector magnificus prof.dr. L.M. Bouter,

in het openbaar te verdedigen ten overstaan van de promotiecommissie

van de faculteit der Geneeskunde op donderdag 2 juli 2009 om 15.45 uur

in de aula van de universiteit, De Boelelaan 1105

door

David Sebastiaan Rebergen

geboren te Ede

promotoren: prof.dr. W. van Mechelen prof.dr. A.J. van der Beek copromotor: dr. D.J. Bruinvels

Contents Chapter 1 p. 9 General Introduction Chapter 2 p. 23 Return to work interventions for adjustment disorders Chapter 3 p. 59 Design of a randomized controlled trial on the effects of Counseling of mental health problems by Occupational Physicians on return to work: the CO-OP-study Chapter 4 p. 81 Guideline-based care of common mental disorders by occupational physicians (CO-OP-study): a randomized controlled trial Chapter 5 p. 95 Cost-effectiveness of guideline-based care for workers with mental health problems Chapter 6 p. 113 Process evaluation of a randomised controlled trial on guideline-based care by occupational physicians counselling mental health problems Chapter 7 p. 131 Adherence to a national guideline on mental health problems by Dutch occupational physicians Chapter 8 p. 143 General Discussion Summary p. 163 Samenvatting p. 169 Dankwoord p. 177 Curriculum Vitae p. 183 List of publications p. 185

CHAPTER 1

GENERAL INTRODUCTION


10

CASE DESCRIPTION A young worker moved to another city to start a new job. In this new situation, his long-term relationship comes to an end and his work stagnates in different ways. Then he starts to doubt everything. He consults his general practitioner (GP) and receives referral to a psychotherapist. In the following months he can’t stop his negative thinking, looses his interest for work and leisure activities, sleeps bad and occasionally feels incompetent. Although the consultations with his psychotherapist are supportive, he doubts the effectiveness and approach, and he quits the treatment. When he reports sick for work, he is invited by the occupational health service to consult the occupational physician (OP) 3 weeks later. Meanwhile, he has informed his employer about his mental health problems and his incapacity for work. The main question of the worker, the employer, the OP and this thesis is: what is the most effective way of treatment for the worker to return to work successfully? INTRODUCTION In the past decades, the relationship between work and mental health has changed enormously. There is a worldwide increase in the experience of stress in the workplace for a range of reasons, including structural changes, changing work contexts, the shift to more knowledge-based work and the continuous introduction of new technology (van der Klink, 2002). Western societies have changed from industrial to service economies, which implies that the work of most of the working population is no longer physical but mental. Meanwhile, workers started to perceive work as an important contributor to quality of life, in which ambitions can be fulfilled. Worldwide, common mental health problems in workers, such as adjustment disorders, depression and anxiety, may affect functioning and often lead to reduced productivity at work and sick leave (Wang et al., 2006; Lerner & Henke, 2008). As the work content and the perception of work have changed, a new type of support was sought by workers from professionals in occupational health care. Nowadays, workers need tools to cope with mental aspects of and in work, besides protection from riskful physical work demands. Therefore, specific interventions need to be initiated in occupational health care of workers on sick leave due to common mental health problems. Primary and secondary care usually focused on recovery of symptoms instead of return to work (RTW). The general practitioner (GP) and the occupational physician (OP) often lack time and skills to optimally deal with these workers, resulting in a minimal approach (Anema et al., 2006; Nieuwenhuijsen, 2004). As an alternative, ways have been sought to encourage OPs to play a more active role. In 2000, the Dutch Society of Occupational Medicine (NVAB) published a new practice guideline, which promotes a more active role of the OP facilitating RTW of the worker, instead of a minimal role. The focus of this thesis will be the evaluation of this practice guideline entitled ‘The management by OPs of workers with common mental health problems’. Aim of this thesis is to contribute to quality improvement of occupational health care for these workers.

This chapter started with a case description of a worker with mental health problems. From here, the introduction continues with an explanation of the main concepts used in this thesis, followed by a description of recent developments in (Dutch) occupational health care. Specifically, the role of the OP is explained in the context of the guideline on the management of workers with mental health problems. The elements of the guideline for the OP can be placed in a widely accepted conceptual model that serves as theoretical framework for this thesis. According to this framework, the objectives and outline of this thesis are described.


11

MAIN CONCEPTS AND SETTING Common mental health problems Common mental health problems, or common mental disorders, reflect a broad term that applies to conditions ranging from experiencing stress symptoms to severe psychiatric disorders, such as psychosis. In working populations, the following common mental disorders constitute the majority of mental health problems: adjustment disorder (stress-related disorder), depression, and anxiety disorder. Adjustment disorder is an accepted diagnosis in DSM-IV and ICD-10, and refers to a maladaptive reaction to an identifiable stressor, occurring within a short time after onset of the stressor (APA, 1994; WHO, 1992; van der Klink & van Dijk, 2003). Such a reaction is characterized by depressive symptoms, anxiety, and/or inappropriate behaviour. Stressors may be related to work or family life. Although generally considered as minor psychiatry, adjustment disorders may have extensive disabling consequences. Most patients with an adjustment disorder recover within two to six months, even when no specific guidance or therapy is given, but there is a considerable risk of prolonged disability. In occupational settings, this diagnosis can be applied to many overlapping stress-related concepts and diagnoses, such as neurasthenia, nervous breakdown, burnout, and surmenage. These concepts and diagnoses have distress symptoms and malfunctioning in one or more social roles in common (van der Klink, 2003).

Depressive and anxiety symptoms are terms to describe minor, and usually mixed, syndromes often seen in primary care, as well as major conditions classified by the DSM-IV (APA, 1993). Common mental disorders do not only share a high incidence, they also show communality of symptoms. While adjustment disorders are often accompanied by depressive and anxiety symptoms, anxiety and depressive disorders in terms of the DSM-IV classification are considered more severe. This notion is reflected by the DSM-IV criteria, which state that an adjustment disorder diagnosis is not allowed if the severity and duration threshold for anxiety or depressive disorder are reached (APA, 1994). In these more severe mental disorders, an individual's vulnerability plays a dominant role, which is relatively independent of the interaction with the environment. There is a gradual transition from adjustment or stress-related disorders that are fully determined by the interaction between individual and a demanding environment up to more severe mental disorders, such as depression and anxiety disorder, which are determined by other factors as well. Factors that constitute a risk for more severe disorders are a hereditary vulnerability, a developmental vulnerability, less adequate coping styles, stressors on several domains of life, a quantitative or qualitative lack of social support, and physical co-morbidity (van der Klink, 2002). This thesis concerns all common mental disorders mentioned above, as they prevail in the occupational health care setting. Mental health problems and work disability The WHO (2002) predicts that by 2020, mental illness will be worldwide the second most important cause for work disability after heart disease. Studies on mental health policies and programs affecting the workforces of Finland, Germany, Poland, UK and USA showed that the incidence of mental health problems is increasing (ILO, 2000). In these studies it was reported that as many as one in 10 workers suffer from depression, anxiety, stress or burnout, with problems leading to unemployment and hospitalisation in some cases. Wang et al. (2006) reports a one-year prevalence of mental health problems in the working population of developed countries ranging from 10 to 18%. Mental health problems often affect functioning to such an extent that they result in sick leave and may


12

lead to chronic disability and loss of work (Nieuwenhuijsen et al., 2003; Anema et al., 2006; Wang et al., 2005).

Mental health problems account for one third of all disability benefits in The Netherlands. The morbidity leading to disability benefits is different in the Netherlands from that in other West European countries, with a relative overrepresentation of less severe mental disorders (van der Klink, 2002; IMF, 2004). It has been shown that a majority of these workers are at risk for chronic disability, while they only suffer from minor reversible psychiatric disorders with stress-related symptoms (Schaufeli & Kompier, 2001).

As Dutch workers are required to visit their OP for independent judgement of sick leave and RTW purposes, the Dutch OP has an optimal opportunity to find the people at risk and to influence RTW (Nieuwenhuijsen et al., 2003). The occupational health care setting in which the Dutch OP has been working, changed dramatically in the last decades (Schaufeli & Kompier, 2001). Since the comprehensive Sickness Insurance Act for workers (WAO) was introduced in the Netherlands in 1967, there has been a strong and ongoing disability debate in Dutch society and politics. Initially, the protective function of the system was of central concern: the system offered insurance against loss of income, with a guaranteed minimum income for everyone. As the number of workers collecting disability benefits increased from 150.000 by the start in 1967, to almost one million in the 1990s (around 10% of the working population), the public debate focused largely on sick leave and work disability. In order to diminish duration of sick leave and work disability rates, the government decided to make all stakeholders more aware of the need to limit the number of claims. Therefore, in 1996 the Sickness Insurance Act was privatised, and by means of the ‘gatekeeper model’ the responsibility of both employers and workers was increased. In this way, employers became responsible for the reintegration of workers in the first year of sick leave (Post, 2005).

The Dutch setting: occupational health services Since 1994, Dutch organisations had to be affiliated with a certified Occupational Health Service (OHS, arbodienst), according to the Working Conditions Act (Arbowet). The main reason for the government to introduce this measure was the obligation to implement European policy stating nation legislation to regulate a level of assistance in prevention of health and safety at work (Post, 2005). However, an extra compulsory assistance of workers on sick leave was instigated by the Dutch government, which had to be provided by OPs. This measure was accompanied by the introduction of the commercial OHS in order to stimulate competition in the OHS market. This led to a formation of about 200 OHSs, a number that due to mergers decreased to approximately 90 OHSs in 2004. During this study the five largest OHSs were responsible for the occupational health care of about 90 % of all workers in The Netherlands.

Each OHS had to employ at least one certified professional from each of the following fields: occupational medicine, occupational safety, occupational hygiene, and work & organisational psychology. These professionals were meant to work together as a team. If reported sick, Dutch workers have been required to visit their OP for independent judgement of sick leave and for rehabilitation purposes (Schaufeli & Kompier, 2001). OHSs have been internally or externally organized. By far most Dutch employers, as the Dutch police force in the presented study, have contracts with independent externally organized OHSs. This should help employers to improve their working conditions and to prevent sick leave and work disability.


13

Recent developments in Dutch disability legislation: The Donner Committees Because of the high number of individuals entering the Occupational Disability Insurance Act (WAO) and because of the high proportion of claims that were due to mental problems, the government formed a committee to address these issues (Donner committee). This Donner committee concentrated on the rise in disability due to mental health problems and made various recommendations to improve the prevention, care, treatment, and RTW in case of sick leave due to mental health problems (CPA, 2001). In 2001, a second Donner committee was formed, to further analyse the problem of sick leave and work disability. This committee advised that only fully and permanently disabled individuals should be eligible for disability benefits. The proposals of the Donner committees have been translated into new plans of the government for redesigning the WAO. In the Gatekeeper Improvement Act (Wet Verbetering Poortwachter), the obligations of employers and workers are stated with respect to their activities aimed at activation and RTW in the first year of sick leave. Also the responsibilities of the institutions involved were more clearly discerned: the OHSs should assist the employer and worker to prevent sick leave and promote RTW, whereas the Workers Insurance Authority (UWV) is responsible for evaluating RTW efforts. Under the current legislation, employers have a higher financial risk. They have to continue to pay the wages for the first year of sickness by themselves, a period that was stretched to two years in January 2004. This should encourage employers to have better policies concerning working conditions. For workers, the level and duration of benefits are currently less favourable. In April 2002, the Gatekeeper Improvement Act came into force. The societal effects of the Gatekeeper Improvement Act seem favourable as sick leave and the number of disability benefits declined (OECD, 2004).

The compulsory assistance of workers on sick leave and the introduction of the commercial OHS resulted in new relationships between employer, worker and OHS. Employers were often inclined to conclude minimum contracts with the commercial OHSs, which do not give OHSs much opportunity of intervention if necessary (Post, 2005). Furthermore, the commercialization of the OHSs can affect the independent position of the OP. Within these complex relationships it was important for the OHS to ensure a good quality of care. In recent years several professional guidelines for OPs have been developed for this purpose as well. These guidelines are based on a time contingent approach towards treatment and RTW, which means that the activities of the worker on sick leave increase according to a prestructured time schedule. This was in contrast to the traditional symptom contingent approach in which the course of the symptoms was the guiding principle towards treatment and RTW. A time contingent approach focuses on the activities, which the worker should be able to despite the symptoms.

Since 2006 the Work Conditions Act has changed in a less regulating law, creating a situation in which employers are not obligated anymore to be affiliated with an OHS. However, nowadays still the majority of employers do have contracts with OHSs, and many OPs still work for OHSs, although some have started their own corporation. OCCUPATIONAL HEALTH CARE OF MENTAL HEALTH PROBLEMS Occupational medicine plays a different role in the provision of occupational health care worldwide (e.g. Ladou, 2005; Grove, 2006). However, some trends can be seen in relation to the development of occupational mental health care of workers with mental health problems. Here, the focus will be on the Dutch situation as this may be seen as a trendsetter in this field, instigated by the Dutch worker’s compensation system and the specific role of the Dutch OP.


14

In 1993, a study by Schröer clarified that the existing style of guidance by OPs on common mental health problems was ineffective. In 1994, Terluin published a study on 'surménage', which is French for mental overload or nervous breakdown. Surmenage is a diagnosis commonly used by GPs and OPs to denote a maladaptive response to psychological stress in everyday life. This diagnostic concept, closely related to the DSM diagnosis Adjustment Disorder, reflected problems that could be handled in primary care (Terluin, 1994). At the same time, van der Klink et al. (1993) published a handbook for OPs on guidance for mental disorders related to work. An activating policy was recommended, aimed at both the individual worker and the work environment. These studies represent a change in attitude from a passive, reactive, non-directive style of guidance to a more activating, proactive, therapeutic style. In the early 1990s, the role of OPs regarding sick leave and work disability was seen as predominantly supporting the RTW of the patient from the moment that a more or less stable state of recovery was established. In the years thereafter, the profession aimed to change the recovery process itself in cases of stagnation, prompted by the high risk of disability. Inspired by the Individual Placement and Support model in the vocational rehabilitation for people with severe mental illnesses (Bond et al., 2008; Michon, 2006), activation became the key concept, based on cognitive behavioural principles and graded activity (van der Klink, 2002).

According to Van der Klink (2002), an individual experiences a lack of control in his or her direct interaction with actual features of the environment. An adjustment disorder may result in depression or anxiety disorders, if the instigating crisis remains unsolved. Therefore, he argued that minimal interventions should focus on the stress-related component, and enhance problem-solving capacities for workers to cope with regular problems in work. In this manner, adequate and early guidance of individuals with adjustment disorders could prevent the development of more serious psychopathology.

In addition, enduring sick leave in itself has secondary consequences that are potentially harmful: loss of daily structure, diminished social contacts, and deterioration of self-esteem. Graded activity may be an important element that can help individuals RTW, as it signifies that individuals resume work partially, in order to enable full RTW (Blonk et al., 2006). Partial RTW can be viewed as a type of gradual exposure to the work situation. This type of exposure may promote full RTW through various mechanisms. By gradual work resumption, experiences of success may be fostered. That is, by performing the tasks that one is able to, individuals may acquire a sense of self-efficacy and control (Bandura & Adams, 1977). Exposure to work may also provide experiences that challenge dysfunctional beliefs (Tryon, 2005). Partial RTW might also help to establish a daily working rhythm, it may offer distraction, and it may promote commitment towards one’s work and colleagues. The importance of gradual work resumption is also advocated in the literature with respect to physical injury (Briand et al., 2007; Durand & Loisel, 2001). Gradual work resumption can be established by steps such as the implementation of worksite accommodations, temporarily assignment to alternative job tasks, and gradual increase of the worker’s hours or involvement in performing increasingly demanding job tasks (Franche et al., 2005; Shrey, 2000).

Until recently, OPs did not have the tools for such specific interventions. In the past it was not even considered a legitimate role for OPs to initiate therapeutic interventions, even though they were in a favourable position to do so. GPs were expected to initiate therapy, but were in a less advantageous position with regard to knowledge of the patient's work and work situation. The last decade, mental ill workers have been referred frequently to specialized psychological care. As the focus of mental health specialists has been mostly symptom-based, instead of work-based, workers may not get the optimal care they need and productivity loss may be higher than necessary. As


15

an alternative, ways have been sought to encourage OPs to play a more active intervening role. Guideline on the management of mental health problems Since 1999, the Netherlands Society of Occupational Medicine (NVAB) has been developing and disseminating evidence-based practice guidelines, as they are one of the most promising and effective tools for improving the quality of occupational health care (Grol & Wensing, 2001; van der Weide et al., 1999; Hulshof et al., 1999). In 2000, the NVAB (2000; van der Klink & van Dijk, 2003) published a practice guideline titled ‘The management by OPs of workers with common mental health problems’. It promotes an active attitude and activating approach of the OP, instead of a minimal role.

Inspired by the mentioned context, professional developments and findings of a literature study, van der Klink (2002) developed and evaluated a brief activating intervention for OPs. This was based on cognitive behavioural principles and included graded activity (van der Klink et al., 2003). The main aim of the intervention was to activate patients to develop and implement problem-solving strategies for daily (working) life problems. In a randomised cluster design, this intervention was compared with ‘usual care’ (UC), a treatment by OPs that was based on “empathic counseling, instruction about stress, lifestyle advice, and discussion of work problems with the patient and company management”. A significant difference was found in the duration until full RTW in favour of the activating intervention. The activating intervention was not superior with respect to psychological complaints; both groups reported a decrease in psychological symptoms over time. However, the intervention was considered to be successful and, together with existing evidence, experience in adjacent fields, and consensus procedures, this resulted in the new guideline. After its publication in 2000, the guideline was mailed to all members of the NVAB and several courses emerged to inform OPs about the content of the guideline.

The guideline was mainly based on the results of a study by Van der Klink et al. (2003). In a cluster randomized trial, the intervention consisted of a training in an activating approach by OPs, in which OPs may operate as counselors using elements of cognitive behavioral therapy (CBT) and facilitating RTW by work interventions. The intervention appeared to be effective in reducing productivity loss (i.e. fastening RTW) for workers with adjustment disorders, if compared to a passive UC. In a prognostic study, Nieuwenhuijsen et al. (2003, 2005) showed that guideline-based care may reduce productive loss by fastening RTW for workers on sick leave with common mental disorders, but reduces treatment satisfaction of the worker.

The guideline focuses on five aspects of the management of mental health problems. First, a problem orientation in which the OP acknowledges the interaction between the disabled worker and his surroundings (work, personal and care). Second, a simplified classification of mental health problems is introduced, with only four categories: adjustment disorder (distress, nervous breakdown, burnout); depression; anxiety; and remaining psychiatric disorders. Third, early and activating interventions by the OP are promoted, in which time contingent RTW is part of the recovery process, even if the mental health problems are not related to work. The OP is stimulated to operate as counsellor, applying cognitive behavioural techniques, of workers with stress- and/or work-related problems. Fourth, time contingent evaluation in which the OP acts as case manager, who intervenes when recovery stagnates. Finally, relapse prevention is an integral aspect of the treatment.


16

CONCEPTUAL MODEL World Health Organization Model of Functioning (ICF) Occupational health care typically has a multidimensional focus on health since it addresses not only physical and mental disorders, but also social functioning. The World Health Organization (WHO) explains in its International Classification of Functioning, Disability, and Health (ICF model) how disease and disability are related (Figure 1). The model considers the influence of disease and its intermediaries on an individual’s participation in society. Diseases or disorders affect the triad of “body structure and function”, “activities”, and “participation”, which lead to either disability or no disability, depending on important conditional factors of environmental origin, such as heavy physical work, and of personal origin, such as personal ideas about disability (WHO, 2001). By using the ICF in the coaching and treatment of workers, a shift from a medical assessment to a broader ‘biopsychosocial’ evaluation is promoted (Heerkens et al., 2004). The ICF model is supported by many studies that have investigated the prognosis for RTW among patients suffering from a variety of diseases. From these studies, it can be concluded that the severity of the disease resulting in impairment of body function or structure usually has the largest influence on the time needed to RTW, but environmental factors and person-related factors play an additional role (Spelten et al., 2003). Looking further into personal factors, it has been found that the expectation of the patient about recovery best predicts the time taken to RTW and is better than those of the physician (Fleten et al., 2004; Nieuwenhuijsen, 2004; Cole et al., 2002). Conceptual model of this thesis Figure 1 shows the ICF model applied to RTW of workers with mental health problems (Sanderson et al., 2008; Nieuwenhuijsen, 2004; Cole et al., 2002). This model can be helpful in understanding the dimensions of sick leave due to mental health problems and factors that influence RTW (Wasiak et al., 2007; Schultz et al., 2007). The model has been used as a conceptual model for this thesis, especially regarding factors that influence RTW, based on the adaptations by Verbeek (2006). With respect to sick leave of workers with mental health problems, the point of departure is a worker who is on sick leave due to his or her health condition. This condition (for instance a depression) has led to an impairment of mental functions, e.g. impaired energy function leading to fatigue (Cieza et al., 2004). This impairment interacted with environmental factors, such as work characteristics, and personal factors, such as coping style, in determining the extent of the limitations in activities and participation (sick leave).

RTW from sick leave is the central outcome of this thesis. In accordance with the multi-factorial view on sick leave, we assume that regaining health does not necessarily result in RTW. When related to RTW, concepts of the ICF model can be specified as following. Impairments in mental functions (‘Body Functions and Structure’) may be made operational as experiencing fatigue, anxiety, or depressive symptoms. Irrational cognitions, which are assumed to be one of the causes of psychological symptoms, can also be categorised as belonging to the concept of impairment in mental functioning. RTW was initially considered within the Activities and Participation domains, which are said to give an indication of the experience of disability (Wasiak et al., 2007). In this thesis, we assume impairments in work performance as activity limitations, while we view RTW as the end of participation limitations (Verbeek, 2006; Nieuwenhuijsen, 2004). Contextual factors, such as work characteristics and supervisory behaviour during the RTW process, may be regarded as environmental factors. Personal factors include demographic variables, such as age, coping style, and recovery expectations of the worker.


17

Figure 1 The ICF model adjusted to this thesis (Rebergen, 2008)

We assume that a reduction of mental impairments and activity limitations will

contribute to RTW of the worker on sick leave. However, environmental factors and personal factors will also influence this process. In workers who are on sick leave due to mental health problems, this notion is substantiated by the finding that symptom reduction is not always immediately followed by RTW and earlier RTW is not always accompanied by less symptoms (van der Klink, 2002).

How do the elements of the guideline fit into this model? In general, in congruence with the ICF-model, the guideline assumes that the OP acknowledges the interaction between the disabled worker and the different stakeholders (work, personal and care) (NVAB, 2000). The model offers three opportunities for interventions by OPs: a) regaining proper mental functioning by means of reduction of impairments; b) restoring activities by means of reduction of activity limitation; c) participation by RTW (Verbeek, 2006). Focussing on the health condition and the first opportunity (a), a simplified problem orientation and diagnostic classification of mental health problems is introduced (1). This aims to improve the diagnostic skills of the OP and to initiate an early and activating guidance, in which RTW is part of the recovery process. The second opportunity (b), regarding environmental factors, advocates process management in a cyclic manner (diagnosis, intervention and evaluation), time contingent evaluation with patient, work and curative care, and workplace adaptations to prevent disability and facilitate RTW (2).

Health Condition

Mental health problems

a) Improvement in mental functioning

e.g. reduction fatigue, depressive symptoms

b) Restoring of activities

e.g. recovery of work performance

c) Improvement in Participation

e.g. return to work

Personal Factors

e.g. age, gender, coping style or recovery expectations

Environmental factors

e.g. supervisor behaviour or work

characteristics

1. Problem orientation Diagnosis

Work focused treatment

2. Process manager Evaluation patient/work/GP

Adapt work environment

3. Individual interventions Counseling stress/work Support employment


18

The third opportunity (c), regarding personal factors, provides OPs with counseling to improve or learn skills to the worker (supported employment; Crowther, 2001) (3). Additionally, OPs can do counseling using cognitive behaviour principles to change irrational recovery expectations of the worker (Nieuwenhuijsen, 2004). Counseling in primary care and workplace counseling on stress prevention has proven to be effective in the reduction of stress symptoms and is not associated with more costs (Huibers et al., 2007; Bower et al., 2005 & 2006; Henderson et al., 2004; Verbeek, 2004; McLeod & Henderson, 2003). The guideline stimulates the OP to counsel workers on work-related issues and workers with adjustment disorders on the individual level. When recovery stagnates and/or when the OP diagnoses a depression, anxiety or remaining psychiatric disorder, the OP needs to communicate with/refer to the GP and/or to a professional in secondary care. The third opportunity regarding counseling work- and stress-related problems is the most renewing element, compared to the management by OPs before the introduction of the guideline (van der Klink & Terluin, 2005). OBJECTIVES The main aim of this thesis is to contribute to quality improvement of occupational health care of workers with common mental health problems. The first way to attain this goal is to evaluate the implementation and (cost-)effectiveness of the guideline on RTW and treatment satisfaction. Therefore, we conducted the CO-OP study, to evaluate the effectiveness of counseling by OPs according to the guideline. In a randomised controlled trial we examined the effects of training of the guideline on RTW and satisfaction of workers with common mental health problems, compared to usual care with a minimal involvement of the OP and frequent and immediate referral to a psychologist. Second, this thesis focuses on the quality of the process of occupational health care among workers with common mental health problems and to relate this process to RTW according to the ICF model (Verbeek, 2006). OUTLINE OF THE THESIS First, in chapter 2 a systematic review is presented on RTW interventions for adjustment or stress-related disorders. These are common mental health problems in occupational health care. Chapter 3 describes the design of the CO-OP study, in which the intervention by OPs, guideline-based care of police workers, is evaluated and compared to care as usual. The results on return to work and treatment satisfaction are described in chapter 4. Chapter 5 presents results of an economic evaluation performed alongside the trial. A process evaluation was done to examine which elements of the treatment explain possible effects on RTW. These results are described in chapter 6. In chapter 7 results of a cross-sectional study are given, which explored the implementation process of the guideline on mental health problems by Dutch OPs. Written as separate articles for scientific journals the chapters have some overlap. The thesis concludes with a general discussion in chapter 8. Finally, this thesis contains a summary in English and Dutch. REFERENCES American Psychiatric Association (APA). Diagnostic and statistical manual of mental disorders DSM-IV. Washington, D.C.: APA, 1994. Anema JR, Jettinghoff K, Houtman ILD, Schoemaker CG, Buijs PC, van den Berg R. Medical care of employees long-term sick listed due to mental health problems: A cohort study to describe and


19

compare the care of the occupational physician and the general practitioner. J Occ Rehabil 2006;16:41-52. Bandura A, Adams NE. Analysis of self-efficacy theory of behavioural change. Cognitive Therapy and Research 1997;1(4):287-310. Bond GR, Drake RE, Becker DR. An update on randomized controlled trials of evidence-based supported employment. Psychiatr Rehabil J. 2008;31(4):280-90. Bower P, Gilbody S. Managing common mental health disorders in primary care: conceptual models and evidence base. BMJ. 2005;9;330(7495):839-42. Bower P, Rowland N. Effectiveness and cost effectiveness of counseling in primary care. Cochrane Database Syst Rev. 2006;3:CD001025. Briand C, Durand MJ, St-Arnaud L, Corbière M. Work and mental health: learning from return-to-work rehabilitation programs designed for workers with musculoskeletal disorders. Int J Law Psychiatry. 2007;30(4-5):444-57. Cieza A, Chatterji S, Andersen C, Cantista P, Herceg M, Melvin J, Stucki G, de Bie R. ICF Core Sets for depression. J Rehabil Med. 2004;(44 Suppl):128-34. Cole DC, Mondloch MV, Hogg-Johnson S. Listening to injured workers: How recovery expectations predict outcomes—A prospective study. CMAJ 2002;166:749–754. Commissie Psychische Arbeidsongeschiktheid (CPA). Guideline approach on work disability due to mental health problems. (In Dutch; Leidraad aanpak verzuim om psychische redenen). The Hague, 2001. Crowther R, Marshall M, Bond G, Huxley P. Vocational rehabilitation for people with severe mental illness. Cochrane Database Syst Rev 2001: CD003080. Durand MJ, Loisel P. Therapeutic Return to Work: Rehabilitation in the workplace. Work 2001;17:57–63. Fleten N, Johnsen R, Forde OH. Length of sick leave—why not ask the sick-listed? Sick-listed individuals predict their length of sick leave more accurately than professionals. BMC Public Health 2004;4:46. Franche RL, Cullen K, Clarke J, Irvin E, Sinclair S, Frank J. Workplace-based return-to-work interventions: A systematic review of the quantitative literature. Journal of Occupational Rehabilitation 2005;15(4):607-31. Grol R, Wensing M. Implementation. Effective change in patient care. (In Dutch: Implementatie. Effectieve verandering in de patiëntenzorg) Maarssen, 2001. Grove B. Common mental health problems in the workplace: how can occupational physicians help?Occup Med (Lond). 2006;56(5):291-3. Heerkens Y, Engels J, Kuiper J, van der Gulden J, Oostendorp R. The use of the ICF to describe work related factors influencing the health of employees. Disability & Rehabilitation 2004;26(17):1060-6. Henderson M, Hotopf M, Wessely S. Workplace counseling. Occup Environ Med. 2003;60(12):899-900. Houtman ILD, Schoemaker CG, Blatter BM, de Vroome EMM, van den Berg R, Bijl RV. Psychological complaints, interventions and rehabilitation to work; the prognostic study of INVENT. (in Dutch). Hoofddorp: TNO Work & Employment, 2002. Huibers MJ, Beurskens AJ, Bleijenberg G, van Schayck CP. Psychosocial interventions by general practitioners. Cochrane Database Syst Rev. 2007;(3):CD003494. Hulshof CT, Verbeek JH, van Dijk FJ, van der Weide WE, Braam IT. Evaluation research in occupational health services: general principles and a systematic review of empirical studies. Occup Environ Med 1999;56:361 -77. International Labor Organization (ILO). Mental Health and Work: Impact, Issues and Good Practices. WHO and ILO, 2000 (www.wfmh.org). International Monetary Fund (IMF). Work absence in Europe. IMF-working paper. Washington: IMF, 2004.


20

Ladou J. Occupational Medicine. The Case for Reform. Am J Prev Med 2005;28(4):396-402. Lerner D, Henke RM. What Does Research Tell Us About Depression, Job Performance, and Work Productivity? J Occup Environ Med. 2008;50(4):401–410. McLeod J, Henderson M. For: does workplace counseling work? Br J Psychiatry. 2003;182:103-4. Michon H. Personal characteristics in vocational rehabilitation for people with severe mental illnesses. PhD-thesis. The Netherlands. University of amsterdam, Trimbos Institute; 2006. Nieuwenhuijsen K. Employees with common mental disorders: from diagnosis to return to work. PhD Thesis. The Netherlands. University of Amsterdam, 2004. Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. Occupational and Environmental Medicine 2003;60:21-5. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, van Dijk FJ. Validation of performance indicators for rehabilitation of workers with mental health problems. Medical Care 2005;43(10):1034-42. NVAB (Netherlands Society of Occupational Medicine). Guideline on the management of employees with mental health problems. (van der Klink, JJL, Ed., in Dutch). Eindhoven, 2000. OECD Economic Surveys: Netherlands. Reform of the sickness and disability benefit schemes. OECD, 2004. Post M. Return to work in the first year of sickness absence. An evaluation of the Gatekeeper Improvement Act. Thesis. University Groningen, 2005. Sanderson K, Nicholson J, Graves N, Tilse E, Oldenburg B. Mental health in the workplace: Using the ICF to model the prospective associations between symptoms, activities, participation and environmental factors.Disabil Rehabil. 2008 ;7:1-9. Schaufeli WB, Kompier MAJ. Managing job stress in the Netherlands. International Journal of Stress Management 2001;8:15-34. Schröer CAP. Absenteeism due to 'overstrain'. A study of the nature of overstrain, therapeutic assistance and absenteeism (PhD thesis, in Dutch). University of Maastricht, 1993. Schultz IZ, Stowell AW, Feuerstein M, Gatchel RJ. Models of Return to Work for Musculoskeletal Disorders. J Occup Rehabil 2007;17(2):327-52. Shrey DE. Worksite disability management model for effective return-to-work planning. Occupational Medicine 2000;15(4):789-801. Spelten ER, Verbeek JH, Uitterhoeve AL, Ansink AC, van der Lelie J, et al. Cancer, fatigue and the return of patients to work—A prospective cohort study. Eur J Cancer 2003;39:1562–1567. Terluin B. Nervous breakdown substantiated. A study of the general practitioner's diagnosis of surmenage. PhD-Thesis. The Netherlands, Utrecht: Detam, 1994. Tryon WW. Possible mechanisms for why desensitization and exposure therapy work. Clinical Psychology Review 2005;25(1):67-95. van der Klink JJL (Ed.). Mental health problems and work. Handbook for an activating guidance by occupational physicians [in Dutch]. Amsterdam: Nederlands Instituut voor Arbeidsomstandigheden (NIA), 1993. van der Klink JJL (Ed.). Mental Disorders and Work. Guideline for GeneralPractitioners and Occupational Physicians (ln Dutch). Utrecht/Amsterdam: KNMG/SKB, 2000. van der Klink JJL. Back in balance. The development and evaluation of an occupational health intervention for work-related adjustment disorders. PhD-Thesis. The Netherlands. University of Amsterdam, 2002. van der Klink JJL, Blonk RWB. Schene AH, van Dijk FJH. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomised controlled design. Occup Environ Med 2003;60:429-37.


21

van der Klink JJL, Terluin B. Psychological problems and work. Manual for an activating guidance by the general practitioner and occupational health worker [In Dutch]. Houten: Bohn Stafleu van Loghum, 2005. van der Klink JJL, Blonk RWB, Schene AH, van Dijk FJH. The benefits of interventions for work related stress. Am J Public Health 2001;91:270-76. van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scandinavian Journal of Work, Environment & Health 2003;29:478–487. van der Weide WE, Verbeek JHAM, van Dijk FJG, Hulshof CTJ. The development and evaluation of a quality assessment instrument for occupational physicians. Occ Env Med 2003;60:l21-l25. Verbeek JH. The evidence for workplace counseling is in Medline. Occup Environ Med. 2004;61(6):558-9. Verbeek JH. How can doctors help their patients to return to work? PLoS Med 2006;3(3):e88. Wang JL. Work stress as a risk factor for major depressive episode(s). Psychological Medicine 2005;35:865-71. Wang JL, Adair CE, Patten SB. Mental health and related disability among workers: A population-based study. American Journal of Industrial Medicine 2006;49:514-22. Wasiak R, Young AE, Roessler RT, McPherson KM, van Poppel MN, Anema JR. Measuring return to work. J Occup Rehabil 2007;17(4):766-81. World Health Organisation (WHO). The ICD-10 classification of mental and behavioural disorders: clinical descriptions and diagnostic guidelines. Geneva: WHO, 1992. World Health Organization (WHO). International Classification of Functioning, Disability and Health (ICF). WHO, 2001. World Health Organization (WHO). The World Health Report. Mental Health: New understanding, New hope. WHO, 2002.

CHAPTER 2

Return to work interventions for adjustment disorders

Submitted for publication to Cochrane Occupational Health Field as

Bruinvels DJ, Rebergen DS, Verbeek JH, Nieuwenhuijsen K,Madan I,Neumeyer-Gromen A. Return to work interventions for adjustment disorders.

Published as Protocol in Cochrane Database of Systematic Reviews 2007 Issue 1. Art. No.: CD006389. DOI: 10.1002/14651858.CD006389.

COCHRANE REVIEW

24

ABSTRACT Background Absenteeism among workers is often caused by emotional and psychological stress related to significant life events. Stress often leads to symptoms and impaired functioning and can result in an acute adjustment disorder. Currently no systematic review on return to work interventions for adjustment disorders of workers on sick leave exists, and is the focus of this review. Objectives 1. To evaluate the effectiveness of interventions aimed at return to work for workers with adjustment disorders. 2. To assess heterogeneity of each intervention and perform a meta-analysis for homogeneous groups of studies. Search methods Electronic database searches of the CCDAN Controlled Trials Register, Cochrane Controlled Trials Register, Cochrane Library CENTRAL Register, Cochrane Occupational Health Field Trials and Review Database, MEDLINE, EMBASE, and PsycINFO were performed. The reference lists of all references that are retrieved as full papers and potentially relevant, as well as relevant systematic reviews and literature reviews, were checked to identify other potentially relevant articles. Finally, abstracts from national and international psychiatry, psychology and occupational health conferences were scrutinised to identify unpublished studies. Selection criteria All randomised controlled trials that evaluate return to work interventions for adjustment disorders were considered. The review focuses on all interventions aimed at return to work, using individual or group approaches, including pharmacologic interventions, cognitive behavioural interventions, relaxation techniques and multimodal programmes. Data collection and analysis Two review authors independently selected suitable studies for inclusion in this review. The methodological quality of the included articles were assessed using the checklist of Downs and Black. Subsequently, the two review authors completed the extraction of data from the papers. Primary outcomes, such as time lost from work, were calculated using the weighted standardised mean difference. Dichotomous data, such as able/not able to return to work, were calculated using Mantel-Haenszel odds ratios. Secondary outcomes were assessed using continuous (for example, changes on psychometric scales), categorical (for example, one of three categories on a quality of life scale, such as 'better', 'worse' or 'no change'), or dichotomous (for example, remission from adjustment disorder vs no remission) measures. Results The initial search in electronic databases identified 3789 publications. Based on title and abstract, 32 eligible publications were identified and the full text of the articles was examined. Only two studies met the inclusion criteria and were included in the review. Additionally, four other studies that met the inclusion criteria were identified by searching relevant web sites on the Internet and checking the references from published reviews. Four of the included studies reported on the effects of cognitive behavioural therapy (CBT)

COCHRANE REVIEW

25

on return to work (RTW). Workers treated with CBT started 17.40 (8.14, 26.65) days faster with partial RTW and 14.69 (1.45, 27.96) days faster with full RTW. Mental health symptoms after 2-4 and 10-18 months, such as distress, depression and anxiety, were also significantly lower in workers treated with CBT. Based on a single study, no evidence was found that solution-focused behavioural therapy facilitates RTW. Based on the remaining single study, no evidence was found that a postal intervention aimed exclusively at work adjustments facilitates RTW of workers with adjustment disorders. Authors' conclusions This review found evidence that CBT may facilitate RTW of workers with adjustment disorders. On average, workers who are offered CBT will start two weeks earlier with partial and full RTW. A second finding of this review is that CBT improves the mental health of workers with adjustment disorders. This finding actually supports the hypothesis that early RTW may be associated with improved mental health. Based on a single study, the third finding is that there is no evidence that solution-focused behavioural therapy facilitates RTW of workers with adjustment disorders. The fourth and final finding of this review is that, based on a single study, there is no evidence that an intervention aimed exclusively at work adjustments facilitates RTW of workers with adjustment disorders Plain language summary The effects of cognitive behavioural therapy on return to work This review found evidence that cognitive behavioural therapy (CBT), a commonly used type of psychotherapy, may facilitate return to work of workers with stress-related mental health problems. On average, workers who are offered CBT will start two weeks earlier with partial and full RTW compared to workers who received care as usual. A second finding of this review is that CBT improves the mental health of workers with adjustment disorders. This finding seems to support the hypothesis that early RTW is associated with improved mental health. BACKGROUND Absenteeism among workers is often caused by emotional and psychological stress related to significant life events. Stress often leads to symptoms and impaired functioning and can result in an acute adjustment disorder. Adjustment disorders are defined in both the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (APA, 1994) and the International Statistical Classification of Diseases and Related Health Problems (ICD-10) (WHO, 1992).

DSM-IV has defined adjustment disorders as a debilitating reaction, usually lasting less than six months, to a stressful event or situation. The development of emotional or behavioural symptoms in response to an identifiable stressor(s) occurring within three months of the onset of the stressor(s). Adjustment disorders can be classified acute if the disturbance lasts less than six months. The following diagnostic criteria for adjustment disorders are defined by DSM-IV: A. The development of emotional or behavioural symptoms in response to an identifiable stressor(s) occurring within three months of the onset of the stressor(s). B. These symptoms or behaviours are clinically significant as evidenced by either of the following: (1) marked distress that is in excess of what would be expected from exposure to stressor

COCHRANE REVIEW

26

(2) significant impairment in social or occupational (academic) functioning C. The stress-related disturbance does not meet the criteria for another specific Axis I disorder and is not merely an exacerbation of a pre-existing Axis I or Axis II disorder. D. The symptoms do not represent bereavement. E. Once the stressor (or its consequences) has terminated, the symptoms do not persist for more than an additional 6 months.

The ICD, a detailed description of known diseases and injuries, is published by the World Health Organisation, is revised periodically (last revision in 1992) and is currently in its tenth edition. The ICD-10 defines the following diagnostic criteria for adjustment disorders: ‘States of subjective distress and emotional disturbance, usually interfering with social functioning and performance, arising in the period of adaptation to a significant life change or a stressful life event. The stressor may have affected the integrity of an individual's social network (bereavement, separation experiences) or the wider system of social supports and values (migration, refugee status), or represented a major developmental transition or crisis (going to school, becoming a parent, failure to attain a cherished personal goal, retirement). Individual predisposition or vulnerability plays an important role in the risk of occurrence and the shaping of the manifestations of adjustment disorders, but it is nevertheless assumed that the condition would not have arisen without the stressor. The manifestations vary and include depressed mood, anxiety or worry (or mixture of these), a feeling of inability to cope, plan ahead, or continue in the present situation, as well as some degree of disability in the performance of daily routine.’

Adjustment disorders are very common in workers. In one survey, 50% of responders reported being extremely, very, or moderately stressed at work (Smith et al., 1998). Additionally, significant life events may also cause stress related disorders. Although adjustment disorders are considered mild compared to psychiatric disorders, at least 20% of patients with such a disorder does not return to work within a year (Schroer, 1993). For example, work disability as a result of mental health problems accounts for 30% of all disability benefits in the Netherlands. A majority (69%) of those employees suffer from minor psychiatric disorders such as adjustment disorders (Lisv, 2000). More than 10% of total claims for occupational diseases are adjustment disorders attributed to stress at work (NCvB, 2004).

Many interventions are available for workers with adjustment disorders. Interventions can be aimed at the individual worker or groups of workers. Examples are pharmacologic interventions, such as antidepressant treatment (Hameed et al., 2005), cognitive coping strategies, such as cognitive behavioural therapy (Blonk et al., 2006a), relaxation techniques, exercise programmes, and employee assistance programs. The outcome of these interventions is often reduction of emotional and behavioural symptoms and adequate coping behaviour of the worker, resulting in return to work.

A first search of publications on adjustment disorders and return to work interventions using MEDLINE revealed more than 4000 published articles. The majority of these publications discussed the impact of stress on workers, screening of workers at risk of adjustment disorders, and preventive interventions. Less than 5% of the articles discussed absenteeism and return to work interventions. Most of these articles were published less than 10 years ago.

Currently no systematic review on return to work interventions for adjustment disorders of workers on sick leave exists, and is the focus of this review. One of the co-authors of this review has conducted a Cochrane review on depressive disorders and return to work (Nieuwenhuijsen et al., 2008). Another Cochrane review on the effectiveness of occupational stress management programmes has been published

COCHRANE REVIEW

27

(Marine et al., 2006), which focuses on the efficacy of interventions aimed at reduction of stress symptoms in health care professionals not diagnosed with mental health problems, but exposed to stress at work. Objectives 1. To evaluate the effectiveness of interventions aimed at return to work for workers with adjustment disorders. 2. To assess heterogeneity of each intervention and perform a meta-analysis for homogeneous groups of studies. METHODS Criteria for considering studies for this review Types of studies All randomised controlled trials that evaluate return to work interventions for adjustment disorders were considered. Types of participants Adults (18 years or older) with work disability related to an adjustment disorder causing absenteeism. Adjustment disorders are defined as acute significant emotional or behaviour problems in response to an identified stressor, as described in DSM-IV (APA, 1994) and ICD-10 (WHO, 1992) criteria. Stressors can be related to work or to significant life events. Patients in all organisational settings and job positions were included in the review, including non-paid workers and apprentices/trainees. Patients with chronic adjustment disorders were excluded. Patients with other common mental health disorders and psychiatric disorders were excluded as well. Types of interventions The review will focuses on all interventions aimed at return to work, using individual or group approaches, including pharmacologic interventions, cognitive behavioural interventions, relaxation techniques and multimodal programmes (Terluin et al., 2005; van der Klink et al., 2001). Interventions aimed at individual workers were categorised into the following groups: 1. pharmacotherapy (eg antidepressant treatment) 2. psychological therapy (eg cognitive behavioural therapy, psychodynamic therapy) 3. relaxation techniques (eg yoga) 4. exercise programmes (eg running or fitness training) 5. employee assistance programs (eg participatory ergonomics). Interventions aimed at the workplace, such as occupational stress management programmes, were managed separately. The main comparisons are: 1. Return to work intervention versus no treatment control. 2. Return to work intervention versus alternative treatment. Types of outcome measures Work-status outcomes (primary outcomes) 1. time lost from work, like loss of working days due to absenteeism 2. time on selected/appropriate/light/modified duties, eg productivity loss on working day

COCHRANE REVIEW

28

3. other reported changes in work status, like a change of function or working location without loss of productivity 4. functional status in relation to job demands expressed in terms of "can perform task" or cannot perform task" 5. able/not able to return to work. Other outcomes (secondary outcomes) 1. clinical status as measured by a psychometric scale like the BDI (Beck & Steer, 1987) or HADS (Zigmond & Snaith, 1983) or structured diagnostic interviews. 2. generic functional status and quality of life as measured by questionnaires such as the SF-36 (Ware, 1992) and EuroQol (EuroQol, 1990) 3. patient compliance to the intervention 4. trial drop-out. Search methods for identification of studies 1. Electronic database searches Initially, the CCDAN Controlled Trials Register (CCDANCTR-Studies and CCDANCTR- References) was searched to identify all potentially eligible studies, and the most recent date that the register was checked was noted. Electronic databases including the Cochrane Controlled Trials Register (CCTR), Cochrane Library CENTRAL Register (recording the issues searched, search strategy, and mistakes made), Cochrane Occupational Health Field (COHF) Trials and Review Database, MEDLINE (1966 to present), EMBASE (1980 to present), and PsycINFO (1887 to present) were searched to identify potentially eligible studies and review articles.

For CCTR and COHF Trials and Review Database a search strategy was used for studies on adjustment disorders (Terluin, 2005) by using the following terms: (adjustment next disorder*) or burnout or (occupational next stress) or (job next stress) or (work next stress) or neurasthenia or (minor next depression) or (emotional next disorder*).

For MEDLINE, EMBASE and PsycINFO databases, a search strategy for studies on adjustment disorders (Terluin et al., 2005) was combined with a search strategy on occupational health intervention studies (Verbeek et al., 2005) and a search strategy to identify RCTs (Robinson & Dickersin, 2003). Search terms are presented in Supplement 1. All relevant foreign language papers were translated. 2. Reference lists The reference lists of all references that are retrieved as full papers and potentially relevant, as well as relevant systematic reviews and literature reviews, were checked to identify other potentially relevant articles. These articles were retrieved and assessed for possible inclusion in the review. 3. Personal communication The lead author of all relevant reports identified were written to in order to ascertain if they know of any additional published or unpublished studies that might be relevant to the review. 4. Conference proceedings Abstracts from national and international psychiatry, psychology and occupational health conferences were scrutinised to identify unpublished studies. These included meetings organised by national and international medical colleges, specialty societies and professional organisations. The authors of these studies were contacted to obtain further

COCHRANE REVIEW

29

details about the study and to enquire if they know of any other unpublished or published relevant work. Data collection and analysis 1. Selection of studies for inclusion 1.1 Two review authors (DB and DR) independently selected suitable studies for inclusion in this review as detailed below. Where the two review authors disagreed about the inclusion of a study, disagreements were resolved by consensus of opinion, and a third review author was consulted if they could not be resolved. Where resolution was not possible the study author was contacted to obtain more information and clarification. The titles and abstracts of studies identified by searching electronic databases were assessed to determine whether each article met the eligibility criteria. If the title and abstract contained sufficient information to determine that the article did not meet the inclusion criteria, then it was rejected. A record of all rejected papers and the reasons for rejection were documented. 1.2 The full papers of all remaining titles and abstracts deemed relevant were retrieved. In addition, all other potentially relevant articles identified by the various search strategies (reference checking, personal communications etc) were reviewed. All papers in languages other than English were translated. All articles were reviewed independently by two of the review authors (DB and DR), who completed a form for each study and scored the eligibility of the study. Disagreements were resolved as mentioned above. On the form, data on the study type, participants, interventions and outcome measures were written down. The reasons for exclusion were documented. Where the same study had more than one article written about the outcomes, all articles were treated as one study and the results were presented only once. 2. Quality assessment of studies The methodological quality of the remaining articles were assessed using the checklist of Downs and Black (1998), which is in accordance with the guidelines in the Cochrane Reviewers Handbook (Alderson et al., 2004). Items in this checklist included the way data are reported, external validity, internal validity (bias and confounding), and power of the study. All articles were reviewed independently by two of the review authors (DB and DR). Where the two review authors disagreed about the quality of a study, disagreements were resolved by consensus of opinion, and a third review author consulted if disagreements could not be resolved. Where resolution was not possible, the study author was contacted to obtain more information and clarification. 3. Data extraction The two review authors completed the extraction of data from the papers on to a form to elicit the following information: General: published/unpublished, title, authors, source, contact address, country,

language of publication, year of publication, duplicate publications Methods: design/allocation, allocation concealment, blinding of patients/treatment

providers/outcome assessors, study duration/follow-up, start/end dates, loss to follow-up, crossovers, co-interventions

Participants: number of participants, setting/type of work, region/country, recruitment, diagnosis, co-morbidity, inclusion/exclusion, age, sex, ethnicity, marital status, education level, social economic status

COCHRANE REVIEW

30

Interventions per treatment group: number of patients, treatment type/content, treatment provider, number of treatment providers, treatment frequency/duration, training/supervision of treatment providers

Outcomes: length of follow-up, absenteeism (timing and return to work), type of analysis for absenteeism, clinical outcomes (timing and recovery), type of analysis clinical outcomes, productivity, treatment compliance

Results: absenteeism (effect measure, SD, test statistic, CI), clinical outcomes (effect measure, SD, test statistic, CI)

A summary of data extracted from included studies was reported. If studies were available that were sufficiently similar and of sufficient quality, we have pooled those that could be grouped together and used the statistical techniques of meta-analysis. The data were managed using Review Manager software. 4. Dealing with missing data Where it was not possible to analyse data quantitatively as reported in published studies, we contacted the first author to obtain the additional data required. Where no further usable data were provided, studies were not included in the meta-analysis, and were listed as missing data. 5. Assessment of heterogeneity Graphical representations of the data were inspected; if the confidence intervals for the results of the study did not overlap, it suggests that the differences were likely to be statistically significant (Walker et al., 1988). In addition, differences between the results of each included trial were checked using the I

2 statistic as a test of heterogeneity. If there

was statistically significant heterogeneity, the data were presented separately rather than pooled. Results were analysed using both the fixed effect and random effects methods. However, where there was significant heterogeneity, a random effects model was used, and the review authors explored the reasons for this heterogeneity. 6. Assessment of reporting biases Data from all identified and selected trials were entered into a funnel plot (size of study versus effect size) (Egger et al., 1997), to attempt to detect the possibility of publication bias. 7. Data synthesis Primary outcomes, such as time lost from work, were calculated using the weighted standardised mean difference (WMD). Dichotomous data, such as able/not able to return to work, were calculated using Mantel-Haenszel odds ratios. Secondary outcomes were assessed using continuous (for example, changes on psychometric scales), categorical (for example, one of three categories on a quality of life scale, such as ‘better’, ‘worse’ or ‘no change’), or dichotomous (for example, remission from adjustment disorder vs no remission) measures.

Continuous data: Many rating scales were available to measure outcomes in psychological trials. These scales varied in the quality of their validation and reliability. Therefore, if a rating scale’s validation had not been published in a peer-reviewed journal, then the data were not included in this review. In addition, the rating scale should be either self-report or completed by an independent observer or relative. Trials that used the same instrument to measure specific outcomes were used in direct comparisons where possible. Where continuous data were presented from different scales measuring the same effect, both sets of data were presented and the general direction of the effect

COCHRANE REVIEW

31

inspected. The mean and standard deviation were reported. Where standard deviations were not reported in the paper, attempts were made to obtain them from the authors or to calculate them using other measures of variation that were reported, such as the confidence intervals. If possible, we pooled data from different scales that measured the same effect using the standardised mean difference (SMD).

Dichotomous data: Continuous outcome measures were converted to dichotomous data where necessary. If the authors of the study used a designated cut-off point for determining clinical effectiveness, the reviewers used this where appropriate. Otherwise, cut-offs on rating scales were identified and participants were divided on basis of whether they were ‘clinically improved’ or ‘not clinically improved’. For dichotomous outcomes, a Mantel-Haenszel odds ratio with its associated 95% confidence intervals (CI) was estimated. As a summary measure of effectiveness, where possible, the number needed to treat statistic (NNT) was also calculated. 8. Subgroup analysis and investigation of heterogeneity Clinical heterogeneity was investigated using sub-group analyses. This review investigated: 1. organisational setting 2. type/level of job undertaken 3. group vs individual therapy 4. setting of treatment providers. 9. Sensitivity analysis Methodological heterogeneity, which may have led to differences between the results of individual studies, was investigated using sensitivity analyses. This review investigated: 1. differences between studies that use self-reported or observer-rated outcome measures 2. differences between analyses involving all studies, and excluding trials of low methodological quality. RESULTS Description of studies Results of the search The initial search in electronic databases identified 3789 publications. Based on title and abstract, 32 elegible publications were identified and the full text of the articles was examined. Only two studies met the inclusion criteria and were included in the review (Nystuen et al., 2003; van der Klink et al., 2003). Additionally, four other studies that met the inclusion criteria were identified by searching relevant websites on the internet and checking the references from published reviews (Bakker et al., 2006; Blonk et al., 2006a; Brouwers et al., 2006; Fleten et al., 2006). Included studies Six studies were included in this review (Supplement 2). Two of the studies were from Norway (Fleten et al., 2006; Nystuen et al., 2003 ) and four from The Netherlands (Bakker et al., 2006; Blonk et al., 2006a; Brouwers et al., 2006; van der Klink et al., 2003). Additional information regarding study details and statistical data was provided by two authors (Bakker, 2006; Blonk, 2006a). Included studies were carried out between 1995 and 2004.

COCHRANE REVIEW

32

Design Of the six included studies, three were randomised controlled trials and three were cluster-randomised trials. Intraclass correlations for the studies were reported to be negligible. Five studies compared two groups of patients and one study compared three groups of patients. To include both intervention groups of the latter study in the quantitative part of this review, two identical references to this study were created (Blonk et al., 2006a; Blonk et al., 2006b). Sample sizes The total number of participants in various intervention groups was 1133, and in control groups 1036. The number of participants in the smallest intervention group was between 40 to 100 in two studies, between 101 and 200 in two studies, and more than 201 in two studies. None of the studies were a priori deemed to have sufficient power to detect a clinically relevant difference in return to work (RTW) of 2 weeks. Setting All studies explored the effects of interventions aimed at return to work, but originated from different healthcare settings. Two studies were in a primary care setting, three studies in a public and occupational healthcare setting, and one study in both a psychological care and a public and occupational healthcare setting. Participants This review focuses at adjustment disorders based on the DSM IV (APA, 1994) and ICD-10 (WHO, 1992) criteria. To diagnose an adjustment disorder in potential participants two studies used the CIDI (WHO 1990), two studies used the ICPC criteria for mental disorders (Brage 1996), one study used the DSM IV criteria (APA, 1994), and one study used the 4DSQ (Terluin et al., 2006). Four studies included only patients with adjustment disorders, and two studies also included patients with musculoskeletal disorders. These patients were excluded from this review and from these studies only data presented on the patients with adjustment disorders were used.

Participants were recruited at general practitioners offices in two studies, national insurance and social security offices in two studies, a private insurance company in one study, and an occupational health service in one study. Participants were employed in five studies, and self employed in one study.

Patients with (severe) psychiatric disorders were excluded from five of the studies. Other reasons to exclude patients were (recent) psychotherapy, recent sick leave, recent pregnancy or childbirth, and terminal illness.

The length of sick leave was also a criterion for inclusion or exclusion in five of the six studies. In two studies only patients with a sick leave of more than two weeks were included, and in one study with a sick leave of more than 7 weeks. In the two remaining studies patients were excluded with a sick leave of more than 3 months.

The average age of the participants ranged between 39 and 42 years. In four studies the percentage of female participants ranged between 58% and 67%. In the other two studies only 19% and 34% of the participants were female. Interventions One of the goals of the review was to identify RCT’s with RTW interventions aimed at individual workers. Only studies with psychological therapy were identiefied. No eligible studies on pharmacotherapy, relaxation techniques, exercise programmes, or employee assistance programs were found.

COCHRANE REVIEW

33

All six studies compared one or two interventions with usual care. Four studies offered cognitive behavioural therapy (CBT) to participants in the intervention group (Bakker et al., 2006; Blonk et al., 2006a; Brouwers et al., 2006; van der Klink et al., 2003), one study offered solution-focused behavioural therapy (SFBT; Gingerich & Eisengart, 2000) to participants (Nystuen et al., 2003), and one study used a minimal postal intervention aiming at work adjustments (Fleten et al., 2006).

In one study the intervention was provided by psychologists, in one study by psychologists or labour experts, in one study by general practitioners, in one study by occupational physicians, and in one study by social workers. In the study with the minimal postal intervention there were no treatment providers. Usual care was provided by general practitioners in five studies, and occupational physicians in one study.

The treatment frequency and duration of the non-postal interventions varied widely between the studies, ranging from less than three consultations of 10-20 minutes in one study to 11 consultations of 45 minutes in another study. The total length of the intervention ranged between a few weeks to 20 weeks. Outcomes Only studies which reported data on return to work (RTW) were included in this review. All six studies provided data on full RTW. Additionally, three studies reported data on partial RTW. The length of the follow-up on RTW was 12 months in four studies, 14-16 months in one study, and 18 months in one study. Eligible studies reporting time on selected/appropriate/light/ modified duties, changes in word status, functional status, and the (in)ability to RTW were not found.

Four of the six studies used a validated instrument to follow-up on the course of the adjustment disorder. Three studies used the Four-Dimensional Symptom Questionnaire (4DSQ; Terluin et al., 2006), and one study used the Depression Anxiety Stress Scale (DASS; Lovibond & Lovibond, 1993). Other psychometric instruments used in the studies were the Dutch version of the Maslach Burnout Inventory (MBI-NL; Maslach & Maslach, 1993; Schaufeli & van Dierendonck, 1994), the Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1993), the Symptom Checklist-90 (SCL-90; Derogatis, 1977; Arrindell & Ettema, 1981, 1986), the Mastery Scale (Pearlin & Schooler, 1978), the Utrecht Coping List (UCL; Schreurs et al., 1984, 1993), and the Dutch Work and Health Questionnaire (DWHQ, Grundemann et al., 1993). The length of follow-up ranged from 10 to 18 months. None of the studies measured generic functional status and quality of life, patient compliance to the intervention, or trial drop-out. Excluded studies 30 studies were excluded from the review using a study eligibility form (Supplement 3). Reasons for excluding the studies were: 1. Not a study but a review (Glicken, 1983; Gómez Sanabria et al., 2003) 2. The study is non-randomised by design (Akagi et al., 2001; Bunce & West, 1994; Cuijpers et al., 2005; Dettmers et al., 2003; Hätinen et al., 2002; Kushnir & Milbauer, 1994, Lehmer & Bentley, 1997; Mino et al., 2000; Natsume et al., 1996; Nieuwenhuijsen et al., 2003; Selishchev et al., 1998) 3. Less then 50% of the participants were on sick leave (Jackson ,1983; Kawakami et al., 1999; Mynors-Wallis et al., 1997; Salmela-Aro et al., 2004; Toivanen et al., 1993) 4. Participants do not have an adjustment disorder (Frank et al., 2002; Huibers et al., 2004) 5. The study does not have a RTW outcome (Bruning & Frew, 1987; Firth & Shapiro, 1986; Lange et al., 2004; Proudfoot et al., 2003; Terranova et al., 1997) 6. RTW-data of intervention and control group are not reported separately (Perski, 2004).

COCHRANE REVIEW

34

Ongoing studies There are two recently finished Dutch randomised controlled trial awaiting classification (de Vente et al., 2008; Rebergen et al., 2007). Results of an ongoing Dutch randomised controlled trials is expected to be published in 2009 (Oostrom et al., 2008,). Risk of bias in included studies To appraise the risk of bias within the individual studies the Downs and Black checklist on methodological quality was used (Downs & Blacks, 1998). The scores details are presented in the quality assessment table (Table 1). The overall Cohen s Kappa of agreement between two authors was 0.40. All differences between the two authors could easily be solved by discussion, without the need of a third author.

The overall mean score of the six included studies was 23.2 (SD 2.1) which is 72% of a maximum attainable score of 32. The applicability of studies was rated using the external validity sub scales. Four studies were rated as highly applicable. All six studies scored higher than 75% on the internal validity sub scales and were therefore rated as high quality studies. A global overview of the quality is given in Figure 1 and Figure 2.

Table 1 Quality assessment

Study ID Reporting External validity

Internal validity

Power Total score

Conclusion

Bakker 2006 10 3 13 0 26 high (100% of internal validity)

Blonk 2006 9 2 10 0 21 high (85% of internal validity)

Brouwers 2006 10 3 11 0 24 high (85% of internal validity)

Fleten 2006 10 3 11 0 24 high (85% of internal validity)

Nystuen 2006 8 2 10 0 20 high (77% of internal validity)

van der Klink 2003

10 3 13 0 26 high (100% of internal validity)

COCHRANE REVIEW

35

Figure 1 Methodological quality graph: review authors' judgments about each methodological quality item presented as percentages across all included studies. Figure 2 Methodological quality summary: review authors' judgments about each methodological quality item for each included study. The risk of bias for an outcome across studies was summarized using the GRADE system (Harbour & Miller, 2001). On average the quality of evidence was low for both outcomes related to return to work (RTW) and outcomes related to mental health. Details on the quality of evidence are presented in Table 2 for cognitive behavioural therapy (CBT), Table 3 for solution-focused behavioural therapy (SFBT), and Table 4 for minimal postal intervention. Allocation In all six studies allocation concealment was judged adequate, based on reported information in five studies, and unpublished information from the author in one study (Blonk et al., 2006a).

COCHRANE REVIEW

36

Blinding Blinding of patients and outcome assessors was reported in only two of the studies (Bakker et al., 2006; van der Klink et al., 2003). In none of the studies the treatment providers were blinded. Selective reporting To appraise the risk of bias due to selective reporting a funnel plot was calculated for one of the primary outcome values, full RTW. No evidence of selective reporting influencing the results of the review was found. Table 2 Quality of evidence CBT compared to usual care

Patient or population: patients with adjustment disorders Settings: primary care, psychological care, and occupational healthcare Intervention: cognitive behavioural therapy (CBT) Comparison: usual care

Outcomes Illustrative comparative risks* (95% CI)

N part icipants (stud ies)

Quality of evidence

(GRADE)

Comments

Assumed risk Correspon ding risk

usual care CBT

Partial RTW Scale from: 1 to 540. (follow-up: 10-18 months)

Mean partial RTW in control groups was 93.70 days

1

Mean Partial RTW intervention groups 16.92 lower (26.3 to 7.55 lower)

485 (3)

⊕⊕⊕ moderate 2

Full RTW Scale from: 1 to 540. (follow-up: 10-18 months)

The mean full rtw in control groups was 135.02 days

1

Mean Full RTW in intervention groups 13.74 lower (27.27 to 0.22 lower)

922 (4)


Stress at 2-4 months 4DSQ (distress)

3

Scale from: 0 to 32. (follow-up: 10-18 months)

Mean stress at 2-4 months in control groups was 13.37 points on 4DSQ distress scale

3

Mean Stress at 2-4 months in the intervention groups was 1.59 lower (2.92 to 0.27 lower)

4

723 (4)


SMD -0.18 (-0.33 to -0.03)

Stress at 10-18 months 4DSQ (distress)

3


Mean stress at 10-18 months in the control groups was 9.49 points on 4DSQ distress scale

3

Mean Stress at 10-18 months in intervention groups was 0.60 lower (1.9 lower to 0.69 higher)

5

685 (4)


SMD -0.07 (-0.22 to 0.08)

Depression at 2-4 months 4DSQ (depression)

3.


Mean depression at 2-4 months in control groups was 1.83 points on 4DSQ depression scale

3

Mean Depression at 2-4 months in the intervention groups was 0.84 lower (2.31 lower to 0.63 higher)

6

728 (4)


SMD -0.24 (-0.66 to 0.18)

COCHRANE REVIEW

37

Anxiety at 2-4 months 4DSQ (anxiety)

3


Mean anxiety at 2-4 months in control groups was 3.85 points on 4DSQ anxiety scale

3

Mean Anxiety at 2-4 months in the intervention groups was 0.79 lower (1.63 lower to 0.06 higher)

7

705 (4)


SMD -0.14 (-0.29 to 0.01)

1 Weighted mean of final values.

2 Lack of blinding: patients and caregivers are aware of the arm to which patients are allocated.

3 Three of the four studies used the 4DSQ.

4 Scores estimated using a standardised mean difference of -0.18 (-0.33 to -0.03).

5 Scores estimated using a standardised mean difference of -0.07 (-0.22 to 0.08).

6 Scores estimated using a standardised mean difference of -0.24 (-0.66 to 0.18).

7 Scores estimated using a standardised mean difference of -0.14 (-0.29 to -0.01).

Table 3 Quality of evidence SFBT compared to usual care

Patient or population: patients with adjustment disorders Settings: primary care, psychological care, and occupational healthcare Intervention: solution-focused behavioural therapy (SFBT) Comparison: usual care

Outcomes Illustrative comparative risks* (95% CI)

N Part icipants (studies)

Quality of evidence (GRADE) Assumed risk Corresponding risk

usual care SFBT

Full RTW Scale from: 1 to 480. (follow-up: 14-16 months)

Mean full rtw in the control groups was 212.02 days

Mean Full RTW in intervention groups was 5.43 higher (56.93 lower to 67.79 higher)

113 (1)

⊕⊕ low

1,2


2 Imprecision: 95% confidence interval includes no effect; upper or lower confidence limit crosses

minimal important difference (MID = 2 weeks) either for benefit or harm.

Table 4 Quality of evidence Postal intervention compared to usual care

Patient or population: patients with adjustment disorders Settings: primary care, psychological care, and occupational healthcare Intervention: Postal intervention Comparison: usual care

Outcomes Illustrative comparative risks* (95% CI) N Part icipants (studies)

Quality of evidence (GRADE)

Assumed risk Corresponding risk

usual care postal intervention

Full RTW Scale from: 14 to 365. (follow-up: 12 months)

Mean full rtw in the control groups was 117.20 days

Mean Full RTW in intervention groups was 5.43 higher (56.93 lower to 67.79 higher)

169 (1)

⊕⊕ low

1,2


2 Imprecision: 95% confidence interval includes no effect; upper / lower confidence limit crosses

minimal important difference (MID = 2 weeks) either for benefit or harm.

COCHRANE REVIEW

38

Effects of interventions The effects of the interventions on both RTW-outcomes and mental health outcomes were assessed in a meta-analysis. A complete overview of these effects is presented in the tables below. 1 CBT vs usual care 1.1 Partial RTW [days] Study ID

CBT Usual Care Mean SD Total Mean SD Total

Blonk 2006 93.94 107.05 65 116 120 34 Brouwers 2006 106 87 98 121 94 96 van der Klink 2003 36 21.31 109 53 44.16 83 1.2 Full RTW [days] Study ID


Bakker 2006 123.26 128.89 225 123.2 137.31 206 Blonk 2006 217.06 128.06 71 252 123 34 Brouwers 2006 153 122 98 157 121 96 van der Klink 2003 69 58.59 109 91 74.37 83 1.3 Stress at 2-4 months Study ID


Bakker 2006 14.26 9.37 174 15.24 8.84 142 Blonk 2006 15.42 8.43 71 16.6 8.2 34 Brouwers 2006 8.39 7.2 76 10.88 8.52 73 van der Klink 2003 10.6 7 85 12.5 9.3 68 1.4 Stress at 10-18 months Study ID


Bakker 2006 10.81 8.91 167 10.49 8.64 139 Blonk 2006 14.26 8.52 71 14.1 9.2 34 Brouwers 2006 5.66 6.03 83 7.69 7.91 74 van der Klink 2003 7.47 7.2 66 8.53 7.6 51 1.5 Depression at 2-4 months Study ID


Bakker 2006 2.54 3.53 174 2.59 3.5 142 Blonk 2006 10.3 9.11 71 14.4 10.3 34 Brouwers 2006 0.85 1.92 79 1.25 2.21 75 van der Klink 2003 0.98 2 85 1.58 2.7 68

COCHRANE REVIEW

39

1.6 Depression at 10-18 months Study ID


Bakker 2006 1.74 2.92 167 1.89 3.04 139 Blonk 2006 10.52 8.92 71 13.3 10.8 34 Brouwers 2006 0.52 1.48 84 0.94 2.53 78 van der Klink 2003 0.89 1.9 66 0.84 2.2 51 1.7 Anxiety at 2-4 months Study ID


Bakker 2006 4.19 5.32 174 4.74 5.61 142 Blonk 2006 8.36 6.78 71 8.9 6.9 34 Brouwers 2006 1.64 3.43 78 2.27 4.4 73 van der Klink 2003 2.03 2.9 65 3.07 4.9 68 1.8 Anxiety at 10-18 months Study ID


Bakker 2006 2.83 4.55 167 3.14 4.54 139 Blonk 2006 7.61 6.99 71 7.1 7.2 34 Brouwers 2006 0.82 2.06 83 1.66 3.81 76 van der Klink 2003 1.33 2.8 66 1.94 4 51 2 SFBT vs usual care 2.1 Full RTW [days] Study ID


Nystuen 2003 217.45 166.7 66 212.02 166.7 47 3 Postal intervention vs usual care 3.1 Full RTW [days] Study ID


Fleten 2006 80.7 124.68 79 117.2 124.68 90 1 CBT vs usual care Outcome or Subgroup

Studies Partici pants

Statistical Method Effect Estimate

1.1 Partial RTW [days]

3 485 Mean Difference (IV, Fixed, 95% CI [days])

-16.92 [-26.30, -7.55]

1.2 Full RTW [days] 4 922 Mean Difference (IV, Fixed, 95% CI [days])

-13.74 [-27.27, -0.22]

1.3 Stress at 2-4 months


-0.18 [-0.33, -0.03]

1.4 Stress at 10-18 months


-0.07 [-0.22, 0.08]

COCHRANE REVIEW

40

1.5 Depression at 2-4 months


-0.24 [-0.66, 0.18]

1.6 Depression at 10-18 months


-0.11 [-0.26, 0.04]

1.7 Anxiety at 2-4 months


-0.14 [-0.29, 0.01]

1.8 Anxiety at 10-18 months


-0.12 [-0.27, 0.03]

2 SFBT vs usual care Outcome or Subgroup



2.1 Full RTW [days] 1 Mean Difference (IV, Fixed, 95% CI [days])

Subtotals only

3 Postal intervention vs usual care Outcome or Subgroup



3.1 Full RTW [days] 1 Mean Difference (IV, Fixed,95% CI [days])

Subtotals only

1. Cognitive behavioural therapy versus usual care 1.1 Primary outcome: return to work Four of the included studies (Bakker et al., 2006; Blonk et al., 2006; Brouwers et al., 2006; van der Klink et al., 2003) reported on the effects of cognitive behavioural therapy (CBT) on return to work (RTW). All studies compared CBT with usual care. Data on the number of days to partial RTW were reported by three of these studies and were used to calculate the weighted mean difference (WMD) in the meta-analysis (Figure 3). Data on the number of days to full RTW were reported by all four studies (Figure 4) and presented in a funnel plot (Figure 5).

Figure 3 Forest plot of comparison: CBT vs usual care; 1.1 partial RTW [days].

Figure 4 Forest plot of comparison: CBT vs usual care; 1.2 full RTW [days].

COCHRANE REVIEW

41

Figure 5 Funnel plot of comparison: 1 CBT vs usual care, 1.2 Full RTW [days]. 1.2 Secundary outcome: mental health Data on mental health outcomes were reported by all four studies that reported data on RTW. Because the studies used different psychometric scales (4DSQ, DASS) on stress, depression and anxiety, standardised mean differences (SMD) were used instead of weighted mean differences (WMD). Regardless of the healthcare setting no significant reduction of stress could be demonstrated at 2-4 months (Figure 6) and 10-18 months (Figure 7). Similar findings were found for depression (Figure 8, Figure 9) and anxiety (Figure 10, Figure 11). Figure 6 Forest plot of comparison: CBT vs usual care; 1.3 stress at 2-4 months.

Figure 7 Forest plot of comparison: CBT vs usual care; 1.4 stress at 10-18 months.

COCHRANE REVIEW

42

Figure 8 Forest plot of comparison: CBT vs usual care, 1.5 depression at 2-4 months.

Figure 9 Forest plot of comparison: CBT vs usual care, 1.6 depression at 10-18 months Figure 10 Forest plot of comparison: CBT vs usual care, 1.7 anxiety at 2-4 months.

Figure 11 Forest plot of comparison: CBT vs usual care, 1.8 anxiety at 10-18 months. 2. Solution-focused behavioural therapy 1.1 Primary outcome: return to work One of the included studies (Nystuen et al., 2003) reported on the effects of solution-focused behavioural therapy (SFBT) on return to work (RTW). No significant effect of SFBT on the number of days to full RTW were found (Figure 12). 1.2 Secundary outcome: mental health The included studies did not report on the effect of SFBT on mental health.

COCHRANE REVIEW

43

Figure 12 Forest plot of comparison: SFBT vs usual care, 2.1 full RTW [days]. 3. Minimal postal intervention 1.1 Primary outcome: return to work The remaining included study (Fleten et al., 2006) reported on the effects of a minimal postal intervention aiming at work adjustments on return to work (RTW). No significant effect of this intervention on the number of days to full RTW were found (Figure 13). 1.2 Secundary outcome: mental health The included studies did not report on the effect of a minimal postal intervention on mental health.

Figure 13 Forest plot of comparison: minimal postal intervention vs usual care, 3.1 full RTW [days]. DISCUSSION Summary of main results This review found evidence that cognitive behavioural therapy (CBT) may facilitate return to work (RTW) of workers with adjustment disorders. On average, workers who are offered CBT will start two weeks earlier with partial and full RTW. A second finding of this review is that CBT improves the mental health of workers with adjustment disorders. This finding actually supports the hypothesis that early RTW may be associated with improved mental health. Based on a single study, the third finding is that there is no evidence that solution-focused behavioural therapy (SFBT) facilitates (RTW) of workers with adjustment disorders. The fourth and final finding of this review is that, based on a single study, there is no strong evidence that an intervention aimed exclusively at work adjustments facilitates RTW of workers with adjustment disorders. Overall completeness and applicability of evidence This review is limited by the fact that all studies were conducted in Norway and the Netherlands. This limitation makes it uncertain how far the findings can be generalised to countries with different welfare structures or attitudes to work. External validity of the studies was appraised using the Downs and Black checklist on methodological quality (Downs & Blacks, 1998). Four of the six studies were highly applicable. Quality of the evidence Only six studies on return to work (RTW) interventions for adjustment disorders were included in this review. However, the quality of these six studies was high, when scored on the internal validity subscales of the Downs and Black checklist on methodological quality (Downs & Blacks, 1998). A major point of concern is the blinding of participants and treatment providers in the studies. Only two studies reported blinding of the participants to the intervention they were receiving. In the four other studies blinding would have

COCHRANE REVIEW

44

been impossible due to the nature and aims of interventions being self-evident. Blinding of the treatment providers was impossible in all six studies. Consequently, when using the GRADE system, the quality of evidence score is lowered one category for all comparisons made in this review. Potential biases in the review process In the summary of main results it was concluded that the healthcare setting may be pivotal in the succes of cognitive behavioural therapy (CBT). This conclusion was based on the findings of only two studies (Blonk et al., 2006a; van der Klink et al., 2003). However, there also are other factors that distinguish these two studies from the rest of the studies. First, the participants in these two studies were predominantly male workers, compared to predominantly female workers in the remaining four studies. Second, the participants in these two studies were either employed by a large company (van der Klink et al., 2003), or were self employed and insured by a large insurance company (Blonk et al., 2006a). Both companies were committed to success of the trial, making it easier for participants to return to work early. In the remaining four studies, participant were working for a mix of companies which were not informed of the trials. Agreements and disagreements with other studies or reviews In 2005 a systematic review on the treatment of nervous breakdown was published (Terluin et al., 2005). One of the conclusions of this review was that an activating treatment, such as cognitive behavioural therapy (CBT), appeared to exert a more powerful effect on restoring social functioning than on symptom reduction. Although return to work (RTW) was not discussed in this review, these findings seem to be in line with the main findings of this review. A Cochrane review on the prevention of occupational stress in healthcare workers (Marine et al., 2006) concluded that limited evidence is available that person directed interventions, such as CBT, and work-directed interventions may reduce stress levels in health care workers. It is possible that CBT has a different effect on healthy workers in a preventive setting compared to workers with mental health problems on sick leave. In a recent Cochrane review on interventions to improve occupational health in depressed people (Nieuwenhuijsen et al., 2008) similar conclusions were drawn. CONCLUSIONS Implications for practice In this review limited evidence is presented that cognitive behavioural therapy (CBT) may facilitate partial and full return to work (RTW) of workers with adjustment disorders. Implications for research This review shows that randomised controlled trials (RCTs) on common mental health problems are feasible and can help to answer questions regarding the recovery of patients with adjustment disorders. It is unfortunate that only RCTs from Norway and the Netherlands were published on this subject. For the general applicability of these findings it is essential that these RCTs are also conducted in other countries. RCTs on the relation of CBT and work adjustments could also be of value to future practice.

COCHRANE REVIEW

45

SOURCES OF SUPPORT External sources of support

Dutch Cochrane Centre NETHERLANDS Internal sources of support

EMGO Institute, Department of Public and Occupational Health, VU University Medical Center NETHERLANDS

REFERENCES Included studies Bakker 2006 [ISRCTN: 43779641] Bakker IM, Terluin B, van Marwijk HW, Gundy CM, Smit JH, van Mechelen W, Stalman WA. Effectiveness of a Minimal Intervention for Stress-related mental disorders with Sick leave (MISS); study protocol of a cluster randomised controlled trial in general practice [ISRCTN43779641]. BMC Public Health 2006;6(124). Bakker IM, Terluin B, van Marwijk HW, van der Windt DA, Rijmen F, van Mechelen W, Stalman WA. A cluster-randomised trial evaluating an intervention for patients with stress-related mental disorders and sick leave in primary care. PLoS Clinical Trials 2007;2(6):e26. Blonk 2006 Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Return to work: a comparison of two cognitive behavioural interventions in cases of work-related psychological complaints among self-employed. Work & Stress 2006;20(2):129-144. Brouwers 2006 Brouwers EP, Tiemens BG, Terluin B, Verhaak PF. Effectiveness of an intervention to reduce sickness absence in patients with emotional distress or minor mental disorders: a randomized controlled effectiveness trial. General Hospital Psychiatry 2006;28(3):223-229. Brouwers EP, de Bruijne MC, Terluin B, Tiemens BG, Verhaak PF. Cost-effectiveness of an activating intervention by social workers for patients with minor mental disorders on sick leave: a randomized controlled trial. European Journal of Public Health 2006;17(2):214-220. Brouwers EPM, Tiemens BG, Terluin B, Verhaak PFM. Effectiveness of an intervention to reduce sickness absenteeism from work in patients with emotional distress or minor mental disorders: a randomised controlled effectiveness trial [De effectiviteit van een interventie door het maatschappelijk werk bij huisartspatiënten die overspannen zijn. Een gerandomiseerd vergelijkend onderzoek]. Huisarts & Wetenschap 2007;50(6):238-244. Fleten 2006 Fleten N, Johnsen R. Reducing sick leave by minimal postal intervention: a randomised, controlled intervention study. Occupational and Environmental Medicine 2006;63(10):676-682. Nystuen 2003 [ISRCTN: 39140363] Nystuen P, Hagen KB. Feasibility and effectiveness of offering a solution-focused follow-up to employees with psychological problems or muscle skeletal pain: a randomised controlled trial. BMC Public Health 2003;3:19. van der Klink 2003 van der Klink JJL, Blonk RWB, Schene AH, van Dijk FJH. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomised controlled design. Occupational and Environmental Medicine 2003;60(6):429-437.

COCHRANE REVIEW

46

Excluded studies Akagi H, Klimes I, Bass C. Cognitive behavioral therapy for chronic fatigue syndrome in a general hospital feasible and effective. General Hospital Psychiatry 2001;23(5):254-260. Bruning NS, Frew DR. Effects of exercise, relaxation, and management skills training on physiological stress indicators: a field experiment. Journal of Applied Psychology 1987;72(4):515-521. Bunce D, West MA. Stress management and innovation interventions at work. Human Relations 1996;49(2):209-232. Cuijpers P, Smit F, Voordouw I, Kramer J. Outcome of cognitive behaviour therapy for minor depression in routine practice. Psychology and Psychotherapy: Theory, Research and Practice 2005;78(Pt 2):179-188. Dettmers C, Stein H, Bock H, Simon U, Slowik M. Neurological outpatient rehabilitation is incomplete until the patient is successfully reintegrated into his/her job [Begleitung des Patienten während der beruflichen Wiedereingliederung komplettiert die neurologische Rehabilitation]. Neurologie und Rehabilitation 2003;9(5):217-225. Firth J, Shapiro DA. An evaluation of psychotherapy for job-related distress. Journal of Occupational Psychology 1986;59(2):111-119. Frank E, Rucci P, Katon W, Barrett J, Williams JW Jr, Oxman T, Sullivan M, Cornell J. Correlates of remission in primary care patients treated for minor depression. General Hospital Psychiatry 2002;24(1):12-19. Glicken MD. A counseling approach to employee burnout. Personnel Journal 1983;62(3):222-228. Gómez Sanabria A, Gala Léon FJ, Guillén Gestoso C, Lupiani Giménez M. Intervention and prevention in occupational stress [Intervención y prevención del estrés laboral]. Psiquis 2003;24(5):22-31. Huibers MJ, Beurskens AJ, Van Schayck CP, Bazelmans E, Metsemakers JF, Knottnerus JA, Bleijenberg G. Efficacy of cognitive-behavioural therapy by general practitioners for unexplained fatigue among employees: Randomised controlled trial. British Journal of Psychiatry 2004;184:240-260. Huibers MJ, Bleijenberg G, van Amelsvoort LG, Beurskens AJ, van Schayck CP, Bazelmans E, Knottnerus JA. Predictors of outcome in fatigued employees on sick leave: results from a randomised trial. Journal of Psychosomatic Research 2004;57(5):443-449. Hätinen M, Kinnunen U, Mauno S, Pekkonen M, Sörensen L, Alen M, Aro A. Job burnout and rehabilitation: short-term effects in different burnout patterns [Tyouupumus ja kuntoutus: Kuntoutuksen lyhytaikaiset vaikutukset erilailla oirehtivien uupumustyypeissa]. Psykologia 2002;37(4):302-314. Hätinen M, Kinnunen U, Pekkonen M, Aro A. Burnout patterns in rehabilitation: short-term changes in job conditions, personal resources, and health. Journal of Occupational Health Psychology 2004;9(3):220-237. Jackson SE. Participation in decision making as a strategy for reducing job-related strain. Journal of Applied Psychology 1983;68(1):3-19. Kawakami N, Haratani T, Iwata N, Imanaka Y, Murata K, Araki S. Effects of mailed advice on stress reduction among employees in Japan: a randomized controlled trial. Industrial Health 1999;37(2):237-242. Kushnir T, Milbauer V. Managing stress and burnout at work: a cognitive group intervention program for directors of day-care centers. Pediatrics 1994;94(6 Pt 2):1074-1077. Lange A, van de Ven JP, Schrieken B, Smit M. 'Interapy' burnout: prevention and therapy of burnout via the internet ['Interapy' Burn-out: Prävention und Behandlung von Burn-out über das Internet]. Verhaltenstherapie 2004;14(3):190-199. Lehmer M, Bentley A. Treating work stress: an alternative to workers' compensation. Journal of Occupational and Environmental Medicine 1997;39(1):63-67. Mino Y, Shigemi J, Tsuda T, Yasuda N, Babazono A, Bebbington P. Recovery from mental ill health in an occupational setting: a cohort study in Japan. Journal of Occupational Health 42;2(66-71).

COCHRANE REVIEW

47

Mynors-Wallis L, Davies I, Gray A, Barbour F, Gath D. A randomised controlled trial and cost analysis of problem-solving treatment for emotional disorders given by community nurses in primary care. British Journal of Psychiatry 1997;170:113-119. Natsume M, Noda T, Sato T, Inui T, Takagaki Y. Studies on 119 employees who visited the stress dock of Osaka Prefectural Mental Health Center. Japanese Journal of Psychosomatic Medicine 1996;36(2):169-174. Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. Occupational & Environmental Medicine 2003;60(Suppl I):i21-i25. Perski A, Grossi G. Treatment of patients on long-term sick leave because of stress-related problems. Results from an intervention study [Behandling av långtidssjukskrivna patienter med stressdiagnoser. Resultat från en interventionsstudie]. Läkartidningen 2004;101(14):1295-1298. Perski A. Rehabilitation of stress-related diseases goes on different phases and is often long-lasting [Rehabilitering av stressjukdomar sker i olika faser och blir ofta lång]. Läkartidningen 2004;101(14):1292-1294. Proudfoot J, Goldberg D, Mann A, Everitt B, Marks I, Gray JA. Computerized, interactive, multimedia cognitive-behavioural program for anxiety and depression in general practice. Psychological Medicine 2003;33(2):217-227. Salmela-Aro K, Näätänen P, Nurmi JE. The role of work-related personal projects during two burnout interventions: a longitudinal study. Work & Stress 2004;18(3):208-230. Selishchev GS, Petchot-Bacque JP, Volkov AK, Bolotina SI, Vein AM, Suter M, Sudakov KV, Alleaume B, Shmyriov VI. An open non-comparative study on the efficacy of an oral multivitamin combination containing calcium and magnesium on persons permanently exposed to occupational stress-predisposing factors. Journal of Clinical Research 1998;1:303-313. Terranova R, Gilotta SM, Luca S. Clinical evaluation of the efficacy of pivagabine in the treatment of mood and adjustment disorders. Arzneimittel Forschung 1997;47(11A):1325-1328. Toivanen H, Helin P, Hänninen O. Impact of regular relaxation training and psychosocial working factors on neck-shoulder tension and absenteeism in hospital cleaners. Journal of Occupational Medicine 1993;35(11):1123-1130. Studies awaiting classification de Vente W, Kamphuis JH, Emmelkamp PM, Blonk RW. Individual and group cognitive-behavioral treatment for work-related stress complaints and sickness absence: a randomized controlled trial. Journal of Occupational Health Psychology 2008;13(3):214-231. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W. Design of a randomized controlled trial on the effects of counseling of mental health problems by occupational physicians on return to work: the CO-OP-study. BMC Public Health 2007;7(147):183. Ongoing studies van Oostrom SH, Anema JR, Terluin B, de Vet HC, Knol DL, van Mechelen W. Cost-effectiveness of a workplace intervention for sick-listed employees with common mental disorders: design of a randomised controlled trial. BMC Public Health 2008;8:12. Additional references Alderson P, Green S, Higgins JPT, editors. Cochrane Reviewers' Handbook 4.2.2 [updated December 2003]. In: Cochrane Database of Systematic Reviews, 1, 2004. Chichester, UK: John Wiley & Sons, Ltd. APA. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV). Washington DC: American Psychiatric Association, 1994.

COCHRANE REVIEW

48

Arrindell WA, Ettema H. Dimensional structure, reliability and validity of the Dutch version of the Symptom Checklist (SCL-90); data based on a phobic and a “normal” population [in Dutch]. Nederlands Tijdschrift voor de Psychologie 1981;36:77–108. Arrindell WA, Ettema H. SCL90, Manual of a multidimensional indicator of psychopathology [in Dutch]. Lisse: Swets & Zeitlinger BV, 1986. Beck AT, Steer R. Beck Depression Inventory: Manual. San Antonio, Texas: Psychological Corporation, 1987. Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Return to work: A comparison of two cognitive behavioural interventions in cases of work-related psychological complaints among the self-employed. Work & Stress 2006;20(2):129{144. Brage S, Bentsen BG, Bjerkedal T, et al. ICPC as a standard classification in Norway. Fam Pract 1996;13:391–6. Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. Journal of Epidemiology and Community Health 1998;52(6):377{84. Derogatis LR. SCL90: Administration, scoring and procedures manual for the (revised) version. Baltimore: Johns Hopkins University school of Medicine, Clinical Psychometrics Research Unit, 1977. Derogatis LR, Cleary PA. Conformation of the dimensional structure of the SCL-90: a study in construct validity. J Clin Psychol 1977;33:981–9. Egger M, Davey Smith G, Schneider M, Minder CE. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315(7109):629{34. EuroQol Group. Euroqol EQ-5D - A new measurement of quality of life. Health Policy 1990;16. Gingerich WJ, Eisengart S: Solution-focused brief therapy: A review of the outcome research Family Process 2000, 39:477-498. Gründemann RMW, Smulders PGW, Winter CR de. DWHQ, the Dutch Work and Health Questionnaire; Manual [in Dutch]. Lisse: Swets & Zetlinger BV, 1993. Hameed U, Schwartz TL,Malhotra TK,West RL, Bertone F. Antidepressant treatment in the primary care office: outcomes for adjustment disorder versus major depression. Annals of Clinical Psychology 2005;17:77{81. Harbour R, Miller J. A new system for grading recommendations in evidence based guidelines. BMJ. 2001;323(7308):334-6. Lisv. Trendrapportage arbeidsongeschiktheid 2000. Amsterdam: Landelijk Instituut Sociale Verzekeringen (Lisv), 2000. Lovibond SH, Lovibond PF. Manual for the Depression Anxiety Stress Scales (DASS). University of New South Wales, 1993. Marine A, Ruotsalainen J, Serra C, Verbeek A. Preventing occupational stress in healthcare workers. In: Cochrane Database of Systematic Reviews, 4, 2006. Maslach C, Jackson SE: MBI: Maslach Burnout Inventory Manual (research edition). Dutch edition. Edited by: 'Utrecht Burnout Scale' by: Schaufeli WB, Van Dierendonck. NCvB; NCOD. Alert Report on Occupational Diseases 2004. Amsterdam, The Netherlands: Netherlands Center for Occupational Diseases, 2004. Nieuwenhuijsen K, Bültmann U, Neumeyer-Gromen A, Verhoeven AC, Verbeek JH, van der Feltz-Cornelis CM. Interventions to improve occupational health in depressed people. Cochrane Database Syst Rev. 2008 Apr 16;(2):CD006237. Review. NVAB. Van der Klink J, ed. Guideline for Mental Health Problems. Revision 1st edition 2000 [in Dutch] Utrecht: NVAB (Netherlands Society of Occupational Medicine), 2007. Pearlin LI, Schooler C. The structure of coping. J Health Soc Behav 1978;19:2–21. Robinson KA,Dickersin K. Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed. International Journal of Epidemiology 2002;31:150{3.

COCHRANE REVIEW

49

Schaufeli WB, van Dierendonck D. (1994). Burnout, een begrip gemeten. De Nederlandse versie van de Maslach Burnout Inventory (MBI-NL) [Burnout, the measurement of a concept. The Dutch version of the Maslach Burnout Inventory (MBI-NL)]. Gedrag en Organisatie: Tijdschrift voor Psychologie en Gezondheid 1994;22:153-172. Schreurs PJG, Tellegen B, Willige G van de. Health, stress and coping: the development of the Utrecht Coping List [in Dutch]. Gedrag, Tijdschrift voor Psychologie 1984;12:101–17. Schreurs PJG, Willige G van de, Brosschot JF, et al. De Utrecht Coping List: UCL; coping with problems and events; revised manual [in Dutch]. Lisse: Swets & Zetlinger BV, 1993. Schroër CAP. Ziekteverzuim wegens overspanning [absenteeism due to `overstrain']. Maastricht: UM, 1993. Smith AP, Johal S,Wadsworth E, Smith GD, Harvey I, Peters T. The scale and costs of occupational stress. Occupational Health Review 1998;May/June:19{22. Terluin B, van Dijk DM, van der Klink JJL, Hulshof CTJ, Romeijnders ACM. De behandeling van overspanning. Een systematisch literatuuroverzicht. Huisarts Wet 2005;48(1):7{12. Terluin B, Van Marwijk HWJ, Ader HJ, de Vet HC, Penninx BW, et al. (2006) The Four-Dimensional Symptom Questionnaire (4DSQ): A validation study of a multidimensional self-report questionnaire to assess distress, depression, anxiety and somatization. BMC Psychiatry 6: 34. van der Klink JJ, Blonk RW, Schene AH, van Dijk FJ. The benefits of interventions for work-related stress. American Journal of Public Health 2001;91(2):270{6. Verbeek J, Salmi J, Pasternack I, Jauhianen, Laamanen I, Schaafsma F, Hulshof C, van Dijk F. A search strategy for occupational health intervention studies. Occupational and Environmental Medicine 2005;62(10):682{7. Walker AM, Martin-Moreno JM, Artalejo FR. Odd man out: a graphical approach to meta-analysis. American Journal of Public Health 1988;78(8):961{6. Ware JE, Sherbourne CD. The MOS36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Medical Care 1992;30(6):473{483. World Health Organization (WHO). Composite international diagnostic interview. Geneva7 World Health Organization; 1990. World Health Organisation (WHO). The ICD-10 Classification of Mental and Behavioural Disorders. Geneva: WHO, 1992. Williamson AM. Managing stress in the workplace: Part II – The scientist basis (knowledge base) for the guide. International Journal of Industrial Ergonomics 1994;14:171{196. Zigmond AS, Snaith RP. TheHospital Anxiety and Depression Scale. Acta Psychiatrica Scand 1983;67:361;70.

SUPPLEMENTS Supplement 1 MEDLINE (PubMed) search terms (adjustment disorders[MeSH Terms] OR burnout, professional[MeSH Terms] OR work stress[Text Word] OR occupational stress[Text Word] OR job stress[Text Word] OR neurasthenia[MeSH Terms] OR minor depression[Text Word] OR ((mixed[All Fields] AND (anxiety[MeSH Terms] OR anxiety[Text Word])) AND ((depressive disorder[MeSH Terms] OR depression[MeSH Terms]) OR depression[Text Word])) OR emotional disorder[Text Word]) AND ("return to work\ OR (occupational AND therap*) OR (occupational AND intervention*) OR "supported employment\ OR employment OR "vocational rehabilitation\ OR "work capacity evaluation\ OR "vocational guidance\ OR absenteeism OR "occupational health serviceS\ OR "occupational health\ OR unemployed OR employed OR unemployment OR "sick leave\ OR "sick absence\ OR "sickness absence\ OR retirement OR "disability pension\ OR occupation* OR job OR vocational) AND ((randomized controlled trial[pt] OR controlled clinical trial[pt] OR randomized controlled trials[mh] OR random allocation[mh] OR double-blind method[mh] OR single-blind method[mh] OR clinical trial[pt] OR clinical trials[mh] OR ("clinical trial\[tw]) OR ((singl*[tw] OR doubl*[tw] OR trebl*[tw] OR tripl*[tw])AND (mask*[tw] OR blind*[tw])) OR ("latin square\[tw]) OR placebos[mh] OR placebo*[tw] OR random*[tw] OR research design[mh:noexp] OR comparative study[mh] OR evaluation studies[mh] OR follow-up studies[mh] OR prospective studies[mh] OR cross-over

COCHRANE REVIEW

50

studies[mh] OR control*[tw] OR prospective*[tw] OR volunteer*[tw] OR time factors[mh] OR treatment outcome[mh] OR time series[tw] OR ITS[tw]) NOT (animal[mh] NOT human[mh])) EMBASE search terms (('adjustment disorder'/exp) OR ('burnout'/exp) OR ('work *1 stress':ab,ti,tn,tt,mn,df, de) OR ('occupational stress':ab,ti,tn,tt,mn,df,de) OR ('job stress':ab,ti,tn,tt,mn,df,de) OR ('neurasthenia'/exp) OR ('minor depression':ab,ti,tn,tt,mn,df,de) OR(('depression'/ exp OR 'depression':ab,ti,tn,tt,mn,df,de) AND (('mixed':ab,ti,tn,tt,mn,df,de) AND ('anxiety'/exp OR 'anxiety':ab,ti,tn,tt,mn,df,de))) OR ('emotional disorder':ab,ti,tn,tt, mn,df,de)) AND (('occupational therapy':ab,ti,tn,tt,mn,df,de) OR ('occupational disease':ab,ti,tn,tt,mn,df,de) OR ('occupational medicine':ab,ti,tn,tt,mn,df,de) OR ('employment':ab,ti,tn,tt,mn,df,de) OR ('vocational rehabilitation':ab,ti,tn,tt,mn,df, de) OR ('work capacity':ab,ti,tn,tt,mn,df,de) OR ('vocational guidance':ab,ti,tn,tt mn,df,de) OR ('absenteeism':ab,ti,tn,tt,mn,df,de) OR ('occupational health':ab,ti,tn, tt,mn,df,de) OR ('unemployment':ab,ti,tn,tt,mn,df,de) OR ('retirement':ab,ti,tn,tt,mn, df,de) OR ('occupation':ab,ti,tn,tt,mn,df,de) OR ('vocation':ab,ti,tn,tt, mn,df,de) OR ('disability evaluation':ab,ti,tn,tt,mn,df,de) OR ('return to work': ab,ti,tn,tt,mn,df, de) OR ('occupational intervention':ab,ti,tn,tt,mn,df,de) OR ('occupational interventions': ab,ti, tn,tt,mn,df,de) OR ('supported employment':ab,ti,tn,tt,mn,df,de) OR ('unemployed': ab,ti,tn,tt, mn,df,de) OR 'employed':ab,ti,tn,tt,mn,df,de) OR ('sick leave': ab,ti,tn,tt,mn,df,de) OR ('sick absence':ab,ti,tn,tt,mn,df,de) OR ('sickness absence': ab,ti,tn,tt,mn,df,de) OR ('disability pension':ab,ti,tn,tt,mn,df,de) OR ('job': ab,ti,tn,tt,mn,df,de) OR ('vocational':ab,ti,tn,tt,mn,df,de) OR ('work'/exp) OR ('work disability':ab,ti,tn,tt,mn,df,de) OR ('occupation disability':ab,ti,tn,tt, mn,df,de) OR ('occupational disability':ab,ti,tn,tt,mn ,df,de) OR ('vocational disability':ab,ti,tn,tt, mn,df,de) OR ('job':ab,ti,tn,tt,mn,df,de)) AND ((('controlled study':de) OR ('clinical trial':de) OR ('major clinical study':de) OR ('randomized controlled trial':de) OR ('double blind procedure': de) OR ('clinical article':de) OR (random*: ab,ti,tn,tt,mn, df,de) OR (compar*:ab,ti,tn,tt,mn,df, de) OR (control*:ab,ti,tn,tt,mn,df,de) OR ('follow up':ab,ti,tn,tt,mn,df,de) OR (((singl*: ab,ti,tn, tt,mn,df,de OR doubl*:ab,ti,tn, tt,mn,df,de OR tripl*:ab,ti,tn,tt,mn,df,de OR trebl*:ab,ti,tn,tt,mn, df,de) AND (blind*: ab,ti,tn,tt,mn,df,de OR mask*:ab,ti,tn,tt,mn,df,de OR dummy:ab,ti,tn,tt,mn, df,de))) OR (placebo*:ab,ti,tn,tt,mn,df,de) OR ((clinic*:ab,ti,tn,tt,mn,df,de AND (trial*:ab,ti,tn, tt,mn,df,de OR study:ab,ti,tn,tt,mn,df,de OR studies*:ab,ti,tn,tt,mn,df,de)))) NOT ((nonhuman: de) NOT ((human:de) AND (nonhuman:de)))) AND [1988-2005]/py PsycINFO search terms #57 (#47 not #51) and (#1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 or #19 or #20 or #21 or #22 or #23 or #24 or #25 or #26 or #27 or #28 or #29 or #30 or #31 or #32 or #33 or #34) and (#53 or #54 or #55) ; #56 #53 or #54 or #55 ; #55 explode "Occupational-Stress\ in MJ,MN ; #54 explode "Occupational-Neurosis\ in MJ,MN ; #53 explode "Adjustment-Disorders\ in MJ,MN ; #52 #47 not #51; #51 #48 not #50 ; #50 #48 and #49 ; #49 (human or inpatient or outpatient) in po ; #48 animal in po ; #47 #36 or #37 or #38 or #39 or #40 or #41 or #42 or #43 or #44 or #45 or #46 ; #46 explode experimental design #45 explode mental health program evaluation ; #44 explode treatment effectiveness evaluation ; #43 explode placebo ; #42 ((clin* or control* or compare* or evaluat* or prospective*) near25 (trial* or studi* or study)) in TI,AB,TC,SU ; #41 allocat* in TI,AB,TC,SU ; #40 assign* in TI,AB,TC,SU ; #39 crossover in TI,AB,TC,SU ; #38 placebo* in TI,AB,TC,SU ; #37 ((singl* or doubl* or trebl* or tripl*) near25 (blind* or dummy or mask*)) in TI,AB,TC,SU ; #36 random* in TI,AB,TC,SU ; #35 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 or #19 or #20 or #21 or #22 or #23 or #24 or #25 or #26 or #27 or #28 or #29 or #30 or #31 or #32 or #33 or #34 ; #34 vocational in TI,AB,TC,SU ; #33 job in TI,AB,TC,SU ; #32 occupation* in TI,AB,TC,SU ; #31 disability pension in TI,AB,TC,SU ; #30 retirement in TI,AB,TC,SU ; #29 sick* absence in TI,AB,TC,SU ; #28 sick leave in TI,AB,TC,SU ; #27 unemployment in TI,AB,TC,SU ; #26 employed in TI,AB,TC,SU ; #25 unemployed in TI,AB,TC,SU ; #24 occupational health in TI,AB,TC,SU ; #23 occupational health services in TI,AB,TC,SU ; #22 absenteeism in TI,AB,TC,SU ; #21 vocational guidance in TI,AB,TC,SU ; #20 work capacity evaluation in TI,AB,TC,SU ; #19 vocational rehabilitation in TI,AB,TC,SU ; #18 employment in TI,AB,TC,SU ; #17 supported employment in TI,AB,TC,SU ; #16 occupational intervention* in TI,AB,TC,SU ; #15 occupational therap* in TI,AB,TC,SU ; #14 return to work in TI,AB,TC,SU ; #13 explode Work-Related-Illnesses ; #12 explode Reemployment- [searched Reemployment] ; #11 Occupational-Therapy ; #10 Occupational-Stress ; #9 explode Occupational-Status ; #8 explode Employee-Absenteeism ; #7 explode Disability-Management; #6 explode Vocational-Rehabilitation ; #5 explode Occupational-Guidance ; #4 explode Job-Satisfaction ; #3 explode Employee-Leave-Benefits ; #2 explode Employability- [searched Employability] ; #1 explode Disability-Evaluation Supplement 2 Characteristics of included studies Bakker 2006 Methods Trial design: randomised controlled trial Randomisation procedure: cluster randomisation on treatment provider level Allocation concealment: randomisation was conducted blindly by the research team using randomised lists Blinding: patients; interviewers; outcome assessors Follow-up: 12 months Inclusion period: September 2003 to October 2004

COCHRANE REVIEW

51

Participants Country: The Netherlands Healthcare setting: primary care in The Netherlands Work setting: employees in The Netherlands Number: Trial intervention: n=227; Comparison intervention: n=206 Age, mean (sd): Trial intervention: 42.0 (8.8) years; Comparison intervention: 39.5 (9.6) years Sex: Trial intervention: 67% female; Comparison intervention: 65% female Recruitment: employees who visited consulting hours of the participating general practitioners were approached by mail by the research team Inclusion: moderately elevated distress level (measured with 3 questions of the 4DSQ distress scale), having paid work and being (partially) on sick leave for no longer than three months Exclusion: severe psychiatric disorders (mania or psychosis), terminal illness or inadequate command of the Dutch language Interventions Trial intervention: - Treatment type: minimal intervention for stress-related mental disorders with sick leave (MISS) for general practice, using the principle of time contingency and parts of more specialised psychological treatments like cognitive behavioural therapy (CBT) and problem solving treatment (PST) - Treatment providers: 24 primary care physicians received a training of two 3.5 hour sessions and two 2 hour follow-up sessions by a primary care physician and an occupational physician over a 6 to 10 week period - Treatment frequency and duration: no more than 3 consultations of 10-20 minutes Comparison intervention: - Treatment type: usual care based on routine care by primary care physicians - Treatment providers: 22 primary care physicians who had received no information or advice about the content of the intervention Outcomes Work-status outcomes: - time to full RTW lasting for period of at least 4 weeks without partial/full relapse into sick leave - self-reported days of sick leave at baseline and after 2, 6, and 12 months follow-up Other outcomes: - Timing of assessments: mailed questionnaires at baseline and after 2, 6, 12 months of follow-up - Four-Dimensional Symptom Questionnaire (4DSQ): baseline measurement and follow-up measurements at baseline and after 2, 6, and 12 months - Compliance patients: dropouts at baseline, 2, 6 and 12 months - Compliance treatment providers: primary care physicians in both groups were asked to fill in structured questionnaire two months after baseline assessment on care provided and any diagnoses or working hypotheses in past 3 months according to their electronic medical record Notes Source of funding: The Netherlands Organisation for Health Research and Development (ZonMw), (grant 4200.0003) Ethics: approved by the medical ethics committee of the VU University Medical Center Risk of bias table Item Judgement Description Adequate sequence generation? Yes Randomisation was conducted blindly by

research team using randomised lists Allocation concealment? Yes A - Adequate Blinding? Yes Patients were blinded Incomplete outcome data addressed? Yes Free of selective reporting? Yes Quality score higher than 75%? Yes Applicability score higher than 75%? Yes 100%

Blonk 2006 Methods Trial design: randomised controlled trial Randomisation procedure: randomisation on patient level Allocation concealment: randomisation was conducted blindly by the research team (based on unpublished information from the author) Blinding: blinding of patients was not reported, but would be impossible due to the large differences between the interventions Follow-up: 360 days Inclusion period: January 2001 to September 2002 (based on unpublished information author) Participants Country: The Netherlands Healthcare setting: private insurance company Work setting: self employed individuals insured for work disability at a private insurance company Number: Trial intervention 1: n=40; Trial intervention 2: n=40; Comparison intervention: n=42

COCHRANE REVIEW

52

Age, mean (sd): 42 (7.9) years Sex: 19% female Recruitment: self-employed individuals who were unable to work owing to psychological complaints and had called upon their insurance company for disability benefits were approached by the research team Inclusion: sick leave caused by adjustment disorders (e.g. burnout and job stress) based on a structured diagnostic telephone interview using a shortened version of the Composite International Diagnostic Interview (CIDI) conducted by experienced psychologists Exclusion: serious psychiatric disorders (e.g. major depression, addictive disorders, posttraumatic disorders, and other anxiety disorders) based on the structured diagnostic telephone interview, or individuals who did not want to postpone their current treatment during the research period Interventions Trial intervention 1: - Treatment type: individual intervention based on cognitive behavioural therapy (CBT) combined with a workplace intervention focusing on stressor reduction at work using a graded activity approach - Treatment providers: 6 labour experts trained in brief CBT-based stress management with follow-up meetings every 3 months during the course of the study - Treatment frequency and duration: five to six sessions of approximately an hour, twice a week, which were held at home or at the workplace of the self-employed Trial intervention 2: - Treatment type: individual cognitive behavioural therapy (CBT) focused on RTW and based on a highly structured protocol for the treatment of burnout or other adjustment disorders - Treatment providers: psychologists - Treatment frequency and duration: 11 two-weekly sessions of approximately 45 minutes per session, where the first six sessions focused on cognitive restructuring and on registration of symptoms and situations, and the following five sessions focused predominantly on a further expansion of cognitive restructuring Comparison intervention: - Treatment type: no-treatment intervention consisting of two brief medical checks of the legitimacy of the work-disability benefit - Treatment providers: general practitioner assigned by the private insurance company - Treatment frequency and duration: a first visit shortly after the initial sick leave and a second visit approximately 4 months later Outcomes Work-status outcomes: - time to partial and full return to work Other outcomes: - Timing of assessments: baseline questionnaires were sent to patients before the intervention, and follow-up questionnaires at 4 and 10 months of follow-up - Depression Anxiety Stress Scales (DASS): baseline measurement and follow-up measurements at 4 and 10 months - Maslach Burnout Inventory (MBI-NL): baseline measurement and follow-up measurements at 4 and 10 months - Compliance patients: dropouts at baseline, 4 and 10 months Notes Source of funding: not reported Ethics: approved by ethical committee of the Netherlands Organisation for Applied Scientific Research (TNO) Risk of bias table Item Judgement Description Adequate sequence generation? Yes Randomisation was conducted blindly by

research team using randomised lists (based on unpublished information from the author)

Allocation concealment? Yes A – Adequate (based on unpublished information from the author)

Blinding? Yes Patients were not blinded Incomplete outcome data addressed? Yes Free of selective reporting? Yes Quality score higher than 75%? Yes 85% Applicability score higher than 75%? No 67% Brouwers 2006 Methods Trial design: randomised controlled trial Randomisation procedure: block randomisation on patient level (block size 4) Allocation concealment: randomisation was conducted by an administrative assistant who was not in contact with the patients, with the aid of a dice (evens being intervention group) Blinding: interviewers (partially blinded); blinding of patients was not reported, but would be impossible due to the large differences between the interventions Follow-up: 18 months Inclusion period: August 2001 to July 2003

COCHRANE REVIEW

53

Participants Country: The Netherlands Healthcare setting: primary care in the city of Almere Work setting: employees living in the city of Almere Number: Trial intervention: n=96; Comparison intervention: n=98 Age, mean (sd): Trial intervention: 39.4 (9.1) years; Comparison intervention: 40.1 (9.3) years Sex: Trial intervention: 58% female; Comparison intervention: 60% female Recruitment: patients who visited their GP were approached by research team Inclusion: emotional distress or minor mental disorders according to GP and self-report; paid employment; sick leave or planning to be on sick leave directly after visit to GP for emotional or mental problems existing less than 3 months; aged 18-60 years; Dutch-speaking Exclusion: patients with moderately severe or severe mood disorders (major depressive disorder and bipolar disorder), agoraphobia, panic disorder and social phobia based on a Composite International Diagnostic Interview (CIDI); patients already receiving psychotherapy Interventions Trial intervention: - Treatment type: individual CBP aimed at return to work, using a graded activity approach and based on a three stage model resembling stress inoculation training - Treatment providers: 11 social workers who received a 3-day training conducted by the researchers, including two follow-up sessions at different times during the study period wherein adherence to the protocol was checked and knowledge was refreshed. - Treatment frequency and duration: five individual 50 minute sessions over 10 weeks Comparison intervention: - Treatment type: usual care based on routine care by general practitioners, which could include medication or counseling, or even referral - Treatment providers: 70 general practitioners Outcomes Work-status outcomes: - time to full return to work - partial and full return to work rate at baseline and 3, 6 and 18 months follow-up Other outcomes: - Timing of assessments: baseline questionnaires were handed to patients at their baseline interview with GP; follow-up questionnaires were sent to patients at 3, 6, 18 months follow-up - Hospital Anxiety and Depression Scale (HADS): baseline measurement and follow-up measurements at 3, 6 and 18 months - Four-Dimensional Symptom Questionnaire (4DSQ): baseline measurement and follow-up measurements at 3, 6 and 18 months - Short Form Health Survey (SF-36): baseline measurement and follow-up measurements at 3, 6 and 18 months; eight individual sub scales, as well as the mental component summary scale score and the physical component summary scale score were computed and used in analyses - Patient satisfaction based on a questionnaire with eight statements developed for this study: measurement at 3 months - Compliance patients: dropouts at baseline, 3, 6 and 18 months Notes Source of funding: The Netherlands Organisation for Health Research and Development (ZonMw), (grant 2200.0100) Ethics: approved by the ethical committee of The Netherlands Institute of Mental Health and Addiction Risk of bias table Item Judgement Description Adequate sequence generation? Yes Randomisation was conducted blindly by

administrative assistant who was not in contact with patients, with aid of dice (evens intervention group)

Allocation concealment? Yes A – Adequate Blinding? No Patients were not blinded Incomplete outcome data addressed? Yes Free of selective reporting? Yes Quality score higher than 75%? Yes 85% Applicability score higher than 75%? Yes 100% Fleten 2006 Methods Trial design: randomised controlled trial Randomisation procedure: cluster randomisation on treatment provider level Allocation concealment: sick-listed persons were assigned consecutive numbers at enrolment and then randomised into the intervention or control group according to a pre-drawn randomisation list

COCHRANE REVIEW

54

Blinding: local National Insurance Offices (partially blinded); blinding of patients was not reported, but would be impossible due to the large differences between the interventions Follow-up: 12 months Inclusion period: October 1997 to November 1997, and March 1998 to April 1998 Participants Country: Norway Healthcare setting: local National Insurance Offices in Tromsø and Harstad Work setting: employees working in the Tromsø and Harstad region Number randomised: Trial intervention: n=499; Comparison intervention: n=501 Number included: Trial intervention: n=495; Comparison intervention: n=495 Age, mean: Trial intervention: 40.9 years; Comparison intervention: 39.9 years Sex: Trial intervention: 61% female; Comparison intervention: 60% female Recruitment: persons on sick leave for more than 14 days were selected on the basis of the diagnosis made by their general practitioner and were sent information about the project Inclusion: sick leave caused by musculoskeletal or mental disorders based on ICPC criteria Exclusion: not reported Interventions Trial intervention: - Treatment type: usual care in combination with a minimal intervention package containing general written information on possible work related measures if sick-listed, and a questionnaire related to the actual sick leave - Treatment providers: general practitioners and National Insurance Offices - Treatment frequency and duration: a minimal intervention package posted 14 days after the start of the current sick leave Comparison intervention: - Treatment type: usual care - Treatment providers: general practitioners and National Insurance Offices Outcomes Work-status outcomes: - time to full return to work - full return to work rate at 12 weeks follow-up Other outcomes: - percentage of patients with a benefit from the National Insurance Service one year after the start of the actual sick leave - Compliance patients: number of patients in the Trial intervention that returned the request for contact with the National Insurance Office and filled in the questionnaire Notes Source of funding: the Royal Ministry of Health and Social Affairs (project no. 13345) Ethics: approved by the Regional Medical Ethics Committee Risk of bias table Item Judgement Description Adequate sequence generation? Yes Randomisation was conducted blindly by the

research team using a pre-drawn randomisation list

Allocation concealment? Yes A – Adequate Blinding? No Patients were not blinded Incomplete outcome data addressed? Yes Free of selective reporting? Yes Quality score higher than 75%? Yes 85% Applicability score higher than 75%? Yes 100% Nystuen 2003 Methods Trial design: randomised controlled trial Randomisation procedure: randomisation on patient level Allocation concealment: randomisation was conducted blindly by the project administrator using a computer-generated randomisation list Blinding: blinding of patients was not reported, but would be impossible due to the large differences between the interventions Follow-up: 14-16 months Inclusion period: January 2001 to December 2001 Participants Country: Norway Healthcare setting: social security offices in Oslo Work setting: employees working in the Oslo region Number randomised: Trial intervention: n=122; Comparison intervention: n=106 Number included: Trial intervention: n=113; Comparison intervention: n=100

COCHRANE REVIEW

55

Age, mean (sd): Trial intervention: 40.7 (10.8) years; Comparison intervention: 40.1 (11.0) years Sex: Trial intervention: 65% female; Comparison intervention: 58% female Recruitment: employees sick-listed at their local social security office for more than seven weeks were selected on the basis of the diagnosis made by their general practitioner and were sent information about the project Inclusion: sick leave caused by psychological problems, general exhaustion and burn-out, or musculoskeletal pain based on ICPC criteria Exclusion: other disorder based on ICPC criteria; self employed; pregnancy; graded sick leave of less than 50%; those awaiting for elective orthopedic surgery; those becoming 66 or more in the present year; foreign born persons in need of interpreter to communicate Interventions Trial intervention: - Treatment type: usual care in combination with individual and/or group based solution-focused therapy aimed at the work situation, with a main focus was on coping strategies, support between the participants and solutions and goals for the future - Treatment providers: 3 psychologists, trained and experienced in solution-focused therapy in individual consultations and group settings - Treatment frequency and duration: "The Road Ahead Course" which comprised of 8 group sessions of 3 to 4 hours, and/or individual counseling sessions Comparison intervention: - Treatment type: usual care consisting of written information from the social security office and where patients were free to visit their regular general practitioner Outcomes Work-status outcomes: - days absent from work in the 14 to 16 months following inclusion Other outcomes: - Compliance patients: uptake rates for the different information elements and the intervention Notes Source of funding: the Royal Ministry of Health and Social Affairs Ethics: approved by the Regional Medical Ethics Committee Risk of bias table Item Judgement Description Adequate sequence generation? Yes Randomisation was conducted blindly by the

project administrator using a computer-generated randomisation list

Allocation concealment? Yes A – Adequate Blinding? No Patients were not blinded Incomplete outcome data addressed? No Free of selective reporting? Yes Quality score higher than 75%? Yes 77% Applicability score higher than 75%? No 67% van der Klink 2003 Methods Trial design: randomised controlled trial Randomisation procedure: cluster randomisation on treatment provider level Allocation concealment: randomisation was conducted blindly by an independent research assistant Blinding: patients, outcome assessors Follow-up: 12 months Inclusion period: May 1995 to July 1996 Participants Country: The Netherlands Healthcare setting: in-company occupational health service Work setting: employees of the Dutch postal service (Royal KPN) Number: Trial intervention: n=109; Comparison intervention: n=83 Age, mean (sd): Trial intervention: 39 (8.0) years; Comparison intervention: 42 (8.8) years Sex: Trial intervention: 34% female; Comparison intervention: 41% female Recruitment: employees who were two weeks on sick leave were referred to their occupational physician (OP) and were asked by their OP to participate in the study Inclusion: first sick leave caused by an adjustment disorder based on DSM IV criteria Exclusion: other disorder based on DSM IV criteria; treatment for adjustment disorder previous year; physical comorbidity with effect absenteeism; pregnancy/child birth previous six months Interventions Trial intervention: - Treatment type: individual CBT aimed at return to work, using a graded activity approach and based on a three stage model resembling stress inoculation training

COCHRANE REVIEW

56

- Treatment providers: 17 occupational physicians, trained during a three day training course by experienced trainers with backgrounds in psychology, occupational medicine, and general practice, and frequently supervised by the trainers during the study - Treatment frequency and duration: four or five consultations in the first six weeks of sick leave with a total length over these sessions of at least 90 minutes Comparison intervention: - Treatment type: usual care based on empathic counseling, instruction about stress, lifestyle advice, and discussion of work problems - Treatment providers: 16 occupational physicians without training in cognitive behavioural therapy, but with a three hour session on the use of the inclusion and exclusion criteria and the recording of their guidance activities - Treatment frequency and duration: there was neither a professional nor a company guideline available for the care of patients with adjustment disorders Outcomes Work-status outcomes: - time to partial and full return to work - time to full return to work corrected for partial return to work - partial and full return to work rate at 3 and 12 months follow-up - incidence of recurrent sick leave in the year following full return to work - time to first recurrent sick leave in the year following full return to work Other outcomes: - Timing of assessments: baseline questionnaires handed to patients at end of their first visit with OP, and follow-up questionnaires were sent to patients at 3 and 12 months of follow-up - Dutch Work and Health Questionnaire (DWHQ): baseline measurement; only the eight scales on work (32 items) were combined to estimate a total score of quality of work life - Utrecht Coping List (UCL): baseline measurement - Four-Dimensional Symptom Questionnaire (4DSQ): baseline measurement and follow-up measurements at 3 and 12 months - Symptom Checklist-90 (SCL-90): baseline measurement and follow-up measurements at 3 and 12 months - Mastery Scale: baseline measurement and follow-up measurements at 3 and 12 months - Compliance patients: dropouts at baseline, 3 and 12 months - Compliance treatment providers: contact duration, use of tools CBT Notes Source of funding: the Occupational Health Service of Royal KPN; The Netherlands Organisation of Scientific Research (NWO); TNO Work and Employment; the Foundation for Quality in Occupational Health (SKB) Ethics: not reported Risk of bias table Item Judgement Description Adequate sequence generation? Yes Randomisation was conducted blindly by

project administrator using a computer-generated randomisation list

Allocation concealment? Yes A – Adequate Blinding? Yes Patients were blinded Incomplete outcome data addressed? Yes Free of selective reporting? Yes Quality score higher than 75%? Yes 100% Applicability score higher than 75%? No 100% Characteristics of studies awaiting classification de Vente 2008 Methods Trial design: randomised controlled trial Randomisation procedure: randomisation on patient level Allocation concealment: randomisation was conducted blindly by an independent person using a computer-generated list of random numbers in blocks of 24. Blinding: blinding of patients was not reported, but would be impossible due to the large differences between the interventions Follow-up: 10 months Inclusion period: not reported Participants Country: The Netherlands Healthcare setting: primary care in The Netherlands Work setting: employees in The Netherlands Number: Trial intervention 1: n=28; Trial intervention 2: n=28; Comparison intervention: n=26

COCHRANE REVIEW

57

Age, mean (sd): Trial intervention 1: 41.6 (9.4) years; Trial intervention 2: 41.5 (10.3) years; Comparison intervention: 40.9 (9.6) years Sex: Trial intervention 1: 39% female; Trial intervention 2: 43% female; Comparison intervention: 35% female Recruitment: employees who were between two weeks and six months on sick leave were recruited through two occupational health services (n 62), general practitioners (n 7), and by self-referral in reaction to advertisements (n 13) Inclusion: the presence of symptoms of neurasthenia, a primary role of work-related stressors in the development of complaints, and the presence of impaired daily functioning as indicated by (partial) sickness absence Exclusion: a primary diagnosis of major depression, social phobia, panic disorder, somatoform disorder other than undifferentiated, posttraumatic stress disorder, obsessive-compulsive disorder, hypomania, or psychotic disorders; severe depressive complaints; a medical condition that might explain fatigue (e.g., diabetes); excessive alcohol or drug use Interventions Trial intervention 1: - Treatment type: individual CBT-based stress management training (SMT) including (a) psycho education, self-assessment of stressors and complaints, lifestyle, and relaxation techniques; (b) cognitive restructuring; (c) time management and goal setting; (d) assertiveness skills; and (e) evaluation and relapse prevention - Treatment providers: 12 experienced therapists with a master s degree in clinical psychology delivered the SMT; therapists were trained in delivering the treatment according to the protocol in 4 training sessions of 1 hour and received at least four supervision sessions per treatment by one of two experienced senior cognitive-behavioural therapists - Treatment frequency and duration: 12 sessions of approximately 1 hour Trial intervention 2: - Treatment type: group CBT-based stress management training (SMT) including (a) psycho education, self-assessment of stressors and complaints, lifestyle, and relaxation techniques; (b) cognitive restructuring; (c) time management and goal setting; (d) assertiveness skills; and (e) evaluation and relapse prevention - Treatment providers: 12 experienced therapists with a master s degree in clinical psychology delivered the SMT; therapists were trained in delivering the treatment according to the protocol in 4 training sessions of 1 hour and received at least four supervision sessions per treatment by one of two experienced senior cognitive-behavioural therapists - Treatment frequency and duration: 12 sessions of approximately 2 hours with 8 participants, conducted by 2 therapists Comparison intervention: - Treatment type: usual care based on routine care by an occupational physician or general practitioner, or a maximum of five treatment sessions by a psychologist or social worker - Treatment providers: OPs, general practitioners, psychologists, social workers Outcomes Work-status outcomes: - time to partial and full return to work Other outcomes: - Maslach Burnout Inventory (MBI-NL): baseline questionnaires were sent to patients before the intervention and follow-up measurements at 4, 7 and 10 months of follow-up - Checklist Individual Strength (CIS): baseline questionnaires were sent to patients before the intervention and follow-up measurements at 4, 7 and 10 months of follow-up - Depression Anxiety Stress Scales (DASS): baseline questionnaires were sent to patients before the intervention and follow-up measurements at 4, 7 and 10 months of follow-up - Treatment satisfaction/ perceived effectiveness with care received assessed by 4questions. Notes Source of funding: The Netherlands Organisation for Health Research and Development (ZonMw), The Netherlands Organization for Scientific Research (NWO; Fatigue at work) Ethics: approved by ethical committee Department of Psychology, University of Amsterdam Rebergen 2007 Study name CO-OP Study See this thesis for more information Characteristics of ongoing studies Oostrom 2008 Study name Adapt study Methods RCT Participants NA Interventions NA Outcomes NA Starting date 2000 Contact information Corresponding author: Sandra H van Oostrom, MSc

COCHRANE REVIEW

58

Supplement 3 Study eligibility form Study ID ID code: First author: Publication year: Type of study Instruction: RCT: study is described as randomized CCT: concurrent control group used with outcome measurements before/after the intervention Interrupted time series: outcome measurements at least 3 times before and 3 time after the intervention in one group

RCT, CCT or interrupted time series other study design exclusion unsure

Type of participants 1. Adults? Yes / no exclusion / unsure 2. Worker population? Yes / no exclusion / unsure 3. At least 50% of study population on sick leave? Yes / no exclusion / unsure 4. Participants with adjustment disorders? Yes / no exclusion / unsure Instruction: Adjustment disorder must be defined as either: - diagnosis of adjustment disorder according to DSM IV - level of stress related symptoms according to validated self-rated or clinician-rated instrument, published in peer reviewed journal 5. Co-morbidity?

none or common mental disorder bipolar or psychotic features exclusion unsure

Intervention Is the study an intervention study? Yes / no exclusion / unsure Instruction: Interventions can be aimed at the workplace (e.g. job re-design) or the individual (e.g. psychotherapeutic interventions) Outcome measure Was sickness absence measured as outcome? Yes / no exclusion / unsure

CHAPTER 3

Design of a randomized controlled trial on the effects of Counseling of mental health

problems by Occupational Physicians on return to work: the CO-OP-study

Published as: Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W

Design of a randomized controlled trial on the effects of Counseling of mental health problems by Occupational Physicians on return to work: the CO-OP-study

BMC Public Health 2007;7:183.

DESIGN

60

ABSTRACT Background Mental health problems often lead to prolonged sick leave. In primary care, the usual approach towards these patients was the advice to take rest and not return to work before all complaints had disappeared. When complaints persist, these patients are often referred to psychologists from primary and specialized secondary care. As an alternative, ways have been sought to activate the Dutch occupational physician (OP) in primary care. Early 2000, the Dutch Association of Occupational Physicians (NVAB) published a guideline concerning the management by OPs of employees with mental health problems. The guideline received positive reactions from employees, employers and Dutch OPs. This manuscript describes the design of a study, which aims to assess the effects of the guideline, compared with usual care. Methods/Design In a randomized controlled trial (RCT), subjects in the intervention group were treated according to the guideline. The control group received usual care, with minimal involvement of the OP and easy access to a psychologist. Subjects were recruited from two Dutch police departments. The primary outcomes of the study are return to work and treatment satisfaction by the employee, employer, and OP. A secondary outcome is cost-effectiveness of the intervention, compared with usual care. Furthermore, prognostic measures are taken into account as potential confounders. A process evaluation will be done by means of performance indicators, based on the guideline. Discussion In this pragmatic trial, effectiveness instead of efficacy is studied. We will evaluate what is possible in real clinical practice, rather than under ideal circumstances. Many requirements for a high quality trial are being met. Results of this study will contribute to treatment options in occupational health practice for employees on sick leave due to mental health problems. Additionally, they may contribute to new and better-suited guidelines and stepped care. Results will become available during 2007. Trial registration Current Controlled Trials ISRCTN34887348

DESIGN

61

BACKGROUND Common mental health problems and productivity loss Common mental health problems can affect functioning to such an extent that they can lead to work absenteeism and presenteeism. These may result in productivity loss [1]. The prevalence of absenteeism due to mental health problems is reported to be between 10 and 18%, which causes extensive societal and financial costs [2-7]. Up to ninety percent of absenteeism is caused by minor, stress-related, mental health problems [3,4,6]. A small, but substantial part (over 20%) of these ‘common’ mental health problems result in long lasting productivity loss. In the Netherlands associated costs are enormous (9.407 billion Euros in 2004)[8,9]. Stress management in Dutch occupational health care For employees with common mental health problems, health care utilization is mainly restricted to primary care. In the Netherlands, primary care is regularly given by general practitioners (GPs) and occupational physicians (OPs). The majority of these OPs are working for commercially operating Occupational Health Services (OHSs). Each Dutch employee has to visit their OP for rehabilitation purposes when they are on sick leave. Therefore, the Dutch OP has a perfect opportunity to play a central role in the diagnosis and treatment of employees with common mental health problems. However, OPs often lack time and skills to deal with these employees [10,11]. Consequently, the approach of Dutch OPs towards employees on sick leave due to mental health problems has been minimal. The usual initial advice given by OPs and GPs towards these patients has been to take rest and only return to work when all complaints have disappeared. When complaints persist, these patients are often referred to psychologists from primary and specialized secondary care. Treatment by these psychologists is mostly symptom based rather than focusing on return to work. Most employees are not insured for these treatments. Recent Dutch, and Scandinavian, studies suggest that an inactive primary care that easily refers to specialized secondary care, may cause ‘referral’ delay in recovery *10-14]. In addition, referrals to specialized secondary care can be expensive for employees and employers, as they have to pay the price. As a consequence, patients may not get the optimal care they need. Stress management by occupational physicians As an alternative to usual care, ways have been sought to activate the Dutch OP in primary care. As the OP is visited by each employee with or at risk of common mental health problems, the OP has a key role to detect them and influence their return to work. Therefore, a renewed position of the OP was introduced in a national evidence based guideline regarding the management by OPs of employees with mental health problems [15,16]. The guideline was published by the Dutch Association of Occupational Physicians (NVAB) in 2000. It promotes an active attitude and activating approach, instead of a minimal role of the OP. The guideline received positive reactions from employees, employers and Dutch OPs. However, there is reason to believe that the actual implementation of the guideline lags behind its acceptance, which questions the effectiveness of the guideline in practice [17,18]. Study rationale/Objective The aim of this study is to examine the effectiveness of the care by Dutch OPs according to the new common mental health guideline. The guideline may improve the effectiveness of occupational rehabilitation among workers with common mental health problems. This

DESIGN

62

study will focus on the effects of training in the guideline on the skills of the OP, resulting in positive effects on return to work and treatment satisfaction. METHODS Study design In a randomized controlled trial (RCT) the effect of the ‘Dutch national guideline on the management of employees with mental health problems by OPs’ was evaluated. The focus of this study is to examine the new, more active role of Dutch OPs according to the new guideline. Therefore, subjects in the intervention group were treated by OPs, who were trained to provide treatment according to the guideline. The control group received usual care, with minimal involvement of the OP and if applicable, access to treatment by a psychologist.

The first hypothesis of this study is that the intervention will lead to health gain for employees on sick leave due to common mental health problems. This will result in faster recovery, less stagnation and less referrals to psychologists. Counseling, instead of symptom based treatment, will result in earlier return to work and consequently a decrease of productivity loss. The second hypothesis is that the intervention will additionally lead to relatively more treatment satisfaction of the employee, the employer and the OP. The third hypothesis is that a decrease of productivity loss and prevention of expensive referrals to secondary care, will reduce costs.

The recruitment of participants for the study started in January 2002 and ended in January 2005. There was a one year follow-up. The study was funded by the Dutch Ministry of Internal Affairs and Kingdom Relations, and the Insurance Agency on Medical Guidance of the Dutch Police (DGVP). The study design, protocol and procedures were approved by the Medical Ethics Review Committee of the VU University Medical Centre. Participants Setting This study was conducted with the cooperation of the Dutch police force, which is an organization with a relatively high incidence of mental health problems. These problems are mainly work related as police work has inevitable risks and stress may develop as ‘part of the job’ *19+. The employer of the Dutch police, the Ministry of Internal Affairs, tried to provide an optimal care and were open to alternative effective treatments. Each Dutch police employee was insured by the insurance company, the DGVP. DGVP tried to provide optimal usual care by partly financing referrals by OPs of police employees with mental health problems to a commercial psychotherapeutic centre as part of a protocol. Therefore, these police departments and their occupational health care provided a representative study population.

This intervention was developed for the occupational health care setting with its typical case load of common mental disorders. Two police departments were chosen because they had contracts with the same private OHS, i.e. Commit. Consequently, uniformity in treatment was more secured. Commit is one of the largest OHSs in the country. The police departments, i.e. Zaanstreek-Waterland and Hollands Midden, were located in the South-West of the Netherlands. Hollands Midden comprises approximately 1700 employees; Zaanstreek-Waterland approximately 800, totalling a source population of 2500 police employees.

Because we wanted to prevent employees with chronic disability to participate in the study, each employee on sick leave due to mental health problems before the start of

DESIGN

63

the study in 2002, was detected by the OHS. They received a treatment by a psychologist in secondary care funded by the DGVP. Recruitment and selection of the participants Regularly, employees were registered on their first day of sick leave by the OHS (figure 1). Since January 2002, each employee on sick leave due to mental health problems was invited to meet with a case manager of the OHS within one week. This case manager informed the employee about the study and planned a consultation with an OP in the first two weeks of sick leave. To enhance recruitment one of the researchers (DB), who was allowed to check the registration system of the OHSs, informed the OP when a potential participant would come for consultation. Each employee who consulted an OP, and was still on sick leave due to mental health problems, was then asked by the OP to participate in the study. After an employee had signed informed consent during this consultation (T0), the OP unsealed a study envelope containing the allocated treatment for the patient, and sent the signed informed consent to the researcher (DR). In the same consultation the employee received the baseline questionnaires and was asked to return this questionnaire to the researcher after completion. In- and exclusion criteria As the guideline focuses on all kinds of mental health problems, we aimed to include employees with a broad range of mental health problems consulting their OP. Employees were included if they met the following inclusion criteria: Mental health problems according to the diagnosis of the OP Sick leave at the moment of inclusion Sick leave period did not start before 2002.

Exclusion criteria were the same as stated in the OP-guideline: Mental health symptoms that were caused by somatic illness Disagreement between OP and employee about the diagnosis Lack of confidence in the relation between OP and employee

The application of the exclusion criteria was dependent on the OPs expert judgement. To prevent selection bias, employees were not included of whom the period of sick leave started before 2002. Randomisation Block randomisation (size 50) was done on the patient level before the start of the study using SPSS. The randomisation results were sealed in 250 consecutive envelopes. The OPs were informed about the study procedure and received sealed numbered envelopes, in which the treatment was stated which they had to provide. They were allowed to open an envelope only after an employee voluntarily signed an informed consent. Then the OP told the participant to which treatment her or she was assigned. To minimize the risk of irregularities by letting OPs open their treatment concealment themselves, randomisation was checked by an independent researcher (AvdB) one year after the start of the study. At the end of the study this procedure was repeated by checking the treatment allocation of all the in- and excluded persons.

DESIGN

64

Figure 1 Flow chart of time line, study design and return to work Blinding Participants, employers and OPs were not blinded for the intervention. The researchers were blinded and did not know the treatment allocation of the employee, to prevent any influence on the study procedure. As this is an effectiveness study researchers were blinded for protocol compliance as well, to make the trial as realistic as possible. Blinding of the gathered sick leave and medical files data was secured, since these measurements

T –1: Start sick leave employee due to mental health problems Registration of employees by OHS as being on sick leave

1st week after start sick leave: Consultation case manager Study information to employee by case manager

2nd week after start sick leave: First consultation OP Recruitment participants study by OP during first consultation

Employee signs informed consent, OP sends this to researcher OP unseals envelop with treatment allocation

T 0: Inclusion in RCT: study population Checking for eligibility criteria by researcher

Usual care: Minimal involvement OP and easy access to psychologist

Intervention: Guidance OP according to guideline

waarbij inclusie in onderzoek,

waarop data-analyse is gebaseerd

T1: Mental health questionnaires Data collection treatment satisfaction


waarop data-analyse is gebaseerd T4:Follow-up 1 year after inclusion

Data collection sick leave

T3: Last consultation OP Data collection treatment satisfaction

T1: Mental health questionnaires Data collection treatment satisfaction

T3: Last consultation OP Data collection treatment satisfaction

T4:Follow-up 1 year after inclusion Data collection sick leave

T2: Second consultation OP Data collection treatment satisfaction

T2: Second consultation OP Data collection treatment satisfaction


DESIGN

65

were gathered from the automated databases of the police constabularies, the insurance agency and the OHSs. Interventions Usual care In this study the aim was to compare usual care of employees with mental health problems with an intervention. Usual care consisted of minimal involvement of the OP and access to treatment by a psychologist, as this represents daily practice. As the aim was to deliver the best optional care to our study population, optimal usual care was provided in this study. This consisted of the advice to OPs to refer to a psychologist, whose treatment was fully funded by the DGVP. The OP could refer to a psychologist working for a commercial multidisciplinary rehabilitation center, i.e. De Gezonde Zaak (DGZ), as this was part of an agreement with the OHS Commit. A patient was only referred if according to the expert judgement of the OP this made sense to the health condition of this person.

DGZ is one of the largest Dutch commercial psychotherapeutic intervention centers, which focuses on return to work of the employee. DGZ is located in different parts of the country. Besides physical therapists, around 100 psychologists are working for this organization. The psychologists are working according to cognitive behavioral principals. The standard therapy offered was based on protocols of the Dutch Institute for Work and Stress [20]. Intervention The intervention consisted of treatment by OPs according to the guideline of employees on sick leave due to mental health problems. The guideline promotes a more active role of the OP as case and care manager facilitating return to work of the employee. The guideline is based on an activating approach, time contingent process evaluation and cognitive behavioral principles. The latter mainly concern stress inoculation training and graded activity and aim to enhance the problem-solving capacity of patients in relation to their work environment.

The guideline focuses on four aspects of the management of mental health problems. First, an early and activating guidance by the OP is promoted, in which return to work is part of the recovery process, even if the mental health problems are not related to work. Second, a simplified classification of mental health problems is introduced, with only four categories: 1) adjustment disorder (distress, nervous breakdown, burnout), 2) depression, 3) anxiety, and 4) other psychiatric disorders. Third, the OP acts as case manager, who is stimulated to be a care manager by counseling employees with adjustment disorders and work-related problems. Fourth, the OP performs a time contingent process evaluation and intervenes when recovery stagnates.

OPs participating in the study received training in the guideline before the study started. During this training, consisting of a three-day course with 10-15 other OPs, knowledge about and practice in working with the guideline were educated and exchanged [21]. The course reflected the guideline by training OPs in multiple cognitive-behavioral prescriptive interventions, to stimulate the patients’ acquisition of problem solving skills, and to structure the patients’ daily activities. Information was given and OPs were trained to differentiate between adjustment disorder and depression, anxiety and other psychiatric disorders. Questionnaires were introduced, which can be helpful in making an accurate diagnosis. In addition, a graded activity treatment approach was introduced, which was based on a three stages model. This treatment approach resembles stress inoculation training, a highly effective form of cognitive behavioral treatment [16]. In the first stage, there is emphasis on information: understanding the origin and cause of

DESIGN

66

the loss of control. Patients are also stimulated to do more non-demanding daily activities. In the second stage, patients are asked to draw up an inventory of stressors and to develop problem solving strategies for the causes of stress. In the third stage, patients put these problem solving strategies into practice and extend their activities to include more demanding ones. The patients’ own responsibility and active role in the recovery process was emphasized and the same goes for the importance of an early start of the intervention aimed at the acquisition of coping skills and at regaining control. Problem solving activities according to a time contingent scheme were educated and practiced. The OPs were free to choose the specific tools for use in each phase of the process. The training course was given by four persons: an experienced OP/psychologist, a psychologist/therapist, an experienced general practitioner/ researcher on emotional distress, and a psychiatrist. Co-interventions Co-interventions cannot always be avoided. In case of post traumatic stress disorders, patients were also referred to a specialized trauma centre according to a special protocol of the DGVP. As the rehabilitation centre DGZ worked with multiple disciplines, it is possible that some patients in the control group received a combined intervention for mental and physical complaints by a psychologist and a physical therapist. In both the intervention and control groups co-interventions were registered in the medical files of the OHS and the database of the DGVP. These data can be used to adjust for co-interventions in the final multivariate analyses. Compliance As this is an effectiveness trial, we tried to mimic a realistic situation in both treatment groups. Therefore, no activities were undertaken to improve the actual treatment compliance by the OP with the allocated treatment. Treatment compliance of the OPs was examined by measuring guideline adherence and assessing the proportion of referrals by the OP in both groups to the psychologist of DGZ. Patient compliance of the treatment was examined by registering no shows of patients during consultations with the OP were, as this may give information about the willingness of the patient to adhere to the treatment. Contamination As randomization was done on patient level, OPs which were trained in the guideline treated all participants. Obviously this situation created a risk of treatment contamination between the groups. The trained OP treated an employee in the intervention group according to the guideline, as far as this happens in practice. The same OP treated an employee in the control group with minimal involvement and if applicable, direct referral to a psychologist. A cross-over learning effect may have happened in the control group, since the OP can adhere to the guideline in this group as well. The other way around, the OP may have referred an employee in the intervention group to a psychologist as well. The guideline promotes this in case of stagnation in recovery or in case of severe mental health problems of the employee. However, we tried to maximize the contrast by creating a situation in which referral to the psychologist in the control group was always granted by the insurance company (DGVP).

DESIGN

67

Primary outcomes Return to work Return to work (RTW) was chosen as the primary outcome in this study [22]. A follow-up time of one year after inclusion was chosen, as effects of the intervention on RTW were expected to happen in this period. The RTW-outcomes are visualised in a time line in figure 2.

The primary outcome measure described in our study is full RTW: i.e. duration of sick leave due to mental health problems in calendar days from the first day of sick leave to full return to work in own or equal earnings (table 1). In addition, the net return to work to own or equal work was measured. This is the net duration of sick leave due to mental health problems in hours of full work absenteeism from the first day of sick leave to full return to work with own or equal earnings. The difference with full RTW is that the hours of partial return to work and % contract working hours (36 hours = 100%) are converted into the number of hours of full work absenteeism [23]. This may result in a more precise measure of RTW, when compared to full RTW. As figure 2 shows, net and full RTW consider the time period between T0 and T3, in which data collection by questionnaires took place (figure 1).

Figure 2 Timeline of measured sick leave data of a potential participant

Another primary outcome variable that is part of the RTW process is first RTW: i.e. the duration of sick leave due to mental health problems in calendar days to first (partial or full) return to own or equal work. Other variables are related to recurrences of sick leave periods in the one year follow-up (T4)(table 1). These variables are the time in calendar days until the first recurrence of sick leave takes place and the number of and days during recurrences [21]. Total days of sick leave during follow-up is a primary outcome variables as well.

There was double registration of sick leave data, as both the employer and the OHS have their own registration system. The aim was to compare data of both systems, with the sick leave data of the employer as the ‘golden standard’. This additional effort

Data collection in one year follow up (figure 1)

1 year before start sick leave

Start sick leave

Inclusion in study

Data collection sick leave in year before study

Full /net RTW

1 year follow-up

Sick leave periods year before study

Recurrent sick leave periods

T -2 T -1 T 0 T 3 T 4

Time until first

recurrence

T2 T1

First (partial) RTW

Sick leave period of study

inclusion

DESIGN

68

was done as reliable sick leave data are hard to get, since there is a known discrepancy with self-reported sick leave [24,25]. Treatment satisfaction Treatment satisfaction is a relevant outcome measure in occupational health care and therefore another primary outcome [26]. Higher patient satisfaction is related to better patient compliance and can improve the quality of OHSs. To date, most researchers assume that patient satisfaction is best defined as a patient’s evaluation of aspects of a health care service based on the fulfilment of patient expectations. Since patients, employers and health care providers (OPs) are all involved stakeholders in the RTW process, it is important to measure the treatment satisfaction of all of these stakeholders. Patient and employer satisfaction were measured using a short version of the Patient Satisfaction with Occupational Health Professionals Questionnaire [26]. This questionnaire was designed specifically for measuring satisfaction with occupational health care. It was designed in previous research on the quality of rehabilitation of cancer survivors [27] and transferred to occupational health care in another study on employees with mental health

problems [18]. The 13 items of this questionnaire refer to (a) satisfaction in general (2 items), (b) interpersonal approach (4 items), (c) communication manner (2 items), (d) professional knowledge (5 items), and (e) total satisfaction of the treatment by the OP (13 items). Respondents answered on thirteen statements on a 5-point Likert scale: ‘totally disagree –disagree – no opinion – agree – totally agree’ (table 2). Because a higher score indicates more treatment satisfaction, item 3,4,9,10 and 12 will be recoded. The patient satisfaction questionnaire was adapted to the situation of the employer, to measure the treatment satisfaction of the supervisor (table 3).

To measure treatment satisfaction of OPs, OPs filled in an evaluation questionnaire for each employee treated. This questionnaire consisted of 6 items, the first 4 referring to possible barriers in the RTW process and the last 2 items referring to the treatment success of their OHS (table 4). Secondary outcome Cost-effectiveness measures Cost-effectiveness of the intervention is a secondary outcome and was evaluated from the employers and the health care insurance company’s perspective (expenditures for the employer and insurance company, respectively), as they are responsible for covering the costs of sick leave and treatment [28]. Direct costs of health care treatment are (table 1): (a) consultations of OPs and other OHS-professionals, (b) consultations of the psychologists of DGZ, (c) consultations of general practitioners, (d) consultations of a psychiatrist and/or psychologist and/or alternative therapist not participating in the study, and (e) medication related to the treatment of mental health problems [29].

Indirect costs are not related to health care, but are costs as a consequence of absence from work because of sickness: sick leave, disability and or death of productive persons. Costs of lost productivity caused by (partial) sick leave due to mental health problems were calculated from the net number of days of sick leave and lost earnings, as provided by the employer. Since our study took place in occupational health care and since most costs were caused by sick leave, extra efforts were made to gather reliable data on sick leave [30].

DESIGN

69

Table 1 Measured data and their instruments and timing of data collection

Data Instrument Base line

Follow up

Data related to primary outcomes T 0 T1 T2 T3 T4

Return to work Full Return to work Database company X X Net return to work Database company X X First (partial or full) return to work Database company X X X X Time until 1st recurrence Database company X X Number and days of recurrences of sick leave

Database company X X

Total days of sick leave during one year follow up

Database company X X X

Treatment satisfaction Treatment satisfaction of employee Questionnaire X X X Treatment satisfaction of employer Questionnaire X X X Treatment satisfaction of the OP Questionnaire X Data related to secondary outcome T 0 T1 T2 T3 T4

Cost effectiveness

Direct costs of treatment

Consultations OP, treatment OHS Medical files OHS X X X X

Consultations of participating psychologist centre

Medical files centre X X X X

Consultation of general practitioner Insurance company X X X X

Consultations psychiatrist/ psychologist/alternative therapist

Insurance company X X X X

Medication Insurance company X X X X

Indirect costs of lost productivity

Net lasting RTW and earnings Database company X X

Replacement Database company X X

Data related to prognostic measures T 0 T1 T2 T3 T4 I) Personal characteristics

Gender, Age Database company X

Severity disorder: DASS/HADS Questionnaires X

Work-relatedness of the disorder Medical files OHS X

Sick leave in year before inclusion Database company X

II) Treatment

Treating OP Medical files OHS X

Diagnosis made by the OP Medical files OHS X

Guideline adherence of the OP Medical files OHS X X X X

III) Work characteristics

Type of function Database company X

Number of working hours Database company X

Police constabulary Database company X

DESIGN

70

Table 2 - Treatment satisfaction questionnaire employee Items relate to last consultation the employee has had with the OP* General satisfaction with the OP 1 I am very satisfied about the contact with the OP 2 In general, contact with the OP made sense Interpersonal approach by the OP 3 The OP can be more respectful to me 4 The OP is more interested in the employer’s, than my point of view 5 The OP seems interested in me as a person 6 The OP treats me in a pleasant manner Communication manner of the OP 7 The OP is good in explaining his or her opinion about returning to work 8 The OP listens well to what I have to say Professional knowledge by the OP 9 The OP forces me to return to work, while this is impossible 10 The OP has no experience with my kind of problems 11 The OP gives me good advice about how to deal with my health complaints 12 The OP does not seem professional to me 13 The OP knows what he/she is talking about 1-13 Total satisfaction of the treatment by the OP (all 13 items mentioned above) * 5-point Likert scale from 1 (totally disagree) to 5 (totally agree) To increase the readability of this article we have translated the questionnaire from Dutch

to English. In the study the Dutch version was used.

Table 3 - Treatment satisfaction questionnaire supervisor Items relate to last consultation your employee has had with the OP* General satisfaction with the OP 1 I am very satisfied about the contact with the OP 2 In general, contact with the OP made sense Interpersonal approach by the OP 3 The OP could be more respectful to me 4 The OP is more interested in employee’s, than the employer’s point of view 5 The OP seems interested in me as supervisor 6 The OP treats me in a pleasant manner Communication manner of the OP 7 The OP is good in explaining his or her opinion about RTW of my employee 8 The OP listens well to what I have to say Professional knowledge by the OP 9 The OP forces my employee to return to work, while this is impossible 10 The OP has no experience with my kind of problems as being a supervisor 11 The OP gives me good advice about how to deal with the health complaints of my employee 12 The OP does not seem professional to me 13 The OP knows what he/she is talking about 1-13 Total satisfaction of the treatment by the OP (all 13 items mentioned above) * 5-point Likert scale from 1 (totally disagree) to 5 (totally agree) To increase the readability of this article we have translated the questionnaire from Dutch

to English. In the study the Dutch version was used.

DESIGN

71

Table 4 Treatment satisfaction questionnaire OP Process evaluation of the treatment by the OP 1 Which was the effect of the treatment given by the OHS on the employee, a) related to recovery?* b) related to return to work?* 2 Was the employee cooperative regarding the treatment? a) No, not cooperative; b) Cooperative, but passive; c) Cooperative and active; d) No idea 3 Was the employer cooperative regarding the treatment? a) No, not cooperative; b) Cooperative, but passive; c) Cooperative and active; d) No idea 4 What was the influence of the following factors on return to work of the employee?* a) Degree of physical work load b) Degree of mental work load c) Degree of physical work ability of the employee d) Degree of mental work ability of the employee e) Support by supervisor f) Support by colleagues g) Support by employer h) Work motivation of the employee i) Job control of the employee j) Relationships at work between employee and employer k) Duration of curative treatment l) Advices of the curative sector m) Waiting lists in the curative sector n) Inadequate sickness behaviour of the employee o) Psychosocial situation of the employee p) Financial situation of the employee q) Home situation of the employee (including care tasks) r) Remaining, not work-related, factors s) Practical (including organizational) options to work accommodations t) Financial circumstances employer u) Other factor, namely… 5 To what extent are you satisfied by the treatment of the OHS, related to a) Treatment effectiveness?** b) Treatment process?** * Response range: 1. obstructive ; 2. no effect or influence ; 3. supportive ** 7-point Likert scale from 0 (totally dissatisfied) to 6 (totally satisfied) To increase the readability of this article we have translated the questionnaire from Dutch to

English. In the study the Dutch version was used.

Prognostic measures Prognostic measures and potential confounders were searched for in the literature [16,31,32]. The following prognostic measures were selected and will be taken into account as potential confounders (table 1): I) personal characteristics: (a) gender, (b) age, (c) disorder severity, based on mental health symptoms (DASS, HADS), (d) work relatedness of sick leave on the moment of inclusion, e) total days of sick leave in the year before the inclusion (figure 1); II) treatment characteristics: (f) treating OP, (g) guideline adherence by the OP, (h) referral behaviour of the OP; III) work characteristics: (i) type of function (executive vs. administrative), (j) working hours (part-time vs. full-time), and (k) police department (Zaanstreek-Waterland vs. Hollands Midden).

DESIGN

72

Depression Anxiety Stress Scale (DASS) To measure mental health complaints at baseline in this study, the Depression Anxiety Stress Scales (DASS) were used [33]. The structure of the DASS seems to support the view that both anxiety disorders and depression need to be distinguished from adjustment disorders in spite of their communality. The psychometric properties of this instrument appear to be sound enough to be applied to both healthy and psychiatric populations. Therefore, the psychometric properties of the DASS are suitable for use in an occupational health care setting. Moreover, convergent and divergent validity have been shown to be satisfactory [34].

The employees participating in this study filled in a self-report questionnaire that comprises the DASS-42, which takes 7 minutes to complete. The DASS-42 consists of 42 symptoms divided into three subscales of 14 items: depression scale, anxiety scale, and stress scale. Participants rated at baseline the extent to which they had experienced each symptom over the previous week on a four point Likert scale ranging from 0 (did not apply to me at all) to 3 (applied to me very much, or most of the time).

Based on the results of their study on employees with mental health problems in occupational health care, Nieuwenhuijsen et al. [34] developed cut-off scores to divide the DASS-rates into four categories: stress, depression, anxiety, and depression/anxiety. The cut-off scores are >12 on symptoms of depression and >5 on symptoms of anxiety. Hospital Anxiety Depression Scale (HADS) The HADS is a 14-item screening scale that measures the presence of anxiety and depressive states [35]. It contains two 7-item subscales: a depression subscale and an anxiety subscale, each item being scored on a four point Likert scale (0–3) that applies to the previous week. The HADS has been developed as a screen for detecting depressive and anxiety disorders in hospitalised patients. Items referring to symptoms that may have a physical cause (for example, weight loss or insomnia) are not included in the scale. Because a higher rate indicates more mental health symptoms, item 1,3,5,6,8,10,11 and 13 will be recoded.

The HADS is easily administered as a self-report measure as it usually takes 3–5 minutes to complete. A total score (out of a possible 21) for each subscale is then calculated. Zigmond et al. [35] recommended cut-off points with scores less than eight on either of the two subscales to be non-cases and scores between eight and ten as borderline cases. Guideline adherence by the OP The aim of this study is to examine the effectiveness of the management by Dutch OPs, under the expectation that (training in) the guideline will lead to additional skills for the OP and consequently to positive outcomes. To explore our hypothesis that the guideline leads to additional skills and outcomes, we examined the performance by the OP according to the guideline (guideline adherence). Guideline adherence by the OP was checked by means of an audit of the medical files. Guideline adherence was defined as the total score on ten validated performance indicators for the treatment of each participant by the OP (table 5)[11,18,36]. For each performance indicator, we used validated criteria. If a criterion was not met, the case was assigned 1 for that performance indicator. If all applicable criteria for a performance indicator were met, the resulting score was 0 for that case. The medical files of all the participants were assessed on if they met the criteria of the different indicators (0=adequate care; 1=deviant care). In this way an average performance rate was obtained for each performance indicator. Furthermore, a total score of all performance indicators was calculated (guideline adherence). Guideline

DESIGN

73

adherence was dichotomized into adequate adherence and deviant adherence. Adherence was considered deviant if three or more performance indicators were assigned a score of 1, and adequate if less than three performance indicators had a score of 1.

Additionally, this audit gave us information about treatment compliance by the OP (guideline adherence) in both groups. In this way contamination between the study groups was studied as well. Adherence in the intervention group was considered compliant if there was adequate adherence. Guideline adherence in the control group was considered compliant if there was deviant adherence.

The performance indicators will be assessed on their criteria by three independent researchers, resulting in a dichotomised score on guideline adherence for each employee (adequate versus deviant). Data collection The participants had to complete the mental health questionnaires (DASS,HADS) on the moment (T1) after they signed the informed consent (T0) (figure 1) (table 1). At the same time a questionnaire had to be filled in about their satisfaction with the treatment of the OP (T1). This questionnaire was sent again to the participant by the researcher (DR) after their second consultation with the OP (T2) and after the last consultation with the OP, at the moment of full RTW (T3). If T2 and T3 happened at the same moment, T2 was considered as T3. The questionnaires were returned to the researcher after completion in pre-stamped envelopes. The same was done for the supervisors of the participants, who received a questionnaire about their treatment satisfaction at the same moments (T1, T2 and T3). The OP received for each participant after the moment of full RTW another questionnaire to assess their treatment satisfaction (T3).

Baseline characteristics of the participants such as gender, age, marital status, work characteristics (type of function, hours working, part/full time), sick leave data and costs of work incapacity of the participants were gathered from records of the police constabularies, the latter after one year follow-up. Data about direct costs of treatment and medication of the employee were obtained after one year follow up from the insurance company of the police, the DGVP. Guideline adherence and the according performance were based on data of the medical files of the participant, gathered from the databases of the participating OHSs. The data of the medical files were made anonymous and were transferred to an Access database, to select the relevant data of the medical files. Study population: sample size and power analysis In order to detect a relevant difference in survival analysis on our primary outcome return to work, nQuery Advisor [37] was used to calculate the sample size. Proportions used to determine the sample size needed, were analysed from sick leave data of the police constabularies in 1999. In 1999, 286 employees were registered as being on sick leave due to mental health problems, which was 6.6 % of the total sick leave registrations. Their duration of sick leave in 1999 was 35.5 % of the total volume of sick leave, with an average of three months per case. With a power of 90%, at a 0.05 level, a two-sided log-rank test for equality of survival curves was done, assuming a difference between the intervention and control group proportion still on sick leave after one year of 0.25. This test indicated that a sample size was needed of 107 in each group. Assuming a dropout rate of 20%, inclusion of a total of 268 patients was necessary to statistically detect a clinically relevant difference.

DESIGN

74

Table 5 Performance indicators guideline adherence and their criteria 1=deviant care, NA= Not applicable PI 1 Assessment of symptoms Score

Criteria: 1. Presence or absence of essential symptoms of anxiety disorder and depressive disorder should be noted in file

2. Presence or absence of distress symptoms (fatigue, concentration problems, sleeping problems, and emotional reactivity) should be noted in file

One or both criteria not met within 2 consultations? PI1=1

PI 2 Correct diagnosis

Criteria: 1. Diagnosis should be noted in file 2. Diagnosis should be correct: - IF adjustment disorder: at least one psychological distress symptom should

be noted in file

- IF depressive disorder: at least one essential symptom AND five depressive symptoms should be noted in file

- IF anxiety disorder: at least one anxiety disorder should be noted in file 3. Diagnosis should not be missed if criteria above apply One or more criteria not met within 2 consultations? PI2=1

PI 3 Evaluation curative care

Criteria: 1. Treatment in the curative sector, or its absence, should be noted in file 2. IF patient receives treatment, THEN the OP should evaluate whether this

treatment is effective


PI 4 Assessment work-related causes

Criterion: 1. work-related causes, or their absence, should be stated in file Criterion not met within 2 consultations? PI4=1

PI 5 Evaluation of work disabilities

Criteria: 1. Functional limitations in home or work environment, or their absence, should be stated in file.

2. Work activities of patient should be noted by OP 3. OP should assess whether patient is limited in his work functioning 4. IF patient has work limitations, THEN OP should assess other

impediments for return to work (such as problems in home situation or with supervisor)


PI6 Interventions targeted at individual

Criterion: 1. Intervention aimed at the individual should be noted or be referred - IF adjustment disorder, THEN OP should start interventions OR should refer patient to psychologist/social worker/GP OR should consult with practitioner giving current treatment - IF anxiety disorder OR depression OR other psychiatric disorder, THEN OP should refer patient to psychologist/social worker/GP OR should consult with practitioner giving current treatment

Criterion not met within 3 consultations? PI6=1

PI7 Interventions targeted at organisation

Criterion: IF work is a causal, eliciting or maintaining factor in the mental health problem, THEN OP should intervene in the work organisation (confer with supervisor/personnel officer)

DESIGN

75

Criterion not met within 3 consultations? PI7=1 IF work is neither causal, eliciting nor maintaining factor in mental health

problem PI7=NA

PI8 Interventions targeted at providers of care in curative sector

Criteria: 1. IF treatment in curative sector is lacking and deemed necessary, THEN OP should start interventions targeted at the individual OR refer patient to psychologist/social worker/general practitioner

2. IF treatment in curative sector is not effective, THEN OP should consult with practitioner giving current treatment

One or both criteria not met within 3 consultations? PI8=1 IF patient receives effective treatment in curative sector PI8=NA

PI9 Advice on return to work

Criteria: 1. Advice on return to work should be provided by OP 2. IF no impediments for return to work are present, THEN OP should advise

full or partial return to work

One or both criteria not met at each consultation? PI9=1 Patient already (partially) returned to work? PI9=NA

PI10 Timing of consultations

Criteria: 1. First consultation should be within 3 weeks from first day of sickness absence

2. IF patient has not yet completely recovered, THEN next consultation should be within 4 weeks from previous one

Criterion 1 not met at first consultation OR criterion 2 not met at consultation 2 or 3?

PI10=1

Data analysis All analyses will be conducted according to the intention-to-treat principle and will be performed on the patient level. To examine the success of randomization, descriptive statistics will be used to compare the baseline measurements of the two groups. If necessary, analyses will be adjusted for prognostic dissimilarities.

The evaluations on the effectiveness of the guideline compared to usual care will be performed with two tailed tests at a significance level of 5% (P < 0.05). To examine differences in the data on RTW, we will use Kaplan Meier’s and the Cox proportional hazard regression for recurrent events. The general idea behind this analysis is that the different time periods are analysed separately adjusted for the fact that the time periods within one patient are dependent. Recurrences of sick leave for any reason during follow-up will be added to the Cox proportional hazards model with the time to event approach, in which only the transitions from no treatment success (sick leave) to treatment success (full RTW) are taken into account [38].

In this model, the state of sick leave until the moment of inclusion will be added as a covariate. Number and days of sick leave periods in the year before inclusion will be added as a potential effect-modifier.

In a linear regression model treatment satisfaction during treatment (T2) and after treatment by the OP (T3) will be measured as respectively a short term and a long term effect and differences will be compared between the groups. Treatment satisfaction at the start of the treatment (T1) will be added as a covariate, even as the treatment group to examine differences in effects between the groups. The levels of treatment satisfaction of both the employees and their supervisors will be compared for the

DESIGN

76

different moments with a chi-square test. Pair-wise correlations will be used to compare treatment satisfaction of the employee and their supervisor on the different moments.

In the Cox and linear regression models potential treatment differences by the OPs and their OHSs will be taken into account by means of nested dummy variables. The police constabulary the employee works for, will be put into the model as a binomial variable. Differences in sick leave patterns in the year before inclusion and in the severity of the mental health problems (DASS/HADS-scores) will be put into the model as potential effect modifiers.

To assess whether protocol deviations will cause bias, the results of the intention-to-treat analyses will be compared to per-protocol analyses. A process evaluation will be done, based on the assessment of guideline adherence by means of performance indicators [18]. For each performance indicator potential effects on our primary outcomes will be measured. Indirect costs can be calculated using the friction cost approach (friction period 122 days) and the human capital approach, based on income as provided by the employer or as derived from function, age and gender [30]. Bootstrapping will be used for pair wise comparison of the mean groups to calculate mean differences and confidence intervals in costs and cost-effectiveness ratios for all interventions. All these analyses will be conducted in SPSS 14.0, Excel and, if necessary, in Strata. DISCUSSION Methodological considerations External validity of study results The study population, Dutch police employees, has a higher risk of getting into stressful situations than regular workers [19]. This is caused by a relatively high workload and emotional pressure, and to a certain extent this reflects that police employees have other occupational risks than the regular worker population. Additionally, the Dutch police workforce underwent two big reorganizations in the last ten years. This resulted in some negative consequences as problems with the internal communication, especially between employee and employer [39]. Because of the Volendam fire in January 2001, the police department Zaanstreek-Waterland had been exposed to extra professional risks on mental health problems and traumas one year before the start of the study [40]. As the study population will not be fully representative of the general working population, external validity of study results may be limited and caution has to be taken in generalizing the results.

Unfortunately, we cannot rule out selection bias. This may have occurred as the OPs were asked to select employees to participate in the study. Because the treatment of the OP depends on the randomization, an OP could have been tempted to forfeit the randomization procedure. As mentioned earlier, a check was made to detect this possible selection bias, and eventual irregularities would or will be noticed. Also, selection bias may have been introduced because participants completed the mental health symptoms questionnaires (DASS and HADS) after the consultation of inclusion. The consultation with their OP might have changed their point of view on their mental health problems and consequently, their response behaviour.

Still, these disadvantages regarding the ability to generalize our study results do not outweigh the advantages of this study population. This project was developed for the occupational health care setting with its typical case load of stress-related mental disorders, and not for specialized care, in which patients have more clearly defined mental

DESIGN

77

disorders. The police is an organization with a relatively high incidence of common mental health problems and is therefore an interesting and representative target population. In addition, the police is a homogenous population that was treated in a confident manner, has a uniform sick leave registration, is connected to one insurance company, and has a well-defined ‘usual care’. Contamination between study groups In this pragmatic study design we examine the effectiveness of occupational health care by OPs who are trained in a practice guideline, compared to usual care. The best situation would be to randomize on physician level and patient level. Due to the limited number of participating OPs this was not possible, hence randomization was done on patient level. Consequently, there was a risk of treatment contamination between the groups. However, contrast between the treatment groups was maximized by free access to the psychologist in usual care as this was granted by the insurance company (DGVP). Treatment adherence will be examined by per-protocol analyses of guideline adherence by the OP. Furthermore, the current method allows us to consider the pragmatic effectiveness and to avoid much interference with daily practice of consulting hours. Evaluation of productivity loss In our cost-effectiveness evaluation we will not consider productivity loss due to sick leave prior and after the episode of sick leave due to mental health problems as proposed by Brouwer et al. [41]. Considering productivity loss prior and after the episode of sick leave can lead to an increase in estimated production losses of about 16%. In this study productivity loss during the RTW process has been measured.

Productivity loss is not only influenced by the cause of sick leave, but also by the type of work. Some jobs can only be performed in case of full functioning. Police employees, for instance, are called off sick leave only when they can perform all necessary tasks. In all other cases they are still on sick leave, for instance if they have a restriction in work on the street or wearing a gun. Since the start of our study in 2002 better methods in calculating costs have become available [30]. The availability of instruments for the measurement of productivity losses in recent years will give a better estimate of costs in this study. Prospect on outcomes In this trial effectiveness instead of efficacy is studied. We will evaluate what is possible in real clinical practice, rather than under ideal circumstances. As a consequence, mental health state may have varied between the participants. Through subgroup analysis on severity of complaints and levels of distress, measured by the DASS and HADS, we will classify possible high or low risk groups for prolonged sick leave within this heterogeneous group. Identification of a high-risk group for non-recovery may lead to better suited guidelines on stepped care and treatment.

Finally, many requirements for a high quality trial are being met. Results of this study will contribute to treatment options in occupational health practice, for employees on sick leave due to mental health problems. In addition, they may contribute to new and better-suited guidelines and stepped care. Results will become available during 2007.

DESIGN

78

ABBREVIATIONS DASS = Depression Anxiety Stress Scale DGVP = Insurance Agency on Medical Guidance of the Dutch Police DGZ = De Gezonde Zaak; Dutch commercial multidisciplinary rehabilitation centre HADS = Hospital Anxiety Depression Scale NVAB = Dutch Professional Organization of Occupational Physicians OHS = Occupational Health Service OP = Occupational Physician RCT = Randomized Controlled Trial RTW = Return to work

REFERENCES 1. Brouwers EPM, Tiemens BG, Terluin B, Verhaak PFM. Effectiveness of an intervention to reduce sickness absence in patients with emotional distress or minor mental disorders: a randomized controlled effectiveness trial. General Hospital Psychiatry 2006;28:223-229. 2. Wang JL, Adair CE, Patten SB. Mental health and related disability among workers: A population-based study. American Journal of Industrial Medicine 2006;49:514-522. 3. OECD Economic Surveys: Netherlands. Reform of the sickness and disability benefit schemes. 2004. 4. National Mental Health Association USA. Mental health facts. www.nmha.org 2004. 5. The Mental Health Foundation UK. Statistics on mental health, the costs of mental health problems. www.mentalhealth.org.uk 2003. 6. Henderson M, Glozier N, Elliott KH. Long term sickness absence - Is caused by common conditions and needs managing. British Medical Journal 2005;330:802-803. 7. Nystuen P, Hagen KB, Herrin J. Solution-focused intervention for sick listed employees with psychological problems or muscle skeletal pain: a randomised controlled trial Scandinavian BMC Public Health 2006;6:69. 8. UWV. Costs of disability benefits in 2004, defined by ICD-10 diagnoses. [in Dutch] UWV (Workers Insurance Authority), 2005. 9. NKAP. Factsheet 2, Facts and figures about work incapacity due to mental health disorders. [in Dutch] Utrecht: NKAP (Dutch Knowledge Centre of Work and Mind), 2004. 10. Anema JR, Jettinghoff K, Houtman ILD, Schoemaker CG, Buijs PC, van den Berg R. Medical care of employees long-term sick listed due to mental health problems: A cohort study to describe and compare the care of the occupational physician and the general practitioner. Journal of Occupational Rehabilitation 2006;16:41-52. 11. Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. Occupational and Environmental Medicine 2003;60:21-25. 12. Bakker IM, Terluin B, van Marwijk HWJ, Gundy CM, Smit JH, van Mechelen W, Stalman WAB. Effectiveness of a Minimal Intervention for Stress-related mental disorders with Sick leave (MISS); study protocol of a clusterrandomised controlled trial in general practice [ISRCTN43779641]. BMC Public Health 2006;6:124. 13. Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Treatment of Work-Related Psychological Complaints: a Randomized Field Experiment among Self-employed. Work & Stress 2006;20(2):129-144. 14. Salmela-Aro K, Näätänen P, Nurmi JE. The role of work-related personal projects during two burnout interventions: a longitudinal study. Work & Stress 2004;18(3):208-230.

http://www.mentalhealth.org.uk/

DESIGN

79

15. NVAB. Van der Klink J, ed. Guideline for Mental Health Problems. [in Dutch] Eindhoven: NVAB (Dutch Association of Occupational Physicians), 2000. 16. van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scandinavian Journal of Work, Environment & Health 2003;29:478–487. 17. Rebergen DS, Hoenen JAHJ, Heinemans AMEC, Bruinvels DJ, Bakker AB, van Mechelen W. Adherence to a national guideline on mental health problems by Dutch occupational physicians. Occupational Medicine 2006;56:461-468. 18. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. Validation of performance indicators for rehabilitation of workers with mental health problems. Medical Care 2005;43(10):1034-42. 19. Houtman I, Jettinghoff K, Brenninkmeijer V, van den Berg R. Work stress in the police force five years later: the effect of sectored agreements on (stress) management. [in Dutch]. TNO, 2005. 20. De Gezonde Zaak. Protocol for provision of services. Edition 1.1. [in Dutch]. Arnhem, 2004. 21. Van der Klink JJL, Blonk RW, Schene AH, Dijk FJ van. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomized controlled design. Occupational and Environmental Medicine 2003;60(6):429-37. 22. Young AE, Roessler RT, Wasiak R, McPherson KM, van Poppel MNM, Anema JR. A developmental conceptualization of Return to Work. Journal of Occupational Rehabilitation 2005;15(4):557-68 . 23. Steenstra IA, Anema JR, Bongers PM, de Vet HCW, van Mechelen W. Cost effectiveness of a multi-stage return to work program for workers on sick leave due to low back pain, design of a population based controlled trial. BMC Musculoskeletal Disorders 2003,4:26. 24. Dasinger LK, Krause N, Deegan LJ, Brand RJ, Rudolph L: Duration of work disability after low back injury: a comparison of administrative and self-reported outcomes. American Journal of Industrial Medicine 1999;35:619-31. 25. van Poppel MN, de Vet HC, Koes BW, Smid T, Bouter LM: Measuring sick leave: a comparison of self-reported data on sick leave and data from company records. Occupational Medicine 2002;52:485-90. 26. Verbeek JHAM, de Boer AG, van der Weide WE, Piirainen H, Anema JR, van Amstel R et al. Patient satisfaction with occupational health physicians, development of a questionnaire. Occupational and Environmental Medicine 2005;62(2):119-23. 27. Spelten ER, Verbeek JH and Uitterhoeve AL et al., Cancer, fatigue and the return of patients to work—a prospective cohort study. European Journal of Cancer 2003;39:562–1567. 28. Drummond MF, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programs. Oxford Medical Publications. Oxford University Press, Oxford, New York, Toronto, 1987. 29. Hlobil H, Uegaki K, Staal JB, de Bruyne MC, Smid T, van Mechelen W. Substantial sick leave costs savings due to a graded activity intervention for workers with non-specific sub-acute low back pain. European Spine Journal 2006;21. 30. Oostenbrink JB, Koopmanschap MA, Rutten FF: Standardisation of costs: the Dutch Manual for Costing in economic evaluations. Pharmacoeconomics 2002;20:443-454. 31. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. Predicting the duration of sickness absence for patients with common mental disorders in occupational health care. Scandinavian Journal of Work, Environment & Health 2006;32(1):67-74. 32. Nystuen P, Hagen KB, Herrin J. Mental health problems as a cause of long-term sick leave in the Norwegian workforce. Scandinavian Journal of Public Health 2001;29:175-182. 33. Lovibond SH, Lovibond PF. Manual for the Depression Anxiety Stress Scales (DASS). University of New South Wales, 1993.

DESIGN

80

34. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. The Depression Anxiety Stress Scales (DASS): detecting anxiety disorder and depression in employees absent from work because of mental health problems. Occupational and Environmental Medicine 2003;60:i77. 35. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica 1983;67:361–70. 36. Van der Weide WE, Verbeek JH, van Dijk FJ. Relation between indicators for quality of occupational rehabilitation of employees with low back pain. Occupational and Environmental Medicine 1999;56:488–93. 37. nQuery Advisor®. http://www.statsol.ie/nquery/nquery.htm 2005. 38. Twisk JWR, Smidt N, Vente W de. Applied analysis of recurrent events: a practical overview. Journal of Epidemiology and Community Health 2005;59:706-710. 39. Houtman I, Bosch CM, Jettinghoff K, van den Berg R. Work stress in the police force. [in Dutch]. TNO, 2000. 40. Welling L, van Harten SM, Patka P, et al. The cafe fire on New Year's Eve in Volendam, the Netherlands: description of events. Burns 2005;31(5):548-54. 41. Brouwer WB, van Exel NJ, Koopmanschap MA, Rutten FF. Productivity costs before and after absence from work: as important as common? Health Policy 2002;61:173-87.

CHAPTER 4

Guideline-based care of common mental disorders by occupational physicians

(CO-OP study): a randomized controlled trial

Published as: Rebergen DS, Bruinvels DJ, Bezemer PD, van der Beek AJ, van Mechelen W.

Guideline-based care of common mental disorders by Occupational Physicians (CO-OP-study): a randomized controlled trial.

J Occup Environ Med. 2009;51(3):305-12.

EFFECTIVENESS

82

ABSTRACT Objective To evaluate the effectiveness of guideline-based care (GBC) of workers with mental health problems, which promotes counseling by the occupational physician (OP) facilitating return to work (RTW). Methods In a randomized controlled trial with police workers on sick leave due to mental health problems (n=240), trained OPs delivered GBC in the intervention group. Time to RTW and recurrences during 1-year follow-up, analyzed using Cox proportional hazards models, were compared to usual care with easy access to a psychologist. Results GBC by OPs did not result in earlier RTW than usual care. Subgroup analysis showed a small effect in favor of GBC for workers with administrative functions and/or ‘minor’ stress-related symptoms. Conclusions GBC did not differ in RTW compared to usual care, but may be beneficial for the majority of workers with ‘minor’ stress-related disorders. Trial registration www.controlled-trials.com; Identifier: ISRCTN34887348.

EFFECTIVENESS

83

INTRODUCTION Common mental health problems or common mental disorders in workers, such as adjustment disorders, depression and anxiety, may affect functioning and often lead to reduced productivity at work and sick leave (1-4). Primary and secondary care usually focus on recovery of symptoms instead of work functioning and return to work (RTW) (5). In The Netherlands, the first points of contact are the general practitioner (GP) and the occupational physician (OP) (6,7). As workers have to visit their OP when they are on sick leave, OPs may play a central role in the treatment of workers with common mental health problems. Both GPs and OPs often lack time and skills to optimally deal with these workers, resulting in a minimal approach (6-9). When complaints persist, workers are often referred to psychologists (10). Recent Dutch and Scandinavian studies suggest that easy referral to specialized secondary care may cause some negative consequences on duration until return to work (RTW) of the worker (11-13). Specialized secondary care even more than primary care, usually focuses on recovery of symptoms instead of work-related solutions. Consequently, GPs and OPs postpone the time to facilitate RTW as workers and their professional in secondary care may hinder this process. Additionally, referrals to specialized secondary care can be expensive. As an alternative to this usual care (UC), ways have been sought to encourage OPs to play a more active role. In 2000, the Netherlands Society of Occupational Medicine (NVAB) published a practice guideline entitled ‘The management by OPs of workers with common mental health problems’ (14,15). The guideline promotes a more active role of the OP facilitating RTW of the worker, instead of a minimal role. The guideline was mainly evidence-based on the results of a study by Van der Klink et al. (16). In a cluster randomized trial, the intervention consisted of a training in an activating approach by OPs, in which OPs were enabled to operate as counselors using elements of cognitive behavioral therapy (CBT) and to facilitate RTW by work interventions. The intervention appeared to be effective in fastening RTW for workers with adjustment disorders, if compared to a passive UC. Although the guideline focuses on workers with more severe disorders as depression and anxiety as well, evidence for this relevant subgroup is scarce and the guideline more consensus-based. In a prognostic study, Nieuwenhuijsen et al. showed that guideline-based care may fasten RTW for workers on sick leave with common mental disorders, but reduces treatment satisfaction of the worker (9,11). Blonk et al. showed that for workers with work-related mental health problems a combined CBT-derived and workplace intervention delivered by occupational experts, appeared to be effective in reducing days until RTW, if compared to a CBT-intervention of psychotherapists, and UC (12). Comparable interventions showed to be (cost-) effective as well on RTW for workers with depression (17,18). In the mentioned studies, the interventions did not result in less reduction of mental health symptoms, than in UC. These results indicate that a combined work and individual intervention by an occupational expert, as proposed in the OP-guideline, could be more effective in facilitating RTW, than passive care with easy referral to secondary mental health care. The aim of this study was to examine the effectiveness of a minimal intervention, a 3-days training course in guideline-based care (GBC) by OPs, using their individual CBT-counseling and occupational expert role in facilitating RTW (19). This study focuses on the effects for workers with common mental health problems on productivity loss (duration of RTW) and treatment satisfaction. The first hypothesis was that GBC will lead to earlier RTW and consequently to less productivity loss, compared to UC. As GBC may lead to a more professional and effective treatment, the second hypothesis was that workers, employers and OPs will be more satisfied with GBC than with UC.

EFFECTIVENESS

84

METHODS Design The present study is a randomized controlled trial (RCT), in which the effect of a minimal intervention in guideline-based care (GBC) was evaluated (19). After a 3-days training course for the participating OPs, workers in the intervention group received GBC by the OP. The control group received usual care (UC), with minimal involvement of the OP and, if applicable, easy access to a psychologist in secondary care. The study was funded by the Dutch Ministry of Internal Affairs and Kingdom Relations, the Health Insurance agency for the Dutch Police force (DGVP) and the VU University Medical Center. The study design, protocol and procedures were approved by the Medical Ethics Review Committee of the VU University Medical Center. Study population and recruitment procedures This study was conducted with the cooperation of the Dutch police force, which is an organization with a relatively high incidence of sick leave due to common mental health problems (20). The participating police departments Zaanstreek-Waterland and Hollands Midden comprised a source population of 2500 police workers. Both participating police departments had contracts with the same occupational health service (OHS), i.e. Commit. Since January 2002, each worker who consulted an OP, and was still on sick leave due to mental health problems, was asked by the OP to participate in the study. After a worker had signed informed consent, the OP unsealed a study envelope containing the allocated treatment for the patient, and sent the signed informed consent to the researcher (DR). In the same consultation the worker received the baseline questionnaires and first treatment satisfaction questionnaire and was asked to return them to the researcher after completion. Recruitment procedures, and inclusion and exclusion criteria are more extensively described in an earlier publication of the study protocol (19). Randomization and blinding Randomization in two groups was done at the patient level, with six blocks of size 50. To minimize the risk of irregularities by letting OPs open their treatment concealment themselves, randomization was checked by an independent researcher (AvdB) one year after the start of and at the end of the study. Participants, employers and OPs were not blinded for the intervention. The researchers were blinded for the treatment allocation and for protocol compliance as well. Interventions Usual care Usual care consisted of minimal involvement of the OP and easy referral to a psychologist, which represents daily practice of the OHSs of the Dutch police force. Psychological treatment in secondary care was fully funded by the Health Insurance agency for the Dutch Police force (DGVP). Intervention The intervention consisted of guideline-based care (GBC) by OPs. OPs participating in the study received a 3-days training course by experienced OPs and psychologists in delivering GBC (19). The guideline is based on an activating approach, time contingent process evaluation and cognitive behavioral principles (14,15). The latter mainly concern stress inoculation training and graded activity and aim to enhance the problem-solving capacity of patients in relation to their work environment. Work-related interventions were

EFFECTIVENESS

85

proposed if the cause of the mental health problems were work-related or resulted in work-disabilities. Proposed work-related interventions were gradual RTW, regular contact with the supervisor, work accommodations, all especially instigated when there was stagnation in RTW. The OPs were encouraged to use specific tools, such as symptom-questionnaires, patient information leaflets on stress, and day structuring exercises. To mimic a realistic situation, no activities were undertaken to improve the implementation of the guideline by the OP. After follow-up, guideline adherence was examined in a process evaluation, by auditing the medical files. Risk of contamination As all participating OPs received the training course and randomization was done on patient level, OPs treated patients from both groups. The advantage was that all participants were diagnosed in the same way. However, this situation created a risk of treatment contamination between the groups. We tried to maximize the contrast by creating a situation in which referral to the psychologist in UC was always granted by the insurance company (DGVP). By this pre-authorization, OPs were instigated to refer immediate to a secondary care psychologist in UC, and initiated to deliver GBC in the intervention group. Outcomes Productivity loss The primary outcome measure in our study was productivity loss, which consisted of first RTW, full RTW and total productivity loss. First and full RTW are defined as the duration of sick leave due to mental health problems in calendar days from the moment of inclusion to first (partial or full) and full RTW respectively in own or equal earnings (21,22). Total productivity loss is the duration of sick leave days until full RTW added with number of days of recurrences on sick leave in the one-year follow-up. Sick leave data were gathered from records of the police departments, which is more accurate than from self-report (23). Treatment satisfaction Treatment satisfaction is a relevant outcome measure in occupational health care and therefore another primary outcome (11,24). Worker and employer satisfaction were measured at baseline (T1), during treatment (T2) and after full RTW (or one-year follow-up if no full RTW yet;T3) using a short version of the Patient Satisfaction with Occupational Health professionals Questionnaire (24). This questionnaire refers to different aspects of the treatment by the OP and was sent to all the participants and their supervisors (19). To measure treatment satisfaction by the OPs themselves, OPs filled in an evaluation questionnaire for each worker treated at T3. Prognostic measures Data about personal, treatment and work characteristics were gathered from the records of the police departments (e.g. type of function) and of the medical files of the OHSs (e.g. work-relatedness of the disorder according to the OP). Severity of depression, anxiety and stress was measured at baseline (T1), by using the Depression Anxiety Stress Scales (DASS-42) (25,26) and the Hospital Anxiety Depression Scale (HADS) (27). Statistical analyses All analyses were conducted according to the intention-to-treat principle and were performed at the individual level. Baseline measurements and RTW characteristics of the

EFFECTIVENESS

86

two groups were compared by using Chi-square and Independent T-tests, after imputing missing by mean values. The evaluations on the effectiveness of the intervention were performed with two-tailed tests at a significance level of 5% (P < 0.05). Two statistical models were used for the different outcomes. First, differences in the data on (first) RTW were examined by using Kaplan Meier curves, in which participants lost to follow-up were censored, and the Cox proportional hazards model. Recurrences of sick leave for any reason during follow-up were added to the Cox proportional hazards model for recurrent events with the time to event approach (28). Second, treatment satisfaction during treatment (T2) and after full RTW or one-year follow-up (T3) were studied in a linear regression model as a long term effect with treatment satisfaction at the start of the treatment (T1) as a covariate. If necessary, analyses were adjusted for prognostic dissimilarities between the baseline measurements of both groups. Gender, OP (as nested dummy variable), severity of the disorder (cut-off scores DASS-depression/anxiety) (26), work relatedness of the disorder, and type of function (executive versus administrative) were considered as a potential effect-modifier and tested on interaction effects. Statistical analyses were performed using SPSS 14.0 and Stata 8.0.

RESULTS Participants and baseline data The recruitment of participants for the study started in January 2002 and ended in January 2005, with a one year follow-up period (figure 1). Since January 2002, 489 workers were registered as being absent from work due to mental health problems by the OHS (T -1, figure 1). Of those, 240 workers who consulted an OP were asked by the OP to participate in the study and signed informed consent (T 0). Figure 1 shows that the mental health symptoms questionnaires (DASS/HADS) and the first worker satisfaction questionnaire (T1) were returned by 213 of the 240 participants (89%). The response for returning questionnaires at T1 was lower for the employer (53%). Questionnaires of T2 were not used in the analyses, nor mentioned in figure 1, as these were not sent to each participant and response was minimal. The response rate at T3 was 80% for both workers and employers, and 91% for OPs. In total 16 participants were lost to follow up. 15 because they left the police force during their RTW-process, one committed suicide. According to the intention-to-treat principle, their RTW data were censored in the analysis to include all potential information. Baseline data are shown in table 1. There were no significant differences in participant characteristics between the groups.

EFFECTIVENESS

87

Figure 1 Flow chart CO-OP study Treatment characteristics and outcomes on productivity loss Table 2 presents the differences between the groups for each of the treatment characteristics and outcomes on productivity loss. No differences were found in the number of consultations with the OP during the sick leave period; 3.3 in UC versus 3.4 in GBC. In UC 94 (82%) workers were referred by the OP to a ‘funded’ psychologist. As was intended by our study protocol, the number of referrals to a ‘funded’ psychologist were higher and faster in UC than in GBC (n=47 (38%)). During one-year follow-up, 98 workers (85%) in UC received psychological treatment, compared to 58 (46%) in GBC. In GBC significant more participants partially returned to work (69%) before full RTW, compared to UC (54%) (p=0.01). The number of recurrences of sick leave periods was higher in GBC, but this did not result in significantly more sick leave days in GBC.

Table 3 presents the median scores and the adjusted hazard ratios on the different productivity loss outcomes of both groups. No clear effects of the intervention were found. Figure 2 shows the adjusted Cox regression curves for GBC and UC on full RTW (one minus survival function at mean of covariates). Ancillary analyses on productivity loss Gender, age, number and days of sick leave periods in the previous year, type of function and severity of the disorder were added as a potential effect-modifier in the adjusted Cox proportional hazard models productivity loss outcomes. The type of function showed

T -1 Workers registered as being absent from

work due to mental health problems n=489

T 0 Inclusion in RCT during consultation OP in

which informed consent was obtained n=240

Control group n=115

Randomisation

Intervention group n=125

T1 Baseline HADS/DASS, n=112

Worker satisfaction, n=109 Employer satisfaction, n=65

T1 Baseline HADS/DASS, n=101

Worker satisfaction, n=97 Employer satisfaction, n=61

T3 Full RTW (or 1-year follow up) Worker satisfaction, n=101

Employer satisfaction n=107 Evaluation by OP, n=117

T3 Full RTW (or 1-year follow up) Worker satisfaction, n=89

Employer satisfaction, n=90 Evaluation by OP, n=101

EFFECTIVENESS

88

Table 1 Characteristics of the study population Patient characteristics GBC n=125 UC n=115 P-value Age in years, mean (sd) 38.8 (8.4) 40.0 (9.5) 0.27 Gender (% male) 51.2 60.5 0.15 Children (%) 66.4 63.2 0.68 Married or cohabiting (%) 66.4 68.4 0.78 Work related characteristics Police department (% PHM) 70.4 64.0 0.33 Mean contract hours a week, mean (sd) 33.6 (6.2) 34.1 (6.0) 0.47 Executive work (%) 59.2 65.8 0.35 Irregular work (%) 59.2 62.5 0.76 Work relatedness mental health disorder (%) 48.8 44.7 0.39 Absenteeism previous year N sick leave periods previous year, mean (sd) 2.7(2.2) 2.5 (1.9) 0.53 Days of sick leave in previous year, mean (sd) 56.9 (61.4) 56.1 (86.0) 0.94 Severity of disorder (symptoms HADS/DASS) n=112 n=101 HADS-Anxiety, mean (sd) HADS-Depression, mean (sd)

11.4 (3.8) 11.5 (4.4)

11.4 (3.8) 11.8 (4.5)

0.92 0.60

DASS-Stress, mean (sd) DASS-Anxiety, mean (sd) DASS-Depression, mean (sd)

9.2 (7.3) 4.2 (5.5) 6.9 (7.4)

9.0 (6.8) 3.9 (5.2) 6.6 (7.4)

0.91 0.69 0.78

Depression and/or Anxiety based on DASS (%)* 34.8 29.0 0.38 * Based on cut-off scores Nieuwenhuijsen et al. (2003): >12 on symptoms depression and >5 on symptoms of anxiety

Table 2 Differences for treatment and productivity loss (RTW) characteristics Treatment characteristics GBC n=125 UC n=115 P-value N consultations OP, mean (sd) 3.4 (2.3) 3.3 (2.3) 0.75 N consultations GP, mean (sd) 1.9 (1.4) 1.9 (1.2) 0.98 Referral ‘funded’ psychologist by OP, n (%) Treatment psychologist during follow-up, n (%)

47 (38) 58 (46)

94 (93) 98 (97)

0.00 0.00

Productivity loss (RTW) characteristics Return to work – process (Immediate) Full RTW (%) Partial RTW (%) Duration partial RTW, mean days (sd)

31 69 53.1 (56.3)

46 54 50.6 (78.4)

0.28

N recurrences, mean (sd) 1.7 (1.9) 1.4 (1.5) 0.08

Duration of recurrences, mean days (sd) 19.4 (39.0) 18.6 (39.1) 0.95

Table 3 Differences with Cox regression for productivity loss outcomes RTW-outcomes GBC n=125 UC n=115 Hazard Ratio P-value Partial RTW, median days (CI) 50 (34-66) 47 (31-63) 0.99 (0.75-1.31)¹ 0.94

Full RTW, median in days (CI) 105 (84-126) 104 (81-127) 0.96 (0.73-1.27)¹ 0.78

Total productivity loss, mean days (SD)

151 (97) 147 (102) 1.21 (0.86-1.71)² 0.28

¹ Adjusted HR for OP, HADS-total, children and n sick leave periods previous year ² Adjusted HR for OP, DASS-depression/anxiety, work relatedness, n sick leave periods previous year

EFFECTIVENESS

89

Figure 2 Adjusted Cox regression curves on full return to work significant interaction with the intervention on differences in full RTW (p=0.03). Workers in administrative functions seemed to benefit more from the intervention than their colleagues with executive functions (‘working on the street’). In the adjusted proportional hazards model for recurrent events a significant interaction effect (p=0.02) was found for the severity of the disorder (DASS-depression/anxiety) with the intervention on total productivity loss. When recurrent work absence periods during follow-up were taken into account, GBC seemed more effective for workers with ‘minor’ stress related symptoms when compared to UC. For this subgroup (n=172), the Hazard Ratio was 1.28 with a p-value of 0.22 (CI=0.85-1.93, adjusted for OP, sick leave year before, OHS, function). After imputing missing by mean DASS-values the subgroup of DASS-stress grew from 145 to 172 workers. This did not influence our results, as the 27 participants who did not return the DASS-questionnaires did not differ in gender, age, and RTW-data of the other participants. For workers with a depressive or anxiety state according to the DASS cut-off-scores UC seemed more effective than GBC. For this subgroup (n=68), the Hazard Ratio was 0.67 with a p-value of 0.21 (CI=0.36-1.26, adjusted for diagnosis OP (stress vs rest), OP, gender). So each of both subgroups was not significant, but the difference was. Outcomes treatment satisfaction Table 4 presents the scores on treatment satisfaction, which were recoded into mean scores on a scale of 0-10 for worker and employer satisfaction and 1-10 for the evaluation of the OP. No significant differences were found in the mean worker and employer satisfaction scores between the groups, nor univariate nor with a linear regression model adjusted for worker and employer satisfaction at T1. A significant reduction in all mean worker satisfaction scores was found with a one sample T-test between the start of the treatment (T1) and after full RTW or one-year follow-up (T3) (n=178; mean Δ T = 1.07; p=0.00). The evaluation scores of the OP about the treatment process were significantly higher in GBC compared to UC (p=0.03).

Days until full return to work

4003002001000

Pro

po

rtio

n f

ull r

etu

rn t

o w

ork

1,0

0,8

0,6

0,4

0,2

0,0

- GBC (125)

- UC (115)

EFFECTIVENESS

90

DISCUSSION Interpretations The intervention did not show an effect on productivity loss in our study sample of police workers on sick leave due to mental health problems who consulted their OP. Ancillary analyses showed a significant interaction between the treatment group and the severity of the disorder on total productivity loss, when taking into account the recurrent sick leave periods during follow-up. For the substantial group of workers with ‘minor’ mental health symptoms (according to the DASS) GBC seemed to be more effective on reducing productivity loss, while UC seemed beneficial for workers who reported more ‘severe’ symptoms of depression and anxiety. Additionally, in ancillary analyses a significant interaction was found between the treatment group and the type of function on full RTW. GBC seemed more effective for the administrative functions compared to executive functions. Treatment satisfaction rated by the OP about the treatment process was significantly in favor of GBC. Satisfaction of the worker about the treatment by the OP diminished significantly in both groups between the start of the treatment and after full RTW. Strengths and limitations In this RCT the effectiveness of a practice guideline was assessed. Although the guideline has been evaluated in a pragmatic setting, many requirements for a high quality trial were met. A representative intervention setting, in which patients were recruited over the same period of time and from the same source population, guarantee external and internal validity. Losses to follow-up and principal confounders were taken into account, and the study had sufficient power to detect a clinically important effect. This study represents a further step in the evaluation of guideline-based care for workers with common mental health problems. The trial was unable to prove the hypothesis that the intervention was more effective on RTW than usual care. However, this may be due to limitations in validity of the guideline, treatment potential of the OPs, and contamination between the intervention and the control group used.

Table 4 Treatment satisfaction worker and employer, evaluation by OP Satisfaction Worker Employer OP Time T1¹ T3² T1¹ T3² T3² Group GBC UC GBC UC GBC UC GBC UC GBC UC n 111 98 110 89 65 61 107 90 117 101 General 6.9 7.2 5.7 6.0 6.4 6.5 6.5 6.5 - - Interpersonal 7.2 7.3 6.1 6.1 6.4 6.7 6.7 6.6 - - Communication 7.0 6.9 6.1 6.3 6.4 6.7 6.7 6.6 - - Knowledge 7.0 7.2 6.0 6.2 6.6 6.7 6.7 6.7 - - Total 7.0 7.2 6.0 6.2 6.5 6.7 6.6 6.7 - - Efficacy - - - - - - - - 7.1 6.7 Process - - - - - - - - 7.6 7.0 ¹ T1 = Treatment satisfaction after the moment of inclusion (start of intervention) ² T3 = Treatment satisfaction after the moment of full RTW of the worker (end of treatment) or after one-year follow-up

EFFECTIVENESS

91

Validity of the guideline The inclusion criteria used in our study might have been too broad, leading to a heterogeneous study sample. Possibly, workers with either almost none or with many, more severe symptoms may not have been sensitive to the intervention. The guideline is mainly based on evidence of a study showing effect on adjustment disorders (16). Therefore, the guideline may not have sufficient external validity to be applied to broad range of symptoms as advocated. A minor, but substantial group of workers, may have had symptoms of a more chronic nature, who needed more extensive care as was given in GBC. As our intention was to study the effects of the complete guideline in practice, we believe that our results give relevant insight in the overall effect of the guideline for a broad study sample. Occupational physicians (OPs) The lack of effect may be due to the treatment potential of the OPs as well. They are the gatekeepers of occupational health care and have extensive workloads. The guideline promotes an intensive form of treatment, which may have been a barrier to implement the guideline in their daily practice (8). Alternatively, the training in the guideline might have been too minimal, or the training hours too short for the OPs to actually learn the necessary skills. However, Smits et al. showed that a problem-based learning program for this guideline, which corresponds with the training in our study, appeared to be effective in improving performance in postgraduate education of OPs (29). Contamination between intervention and usual care As randomization was done at the individual level, OPs who were trained in the guideline treated all participants. Obviously this situation created a risk of treatment contamination between the groups. A cross-over learning effect may have happened in UC, since the OP can deliver GBC in UC as well. In the GBC group, the OP may have referred a worker to a psychologist as well. The guideline promotes this in case of stagnation in recovery or in case of severe mental health problems of the worker (14,15). However, we tried to maximize the contrast by creating a situation in which referral to the psychologist in UC was always granted and pre-authorized by the health insurance company (DGVP). The results indicate that there is a lack of contrast as a result of contamination, which may have negatively influenced the impact of the intervention which was evaluated. Generalization The study population, Dutch police workers, has a higher risk of getting into stressful situations than other workers (20,30). To a certain extent this reflects that police workers have other occupational risks than the general working population. As the study population will not be fully representative of the general working population, external validity of study results may be limited and caution has to be taken in generalizing the results. This disadvantage does not outweigh the advantages of this study population. The police is an interesting target population, as it has a relatively high incidence of common mental health problems. Other advantages are a uniform sick leave registration and a well-defined ‘usual care’. Overall evidence This study is the first RCT to evaluate the effectiveness of GBC in primary and occupational health care for mental health problems with productivity loss as a primary outcome. Earlier publications about a minimal intervention strategy to enhance treatment in primary care of patients with common mental health problems did not result in a

EFFECTIVENESS

92

reduction of productivity loss (31-33). Recent Cochrane reviews of RTW interventions on adjustment disorders and depression show that a combined intervention on individual and work aspects, conducted in an occupational health care setting, is effective in fastening RTW (34,35). A minimal combined intervention on individual and work aspects, represented in this study by GBC of OPs, seems promising for workers with ‘minor’ stress-related disorders. Workers with ‘more severe’ symptoms seem to benefit from early detection and a fast and accurate way to get sufficient mental health care. Wang et al. showed that this reduces productivity loss of depressive workers in a company setting in the United States (36). More studies are needed, especially in other countries than The Netherlands, to confirm the effectiveness of (preventive) combined individual and work-related interventions in guideline-based care. Conclusions Counseling of common mental health problems by OPs using a guideline after a 3-days training course did not clearly differ in reducing productivity loss and treatment satisfaction compared to usual care. However, workers with ‘minor’ stress-related disorders may benefit from guideline-based care. Results of this study contribute to the further development of effective evidence-based guidelines and collaborative occupational health care for workers with common mental disorders. ABBREVIATIONS CI = Confidence Interval CBT = Cognitive Behavioral Therapy DASS = Depression Anxiety Stress Scale DGVP = Health Insurance agency for the Dutch Police force GBC = Guideline-based care (intervention group) GP = General practitioner HADS = Hospital Anxiety Depression Scale NVAB = Netherlands Society of Occupational Medicine OHS = Occupational Health Service OP = Occupational Physician RTW = Return to work UC = Usual care (control group) REFERENCES 1. Wang JL, Adair CE, Patten SB. Mental health and related disability among workers: A

population-based study. Am J Ind Med. 2006;49:514-522. 2. Lerner D, Henke RM. What Does Research Tell Us About Depression, Job Performance, and

Work Productivity? J Occup Environ Med. 2008;50(4):401–410. 3. National Mental Health Association USA. Mental health facts. www.mentalhealthamerica.net

(formerly known as the National Mental Health Association) 2007. 4. NKAP [Dutch Knowledge Center of Work and Mind]. Factsheet 2, Facts and figures about work

incapacity in The Netherlands due to mental health disorders. [in Dutch] The Netherlands, Utrecht: NKAP; 2004.

5. van der Klink JJ, Blonk RW, Schene AH, van Dijk FJ. The benefits of interventions for work-related stress. Am J Public Health. 2001;91(2):270-6.

EFFECTIVENESS

93

6. Buijs P, Anema H, Evers M, Van Dijk F, van der Klink J. How general practitioners can manage work-related psychological complaints. Design and pilot of a guideline: a contribution towards solving a huge problem. Eur J Gen Pract. 2006;12(3):138-41.

7. Anema JR, Jettinghoff K, Houtman ILD, Schoemaker CG, Buijs PC, van den Berg R. Medical care of employees long-term sick listed due to mental health problems: A cohort study to describe and compare the care of the occupational physician and the general practitioner. J Occup Rehabil. 2006;16:41-52.

8. Rebergen DS, Hoenen JAHJ, Heinemans AMEC, Bruinvels DJ, Bakker AB, van Mechelen W. Adherence to a national guideline on mental health problems by Dutch occupational physicians. Occup Med (Lond). 2006;56:461-468.

9. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. Validation of performance indicators for rehabilitation of workers with mental health problems. Med Care. 2005;43(10):1034-42.

10. Houtman ILD, Schoemaker CG, Blatter BM, de Vroome EMM, van den Berg R, Bijl RV. Mental health problems, interventions, and work resumption: The prognoses Study INVENT [In Dutch]. TNO Quality of Life. The Netherlands, Hoofddorp: 2002.

11. Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. Occup Environ Med. 2003;60:21-25.

12. Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Treatment of Work-Related Psychological Complaints: a Randomized Field Experiment among Self-employed. Work & Stress. 2006;20(2):129-144.

13. Nystuen P, Hagen KB, Herrin J. Solution-focused intervention for sick listed employees with psychological problems or muscle skeletal pain: a randomised controlled trial. BMC Public Health. 2006;6:69.

14. NVAB [Netherlands Society of Occupational Medicine]. Van der Klink J, ed. Guideline for the management of mental health problems by occupational physicians. [in Dutch] Eindhoven: NVAB; 2000.

15. van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scand J Work Environ Health. 2003;29:478-487.

16. van der Klink JJL, Blonk RW, Schene AH, Dijk FJ van. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomized controlled design. Occup Environ Med. 2003;60(6):429-37.

17. de Vries G, Kikkert MJ, Schene AH et al. Does occupational therapy work for patients with depression? [In Dutch] Nederlands Tijdschrift voor Ergotherapie 2003;31:103-8.

18. Schene AH, Koeter MWJ, Kikkert MJ, Swinkels JA, McCrone P. Adjuvant occupational therapy for work-related major depression works: randomised trial including economic evaluation. Psychol Med. 2007;37(3):351-62.

19. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W. Design of a randomized controlled trial on the effects of Counseling of mental health problems by Occupational Physicians on return to work: the CO-OP-study. BMC Public Health. 2007;7(147):183.

20. Houtman I, Jettinghoff K, Brenninkmeijer V, van den Berg R. Work stress in the police force five years later: the effect of sectored agreements on (stress) management. [in Dutch]. TNO Quality of Life. The Netherlands, Hoofddorp; 2005.

21. Young AE, Roessler RT, Wasiak R, McPherson KM, van Poppel MNM, Anema JR. A developmental conceptualization of Return to Work. J Occup Rehabil. 2005;15(4):557-68.

22. de Vet HC, Heymans MW, Dunn KM, et al. Episodes of low back pain: a proposal for uniform definitions to be used in research. Spine. 2002;27(21):2409-16.

EFFECTIVENESS

94

23. van Poppel MN, de Vet HC, Koes BW, Smid T, Bouter LM. Measuring sick leave: a comparison of self-reported data on sick leave and data from company records. Occup Med (Lond). 2002;52:485-90.

24. Verbeek JHAM, de Boer AG, van der Weide WE, et al. Patient satisfaction with occupational health physicians, development of a questionnaire. Occup Environ Med. 2005;62(2):119-23.

25. Lovibond SH, Lovibond PF. Manual for the Depression Anxiety Stress Scales (DASS). University of New South Wales; 1993.

26. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. The Depression Anxiety Stress Scales (DASS): detecting anxiety disorder and depression in employees absent from work because of mental health problems. Occup Environ Med. 2003;60:i77.

27. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983;67:361-70.

28. Twisk JWR, Smidt N, Vente W de. Applied analysis of recurrent events: a practical overview. J Epidemiol Community Health. 2005;59:706-710.

29. Smits PB, de Buisonjé CD, Verbeek JH, van Dijk FJ, Metz JC, ten Cate OJ. Problem-based learning versus lecture-based learning in postgraduate medical education. Scand J Work Environ Health. 2003;29(4):280-7.

30. Slottje P, Smidt N, Twisk JW, et al. Use of health care and drugs by police officers 8.5 years after the air disaster in Amsterdam. Eur J Public Health. 2008;18(1):92-4.

31. Leone SS, Huibers MJ, Kant I, et al. Long-term efficacy of cognitive-behavioral therapy by general practitioners for fatigue: a 4-year follow-up study. J Psychosom Res. 2006;61(5):601-7.

32. Brouwers EPM, Tiemens BG, Terluin B, Verhaak PFM. Effectiveness of an intervention to reduce sickness absence in patients with emotional distress or minor mental disorders: a randomized controlled effectiveness trial. Gen Hosp Psychiatry. 2006;28:223-229.

33. Bakker IM, Terluin B, van Marwijk HW, et al. A cluster-randomised trial evaluating an intervention for patients with stress-related mental disorders and sick leave in primary care. PLoS Clin Trials. 2007;2(6):e26.

34. Bruinvels DJ, Rebergen DS, Verbeek J, Nieuwenhuijsen K, Madan I, Neumeyer-Gromen A. Return to work interventions for adjustment disorders. (Protocol) Cochrane Database Syst Rev. 2007, Issue 1. Art. No.: CD006389. DOI: 10.1002/14651858.CD006389.

35. Nieuwenhuijsen K, Bültmann U, Neumeyer-Gromen A, Verhoeven AC, Verbeek JH, van der Feltz-Cornelis CM. Interventions to improve occupational health in depressed people. Cochrane Database Syst Rev. 2008;16(2):CD006237.

36. Wang PS, Simon GE, Avorn J, et al. Telephone screening, outreach, and care management for depressed workers and impact on clinical and work productivity outcomes: a randomized controlled trial. JAMA. 2007 26;298(12):1401-11.

CHAPTER 5

Cost-effectiveness of guideline-based care for workers with mental health problems

Published as: Rebergen DS, Bruinvels DJ, van Tulder MW, van der Beek AJ, van Mechelen W.

Cost-effectiveness of guideline-based care for workers with mental health problems. J Occup Environ Med. 2009;51(3):313-22.

COST-EFFECTIVENESS

96

ABSTRACT Objective To evaluate the cost-effectiveness of an activating guideline-based care by occupational physicians (OPs). Methods An economic evaluation was conducted in a randomized controlled trial with police workers on sick leave due to mental health problems (n=240). In the intervention group trained OPs provided guideline-based care, compared to usual care with easy access to a psychologist. Sick leave data and health care costs were gathered after one-year follow-up. Analyses comprised bootstrap techniques, cost-effectiveness planes and acceptability curves. Results Health care utilization costs (€574.532 in total) were significantly lower in the intervention group (mean difference -€520; 95% CI: -€980, -€59), while there were no significant differences in days of sick leave and productivity loss costs. Conclusions Guideline-based care could be cost-effective, as lower direct costs lead to equal treatment outcomes of workers with common mental disorders. Trial registration www.controlled-trials.com; Identifier: ISRCTN34887348.

COST-EFFECTIVENESS

97

INTRODUCTION High social and economic costs of poor mental health have led to growing recognition of the need to develop effective interventions, especially in primary and occupational health care (1,2). Across the U.S.A. and the European Union evidence has grown that mental health problems increasingly lead to sick leave, early retirement and productivity loss (3-5). In the Netherlands, one-third of all long-term sick leave or work disability cases are attributable to common mental health problems (6,7). Although several trials have been conducted on the clinical effectiveness of interventions on mental health problems, economic evaluations are scarce and results hardly usable in mental health policy (8,9). In the Netherlands, the first points of contact for workers with mental health problems, are the general practitioner (GP) and the occupational physician (OP) (10). However, both GPs and OPs often lack time and skills to optimally deal with these workers, resulting in a minimal approach (11,12). When after a few weeks complaints persist, GPs and OPs often refer workers to psychologists. This ‘referral delay’ is often associated with delayed return to work. In addition, specialized mental health care usually focus on symptoms instead of work, which may hinder early return to work (12-14). Consequently, workers may not get the optimal care they need, and productivity loss and treatment costs may be higher than necessary. As an alternative to usual care, the Netherlands Society of Occupational Medicine (NVAB) has published a guideline on ‘The management by OPs of workers with mental health problems’ in 2000 (15,16). This guideline promotes a more active role of the OP as counselor facilitating return to work of the worker. This intervention appeared to be effective in an in-company setting in fastening RTW for workers with adjustment disorders, if compared to a passive UC (17).

To evaluate the effectiveness of guideline-based care (GBC) by OPs compared to usual care (UC) for a broad range of workers with common mental health problems, a randomized controlled trial with a 12-month follow-up was conducted in an occupational health care setting (18). GBC was not more effective than UC in reducing productivity loss, but may result in economic benefits. The objective of this study was to conduct an economic evaluation of GBC in reducing productivity loss costs, from both a societal and a company perspective, compared to UC. The hypothesis was that a reduction in health care utilization could be achieved, as a consequence of GBC with fewer referrals to psychologists and increased self-management, compared to UC. METHODS Design and hypothesis An economic evaluation from a societal and a company perspective was conducted within a randomized controlled trial. Subjects in the intervention group received guideline-based care (GBC) by OPs. The control group received usual care (UC), with minimal involvement of the OP and easy access to a psychologist. The study was funded by the Dutch Ministry of Internal Affairs and Kingdom Relations, and the Health Insurance of the Dutch police (DGVP). The study design, protocol and procedures were approved by the Medical Ethics Review Committee of the VU University Medical Center. Details of the study design and the guideline have been reported elsewhere more extensively, but a brief summary will be presented below (18).

COST-EFFECTIVENESS

98

Study population This randomized controlled trial was conducted with the cooperation of two departments of the Dutch police force, comprising a source population of 2500 police workers. The Dutch police force is an organization with a relatively high incidence of stress-related sick leave (19). Recruitment of police workers on sick leave due to common mental disorders started in 2002. Police workers were included by the OP when they visited the occupational health service (OHS). Randomization and blinding Randomization into two groups was achieved by random number generation, balancing after every 50 participants. Participants, employers and OPs were not blinded for the intervention. The researchers were blinded for both the treatment allocation and protocol compliance. Usual care UC consisted of minimal involvement of the OP and easy access to counseling by a psychologist. Counseling was fully funded by the health insurance company of the police (DGVP). Intervention GBC in the intervention group consisted of treatment by OPs according to the NVAB guideline of workers with mental health problems (18). Before the study started, OPs participated in a three-day course. The course focused on an early start of the intervention by OPs, in which they operate as an activating counselor using cognitive behavioral elements aiming to enhance the problem-solving capacity of workers, especially in relation to their work environment (17). Economic evaluation An economic evaluation was performed from both societal and employers’ perspectives (20). From a societal perspective, a cost-effectiveness analysis (CEA) was conducted combining differences in costs of the intervention and other health care, with days of sick leave until full return to work. Additionally, a cost-benefit analysis (CBA) from the perspective of the employer was performed to compare the costs of the intervention and other health care with the monetary benefits of reduced productivity loss. Health care costs Health care costs of participants were collected over a 12-month period, starting from the date of inclusion. As mental health problems may affect physical functioning in addition to mental functioning, all health care related costs were collected during follow-up. Healthcare costs related to the use of primary, psychological and hospital care were extracted from the computerized records of the insurance company of the Dutch police force (DGVP). Data on the use of occupational healthcare were extracted from the computerized medical records of the OHSs of the police departments. The costs associated with occupational health care were based on the contract tariffs of the OHSs and costs found in the literature (21,22). Costs related to counseling by ‘private’ psychologists and social workers were estimated and valuated using cost prices recommended in the Dutch manual for costing (23-25). The cost index of 2003 was used. The training costs of the OPs in GBC were determined bottom-up and compared to the real costs of the training (supplement 1). This resulted in an average training cost of € 40 per participant for the GBC group. Healthcare costs from the perspective of the employer

COST-EFFECTIVENESS

99

comprised the costs of occupational health care and psychological care paid by the employer and DGVP. Productivity loss The primary outcome measure described in this study was sick leave, which is used as a proxy measure of productivity loss. The cumulative number of sick leave days during the one year follow-up was used in the CEA. Sick leave was defined as the duration in calendar days of work absence to partial or full return to work in own or equal earnings, preceded and followed by a period of at least 1 day at work (26,27). The outcome was expressed in two ways: gross and net sick leave (22). Gross sick leave was defined as the total number of calendar days that subjects were completely or partially on sick leave. In cases of net sick leave, it was assumed that subjects were productive during the hours of partial sick leave. Sick leave data were gathered from records of the police departments after one year follow-up, which is more accurate than from self-report (28).

Costs associated with productivity loss were estimated from a company perspective by both the Friction Cost Method (FCM) and Human Capital Approach (HCA) (20,24). According to the FCM, productivity losses are limited to the time needed to restore production back to its initial level. This means that the Friction Cost Period is the estimated time it takes a company to find replacement of the sick worker (7,24). Productivity loss costs according to the FCM were estimated by multiplying an average daily wage by the gross number of sick leave days, up to a standard Friction Cost Period maximum of 154 days (24). Productivity loss costs according to the HCA were estimated by multiplying an average daily wage by the total gross number of sick leave, using the entire period of sick leave. Baseline measures Data regarding personal, treatment and work characteristics were gathered from the records of the police departments and of the medical files of the OHSs. Severity of depression, anxiety and stress were measured at baseline using the Depression Anxiety Stress Scales (DASS-42) and the Hospital Anxiety Depression Scale (HADS) (29-32). Data analysis The economic evaluation was performed according to the intention-to-treat principle. No imputation was done, as less than 10% of the study population was lost to follow-up and as lost to follow-up data were equally spread over both groups. For all analyses 95% confidence intervals (CIs) of the difference in mean costs in both groups were computed by bias corrected and accelerated (Bca) bootstrapping with 2000 replications (33).

From a societal perspective, a cost effectiveness analysis (CEA) was conducted. To assess the cost-effectiveness of GBC, the difference in mean costs (ΔC: incremental costs) between the groups was divided by the difference in mean effects (ΔE: incremental effects). In the CEA ΔC included the difference in the total direct healthcare costs, which was divided by ΔE with net sick leave as effect measure. To avoid double-counting, productivity loss costs due to sick leave were not included in the costs, as the difference in sick leave was the effect measure.

From the company perspective a cost benefit analysis (CBA) was performed, using the net benefit framework (34). The mean net monetary benefit (NMB) of GBC compared with UC was estimated, using the following equation: NMB = ΔE*λ – ΔC > 0 (34). ΔE was the mean difference in total net sick leave days between GBC and UC, using the Human Capital approach. ΔC was the mean difference in direct healthcare costs for the company.

COST-EFFECTIVENESS

100

Lambda (λ) represented the mean salary per day of our population, based on a 36 hours working week per person.

For both CEA and CBA, the 95% CIs around the incremental cost-effectiveness ratios (ICERs) and the NMB were computed by a Bca-bootstrapping procedure with 5000 replications (33). Uncertainty around the pooled mean ICERs and NMB was estimated by 95% confidence intervals, using Brigg’s model (35). Bootstrapped cost-effect pairs were plotted on a cost-effectiveness plane and acceptability curves were generated (36). Sensitivity analysis The effect of partial return to work on productivity was determined with a sensitivity analysis on the CEA. Therefore, the CEA was repeated with the same ΔC, but a different ΔE. Under the assumption that workers who partially resumed work were completely productive (instead of partially productive), ΔC was divided by ΔE with gross sick leave as effect measure.

To assess the robustness of the CBA, two additional sensitivity analyses were conducted. First, an estimation was made of the impact of using total costs that contained productivity loss costs calculated by the FCM, instead of the HCA. Second, the CBA was performed with another λ representing the society’s maximum willingness-to-pay for a day less of sick leave (37). This new λ was selected to be € 255, which equals the Dutch average cost price for a day of work (24), and replaced the mean salary per day of our study population. RESULTS Participants Between January 2002 and January 2005, 489 workers were registered by the OHS as being absent from work due to mental health problems (figure 1). Of those, 240 signed an informed consent and were randomized. A total of 125 subjects was assigned to GBC and 115 to UC. Baseline data are presented in table 1. There were no significant differences in participant characteristics between the groups. In total, 16 subjects were lost to follow up; 15 because they left the police force during follow-up and one because of death. Mean salary per day was € 125. Health care costs The mean costs of health care utilization and productivity loss are presented in table 2 with standard deviations (SD) per group and mean cost differences, in Euros, over the 12-month follow-up, based on complete cases. Resource use and valuation are shown in supplements 2 and 3. Main cost driver is psychological care, followed by primary care, occupational health care, and hospital care. Total health care costs were significantly higher in UC compared to GBC. The mean costs of psychological care were less in GBC, while the mean costs associated with occupational health care were lower in the UC. This was mainly caused by the intervention training costs of OPs and a higher utilization rate of guidance by a company social worker. Mean costs of primary care were higher in GBC, as there was a higher utilization rate of medication and multidisciplinary physical therapy.

Health care costs from the employers’ perspective were significantly higher in UC compared to GBC, due to higher costs of psychological care.

COST-EFFECTIVENESS

101

Figure 1 Flow chart CO-OP study Table 1 Characteristics of the study population Patient characteristics GBC (n=125) UC (n=115) Age in years, mean (sd) 38.8 (8.4) 40.0 (9.5) Gender (% male) 51.2 60.5 Children (%) 66.4 63.2 Married or cohabiting (%) 66.4 68.4 Work related characteristics Police department (% PHM) 70.4 64.0 Mean contract hours a week, mean (sd) 33.6 (6.2) 34.1 (6.0) Executive work (%) 59.2 65.8 Irregular work (%) 59.2 62.5 Work relatedness mental health disorder (%) 48.8 44.7 Absenteeism previous year Days of sick leave in previous year, mean (sd) 56.9 (61.4) 56.1 (86.1) Severity of disorder (symptoms HADS/DASS) n=112 n=101 HADS-Anxiety, mean (sd) HADS-Depression, mean (sd)

11.4 (3.8) 11.5 (4.4)

11.4 (3.8) 11.8 (4.5)

DASS-Stress, mean (sd) DASS-Anxiety, mean (sd) DASS-Depression, mean (sd)

9.2 (7.3) 4.16 (5.5) 6.9 (7.4)

9.0 (6.8) 3.9 (5.2) 6.6 (7.4)

Depression and/or Anxiety based on DASS (%)* 34.8 29.0 * Cut-off scores (36): >12 on symptoms depression and >5 on symptoms of anxiety

Workers registered as being absent from work due to mental health problems n=489

Inclusion in RCT during consultation OP in which informed consent was obtained n=240

Control group n=115

Randomization

Intervention group n=125

One year follow-up n=125

Productivity loss (sick leave days) Health care utilization


Productivity loss (sick leave days) Health care utilization

COST-EFFECTIVENESS

102

Productivity loss No significant differences between the groups were found in mean sick leave days between the groups: Net sick leave HCA (GBC: 113 (Standard deviation (SD) = 83); UC: 114 (SD = 87)); Gross sick leave HCA (GBC: 151 (SD = 97); UC: 146 (SD = 103)). Productivity loss costs were found to be, depending on the way they were measured, 6 to 9 times higher than the total health care costs. No significant differences were found between the groups for any of the four aggregated total costs associated with productivity loss. Cost-effectiveness and cost-benefit analyses Table 3 shows the cost and effect differences and incremental cost and effect ratios (ICERs) from the main cost-effect analysis (CEA) and corresponding sensitivity analysis (SA1), and from the main cost-benefit analysis (CBA) and corresponding sensitivity analyses (SA2, SA3). The cost-effectiveness plane (CE-plane) from the meain CEA is shown in Figure 2 representing the uncertainty around the ICER for the mean difference in total costs divided by the mean difference in net sick leave (HCA) for the total group. The ICER was –736, meaning that the costs of GBC were lower (table 3). Figure 3 shows that regardless the amount one is willing to pay per day of sick leave, the probability that the intervention was cost-effective did not exceed 50%.

Table 2 Total and mean component costs Component costs of resource use

Total costs

Mean costs (SD) Mean cost difference (95% CI)

GBC (N = 125)

UC (N = 115)

Primary care 173692 810 (1253) 630 (814) 180 (-88; 410) General practitioner 31068 127 (123) 132 (130) -5 (-35; 26) Diagnostic tests 20862 83 (184) 91 (199) -8 (-52; 44) Medications 56806 249 (677) 223 (382) 26 (-103; 153) Allied health professionals 23079 93 (176) 99 (213) -6 (-57; 47) Multidisciplinary therapy 41878 258 (880) 84 (461) 173 (-5; 313) Occupational health care 124900 552 (280) 486 (277) 66 (2; 151) Occupational physician 73500 310 (198) 303 (196) 7 (-47; 53) Intervention training costs 5000 40 (--) -- -40 (- ; -) Case manager 19472 81 (79) 81(73) -1 (-20; 18) Company social worker 26928 122 (136) 102 (132) 20 (-14; 55) Psychological treatment 208654 534 (929) 1233 (846) -698 (-987; -420) Psychologist (paid by employer) 196870 463 (903) 1208 (848) -745

(-1027; -461) Psychologist (paid by community) 11784 72 (228) 25 (135) 47 (-4; 88) Hospital 67286 248 (841) 316 (1011) -68 (-283; 188) Total health care costs

574532 2145 (2037) 2664 (1592) - 520 (-980; 17)

Productivity loss gross HCA † 4442152 18801 (12112) 18192 (12797)

609 (-2579; 3638)

Productivity loss net HCA † 3397555 14114 (10306) 14202 (10890)

-88 (-3020; 2404)

Productivity loss net FCM * 2784249 11691 (6448)

11503 (6472)

188 (-1402; 1987)

† HCA = Human Capital Approach * FCM = Friction Cost Method

COST-EFFECTIVENESS

103

Results of the cost-benefit analyses showed that the mean cost difference was still in favor of GBC, although the difference between both groups became smaller. According to the CBA, the estimated NMB of GBC in terms of reducing productivity loss costs, for the measured mean salary per day (λ) of € 125, was € 3,582.

Sensitivity analyses The results of the CEA sensitivity analysis involving gross sick leave (SA 1) showed that the mean effect difference became negative, causing a reversal of the direction of the mean ICER from quadrant II into quadrant III (table 3). The first sensitivity analysis of the CBA involving productivity loss measured by FCM (SA 2) indicated that there was a reversal in the direction of the mean effect difference but no change in mean difference in treatment costs, whereby a shift in the mean ICER from quadrant II to III occurred. The second sensitivity analyses of the CBA (SA3) with a different mean salary per day (λ) showed a positive effect again, as this was measured by the HCA approach. With respect to the CEA and the CBA analyses, the results based on the alternative calculation of sick leave days and productivity loss costs changed the mean effect differences from positive to negative. As these mean effect differences remained small, while the mean cost differences stayed significant in benefit of the intervention, the overall results were similar to the main analyses.

* In the CEA analysis, ΔE = mean difference in sick leave days ‘net’ calculated under the assumption that subjects who partially resume work during a sick leave period are 100% productive during those hours, ΔC = mean difference in total treatment costs; in SA1, ΔE = mean difference in sick leave days ‘gross’ calculated under an alternative assumption that subjects who partially resume work are completely unproductive during those hours. In the CBA analysis, ΔE = mean difference in productivity loss costs (based on sick leave days ‘net’) estimated by the HCA, ΔC = mean difference in total costs from the employer’s perspective; in SA2, ΔE = mean difference in productivity loss costs (‘based on sick leave days ‘net’) estimated by the FCM; in SA3, ΔE = mean difference in productivity loss costs (‘based on sick leave days ‘net’) estimated by the HCA with NMB = λ * ΔE – ΔC, where λ = € 255. † Refers to the northeast quadrant of the CE-plane, which indicates that the intervention is more effective and more costly than UC. ‡ Refers to the southeast quadrant of the CE-plane, which indicates that the intervention is more effective and less costly than UC. § Refers to the southwest quadrant of the CE-plane, which indicates that the intervention is less effective and less costly than UC. ** Refers to the northwest quadrant of the CE-plane, which indicates that the intervention is less effective and more costly than UC.

Table 3 Mean cost and effect differences (ΔC and ΔE), incremental cost-effect ratios (ICERs), and CE-plane distributions

Analysis N ΔC (95% CI) ΔE (95% CI) ICER Distribution CE-plane

CEA GBC UC Euros Days I † II ‡ III § IV*

Main (net) 125 115 -520(-980; -59) 1(-21;22) -736 1% 52% 46% 0%

SA1 (gross) 125 115 -520(-980; -59) -5 (-40; 30) 106 0% 28% 70% 2%

CBA Euros Euros

Main (HCA) 125 115 -219 (-385; -54) 88 (-2600; 2776) -2.49 1% 53% 6% 0%

SA2 (FCM) 125 115 -219 (-385; -54) -188 (-1824; 1449) 1.17 0% 36% 64% 1%

SA3 125 115 -219 (-385; -54) 174 (-5124; 5471) -1.26 1% 53% 64% 0%

COST-EFFECTIVENESS

104

Figure 2 Cost-effectiveness plane Figure 3 Cost-effectiveness acceptability curve of the CEA (net sick leave HCA) DISCUSSION Interpretations From an economical point of view, activating counseling of workers with common mental health problems by OPs using guideline based care (GBC) may be preferable to usual care (UC) with easy access to a psychologist. These results suggest that society, especially

-1500

-1250

-1000

-750

-500

-250

0

250

500

-50 -40 -30 -20 -10 0 10 20 30 40 50

Difference in sick leave

Inc

rem

en

tal co

sts

(E

uro

)

0 % 1 %

52 %46 %

more effective/ more

expensive

more effective/ less

expensive

less effective/ less

expensive

less effective/ more

expensive

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1

0 1000 2000 3000 4000 5000

Value of ceiling ratio (EURO)

Pro

ba

bil

ity

co

st-

eff

ec

tiv

e

COST-EFFECTIVENESS

105

employers, would experience a positive return on investment from a minimal, evidence-based, intervention in occupational healthcare.

In this pragmatic trial, effectiveness instead of efficacy has been studied. The pragmatic study design with broad inclusion criteria allowed variation in context, diagnosis and treatment. As this was reflective of clinical practice in the occupational health care setting, external validity of our results was enhanced. However, caution is advised because these findings may be attributable to our study population, unrealistic contrast between the treatment groups or methodological issues. Study population The guideline is mainly based on evidence of a study showing effect on adjustment disorders (17). Therefore, the aim of the guideline is to activate the OP especially for this subgroup. However, the study population, e.g. Dutch police workers, has a higher risk of getting into stressful situations than many other workers (19). This is caused by a relatively high workload and emotional pressure, and to a certain extent this reflects that police workers have other occupational risks than the general working population. As the study population will not be fully representative of the general working population, external validity of study results may be limited and caution has to be taken in generalizing the results. Contrast between the groups As randomization was done at the individual level, OPs who were trained in the guideline treated the participants of the intervention and control group. Obviously, this situation created a risk of treatment contamination between the groups. However, we tried to maximize the contrast by creating a situation in which referral to the psychologist in UC was always granted by the health insurance company (DGVP). These psychologists may not have been fully representative for UC as they were member of a multidisciplinary rehabilitation center with a focus on return to work.

Attribution of the cost effectiveness of GBC to the intervention is questionable. Although referral patterns to psychological care in our treatment groups differ significantly, treatment in the GBC by OPs did not differ from UC in terms of consultations with workers, employers and curative care. In GBC there were more referrals to company social workers, in UC there were more referrals to psychologists. These results point out that the training provided to the OPs may not have led to another treatment pattern, compared to UC. The training may have been insufficient for learning of GBC, or may have resulted in GBC in UC as well. Methodological considerations Results of this study may have been influenced by three methodical issues. First, incremental costs by training OPs in the GBC group appeared to be lower than the costs of referrals to the psychologist in UC. Training costs were hard to assess, but will not have been much higher, as these costs reflect a minimal intervention. Costs of referrals to psychologists in UC may have been an overestimation if compared to real circumstances, as referral patterns were higher than ‘usual’ and psychological care by the participating provider was rather expensive. This may have resulted in lower costs of the GBC compared to UC. Second, data from patient and family health care costs were not included. This may have led to an over- or underestimation of the total health care costs. As these costs are regarded to contain only a small part of the total costs (7), bias of our overall findings is not likely.

COST-EFFECTIVENESS

106

Third, by using sick leave days as proxy for productivity loss, we did not take into account effects on presenteeism. Mental health problems can influence these aspects of work productivity, as work performance may be suboptimal before and after periods of sick leave (38-41). GBC and UC may have had a differential effect on these aspects of productivity loss. As this may have biased our findings, future research should focus on such productivity measures as well. Overall evidence Economic evaluations on the cost-effectiveness of interventions on mental health problems, are scarce, especially cost-benefit analyses (8,9). Most economic evaluations on common mental disorders have been conducted alongside trials in primary care settings and do not seem promising (42,43). An activating intervention by social workers compared to UC by GPs in primary care patients with minor mental disorders on sick leave, was not cost-effective compared with GP routine care (25). A recent study found that a minimal intervention for workers with stress-related sick leave was not associated with superior clinical or economic impact than usual GP care (7,44). Although some types of healthcare utilization may be reduced, counseling in a primary setting does not seem to reduce overall healthcare costs compared to regular GP care (45,46).

Although guideline-based depression care has proven to be effective, employers have been slow to adopt evidence-based recommendations as they lack evidence for cost-effectiveness from their perspective (45). Wang et al. (39) developed and evaluated a systematic program to identify depression on the work floor and promote effective treatment of depression by providing psychotherapy. The program appeared to improve not only clinical, but also work productivity outcomes. In occupational healthcare, economic evaluations are a new challenge, and are conducted increasingly alongside trials (47-49). Until now, occupational healthcare interventions have been focusing on physical problems as low back pain (21,22,50). Recently, Schene et al. (51) showed that addition of occupational therapy to good clinical practice improves productivity without increasing work stress and is superior to usual care in terms of cost-effectiveness. Additionally, Taimela et al. (52) showed that an occupational health intervention for workers with high risk of sickness absence is a cost effective use of healthcare resources. These results confirm our findings that a minimal intervention in an occupational health care setting, in contrast to a primary care setting, has more potential to be cost-effective by enhancing productivity. This is the first randomized controlled trial that proves that guideline-based care of workers with mental health problems could be cost-effective in an occupational health care setting. Further research should focus on the development of cost-conscious guidelines and improving the efficiency of implementation strategies, including monitoring, of guideline-based care (53-56). Conclusion Guideline-based care of workers with common mental disorders by an OP proves to be economically in favor compared to usual care with easy access to a psychologist, as lower direct costs lead to equal treatment outcomes. These results suggest that from both society and employers perspective, guideline-based care in an occupational health care setting could be cost-effective.

COST-EFFECTIVENESS

107

ABBREVIATIONS Bca = Bias corrected and accelerated bootstrapping CI = Confidence Interval CBA = Cost-benefit analysis CEA = Cost-effectiveness analysis CE-plane = Cost-effectiveness plane DASS = Depression Anxiety Stress Scale DGVP = Health insurance company of the Dutch police force GP = General practitioner HADS = Hospital Anxiety Depression Scale GBC = Guideline-based care ICER = Incremental cost-effectiveness ratio NMB = Net monetary benefit NVAB = Netherlands Society of Occupational Medicine OHS = Occupational Health Service OP = Occupational Physician SD = Standard Deviation UC = Usual care REFERENCES 1. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJL. Global burden of disease and risk factors. New York/Washington, The World Bank/Oxford University Press; 2006. 2. WHO [World Health Organization]. Mental Health Action Plan for Europe. Facing the Challenges, Building Solutions. Copenhagen, WHO; 2005. 3. McDaid D, Curran C, Knapp M. Promoting mental well-being in the workplace: a European policy perspective. Int Rev Psychiatry. 2005;17:365-73. 4. Wang J, Adair CE, Patten SB. Mental health and related disability among workers: a population-based study. Am J Ind Med. 2006;49:514-22. 5. Lerner D, Henke RM. What Does Research Tell Us About Depression, Job Performance, and Work Productivity? J Occup Environ Med. 2008;50(4):401–410. 6. Workers Insurance Authority (UWV). Costs of disability benefits in The Netherlands in 2004 defined by ICD-10 diagnoses. 2005. 7. Uegaki K, Bakker IM, De Bruijne M, et al. Cost-effectiveness of a minimal intervention strategy for stress-related sick leave in general practice: results of an economic evaluation alongside a pragmatic randomized controlled trial. In press. 8. Zechmeister I, Kilian R, McDaid D, The Mheen Group. Is it worth investing in mental health promotion and prevention of mental illness? A systematic review of the evidence from economic evaluations. BMC Public Health. 2008 22;8:20. 9. Evers S, Salvador-Carulla L, Halsteinli V, McDaid D, The Mheen Group. Implementing mental health economic evaluation evidence: Building a bridge between theory and practice. J Mental Health. 2007 16(2):223-241. 10. Houtman ILD, Schoemaker CG, Blatter BM, de Vroome EMM, van den Berg R, Bijl RV. Psychological complaints, interventions and rehabilitation to work; the prognostic study of INVENT. [in Dutch] Hoofddorp: TNO Work & Employment, 2002. 11. Anema JR, Jettinghoff K, Houtman I, et al. Medical care of employees long-term sick listed due to mental health problems: a cohort study to describe and compare the care of the occupational physician and the general practitioner. J Occup Rehabil. 2006;16:41-52.

COST-EFFECTIVENESS

108

12. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. Validation of performance indicators for rehabilitation of workers with mental health problems. Med Care. 2005;43(10):1034-42. 13. Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Treatment of Work-Related Psychological Complaints: a Randomized Field Experiment among Self-employed. Work & Stress. 2006;20(2):129-144. 14. Nystuen P, Hagen KB, Herrin J. Solution-focused intervention for sick listed employees with psychological problems or muscle skeletal pain: a randomised controlled trial. BMC Public Health. 2006;6:69. 15. NVAB [Netherlands Society of Occupational Medicine]. van der Klink J, ed. Guideline for the management of mental health problems by occupational physicians. [in Dutch] Eindhoven: NVAB; 2000. 16. van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scand J Work Environ Health. 2003;29:478-487. 17. van der Klink JJL, Blonk RW, Schene AH, Dijk FJ van. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomized controlled design. Occup Environ Med. 2003;60(6):429-37. 18. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W. Design of a randomized controlled trial on the effects of Counseling of mental health problems by Occupational Physicians on return to work: the CO-OP-study. BMC Public Health. 2007;7(147):183. 19. Houtman I, Jettinghoff K, Brenninkmeijer V, van den Berg R. Work stress in the police force five years later: the effect of sectored agreements on (stress) management. [in Dutch]. Hoofddorp: TNO Work and Employment, 2005. 20. Drummond MF, Sculpher MJ, Torrance GW, et al. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press; 2005. 21. Steenstra IA, Anema JR, Bongers PM, et al. Cost effectiveness of a multi-stage return to work program for workers on sick leave due to low back pain, design of a population based controlled trial. BMC Musculoskelet Disord. 2003,4:26. 22. Hlobil H, Uegaki K, Staal JB, de Bruyne MC, Smid T, van Mechelen W. Substantial sick leave costs savings due to a graded activity intervention for workers with non-specific sub-acute low back pain. Eur Spine J. 2007;16:919–924. 23. Oostenbrink JB, Bouwmans CAM, Koopmanschap MA, Rutten FFH. Dutch manual for costing: methods and standard cost for economic evaluations in health care. Diemen: Health Insurance Council (CVZ); 2004. 24. Oostenbrink JB, Koopmanschap MA, Rutten FF: Standardisation of costs: the Dutch Manual for Costing in economic evaluations. Pharmacoeconomics. 2002;20:443-454. 25. Brouwers EP, de Bruijne MC, Terluin B, et al. Cost-effectiveness of an activating intervention by social workers for patients with minor mental disorders on sick leave: a randomized controlled trial. Eur J Public Health. 2007;17(2):214–220. 26. Young AE, Roessler RT, Wasiak R, et al. A developmental conceptualization of Return to Work. J Occup Rehabil. 2007;17:766–781. 27. De Vet HC, Heymans MW, Dunn KM, et al. Episodes of low back pain: a proposal for uniform definitions to be used in research. Spine. 2002;27(21):2409-16. 28. Van Poppel MN, de Vet HC, Koes BW, Smid T, Bouter LM. Measuring sick leave: a comparison of self-reported data on sick leave and data from company records. Occup Med (Lond). 2002;52:485-90. 29. Lovibond SH, Lovibond PF. Manual for the Depression Anxiety Stress Scales (DASS). University of New South Wales; 1993.

COST-EFFECTIVENESS

109

30. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. The Depression Anxiety Stress Scales (DASS): detecting anxiety disorder and depression in employees absent from work because of mental health problems. Occup Environ Med. 2003;60:i77. 31. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983;67:361–70. 32. Andrews G, Issakidis C, Sanderson K, Corry J, Lapsley H. Utilizing survey data to inform public policy: comparison of the cost-effectiveness of treatment of ten mental disorders. Br J Psychiatry. 2004;184:526-533. 33. Efron B, Tibshirani RJ. An introduction to the bootstrap. New York: Chapman & Hall; 1993. 34. van Hout BA, Al MJ, Gordon GS, Rutten FF. Costs, effects and C/E-ratios alongside a clinical trial. Health Econ. 1994;3:309-19. 35. Hoch JS, Briggs AH, Willan AR. Something old, something new, something borrowed something blue: a framework for the marriage of health econometrics and cost-effectiveness analysis. Health Econ. 2002;11:415-30. 36. Briggs A, Fenn P. Confidence intervals or surfaces? Uncertainty on the cost-effectiveness plane. Health Econ. 1998;7:723-40. 37. Hoch JS, Dewa CS. Lessons from trial-based cost-effectiveness analyses of mental health interventions: why uncertainty about the outcome, estimate and willingness to pay matters. Pharmacoeconomics. 2007;25(10):807-16. 38. Uegaki K, de Bruijne MC, Anema JR, et al. Consensus-based findings and recommendations for estimating health-related productivity loss from a company's perspective. Scand J Work Environ Health. 2007;33:122-130. 39. Wang PS, Simon GE, Avorn J, et al. Telephone screening, outreach, and care management for depressed workers and impact on clinical and work productivity outcomes: a randomized controlled trial. JAMA. 2007 26;298(12):1401-11. 40. Lim D, Sanderson K, Andrews G. Lost productivity among full-time workers with mental disorders. J Ment Health Policy Econ. 2000;3:139-46. 41. Brouwer WB, van Exel NJ, Koopmanschap MA, Rutten FF. Productivity costs before and after absence from work: as important as common? Health Policy. 2002;61:173-87. 42. Mynors-Wallis L, Davies I, Gray A, et al. A randomized controlled trial and cost analysis of problem-solving treatment for emotional disorders given by community nurses in primary care. Br J Psychiatry. 1997;170:113-9. 43. Kendrick T, Simons L, Mynors-Wallis L, et al. Cost-effectiveness of referral for generic care or problem-solving treatment from community mental health nurses, compared with usual general practitioner care for common mental disorders: Randomized controlled trial. Br J Psychiatry. 2006;189:50-9. 44. Bakker IM, Terluin B, van Marwijk HW, et al. A cluster-randomised trial evaluating an intervention for patients with stress-related mental disorders and sick leave in primary care. PLoS Clin Trials. 2007;2(6):e26. 45. Bower P, Rowland N. Effectiveness and cost effectiveness of counseling in primary care. Cochrane Database Syst Rev. 2006;3: CD001025. 46. Gilbody S, Bower P, Whitty P. Costs and consequences of enhanced primary care for depression. Systematic review of randomised economic evaluations. Br J Psychiatry. 2006;189:297-308. 47. Burdorf A. Economic evaluation in occupational health – its goals, challenges, and opportunities. Scand J Work Environ Health. 2007;33(3):164-166. 48. Proper KI, de Bruyne MC, Hildebrandt VH, van der Beek AJ, Meerding WJ, van Mechelen W. Costs, benefits and effectiveness of worksite physical activity counseling from the employer's perspective. Scand J Work Environ Health. 2004;30(1):36-46.

COST-EFFECTIVENESS

110

49. Ramsey S, Willke R, Briggs A, Brown R, Buxton M, Chawla A, Cook J, Glick H, Liljas B, Petitti D, Reed S. Good Research Practices for Cost-Effectiveness Analysis Alongside Clinical Trials: The ISPOR RCT-CEA Task Force Report. Value Health. 2005;8(5):521–33. 50. van der Roer N, Goossens M, Evers S, van Tulder M. What is the most cost-effective treatment for patients with low back pain? A systematic review. Best Pract Res Clin Rheumatol. 2005;19(4):671-684 51. Schene AH, Koeter MWJ, Kikkert MJ, Swinkels JA, McCrone P. Adjuvant occupational therapy for work-related major depression works: randomized trial including economic evaluation. Psychol Med. 2007;37(3):351-62. 52. Taimela S, Justén S, Aronen P, et al. An occupational health intervention programme for workers at high risk for sickness absence. Cost effectiveness analysis based on a randomised controlled trial. Occup Environ Med. 2008;65:242-248. 53. Grimshaw JM, Thomas RE, MacLennan G, Fraser C, Ramsey CR, Vale L, et al. Effectiveness and efficiency of guideline dissemination and implementation strategies. Health Technol Assess. 2004;8(6). 54. Hutchinson A, McIntosh A, Cox S, Gilbert C. Towards efficient guidelines: how to monitor guideline use in primary care. Health Technol Assess. 2003;7(18):iii,1-97. 55. Eccles M, Mason J. How to develop cost-conscious guidelines. Health Technol Assess. 2001;5(16):1-69. 56. Niessen LW, Grijseels E, Koopmanschap M, Rutten F; Dutch Ministry of Health. Economic analysis for clinical practice--the case of 31 national consensus guidelines in the Netherlands. J Eval Clin Pract. 2007;13(1):68-78. SUPPLEMENTS Supplement 1 Overview bottom-up calculation training costs GBC Resources Description Aggregated costs Trainer costs Two trainers for all three training days, including

preparation. 3 (days) x 6 (hours) at € 140 per hour = € 2540 preparation is 0,5 x € 2540 = € 1220 total = € 2540 + € 1220 = € 3760 x 2 = 75200

€ 7,500.00

Administration Time invested in training of each OP € 1,000.00 Room/equipment/ refreshment costs

For rental of meeting room at NSPOH Amsterdam. Equipment includes projector, screen & flip board. Refreshments breaks.

€ 1,250.00

Study material 5 OPs Binder, dividers & 75 printed pages € 230.00 Total training costs for 10 OPs € 10,000.00* Average cost per intervention subject (total training costs/125) € 40.00 * Training cost components were staff, administration, room, catering and study materials, totaling a cost price of € 10,000.00. Costs of OP attendance were not taken into account, as this is part of their education obligations, converted in the work agreement with the OHSs and their contracts with employers. The training costs considered a course delivered to 10 OPs, therefore these costs were divided by two for the five OPs participating in our trial. The resulting € 5,000.00 were divided by 125 (n subjects in the intervention group), resulting in an average training cost of € 40.00.

COST-EFFECTIVENESS

111

Supplement 2 Cost prices used for valuation of resource use in the economic evaluation Units [Units of measurement] Total costs Mean

(total) Mean Cost price

Health care sector Primary care € 172,442.00 GP consultation worker [No.] € 31,068.00 3.5 (838) € 37.00 Diagnostic tests [No.] € 20,862.00 1.2 (285) € 73.00 Medications [per medication ] € 56,806.00 6.8 (1623) € 35.00 Allied health professionals [No.] € 23,079.00 1.3 (314) € 73.50 Multidisciplinary physical therapy center € 35,451.00 0.1 (15) € 2363.40 Occupational health care € 124,900.00 OP consultation worker [No.] € 64,750.00 3.9 (925) € 70.00 ‡ OP contact employer [No.] € 6,250.00 1.0 (250) € 25.00 ‡ OP contact curative care [No.] € 2,500.00 0.4 (100) € 25.00 ‡ Contact case manager worker [No.] € 16.785,00 1.6 (353) € 45.00 ‡ Contact case manager employer [No.] € 2,485.00 0.6 (142) € 17.50 ‡ Contact case manager curative care [No.] € 122.50 0.0 (7) € 17.50 ‡ Administrative contact worker [No.] € 80.00 0.0 (8) € 10.00 ‡ Company Social worker [No. of sessions] € 26,928.00 0.4 (99) € 68.00 ‡ Psychological treatment € 208,654.00 Psychological treatment paid by community € 196,870.00 0.7 (147) € 1361.00 Psychological treatment paid by employer € 11,784.00 0.1 (16) € 77/125 *

Hospital € 67,286.00 Medical specialist [No. of consultations] € 10,319.00 0.8 (194) € 52.00 Medical operations [No.] € 16,259.00 0.1 (32) € 508.00 Hospitalization [No. of days] € 40,707.00 0.2 (54) € 754.00

Productivity losses Sick leave per paid work [per day] hca gross € 4,442,152.00 148 (240) € 124.73 ¶ Sick leave per paid work [per day] hca net € 3,397,554.00 Sick leave per paid work [per day] fcm net € 3,284,745.00 Intervention costs Training costs for MISS [per MISS subject] € 5,000.00 0.5 (125) € 40.00 ** Cost price sources: * Dutch Manual for Costing; ‡ Respective providers or professional organizations; ¶ Mean average gross salary participants, converted to per day cost price assuming a 36-hour work week; ** Determined via a bottom-up calculation.

COST-EFFECTIVENESS

112

Supplement 3 Means and standard deviations (S.D.) of resource use and utilization (n) group Units [Units of measurement] GBC UC Health care sector Mean S.D. n Mean S.D. n Primary care GP consultation employee [No.] 3.9 2.9 110 4.0 2.5 103 Diagnostic tests (GP-practice/hospital) [No.] 3.0* 2.6 53 2.2* 1.5 58 Medications [per medication ] 7.5 7.0 107 9.1 9.6 90 Physical therapist/Allied health professionals 35 36 Multidisciplinary physical therapy center 12 5 Occupational health care OP consultation employee [No.] 4.0 2.5 123 4.0 2.4 110 OP contact employer [No.] 2.2 1.4 61 2.0 1.3 58 OP contact curative care [No.] 1.6 0.9 33 1.8 1.0 27 Contact case manager employee [No.] 2.5 1.5 79 2.3 1.4 76 Contact case manager employer [No.] 1.6 0.7 44 1.5 0.8 47 Contact case manager curative care [No.] 1.0 - 1 1.2 0.4 5 Company Social worker [No. of sessions] 4 56 4 43 Psychological treatment Psychological treatment paid by employer 47* 107* Psychological treatment paid by community 8 12* 8 4* Hospital Medical specialist [No. of consultations] 2.8 1.7 30 2.6 1.7 42 Medical operations [No.] 1.4 1.0 10 1.4 1.0 13 Hospitalization [No. of days] 1.8 1.0 11 2.8 2.3 12 Productivity losses Mean S.D. n Mean S.D. n Net days of sick leave HCA 113.2 82.6 125 113.9 87.3 115 Gross days of sick leave HCA 150.7 97.1 125 146.9 102.4 115 * Significant difference between groups (t-test) for p<0.05

CHAPTER 6

Process evaluation of a randomized controlled trial on guideline-based care by occupational physicians counseling mental

health problems

Submitted for publication as: Rebergen DS, Bruinvels DJ, Bos CM, van der Beek AJ, van Mechelen W.

Process evaluation of a randomised controlled trial on guideline-based care by occupational physicians counseling mental health problems.

PROCESS EVALUATION

114

ABSTRACT Introduction Early 2000 the Netherlands Society of Occupational Medicine (NVAB) published a guideline on the management by occupational physicians (OPs) of workers with mental health problems. This guideline promotes a more active role of the OP by counseling and facilitating return to work (RTW). The aim of this study was to examine guideline adherence by Dutch OPs, as part of the process evaluation of a trial on the effectiveness of guideline-based care (GBC). Methods In a randomised controlled trial, participating OPs had to provide GBC in the intervention group. In the control group OPs had to provide usual care (UC), including minimal involvement and easy access to a psychologist. For the process evaluation, guideline adherence was assessed by means of an audit of medical files, using 20 guideline-based performance indicators (PIs). Mean performance rates on guideline adherence were compared between GBC and UC. In a per-protocol analysis (PPA), selected medical files from the GBC group with the highest performance rates were compared with UC files on their association with RTW using a Cox proportional hazards model. Results 240 police workers on sick leave with common mental health problems were included. Mean performance rates in guideline adherence by the OP were 50% and did not significantly differ between GBC and UC. PPA was not able to identify significant determinants on RTW. Conclusion In this process evaluation of a trial on GBC, guideline adherence by OPs was rated as average and no contrast was found between GBC versus UC. These results may be explained by the fact that the guideline was partially practice-based and contained many elements of UC. A possible explanation for the lack of results in the PPA may be that even in selected medical records with high performance rates, essential elements of the guideline were not applied.

PROCESS EVALUATION

115

INTRODUCTION Common mental health problems in workers, such as adjustment disorders, depression and anxiety, increasingly affect functioning to such an extent that they lead to sick leave, early retirement and productivity loss.[1,2] Primary and occupational health care of workers with mental health problems usually focus on recovery of symptoms instead of on return to work (RTW).[3] In the Netherlands, workers have to visit an occupational physician (OP) for RTW purposes when they are on sick leave.[4] Therefore, OPs play a central role in the treatment of workers with common mental health problems. In 2000, the Netherlands Society of Occupational Medicine (NVAB) published a guideline on ‘The management by OPs of workers with mental health problems’.*4, 5+ This guideline promotes a more active role of the OP as a counsellor facilitating RTW of the worker. Guideline-based care (GBC) aims to activate the OP as an alternative to usual care, in which OPs usually have a minimal role. In this minimal role, they often refer workers to psychologists if complaints persist.*6+ This may lead to a ‘referral’ delay, often associated with delayed RTW.[7-9] Consequently, workers may not get the optimal care they need. A randomised controlled trial (RCT) with a 12-month follow-up (CO-OP study) was conducted to evaluate the effectiveness of the guideline for OPs (GBC), compared to usual care.[10] GBC did not result in better outcomes on RTW.[11] However, GBC appeared to be cost-effective, as health care costs were lower in the GBC group compared to the higher costs of psychological treatments in UC.[12] To determine actual use of the guideline in GBC, a process evaluation was performed in which guideline adherence by OPs was assessed and related to RTW.[13-15] Although other studies have presented data on the relationship between guideline adherence and RTW, no firm relation has been established.[3,9,16-20] These studies used a set of guideline-based performance indicators (PIs) to measure guideline adherence. Performance rates on PIs were interpreted as markers of quality of care, here defined as guideline adherence.

A first set of 10 PIs and corresponding criteria was developed by Smits et al.[16] Nieuwenhuijsen et al. tested these PIs with an audit of medical records of workers with adjustment disorders.[9] For validation of the PIs they adapted the set for common mental disorders, including depression and anxiety.[17] This validated set evaluated a substantial part of the guideline regarding problem orientation, diagnosis and interventions. This set of PIs is in this paper referred to as the ‘initial’ set of 10 PIs. However, some specific constructs of the guideline such as progress evaluation and relapse prevention were not covered. Therefore, based on this initial set of 10 PIs, an extended set of 20 PIs was developed to assess guideline adherence. This new set of PIs also focused on the counseling role by OPs, evaluation of and interventions on stagnation in recovery of the worker, and on relapse prevention. The aim of this study was to examine guideline adherence by Dutch OPs as part of a process evaluation of the trial on the effectiveness of GBC (CO-OP study). The first hypothesis was that guideline adherence in the GBC group will be higher, compared to UC. The second hypothesis was that better guideline adherence by the OP will be associated with earlier RTW. METHODS Design The present study is a RCT, in which the effect of the ‘Dutch national guideline on the management of workers with mental health problems by OPs’ was evaluated (CO-OP study).[10] Subjects in the intervention group were treated by OPs trained to provide

PROCESS EVALUATION

116

guideline-based care (GBC). The control group received usual care (UC), with minimal involvement of the OP and, if applicable, easy access to treatment by a psychologist. The study was funded by the Dutch Ministry of Internal Affairs and Kingdom Relations, the Health Insurance agency of the Dutch police force (DGVP), the occupational health service (OHS) Commit, and the VU Medical Centre. The study design, protocol and procedures were approved by the Medical Ethics Review Committee of the VU University Medical Centre. Details of the study design and the guideline have been reported elsewhere.[10] A brief summary is presented below.

Study population and randomisation This study was conducted with the cooperation of the Dutch police force, which is an organisation with a relatively high incidence of stress-related sick leave.[21] Two police departments, comprising a source population of 2500 police workers, were chosen because they had contracts with the same OHS (Commit). With the start of the study in 2002, police workers were included by the OP when consulting for sick leave due to common mental disorders. Randomisation in two groups (GBC and UC) was done at the individual level of participants. Participants, employers and OPs were not blinded for the intervention. Researchers were blinded for treatment allocation and protocol compliance. In 2003 a second OHS (KLM Health Services) was introduced to the study, when one of the two police departments changed their OHS. Usual care Usual care (UC) consisted of minimal involvement of the OP and easy access to a psychologist, who’s treatment was mainly based on cognitive behavioural therapy principles. This psychological treatment was fully funded by the DGVP. Intervention The intervention (GBC) consisted of treatment by OPs according to the Dutch guideline of workers with mental health problems.[4, 5, 10] The guideline focuses on five constructs in the management of mental health problems. First, a problem orientation in which the OP acknowledges the interaction between the disabled worker and his surroundings (work, personal and care). Second, a simplified classification of mental health problems is introduced, with only four categories: adjustment disorder (distress, nervous breakdown, burnout); depression; anxiety; and other psychiatric disorders. Third, early and activating interventions by the OP are promoted, in which time contingent RTW is part of the recovery process, even if the mental health problems are not related to work. The OP is stimulated to operate as counsellor, applying cognitive behavioural techniques, of workers with stress and/or work-related problems. Fourth, time contingent evaluation is done in which the OP acts as case manager, who intervenes when recovery stagnates. Finally, relapse prevention is an integral aspect of the treatment.

OPs participating in the study received training in the guideline before the study started. During this training, consisting of a three-day educational course, 10-15 OPs, were trained in the use of the guideline.*22+ The workers’ own responsibility in the recovery process was emphasized, as was the importance of an early start of the intervention aimed at the acquisition of coping skills and at regaining control. The OPs were encouraged to use specific tools, such as the four dimensional symptom questionnaire (4DSQ)[23], patient information leaflets on stress, and day structuring exercises. The OPs were free to choose specific tools in each phase of the process.

PROCESS EVALUATION

117

Treatment contamination As this is an effectiveness trial, it aimed to mimic a realistic situation in both treatment groups. Therefore, no activities were undertaken to improve the actual treatment compliance by the OP with the allocated treatment. As randomisation was done at the individual level of participants, OPs who were trained in the guideline treated the participants of both the GBC and UC group. Obviously, this situation created a risk of treatment contamination between the groups. However, the contrast between groups was maximized by creating a situation in which referral to a psychologist in UC was always granted by the insurance agency (DGVP) and thus inviting the OP to play a minimal role. Outcomes Guideline adherence by the OP Guideline adherence by the OP was assessed by an audit of the medical records of the occupational health service (OHS). This means that guideline adherence by the OP was assessed according to guideline-based performance indicators (PIs).

The initial set of 10 PIs and corresponding criteria was validated by Nieuwenhuijsen et al. [17] As this set did not cover the complete treatment process, 10 additional PIs were developed and added to the initial set. The 10 additional PIs were based on recent scientific literature.[3, 9, 16-20, 24, 25] This extended set of 20 PIs was categorised according to the five core constructs of the guideline (Table 1). The initial set of 10 PIs consisted of PIs 1-5, 7, 9, and 15-17.

For each PI criteria were used, which were based on an if-then logic. For example: IF the treatment in curative care is not effective THEN the OP should contact the practitioner. If criteria of a PI were met or not applicable, the score for that PI was 1 corresponding with maximal guideline adherence. If criteria were not met, the score was 0. In this way an average performance rate was obtained for each PI. A higher performance rate reflects higher guideline adherence on the specific related topic.

Guideline adherence for the complete treatment process was measured on the initial set of 10 PIs, and on the extended set of 20 PIs. For both sets, the mean performance rate was calculated as the mean PI score divided by the total number of PIs.

Guideline adherence for the 20 PI-set in the GBC group was dichotomised into ‘higher’ and ‘lower’ performance rates by applying a cut-off point based on the median. In the per-protocol analysis, PIs of medical records with high performance rates in the GBC group were compared with the PIs in the UC group. Table 1 - Performance indicators guideline adherence and their criteria I Problem orientation PI 1* Assessment of symptoms Criteria: 1. Presence or absence of essential symptoms of anxiety disorder and depressive

disorder should be noted in file 2. Presence or absence of distress symptoms (fatigue, concentration problems, sleeping

problems, emotional reactivity) should be noted in file IF one or both criteria not met within 2 consultations, THEN PI1= 0 ELSE = 1 PI 2 Use of 4DSQ Criterion: 1. The OP uses the four dimensional symptom questionnaire (4DSQ) to determine the

correct diagnosis IF criterion not met within 2 consultations, THEN PI2= 0 ELSE = 1 PI 3* Evaluation of curative care Criteria: 1. Treatment in the curative sector, or its absence, should be noted in file 2. IF patient receives treatment, THEN the OP should evaluate whether this treatment is

PROCESS EVALUATION

118

effective IF one or both criteria not met within 2 consultations, THEN PI3= 0 ELSE = 1

PI 4* Assessment of work-related causes Criteria: 1. The work-related causes, or their absence, should be stated in file 2. The OP should note whether colleagues have the same problems IF one or both criteria not met within 2 consultations, THEN PI4= 0 ELSE = 1 PI 5* Evaluation of work disabilities Criteria: 1. Functional limitations in home or work environment, or their absence, should be

stated in file 2. Work activities of patient should be noted by OP 3. OP should assess whether patient is limited in his work functioning 4. IF patient has work limitations, THEN OP should assess other impediments for RTW IF one or both criteria not met within 2 consultations, THEN PI5= 0 ELSE = 1 PI 6 Patient expectations on RTW Criterion: 1. The patients goals and expectations, or absence, are noted in the medical file IF one or both criteria not met within 2 consultations, THEN PI6= 0 ELSE = 1 II Diagnosis PI 7* Correct diagnosis Criteria: 1. Diagnosis should be noted in file 2. Diagnosis should be correct: - IF adjustment disorder: at least one psychological distress symptom noted in file - IF depressive disorder: at least one essential symptom AND five depressive symptoms

noted in file - IF anxiety disorder: at least one anxiety disorder noted in file 3. Diagnosis should not be missed if criteria above apply IF one or both criteria not met within 2 consultations, THEN PI7= 0 ELSE = 1 PI 8 Comparison of diagnosis with DASS Criterion: 1. Diagnosis OP first two consultations is congruent with diagnosis DASS IF criterion not met within 2 consultations, THEN PI8= 0 ELSE = 1 III Interventions PI 9* Interventions targeted at individual Criterion: 1. Intervention aimed at the individual should be noted or be referred

- IF adjustment disorder, THEN OP should start interventions OR should refer patient to psychologist/social worker/general practitioner (GP) OR should consult with practitioner giving current treatment - IF anxiety disorder OR depression OR other psychiatric disorder, THEN OP should refer patient to psychologist/social worker/GP OR should consult with practitioner giving current treatment

IF one or both criteria not met within 3 consultations, THEN PI9= 0 ELSE = 1 PI 10 Counseling Criterion: OP should start with activating counseling regarding work- or stress-related aspects IF criterion not met within 3 consultations, THEN PI10= 0 ELSE = 1 PI 11* Interventions targeted at providers of care in curative sector Criteria: 1. IF treatment in curative sector is lacking and deemed necessary, THEN OP should

start interventions targeted at the individual OR refers patient to psychologist/social worker/general practitioner

2. IF treatment in curative sector is not effective, THEN OP should consult with practitioner giving current treatment

IF one or both criteria not met within 3 consultations, THEN PI11= 0 ELSE = 1

PROCESS EVALUATION

119

PI 12 Referral to secondary care Criteria: 1. IF diagnosis is adjustment disorder AND there is stagnation AND no (partial) RTW in

13 weeks OR not at least 80% RTW in 26 weeks after start of sick leave, THEN OP contacts GP and refers the patient to secondary care

IF criterion not met, THEN PI12= 0 ELSE = 1 PI 13 Contact general practitioner Criterion: 1. The OP consults with the general practitioner during treatment IF criterion not met, THEN PI13= 0 ELSE = 1 PI 14 Contact work system Criterion: 1. The OP consults with the employer during treatment each 6 weeks IF criterion not met, THEN PI14= 0 ELSE = 1 PI 15* Interventions targeted at organisation Criterion: IF work is a causal, eliciting or maintaining factor in the mental health problem, THEN

OP should intervene in the work organisation IF one or both criteria not met within 3 consultations, THEN PI15= 0 ELSE = 1 PI 16* Advice on RTW Criteria: 1. Advice on RTW should be provided by OP 2. IF no impediments for RTW are present, THEN OP should advise full or partial RTW IF one or both criteria not met within 3 consultations, THEN PI16= 0 ELSE = 1 IV Evaluation PI 17* Timing of consultations Criteria: 1. 1st consultation should be within 3 weeks from 1st day of sickness absence 2. IF patient has not yet completely recovered, THEN next 2nd or 3rd consultation

should be within 4 weeks from previous consultation IF one or both criteria not met within 3 consultations, THEN PI17= 0 ELSE = 1 PI 18 Evaluation of early stagnation Criterion: 1. Evaluation of early stagnation should be noted in file in two months after sick leave

AND/OR in 6 weeks after the consultation of inclusion IF criterion not met, THEN PI18= 0 ELSE = 1 PI 19 Intervention when any kind of stagnation occurs Criterion: IF a stagnation during the treatment period occurs - THEN the environment which causes the stagnation should be mentioned (stagnation

in organization AND/OR individual (home) situation AND/OR curative care) AND - THEN OP considers new problem definition or treatment policy AND - THEN OP discusses this with the patient and the environment IF criterion not met, THEN PI19= 0 ELSE = 1 V Relapse prevention PI 20 Relapse prevention by the OP Criterion: 1. After full RTW the OP has at least 1 consultation with the patient IF criterion not met, THEN PI20= 0 ELSE = 1 * Initial 10 PI-set by Nieuwenhuijsen et al.[17]

Assessment of guideline adherence The PI-sets reflecting guideline adherence were assessed independently by two researchers (CB and DB) using the medical records of the OHS. If the ratings were not congruent, a third researcher (DR) was bound to make a decision. PIs were assessed using data of the medical records of the participant, gathered from the databases of the participating OHSs. To guarantee blinding of the outcome assessors, the data of the medical records were stripped of non-essential information and were transferred to a

PROCESS EVALUATION

120

Microsoft Access database. A digital score list was created in the program SPSS Data Entry Builder version 3.0, in which individual criteria of the PIs could be scored. The criteria of the PIs were converted into PI-scores using SPSS syntaxes, enabling computerised appraisal of guideline adherence. Return to work The primary outcome measure was RTW. Our definition of RTW was the duration in calendar days of absence from work due to mental health problems until RTW in own or equal earnings.[26, 27] Sick leave data were gathered from records of the police departments. Prognostic measures Data about personal, treatment and work characteristics, were gathered from the records of the police departments and from the medical files of the OHSs. Severity of depression, anxiety and stress was measured at baseline, using the Depression Anxiety Stress Scales (DASS-42).[28, 29] Statistical analyses Analyses were conducted at the individual level, according to the intention-to-treat principle. The first hypothesis was that guideline adherence in the GBC group will be higher, compared to UC. Therefore, mean performance rates on guideline adherence of both groups were compared by Chi²-tests. This was done for each PI separately, and for the 10- and 20 PI-sets.

The second hypothesis was that better guideline adherence by the OP is associated with earlier RTW. To answer the second hypothesis, a per-protocol analysis was conducted. In this analysis was tested if higher guideline adherence by the OP in the GBC group leads to a significant earlier RTW compared to UC. Higher guideline adherence by the OP in the GBC group was defined as a performance rate above the median in the 20 PI-set. To test our hypothesis, PIs of medical records with high performance rates in the GBC group were compared with UC in a Cox proportional hazards model with two-tailed tests at a significance level of 5% (P < 0.05). Statistical analyses were performed using SPSS 15.0. RESULTS Participants and provided care Between January 2002 and January 2005, 489 workers were registered as being absent from work due to mental health problems by the OHS (figure 1).[11] Of those, 240 workers met the inclusion criteria and signed an informed consent form. By randomisation 125 subjects were assigned to the GBC group and 115 subjects were assigned to UC. Although not all the questionnaires were returned by the subjects, nobody was lost to follow-up regarding guideline adherence as performance rates were based on the medical files. RTW-data of workers who quitted their jobs or died (n=16), or did not RTW during the one-year follow-up (n=24), were censored. Baseline data, treatment and RTW characteristics are shown in table 2. Subjects received significantly more psychological treatments in UC and significantly more treatments by company social workers in GBC.

PROCESS EVALUATION

121

Figure 1 Flow chart CO-OP study Guideline adherence by OPs Table 3 shows the mean performance rates for both GBC and UC. No significant differences were found between the groups on most performance rates. There was a significantly higher performance rate on the use of the 4DSQ in the GBC group to assist in the classification of the mental health problem (22% versus 10%, p=0.02). Early and activating interventions aimed at individual workers, such as referral to a psychologist or social worker differed significantly between GBC and UC (80% versus 93%, p<0.01). In the GBC group, the OP contacted the GP more often (7% versus 1%, p<0.01).

The mean performance rates on guideline adherence for both the initial 10 PI-set and the extended 20 PI-set did not significantly differ between the two groups. 44 workers (37%) of the GBC-group received treatment with high guideline adherence, which means a mean performance rate on the 20 PI-set that was higher than the median of 10. The mean performance rate of the group selected medical files with high guideline adherence in GBC was 11.80 (sd=1.05, n=44), corresponding with 60% guideline adherence. This was significantly higher than the 10.05 (sd=1.65, n=115) in UC. Per-protocol analysis The data of these 44 medical records from the GBC-group were compared with all 115 medical records of the UC-group in a per-protocol analysis. Therefore, the data of 159 workers were entered in a Cox proportional hazards model, of which 18 were censored. After adjusting the model for the potential confounders such as OP, severity of the disorder, children, sick leave in the year before inclusion, and type of function, GBC was no significant predictor of RTW in the Cox proportional hazards model (HR=1.017, p=0.93). This is graphically shown by the Cox regression curves for both treatment groups in figure 2.

Workers registered as being absent from work due to mental health problems n=489

Inclusion in RCT during consultation OP in which informed consent was obtained n=240

Usual care n=115

Randomisation

Guideline-based care n=125


Return to work (sick leave days) Guideline adherence (PIs)


Return to work (sick leave days) Guideline adherence (PIs)

PROCESS EVALUATION

122

Table 2 Characteristics of the study population and provided care GBC (n=125) UC (n=115) P-value Patient characteristics Age in years, mean (sd) 38.8 (8.4) 40.0 (9.5) 0.27 Gender (% male) 51.2 60.0 0.19 Children (%) 66.4 62.6 0.59 Work related characteristics Police department (% PHM) 70.4 64.3 0.34 Mean contract hours a week, mean (sd) 33.6 (6.2) 34.1 (6.0) 0.67 Executive work (%) 59.2 66.1 0.29 Work relatedness mental health disorder (%) 48.8 44.7 0.39 Absenteeism previous year N sick leave periods previous year, mean (sd) 2.7 (2.2) 2.5 (1.9) 0.53 Days of sick leave previous year, mean (sd) 56.9 (61.4) 56.1 (86.1) 0.94 Severity of disorder (symptoms DASS) n=112 n=101 DASS-Stress, mean (sd) DASS-Anxiety, mean (sd) DASS-Depression, mean (sd)

9.2 (7.3) 4.2 (5.5) 6.9 (7.4)

9.0 (6.8) 3.9 (5.2) 6.6 (7.4)

0.91 0.69 0.78

Depression and/or Anxiety DASS (%)* 35.1 29.0 0.38 Diagnosis OP, Depression and/or Anxiety (%) 16.4 13.4 0.73 Treatment characteristics N consultations OP, mean (sd) 3.4 (2.3) 3.3 (2.3) N consultations case manager, mean (sd) 2.3 (1.3) 2.2 (1.4) N consultations GP, mean (sd) 1.9 (1.4) 1.9 (1.2) Contact OP & employer, mean (sd) 1.2 (1.4) 1.1 (1.4) Contact OP & GP, mean (sd) 0.2 (0.4) 0.0 (0.1) OHS, Commit (%) 50.4 53.0 0.70 Treatment psychologist, n (%) 59 (46%) 98 (85%) <0.01 Company social worker, n (%) 56 (45%) 42 (37%) <0.01 RTW-characteristics Partial RTW before full RTW (%) 68 54 0.01 Duration partial RTW period, mean days (sd) 53 (56) 51 (78) 0.28 Full RTW, median in days (CI) 105 (84-126) 104 (81-127) 0.78 Duration of recurrences, mean days (sd) 19 (39) 19 (39) 0.95 * Based on cut-off scores (30): >12 on symptoms depression and >5 on symptoms of anxiety

Table 3 - Optimal performance in GBC and UC

PI Content GBC UC P-value I Problem orientation N=125 N=115

1† Assessment of symptoms 33 (26%) 28 (24%) 0.77

2 Use of 4DSQ 27 (22%) 12 (10%) 0.02* 3

† Evaluation of curative care 70 (56%) 66 (57%) 0.90

4† Assessment of work-related causes 62 (50%) 56 (49%) 0.90

5† Assessment of work disabilities 69 (61%) 70 (55%) 0.43

6 Patient expectations on RTW 1 (1%) 5 (4%) 0.11 II Diagnosis 7

† Correct diagnosis 77 (62%) 77 (67%) 0.42

PROCESS EVALUATION

123

8 Comparison of diagnosis with DASS 84 (67%) 77 (67%) 0.99 III Interventions 9

† Interventions targeted at individual 100 (80%) 107 (93%) <0.01*

10 Counseling 25 (20%) 16 (14%) 0.23 11

† Interventions targeted at curative care 85 (68%) 73 (64%) 0.50

12 Referral to secondary care 108 (86%) 108 (94%) 0.06 13 Contact general practitioner 16 (7%) 2 (1%) <0.01* 14 Contact work system 55 (44%) 55 (48%) 0.61 15

† Interventions targeted at organisation 122 (98%) 111 (97%) 0.71

16† Advice on RTW 110 (88%) 105 (91%) 0.53

IV Evaluation 17

† Timing of consultations 26 (21%) 30 (26%) 0.36

18 Evaluation of early stagnation 76 (61%) 76 (66%) 0.42 19 Intervention when stagnation occurs 78 (62%) 68 (59%) 0.69 V Relapse prevention 20 Relapse prevention by the OP 16 (13%) 14 (12%) 0.99

†10 PI-set Guideline adherence, mean(sd)

Performance rate 10 PI-set [17] 6.03 (1.4) 6.29 (1.4) 0.16

20 PI-set Guideline adherence, mean (sd) Performance rate 20 PI-set

9.92 (1.8) 10.05 (1.7) 0.55

20 PI-set GBC >10

High guideline adherence GBC, mean (sd) n=159 20 Piset > 10

11.80 (1.0) n=44

10.05 (1.7) 0.00*

* Significant result, p ≤ 0.01

Figure 2 Cox regression curves based on the Cox proportional hazards model (HR=1.017, p=0.93)

Days until full RTW

4003002001000

Pro

po

rtio

n f

ull R

TW

1,0

0,8

0,6

0,4

0,2

0,0

- GBC (44) - UC (115)

PROCESS EVALUATION

124

DISCUSSION Interpretations A randomised controlled trial (RCT) was performed to examine the effectiveness of guideline-based care (GBC) compared to usual care (UC) by Dutch occupational physicians (OPs) on return to work (RTW) of workers with common mental health problems.[10] Although no positive effect was found on RTW, GBC appeared to be cost-effective compared to usual care.[11, 12] This paper describes the results of a process evaluation on guideline adherence by the OP in both groups, including a per-protocol analysis on the association between guideline adherence and RTW.

Guideline adherence was measured by a set of 20 performance indicators (PIs), including 10 PIs developed by Nieuwenhuijsen et al.[9, 17] The mean performance rates for these two sets of PIs did not differ significantly between GBC and UC and were 50% for the extended 20 PI-set and 60% for the initial 10 PI-set. The performance rates for the individual PIs of the initial set were comparable to results of earlier studies.[3, 9, 16-20] Specific and important elements of the evaluated guideline were the use of the four dimensional symptom questionnaire (4DSQ), counseling techniques by the OP, and contact between OP and GP. Although PIs on the use of the 4DSQ and contact between OP and GP showed significant higher performance rates in GBC, they were still performed at low rates in both groups.

A per-protocol analysis was conducted, in which selected cases with high guideline adherence in the GBC group were compared to the total UC group on their relations with RTW. No differences were found on RTW, between the subgroup with higher guideline adherence in GBC and the UC group.

These results may explain the lack of positive effects of guideline-based care on RTW in our trial.[11] First, there is not enough contrast of guideline adherence between GBC and UC. Secondly, high guideline adherence of on average 60% in the GBC group did was insufficient to assess the effects on RTW. We will elaborate on these interpretations in the paragraphs presented below. Lack of contrast between guideline-based care and usual care No activities were undertaken to improve the actual treatment compliance by the OPs with the allocated treatment, to make the trial realistic as possible in both treatment groups. All participating OPs were trained in the guideline and treated participants of both GBC and UC group. To avoid the risk of treatment contamination, a situation was created in which referral to a psychologist in UC was always granted by the insurance company (DGVP).

Results on referral patterns indicate that this indeed created a difference between the GBC and UC groups. Psychological treatments and the performance rate on PI12 were significantly higher in UC, referral patterns to a company social worker were significantly higher in GBC. Thus, we might conclude that our randomisation design was successful in increasing contrast between GBC and UC in the use of additional psychological treatment, and can not explain the lack of contrast observed in the results.

Another possible explanation for the lack of contrast may be the nature of the guideline itself. Many practice guidelines, including this guideline, are partially practice based. This may explain the 50% guideline adherence which we found in the UC group. This percentage is comparable to findings of other studies [3, 9, 16-20].

In addition, training of OPs in the guideline did not result in better guideline adherence. In our study OPs were enrolled in a 3-days training course, which is comparable to other postgraduate education of occupational physicians on other Dutch

PROCESS EVALUATION

125

guidelines. However, the training may have been ineffective in raising guideline adherence beyond 50%. Insufficient guideline adherence to assess the effects of guideline-based care on RTW To conduct a per-protocol analysis, selected cases with high guideline adherence in the GBC group were compared to the total UC group on their relations with RTW. These cases with high guideline adherence had performance rates of only 60%, compared to 50% in the UC group. The consequence is that even in selected medical records with ‘high’ performance rates, essential elements of the guideline were still not applied. Among these elements were assessment of symptoms, use of the 4DSQ, inquiring about the expectations of patients on RTW, counseling of patients by the OP, contact of the OP with the GP, timing of consultations, relapse prevention after RTW by the OP. The infrequent use of these essential elements of the guideline may be the result of difficulties in implementing this guideline by OPs. Consequently, no relationship between guideline adherence and RTW could be proven. Strengths and limitations of this study When assessing guideline adherence by physicians, one can distinguish between direct and indirect methods of performance assessment.[30] While direct methods observe clinical performance, indirect methods use information from either medical records or self-reported data of physicians. An audit of medical files, as used in this study to assess guideline adherence and perform a process evaluation, has the advantage that physicians are unaware that their notes will be used for research purposes. However, an audit of medical records is, as any observational study, susceptible to bias.[31]

A first possible source of bias is that OPs did not register all their findings systematically for use in research. Negative findings and routine activities may not have been recorded, thereby resulting in an underestimation of the true performance. Exception was the PI ‘continuity of care’. The OHS routinely lists the date of each consultation, which rules out the possibility of inaccurate registration. A second source of bias is related to this last notion, as the two participating OHSs used different registration systems, which may have influenced performance rates of the PI. For example, at one OHS it was more common to refer to a company social worker, instead of counseling as OP.

One of the strengths of this study is that a new set of PIs has been developed and rated in a new study population by independent researchers. An initial set of 10 PIs was used and extended by adding new PIs, based on recent findings. Use of the initially validated set of 10 PIs generated some problems, as this set was validated using registration forms, which allowed more stringent criteria.[13] A comparable set of PIs was used in auditing medical files, but only focused on adjustment disorders.[9] Both sets only focus on the first three consultations by the OP, while we aimed to incorporate the entire treatment process, including the (evaluation of) stagnation. Positively, the initial set of 10 PIs combines information from experts, OPs and patients. This has led to a comprehensive appraisal of the validity of the initial 10 PI-set.[17] In addition to the initial 10 PI-set, an extended 20 PI-set has been developed to be able to assess a contrast in guideline adherence between GBC and UC, to cover the whole treatment process, and to incorporate recent findings. For example, the initial PI regarding ‘interventions aimed at the individual’ contained broad criteria and did not discriminate in interventions performed by the OP (cognitive behavioural counseling techniques) and referral to other professionals such as psychologists. Therefore, this PI was split into different items with corresponding criteria. The same goes for the initial PI ‘interventions aimed at providers of curative care’, which does not make any distinction in consultation of the GP and referral

PROCESS EVALUATION

126

to other professionals. As this PI was not validated, like the initial PI on the intervention role by the OP, their use was not recommended.[16] We replaced these PIs by other, more specified, PIs and criteria to gather more information about the treatment content. However, this may be an explanation of the relatively high performance rate by the OP in UC. Additionally, specifying the PIs created a higher risk on a lack of variability in the performance rates on the new PIs. For example, the performance rate on PI6 ‘patient expectations on RTW’ scored under 5%, although this specific recommendation is clearly mentioned in the guideline.

Thus, use of an audit of medical files may have caused some bias in measuring guideline adherence. A systematic approach was conducted on the development of an extended set of validated PIs, but may have resulted in a lack of variability of the performance rates. Although performance rates in this study were based on a profound, consistent and democratic way of scoring by a multidisciplinary team of three researchers (OP, psychologist and health scientist), we must conclude that validity of the 20 PI-set remains insecure. This may have caused a lack of contrast between the group of workers that received high guideline adherence versus low guideline adherence in the GBC-group. Implementation and content of the guideline Lack of contrast between GBC and UC in guideline adherence by the OP in this study, might be attributed to insufficient training in treatment according to the guideline, or a possible inapplicability of the trained skills. Multifaceted interventions have proven to be most effective in implementation of evidence based guidelines.[32-35] The intervention in this study, the 3-days training course in use of the guideline, might have been too minimal for the OPs to actually learn the necessary skills. However, Smits et al. [16] showed that a comparable training of the guideline for OPs-in training resulted in an improved guideline adherence. The same training course resulted in an effective intervention in an earlier trial, but this was in an in-company setting and with a population of workers with adjustment disorders.[22]

Optimal implementation of the guideline in occupational health care seems unrealistic. The changing work environment in the last decades has created a difficult position for OPs, as patients, employers and the management of OHSs put pressure on OPs with different interests. The lack of effect of the training on guideline adherence, compared to UC, may be due to this work conditions and perceived behavioural control.[19, 30] The guideline promotes an intensive form of treatment, for instance in prescribing regular consultation by the OP with the GP, which is not applicable in practice.[3, 18, 20] Although the (training in the) guideline considers the time constraints under which OPs work, OPs may have been just too busy to carry out the full intervention.

Recently, the guideline for OPs has been revised.[36] A critical reappraisal of the guideline appeared to be necessary since it was more consensus-based than evidence-based[5, 37, 38]. Another reason was that the guideline seemed to be too ambitious in changing the OP’s performance in the current setting. Therefore, the revised guideline is less complex and less ambitious.[36] It promotes some elements that were measured by the extended 20 PI-set. These are: more use of the 4DSQ, continuous evaluation with patient and employer to facilitate RTW, evaluation of and intervention on stagnation, consultation of the GP after 2 months, and (relapse) prevention of mental health problems.

Future research should investigate how guideline adherence can be improved.[37] Answering the question of why guideline adherence is low and acting on this answer, could be of greater influence on the effectiveness of the guideline than changing its contents alone.[39] Additionally, measurement of guideline adherence needs

PROCESS EVALUATION

127

to be further developed, and complete reporting of information relevant to process evaluation should be enforced and facilitated.[40] Conclusions In this study guideline adherence by Dutch OPs was assessed, as a process evaluation of a trial on the effectiveness of GBC of workers with mental health problems, compared to UC. The intervention, a 3-days training course, did not lead to better guideline adherence, when compared to UC. Per-protocol analyses did not show a significant difference on RTW in favour of the high guideline adherence GBC group. These results may point to a lack of effect of GBC in the trial.

However, it remains unclear if this lack of effect is due to the content of the guideline, or the actual implementation of the guideline. Therefore, future research should investigate which elements of the guideline predict faster RTW and how implementation of the guideline can be improved. This study represents a next step in the development of evidence-based occupational health care for the treatment of workers with common mental health problems. ABBREVIATIONS CI = Confidence Interval DASS-42 = Depression Anxiety Stress Scale with 42 items DGVP = Health insurance company of the Dutch police force GP= General practitioner GBC = Guideline-based care in the intervention group; activating counseling by OP using cognitive behavioural elements and facilitating RTW of the worker HADS = Hospital Anxiety Depression Scale NVAB = Netherlands Society of Occupational Medicine OHS = Occupational Health Service OP = Occupational Physician PI = Performance Indicator RCT = Randomised Controlled Trial RTW = Return to work UC = Usual care in the control group; minimal involvement of the OP and easy access to a psychologist REFERENCES 1. McDaid (Ed). Mental Health in Workplace Settings. Consensus paper. Luxembourg: European Communities. 2008. 2. Lerner D, Henke RM. What Does Research Tell Us About Depression, Job Performance, and Work Productivity? J Occup Environ Med 2008;50(4):401–410. 3. Anema JR, Jettinghoff K, Houtman ILD, Schoemaker CG, Buijs PC, van den Berg R. Medical care of employees long-term sick listed due to mental health problems: A cohort study to describe and compare the care of the occupational physician and the general practitioner. J Occup Rehabil 2006;16:41-52. 4. NVAB (Netherlands Society of Occupational Medicine). van der Klink J, ed. Guideline for Mental Health Problems. [in Dutch] The Netherlands: Eindhoven, 2000. 5. van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scand J Work Environ Health 2003;29:478–487.

PROCESS EVALUATION

128

6. Houtman ILD, Schoemaker CG, Blatter BM, de Vroome EMM, van den Berg R, Bijl RV. Mental health problems, interventions, and work resumption: The prognoses Study INVENT [In Dutch]. TNO Quality of Life. The Netherlands: Hoofddorp, 2002. 7. Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Treatment of Work-Related Psychological Complaints: a Randomized Field Experiment among Self-employed. Work & Stress 2006;20(2):129-144. 8. Nystuen P, Hagen KB, Herrin J. Solution-focused intervention for sick listed employees with psychological problems or muscle skeletal pain: a randomised controlled trial. BMC Public Health 2006;6:69. 9. Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. Occup Environ Med 2003;60:21-25. 10. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W. Design of a randomized controlled trial on the effects of Counseling of mental health problems by Occupational Physicians on return to work: the CO-OP-study. BMC Public Health 2007;7:183. 11. Rebergen DS, Bruinvels DJ, Bezemer PD, van der Beek AJ, van Mechelen W. Guideline-based care of common mental disorders by Occupational Physicians (CO-OP-study): a randomized controlled trial. Accepted for publication in J Occup Environ Med. 12. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Tulder MW, van Mechelen W. Cost-effectiveness of guideline-based care for workers with mental health problems. Accepted for publication in J Occ Environ Med. 13. Hulshof CT, Verbeek JH, van Dijk FJ, van der Weide WE, Braam IT. Evaluation research in occupational health services: general principles and a systematic review of empirical studies. Occup Environ Med 1999;56:361 -77. 14. Green J. The evolving randomised controlled trial in mental health: studying complexity and treatment process. Advances in Psychiatric Treatment 2006;12:268–279. 15. Oakley A, Strange V, Bonell C, et al. Process evaluation in randomised controlled trials of complex interventions. BMJ 2006;332:413-6. 16. Smits PB, de Buisonjé CD, Verbeek JH et al. Problem-based learning versus lecture-based learning in postgraduate medical education. Scand J Work Environ Health 2003;29(4):280-287. 17. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. Validation of performance indicators for rehabilitation of workers with mental health problems. Med Care 2005;43(10):1034-42. 18. Hulshof CTJ, Broersen JPJ, de Haan S. Primary care cooperation guideline on mental complaints and work. (In Dutch: Samenwerkingsrichtlijn 1e lijns handelen bij psychische klachten en arbeid. Evaluatie fase 2: praktijktest). PARAG/SKB. Utrecht/Amsterdam, 2002. 19. Rebergen DS, Hoenen JHAJ, Heinemans AMEC, Bruinvels DJ, Bakker AB, van Mechelen W. Adherence to mental health guidelines by Dutch occupational physicians. Occup Med 2006;56(7):461-468. 20. Buijs PC, van Dijk FJH, Evers M, van der KLINK JJL, Anema H (2007) Managing Work-Related Psychological Complaints by General Practitioners, in Coordination with Occupational Physicians: A Pilot Study. Ind Health 2007;45:37-43. 21. Houtman I, Jettinghoff K, Brenninkmeijer V, van den Berg R. Work stress in the police force five years later: the effect of sectored agreements on (stress) management. [in Dutch] TNO Quality of life. The Netherlands: Hoofddorp, 2005. 22. van der Klink JJL, Blonk RW, Schene AH, Dijk FJ van. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomized controlled design. Occup Environ Med 2003;60(6):429-37.

PROCESS EVALUATION

129

23. Terluin B, van Marwijk H.W., Adèr H.J., et al. The Four-Dimensional Symptom Questionnaire (4DSQ): a validation study of a multidimensional self-report questionnaire to assess distress, depression, anxiety and somatization. BMC Psychiatry 2006;6(34). 24. van der Klink JJL, Terluin B. Mental health problems and work: guide for an activating management by general practitioner and occupational physician. [in Dutch] The Netherlands: Houten, 2005. 25. CPA (Commissie Psychische Arbeidsongeschiktheid). Guideline approach on work disability due to mental health problems. [In Dutch] The Netherlands: The Hague, 2001. 26. de Vet HC, Heymans MW, Dunn KM, Pope DP, van der Beek AJ, Macfarlane GJ, Bouter LM, Croft PR. Episodes of low back pain: a proposal for uniform definitions to be used in research. Spine 2002;27(21):2409-16. 27. Wasiak R, Young AE, Roessler RT, McPherson KM, van Poppel MN, Anema JR. Measuring return to work. J Occup Rehabil 2007;17(4):766-81. 28. Lovibond SH, Lovibond PF. Manual for the Depression Anxiety Stress Scales (DASS). University of New South Wales, 1993. 29. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. The Depression Anxiety Stress Scales (DASS): detecting anxiety disorder and depression in employees absent from work because of mental health problems. Occup Environ Med 2003;60:i77. 30. Schaafsma FG. Evidence based medicine in occupational health care. Thesis. University of Amsterdam, Amsterdam Medical Centre. The Netherlands, 2007. 31. Jamtvedt G, Young JM, Kristoffersen DT, O'Brien MA, Oxman AD. Audit and feedback: effects on professional practice and health care outcomes. Cochrane Database Syst Rev 2006;19(2):CD000259. Update of: Cochrane Database Syst Rev 2003;(3):CD000259. 32. de Koning JS, Smulders AW, Klazinga NS. Appraisal of Indicators through Research and Evaluation (AIRE) Instrument. The Netherlands Amsterdam, Academic Medical Center. 2006. 33. Grimshaw JM, Thomas RE, MacLennan G, et al. Effectiveness and efficiency of guideline dissemination and implementation strategies. Health Technol Assess 2004;8:iii–iv. 1-72. 34. Hugenholtz NIR. Managing knowledge in occupational health care. Thesis. University of Amsterdam, Amsterdam Medical Centre. The Netherlands, 2008. 35. Dzewaltowsk DA, Glasgow RE, Klesges LM, Estahrooks, PA, Brock, BS. RE-AIM: Evidence-Based Standards and a Web Resource to Improve Translation of Research Into Practice. Ann Behav Med 2004;28(2):75-80 36. NVAB. Van der Klink J, ed. Guideline for Mental Health Problems. Revision 1

st edition 2000 [in

Dutch] Utrecht: NVAB (Netherlands Society of Occupational Medicine), 2007. 37. Cates JR, Young DN, Bowerman DS, Porter RC. An Independent AGREE Evaluation of the Occupational Medicine Practice Guidelines. Spine J 2006;6(1):72-7. 38. Madan I, Harling K. The NHS Plus evidence-based guideline project. Occup Med 2007;57(5):307-10. 39. Eccles MP, Grimshaw JM, Johnston M, et al. Applying psychological theories to evidence-based clinical practice: Identifying factors predictive of managing upper respiratory tract infections without antibiotics. Implement Sci 2007;2:39. 40. Murta SG, Sanderson K, Oldenburg B. Process evaluation in occupational stress management programs: a systematic review. Am J Health Promot 2007;21(4):248-54.

CHAPTER 7

Adherence to mental health guidelines by Dutch occupational physicians

Published as: DS Rebergen, Hoenen JAHJ, Heinemans AMEC, Bruinvels DJ, Bakker AB, van Mechelen W.

Adherence to a national guideline on mental health problems by Dutch occupational physicians.

Occup Med 2006;56:461-468

GUIDELINE ADHERENCE

132

ABSTRACT Background In 2000, the Netherlands Society of Occupational Medicine (NVAB) published a national guideline for the management of employees with mental health problems. Objectives To examine predictors of adherence to this guideline by Dutch occupational physicians (OPs). Methods Using the Theory of Planned Behaviour, a questionnaire was developed about self-reported guideline adherence of OPs and possible predictors of this behaviour. A total of 165 OPs were approached to complete the questionnaire and registration forms of first consultations of workers with mental health problems. Performance indicators based on the guideline were developed to calculate performance rates of guideline adherence by OPs. Results Eighty of 165 (48%) OPs approached completed the questionnaire. Fifty-six OPs returned one or more registration forms, totalling 344 consultations. On a five-point Likert scale, ranging from never (1) to always (5), the mean score on self-reported guideline adherence was 2.35, compared to a mean score of 4.06 on the intention to comply with the guideline. The mean performance rate of OPs ranging from 0 to 2 was 1.27 on diagnosis and 0.60 on guidance. No relation was found between self-reported guideline adherence and performance rates. Self-reported guideline adherence correlated significantly with perceived behaviour control (r = 0.48, P < 0.05), subjective norms (r = 0.33, P < 0.05) and positive job stress (r = 0.35, P < 0.05). Conclusions Guideline adherence by Dutch OPs lags behind its acceptance. Further implementation efforts need to focus on diminishing barriers and enhancing social norms of OPs to work according to the guideline.

GUIDELINE ADHERENCE

133

INTRODUCTION Since 1999 the Netherlands Society of Occupational Medicine (NVAB) has been developing and disseminating evidence-based practice guidelines, as they are one of the most promising and effective tools for improving the quality of occupational healthcare.[1-3] In 2000 a national guideline was published regarding the management of employees with mental health problems.[4,5] A main reason for addressing this topic in a practice guideline was that mental health problems account for one third of all disability benefits in the Netherlands. It is shown that a majority of these employees are at risk for chronic disability, while they suffer from minor reversible psychiatric disorders.[6,7] As Dutch employees are required to visit their occupational physician (OP) for rehabilitation purposes, the Dutch OP has an optimal opportunity to find the people at risk and influence recovery.[5,7] The current practice guideline gives OPs recommendations on rehabilitation activities which are successful in facilitating return to work [8-10]. This suggests that rigorous implementation of the guideline could improve the effectiveness of occupational rehabilitation among workers with mental health problems.

After its publication in 2000, the guideline was sent to all members of the NVAB and several courses emerged to inform OPs about the content of the guideline. However, since 1994 Dutch Occupational health services (OHS) are commercial enterprises that sell their services to companies in contracts that only account for a minimum level of services. This competitive work environment has created a difficult position for OPs, as employees, employers and the management of OHS put pressure on them with different interests. This is reflected by a relatively high level of negative job stress among Dutch OPs, and the perception that they are not able to deliver the necessary quality of care.[7,8,11]

The conditions discussed above show the practical relevance of an evaluation of the actual state of guideline adherence and factors that can be changed in the implementation process. Theoretically this is relevant because research is known about predicting guideline adherence in occupational medicine is lacking. In implementation research guideline adherence has been examined by transforming the Theory of Planned Behaviour (TPB), which predicts the occurrence of a specific behaviour provided that the behaviour is intentional.[12,13] Figure 1 shows the TPB-model and represents three variables which theoretically predict the intention to perform behaviour. To predict whether a person intends to do something, we need to know whether that person is in favour of doing it (attitude); how much social pressure that person experiences (subjective norm) and whether that person feels in control of the action in question (perceived behavioural control). These manifest variables should be related to appropriate sets of salient behavioural, normative, and control beliefs about the behaviour. These intentions, together with perceived behaviour control, are precursors of the actual behaviour. Guideline adherence of OPs is an example of intentional behaviour and can be measured by asking OPs how they assess their own guideline adherence (self-report).[14,15] A theoretically new and more objective way to do this is by monitoring guideline adherence using performance indicators (PIs).[10,16] In addition to the variables in the TPB, job stress of OPs can be used as a possible predictor of guideline adherence, as OPs are at risk of negative job stress. [11,17] Positive job stress can affect performance as well and therefore job stress is added to the theoretical framework of the TPB for guideline adherence in figure 1.[18]

In summary, the aim of this study is threefold: 1) to examine guideline adherence by Dutch OPs by applying the theoretical framework of the TPB; 2) to evaluate OPs job stress as a possible predictor of guideline adherence; 3) to assess OPs guideline adherence by means of newly developed PIs and compare this to self-reported guideline adherence.

GUIDELINE ADHERENCE

134

METHODS One hundred and sixty five occupational physicians (OPs) were approached between April and October 2001 to participate in a cross sectional study. The OPs were working for 11 units, belonging to six major occupational health services (OHS), in which a central key figure encouraged OPs to participate in the study. To each OP a questionnaire and fifteen patient registration forms were sent with the request to complete and return them within a month. OPs were asked to complete one registration form for each employee with a mental health problem after their first consultation. Face-validity and feasibility of the forms were tested in daily practice by five OPs in a small pilot. During the registration period, OPs were not informed about specific performance indicators (PIs) and corresponding criteria used in the study. Returned registration forms were only included in the study if the OP diagnosed a psychological disorder into one of the four categories provided by the guideline.

A questionnaire (92 items) was used to collect data on demographic characteristics of the OPs and their experience with the guideline (22 items), on self-reported guideline adherence and its possible predictors according to the TPB (40 items) and on job stress of the OP (30 items).

To transform the items of the TPB for guideline adherence, Ajzen’s most extensive coverage of applying TPB was used.[12] One example of the used items, the number of items and the internal consistency of the constructs are shown in table 1. Some items could only be answered if the respondent had some knowledge of the guideline content. In those cases the option ‘no opinion’ was added. For each construct a mean score was calculated.

Guideline adherence of OPs was measured by two items about use and application of the guideline. Intention represented their plan, desire and prediction to start or keep using the guideline.

Attitude towards guideline adherence contained seven evaluative dimensions on the item: ‘I consider working in accordance with the guideline meaningful / intelligent / fine / skilful / good / necessary / important’.*12+ Behavioural beliefs consisted of expectations of guideline adherence on quality of care, transparency of professional behaviour, evidence based practice and job satisfaction.

Subjective norm represents OPs expectations whether managers, colleagues, employees and employers want them to comply with the guideline. Normative beliefs consisted of the general motivation to comply with these expectations. Perceived behaviour control represents the OPs perception of the ability to perform a given behaviour and is determined by control beliefs about the presence of factors that may facilitate or impede performance of the behaviour. These factors were time, training, work organisation and contracts with employers (companies).

Negative job stress was measured by the Utrecht Burnout Scale (UBOS), which measures three negative job stress factors: emotional exhaustion, cynicism and diminished professional efficacy.[19,20] Positive job stress was measured by the Utrecht Engagement Scale (UBES), which measures three positive job stress dimensions: vigour, dedication and absorption. Each dimension consists of 5 items, describing a state of mind which refers to the work situation (table 1). One must fill in how often this state of mind is experienced on a 7-point Likert scale ranging from never (0) to daily (6). For each job stress dimension a mean score was calculated. Internal reliability of each dimension was satisfactory (Cronbach alpha (α) > 0.75), except the one representing absorption (α=0.65). Factor analysis showed that (diminished) professional efficacy in the UBOS measured the

GUIDELINE ADHERENCE

135

same as the positive job stress dimensions in the UBES. Therefore this dimension was transferred from the negative to the positive job stress construct (table 1).

Table 1 Constructs of the Theory of Planned Behaviour (TPB) and job stress TPB item example n Guideline adherence Do you use the guideline in the management of patients?**

How often do you apply the guideline in practice?** 2 0.93

Intention I intend to use or keep using the guideline in the near future.* 3 0.71 Attitude towards the behaviour

I consider working in accordance with the guideline necessary.*

7 0.73

Behavioural beliefs I expect that care for patients with mental health problems will be improved with this guideline.*

7 0.71

Subjective norm Close colleagues expect me to comply with the guideline.* 5 0.86 Normative beliefs I want to work according to what my colleagues expect me to

do.* 4 0.88

Perceived behaviour control

I am able to organise my work in such a way that guideline adherence is possible.*

7 0.78

Control beliefs Learning to adhere to the guideline takes more time than I have.*

5 0.76

Negative job stress item example n Exhaustion I feel my self exhausted by my work. *** 5 0,85 Cynicism I have became more cynical about the effects of my work. *** 5 0,75 Positive job stress item example n Professional efficacy I know how to solve problems in my work. *** 5 0,82 Vigour In the morning I look forward to going to my job. *** 5 0,81 Dedication My work inspires me. *** 5 0,90 Absorption When I work I forget everything around me. *** 5 0,65 * Likert scale: 1 – 5: Totally disagree – Totally agree ** Likert scale: 1 – 5: Never – Always *** Likert scale: 0 – 6: Never – Daily

Registration forms (47 items) were developed to gather more objective data on

guideline adherence by OPs. The registration forms were constructed in such a way that information of two central elements of the guideline was gathered and could be evaluated. To evaluate this information in an objective way, four performance indicators (PIs) were derived from these two elements.

For each PI, criteria were developed based on if-then logic, which are shown in table 2.[16] These criteria depended on the diagnostic classification by OPs into one of the categories of the guideline. On each of the four PIs, one point was scored if the criteria of the guideline were met according to logical decision rules designed by the researchers. For each consultation with a different patient an OP could thus score two points on diagnosis and two points on management, resulting in a total score between 0 (no compliance) and 4 points (maximum compliance). For each OP the mean score of PIs was calculated by dividing the sum of the scores of the registration forms by the number of completed registration forms.

GUIDELINE ADHERENCE

136

Table 2 Performance Indicators for guideline adherence Accuracy of diagnosis PI 1 Assessment of symptoms Criteria - IF Adjustment disorder: at least one psychological distress symptom noted

- IF Depression: at least one essential symptom and 5 depressive symptoms noted - IF Anxiety: at least one anxiety disorder should be noted on registration form

PI 2 Problem orientation Criteria - All: Evaluation of work related problems and curative care should be noted

- IF Adjustment disorder: - Distress: distress within 3 months after stressor, work functioning still intact - Nervous breakdown: within 3 months impairment in social/occupational

functioning - Burnout: one year between onset of chronic stressor and ultimate crisis

- IF Depression: mood disorder, not related to somatic disorder or drug use - IF Anxiety: anxiety disorder, not related to somatic disorder or drug use - IF Remaining psychiatric disorder: personality disorder, no somatic disorder

Quality of management PI 3 Role and focus Criteria - IF Adjustment disorder: OP may act as care manager if problems are work related

- IF Depression, Anxiety or Remaining psychiatric disorder: OP acts as case manager when recovery stagnates

PI 4 Intervention Criteria - IF Adjustment disorder: OP does individual counseling or contacts/refers to general

practitioner (GP) when recovery stagnates - IF Depression or Anxiety: OP refers to GP and may deal with work related stress - IF Remaining psychiatric disorder: OP refers to GP or curative care

The first step of the data analysis included descriptive statistics of the different

variables. The second step was to calculate the objective measured performance rates (registration forms) for each OP and to relate them to the scores on self-reported guideline adherence (questionnaire) by means of Pearson correlation. The third step was testing the constructs of the TPB as predictors of guideline adherence by means of Pearson correlation. A stepwise regression analysis was performed, firstly for guideline adherence on intention and perceived behaviour control and secondly for intention on attitude, subjective norm and perceived behaviour control to test the TPB for guideline adherence. The forth and final step was a regression analysis for guideline adherence on job stress and if this one was significant, for intention and perceived behaviour control on job stress to measure their expected mediating effects. RESULTS 80 Of the 165 OPs returned the questionnaire corresponding to a response rate of 48%. 48% of the respondents was male, the mean age was 43 years. On average the respondents were working as OP for 9 years and working 33 hours a week, of which 60% consisted of patient consultations (40 consultations a week). 58% were registered as OPs and 29% were in training to become an OP. 84% was member of the professional organisation (NVAB), which was disseminating the guideline. Since the publication of the guideline, 68% of the respondents had followed courses or seminars on mental health

GUIDELINE ADHERENCE

137

problems. Of this group, 80% had spent less and 20% had spent more than 17 hours on these courses.

Table 3 shows the mean scores on the different constructs of the Theory of Planned Behaviour (TPB) and job stress. The mean score on self-reported guideline adherence was relatively low compared to the mean score on intention. The mean score on attitude was relatively high compared to the mean scores on subjective norm and perceived behaviour control. The mean scores on the dimensions of positive job stress are relatively high compared to the mean scores on negative job stress.56 of the 80 participating OPs (70%) returned 344 registration forms of consultations in which a mental health problem was classified, thus making it possible to calculate PIs. These 56 OPs had a significantly higher score on NVAB membership (F=27.6, p=0.01), absorption (F=0.02, p= 0.05), attitude (F= 8.5, p<0.01) and intention (F=8.2, p=0.05), than the 24 OPs that did not return any registration form.

Table 3 Mean scores on different constructs of the TPB and job stress regarding guideline adherence Theory of Planned Behaviour n mean sd range Guideline adherence 78 2.35 0.97 1.00 - 4.00 Intention 73 4.06 0.53 2.00 - 5.00 Attitude towards the behaviour 77 3.85 0.49 2.57 - 4.86 Behavioural beliefs 74 3.75 0.51 2.25 - 5.00 Subjective norm 78 2.66 0.83 1.00 - 4.40 Normative beliefs 78 2.79 0.90 1.00 - 4.00 Perceived behaviour control 75 3.07 0.76 1.60 - 4.80 Cognitive beliefs 75 2.91 0.69 1.60 - 4.60 Negative job stress n mean sd range Exhaustion 80 2.07 1.11 0.00 – 5.00 Cynicism 80 1.98 1.10 0.25 – 5.25 Positive job stress n mean sd range Professional efficacy 80 4.21 0.82 2.00 – 5.83 Vigour 80 4.01 0.97 1.40 – 6.00 Dedication 80 4.13 1.00 1.60 - 5.80 Absorption 80 3.80 0.82 1.80 - 6.00

The median of returned registration forms was 10 (range 1-15). On 225 (65%) of

the registration forms the diagnosis adjustment disorder was made (27% distress, 29% nervous breakdown, 9% burnout), on 19% depression, on 9% anxiety disorder and on 7% remaining psychiatric disorder. Table 4 shows the results of the different PIs on guideline adherence (left side of the table) and the number of registration forms with the same score. The mean score on the accuracy of the diagnosis was twice as high as the mean score on quality of management.

The mean scores for each OP on self-reported guideline adherence and the performance rates of the registered guideline adherence showed no significant correlation (r=-0.05). No significant relationship was found between the various predictors and guideline adherence measured by the performance rates.

GUIDELINE ADHERENCE

138

Table 4 Performance rates on guideline adherence (n=56)

Number of registration forms with the same score (n=344)

Performance Indicators mean (sd) median score n (%) Accuracy of diagnosis (Range)

1.27 (0.35) (0-2)

1.20 0 1 2 3 4

8 (2) 130 (38) 148 (43) 52 (15)

6 (2)

Quality of management (Range)

0.60 (0.31) (0-2)

0.50

Total performance 1.76 (0.47) 1.79

Figure 1 Pair-wise correlations (r) in the theoretical model of planned behaviour and job stress regarding guideline adherence; *P , 0.05, **P , 0.01. Figure 1 shows the pair-wise Pearson correlations between the various predictors in the theoretical model of Planned Behaviour and job stress regarding self-reported guideline adherence. Most of the expected relationships according to the TPB showed a significant correlation. No significant correlation was found between guideline adherence and intention, intention and subjective norm and between intention and perceived behaviour control. A significant positive correlation was found between the mean score of the four dimensions of positive job stress and guideline adherence (r=0.35, p<0.05).

Table 5 shows the results of the stepwise regression analysis, executed to examine the causality of the expected relationships in TPB. In the left column the predicting constructs are mentioned. The third column shows that the regression coefficient for intention on attitude is significant. The right column shows that the regression coefficient for guideline adherence on perceived behaviour control is

GUIDELINE ADHERENCE

139

significant. In addition to the expectations of the TPB, significant regression coefficients were found for guideline adherence on attitude and subjective norm.

Table 5 also shows the results of the pair-wise correlations in the TPB between the different determinants of the TPB and job stress with guideline adherence. Negative job stress was significantly negatively, and positive job stress was significantly positive correlated with perceived behaviour control. Additionally, a mediating effect of perceived behaviour control in the relationship between positive job stress (professional efficacy, vigour, dedication) and guideline adherence was found. Table 5 Pair-wise correlations (r) and regression coefficients (β) for the determinants of guideline adherence Theory of Planned Behaviour

Perceived behaviour control

Intention Guideline adherence

Attitude towards the behaviour

0.26* 0.43** ( =0.45**) 0.34**

Subjective norm 0.21 0.15 ( =0.08) 0.33** Perceived behaviour control

- 0.03 ( =0.10) 0.49** ( =0.45**)

Intention 0.03 - - 0.02 ( = -0.01) Negative job stress Perceived

Behaviour Control Intention Guideline

Adherence Exhaustion -0.28* -0.09 0.00 Cynicism -0.28* -0.20 -0.07 Positive job stress Perceived

Behaviour Control Intention Guideline

Adherence Professional efficacy 0.34** 0.20 0.29* Vigour 0.35** 0.21 0.25* Dedication 0.30** 0.28* 0.30** Absorption 0.29* 0.26* 0.40** *P < 0.05; **P < 0.01

DISCUSSION The present study found that 1 year after the publication of national guidelines on the management of mental health problems (in 2000), Dutch OPs had a positive attitude towards the guidelines and in general intended to apply them in practice but compliance with the guidelines appeared to be minimal. In most consultations, OPs did not seem to work in accordance with the recommendations of the guidelines. No relationship existed between guideline adherence measured by the questionnaire and independently measured PIs on diagnosis and management.

This study shows that Ajzen’s theoretical model of Planned Behaviour is useful in assessing and explaining the actual state of guideline implementation. The constructs of the model were consistent, most expected relationships of the model were confirmed and the constructs gave a clear indication of possible barriers in the implementation process.

The theoretical value of this model of behaviour change seems limited, because ‘intention’, the central element of the model, had no significant relationship with the studied behaviour. Possibly, the ASES model, in which barriers and knowledge are

GUIDELINE ADHERENCE

140

expected to moderate the relationship between intention and ‘behaviour’ might open new perspectives [21,22].

The results of this study may have been biased by the low response rate, which may be due to negative expectations of a time-consuming registration procedure. Participating OPs, however, never complained of this aspect of the study and on the contrary, they commented positively on the structuring effects of registering consultations. This response bias may therefore have led to positive selection. OPs who experienced positive job stress and experienced control in their work may have been more willing to participate in the study. This confounding effect may even have been stronger for the 56 OPs who returned the registration forms. These OPs scored significantly higher on the intention to apply the guideline. This may have influenced the external validity of this study, so caution is needed in generalizing the results.

Surprisingly, no relationship was found between guideline adherence measured by the questionnaires and performance based on the registered consultations, although both showed a minimal amount of guideline adherence. The limited number of items used to measure self-reported guideline adherence may have been a source of bias. The data which were collected on guideline adherence by registration forms were limited to central elements of the guideline and probably did not fully represent the consultation behaviour in practice. The criteria used to calculate performance rates may have been too rigid, especially for the registered quality of management. This may have caused a lack of internal and external validity. At the time of our study, other researchers developed and validated different PIs to measure quality of occupational rehabilitation for workers with mental health problems [8,10,23]. The PIs developed in this study have been useful tools in the further development of indicators to measure implementation and effects of guidelines in the field of occupational or public mental health [24,25].

Apart from these methodological considerations, the results of this study show that the implementation process of the guideline is not completed and should be continued, elaborated and improved in the near future. This study gives reason to be optimistic if new interventions are multifaceted, personalized and aimed at the actual barriers perceived by OPs in applying the guidelines. Such implementation strategies have been proven to be effective [26,27]. Interventions for further implementation should focus on (i) improving working conditions and offering personal education for OPs to improve their perceived control in guideline adherence; (ii) creating a social norm enhancing atmosphere: public attention and reminders on guideline adherence by OHS, professional debate in local consensus groups [27] and (iii) audit of medical files for monitoring guideline adherence by means of indicators.

Besides adapting the practice of OPs on the guideline, the content of the guideline itself should be critically reappraised. Clear action lists for each diagnosis, sharply defined referral criteria, attention for collaborating professions and a better layout could facilitate the application of the guideline. At the moment, a new committee is working on a revision of the guideline in the Netherlands, but in other countries such as the United Kingdom initiatives emerge as well [28]. It is important that future research is carried out to prove that implementation of a guideline leads to better outcomes in terms of quality of care and return to work.

GUIDELINE ADHERENCE

141

REFERENCES 1 Grol R. Successes and failures in the implementation of evidence-based guidelines for clinical practice. Medical Care 2001;39:II46-II54. 2 van der Weide WE, Verbeek JH, van Dijk FJ. Relation between indicators for quality of occupational rehabilitation of employees with low back pain. OEM 1999; 56(7):488-93. 3 Hulshof CT, Verbeek JH, van Dijk FJ, van der Weide WE, Braam IT. Evaluation research in occupational health services: general principles and a systematic review of empirical studies OEM 1999;56;361-377 4 NVAB. Guideline for Mental Health Problems. (Ed. J. van der Klink, in Dutch). NVAB (Dutch Association of Occupational Physicians), 2000. 5 Van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scand J Work Environ Health 2003; 29(6):478-87. 6 NKAP. Factsheet 2, Facts and figures about work incapacity due to mental health disorders. Utrecht: NKAP (Dutch Centre of Knowledge of Work and Mind), 2004. 7 Schaufeli WB, Kompier MAJ. Managing job stress in the Netherlands. International Journal of Stress Management 2001;8:15-34. 8 Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, van Dijk FJ. Validation of performance indicators for rehabilitation of workers with mental health problems. Med Care 2005;43(10):1034-42. 9 Van der Klink JJL, Blonk RW, Schene AH, van Dijk FJ. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomised controlled design. OEM 2003;60(6):429-37. 10 Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. OEM 2002;59:0-4. 11 Straaten RP van, Lamme S. Psychosocial workload and feelings of burnout of the Dutch occupational physician. NSPOH (Netherlands School of Public and Occupational Health) 1998. 12 Ajzen I. The Theory of Planned Behaviour. Organizational behaviour and human decision processes 1991;50:179-211. 13 Grol R, Wensing M, Eccles M. Improving patient care. The implementation of Change in Clinical Practice. Elsevier 2005. 14 Adams AS, Soumerai SB, Lomas J, Ross-Degnan D Evidence of self-report bias in assessing adherence to guidelines. Int J Qual Health Care 1999;11(3):187-92. 15 O'Boyle CA, Henly SJ, Larson E. Understanding adherence to hand hygiene recommendations: Theory of Planned Behavior. Am J Infect Control 2001;29(6):352-60. 16 Van der Weide WE, Verbeek JHAM, Dijk FJG van, Hulshof CTJ. The development and evaluation of a quality assessment instrument for occupational physicians. OEM 1998;55(6):375-82. 17 Jex SM. Stress and job performance: theory, research and implications for managerial practice. Sage publications 1998. 18 Orton P. Stress and family physicians. Canadian Family Physician 1995;41. 19 Schaufeli WB, Salanova M, González-Romá V, Bakker AB. The measurement of burnout and engagement: A confirmatory factor analytic approach. Journal of Happiness Studies 2002;3:71-92. 20 Schaufeli WB, Bakker AB. The Utrecht Work Engagement Scale (UWES). Test manual. University Utrecht, Department of Social & Organizational Psychology 2003. 21 Engels JA, Tigchelaar AP, Gulden JWJ van der. Process evaluation of five projects designed to improve cooperation between general practitioners and occupational physicians. Journal of Social Medicine 2003;3:148-153. 22 Meertens R, Schaalma H, Brug j, Vries N de. Determinanten van gedrag. (Determinants of behaviour). In Dutch. Assen 2000.

GUIDELINE ADHERENCE

142

23 Smits PBA, Buisonje CD de, Verbeek JHAM, et al. Is problem-based learning more effective than lecture-based learning in postgraduate medical education. A randomized controlled trial. Scand J Work Environ Health 2003;29(4):280-7. 24 Davis DA, Thomsom MA, Oxman AD, et al. Changing physician performance. A systematic review of the effect of continuing medical education strategies. JAMA 1995; 274:700-5. 25 Grimshaw JM, Shirran L, Thomas R, et al. Changing provider behaviour: an overview of systematic reviews of interventions. Medical Care 2001;39:8. 26 Grol R, Grimshaw JM. From best evidence to best practice: effective implementation of change in patients' care. Lancet 2003 11;362(9391):1225-30. 27 Baker R, Reddish S, Robertson N, Hearnshaw H, Jones B. Randomised controlled trial of tailored strategies to implement guidelines for the management of patients with depression in general practice. Br J Gen Pract 2001 51(470):737-41 28 Waddell G, Burton AK. Concepts of rehabilitation for the management of common health problems. 2004, The Stationery Office. ISBN 0117033944.

CHAPTER 8

GENERAL DISCUSSION

GENERAL DISCUSSION

144

INTERPRETATIONS Summarizing the main findings

- There is limited evidence that cognitive behavioural therapy (CBT) may facilitate return to work (RTW) and improves the mental health of workers with adjustment disorders (stress-related mental health problems).

- Guideline-based care (GBC) of workers with common mental health problems seems cost-effective compared to usual care with easy access to a psychologist, but does not result in better outcomes in terms of RTW and treatment satisfaction.

- Workers with ‘minor’ stress-related disorders may benefit from guideline-based care by occupational physicians (OPs).

- Workers with ‘more severe’ mental disorders, such as depression and/or anxiety, may benefit from early detection and easy access to a psychologist.

- There is no difference in guideline adherence between OPs providing GBC, and OPs providing care with easy access to a psychologist. In both scenario’s, guideline adherence is equally low, i.e. 50 %.

- Optimal guideline adherence, measured by performance indicators, is not related to RTW.

- Further implementation efforts of guideline-based care need to focus on reducing barriers and enhancing positive job stress (engagement) and social norms of OPs to work according to the guideline.

Interpretations of the CO-OP study results In the CO-OP study, the randomised controlled trial (RCT) described in this thesis, the effects of guideline-based occupational health care (GBC) for mental health problems were evaluated. Specifically, the guideline of Dutch OPs was examined, with RTW as a primary outcome. In this trial effectiveness, instead of efficacy was studied, as there were no ideal circumstances and the two research conditions were not highly contrasting. Therefore, the ‘triad’ of effectiveness was used to interpret the results, which is a conceptual framework for the interpretation of trials. According to the ‘triad’ of effectiveness, findings always result from the interaction between the intervention-giver (GBC-OP), the intervention-receiver (GBC-worker), and the intervention itself (Huibers et al., 2004; Jellema et al., 2006). The lack of effectiveness in our trial can be explained by a number of findings related to this triad.

Firstly, participating OPs were not able to deliver the intervention (GBC) properly. A process evaluation showed that guideline adherence was low (50%) in the intervention group, and was not different than guideline adherence in ‘usual care’ (UC). Secondly, GBC as delivered in our intervention, seems cost-effective but was not effective in the treatment of workers with mental health problems on RTW and treatment satisfaction. Thirdly, a per-protocol analysis showed that high guideline adherence by the OP in the GBC group did not result in faster RTW of this subgroup of workers, if compared to UC.

GENERAL DISCUSSION

145

Thus, GBC is cheaper but not more effective than UC. Training in GBC does not lead to higher guideline adherence compared to UC, and optimal guideline adherence does not fasten RTW. UC, which is more expensive mainly due to more referrals to secondary care psychologists, seems more effective than GBC for workers with ‘severe’ mental health problems (depression/anxiety) and/or ‘executive’ functions (on street in stressful situations). These results do contribute to a more optimal treatment policy for guideline-based occupational health care for workers with mental health problems. However, the theoretical contribution of GBC (e.g. the CBT-elements) remains unclear as the intervention did not determine higher guideline adherence or treatment outcomes compared to UC.

These findings should be interpreted in relation to our choice of design. An advantage of our design is that we were able to deliver realistic treatment outcomes, including an economic evaluation. A disadvantage of effectiveness studies is that they are not informative if the results are negative: if it does not work in daily practice, it might still be likely that the intervention will be of value under ideal circumstances. Hence, the results of an efficacy study might still be positive. Overall evidence Until 2008, eight RCTs were published about the effects on RTW of interventions in primary and occupational health care on stress-related mental health problems (Bruinvels et al., submitted). The first one was the publication by van der Klink et al. (2003), a cluster-randomised controlled trial on which the guideline of OPs is based. This study involved an activating approach by OPs in one company, which appeared to be effective for workers with adjustment disorders when compared to a passive UC. Differences in effects with our study can be explained by the cluster-randomised design. It is fair to assume that there was less contamination between the two research conditions, as care by trained OPs was contrasted with passive usual care of different OPs. There were more ideal circumstances, as there was an ‘in-company setting’ of the occupational health service (OHS), the OPs and the intervention and there was no easy access to a psychologist in UC. Another factor might be the time setting of the study, as in 1996 the activating element focussing on RTW was completely new. Additionally, their study sample consisted of less ‘severe’ symptoms (only adjustment disorders), while our study population consisted of workers with depression and anxiety as well. Results of our pragmatic study may have a better generalization to routine practice. Our results do support the implementation of the guideline, because GBC appeared to be economically in favour of UC. Comparable to the results of van der Klink, GBC seemed to be more effective than UC for workers with ‘minor’ mental health problems. However, UC seemed more effective for workers with ‘more severe’ mental health problems, such as depression and/or anxiety.

Results of another randomised controlled trial by Blonk et al. (2006) puts the UC of our study into another perspective. In a 3-arm study the effectiveness of different treatment options were compared. The first group received extensive cognitive behavioural therapy (CBT), delivered by psychologists working for a commercial centre, which was comparable to the psychologists in our control group. The second group received a combined intervention of brief CBT and workplace interventions and was delivered by ‘labour experts’. Both treatments were compared to UC, which consisted of two brief sessions with a GP. Significant effects on RTW were found in favour of the combined intervention by labour experts. These results indicate that an intervention focussing on both the individual worker and the workplace could be more effective than both our treatment groups. Still, we have to take into consideration that the study population of Blonk et al., self-employed, received different occupational health care and

GENERAL DISCUSSION

146

that only workers with work-related complaints were included. Recently, de Vente et al. (2008) did not find significant effects on RTW of a CBT-intervention provided individually or in a group by a psychologist, compared to UC, among workers on sick leave due to work-related stress complaints.

Meanwhile, another trial in primary care showed that a combined intervention by social workers supporting general practice was not effective (Brouwers et al. 2006). Two other trials delivering minimal interventions by GPs on patients with unexplained fatigue symptoms and common mental disorders, respectively, appeared to be ineffective (Leone et al. 2006, Bakker et al. 2007). However, the last one seemed effective when the GP diagnosed a stress-related mental disorder.

Also, the results of two Norwegian trials on interventions by the national social security offices were published (Nystuen et al. 2006, Fleten et al. 2006). A voluntary solution-focused follow-up offered by a local social security office and psychologists did not result in a substantial participation and beneficial results on RTW. However, a minimal postal intervention sent to the sick-listed persons, appeared to be effective on RTW for the subgroup suffering from mental health problems. The minimal postal intervention consisted of a letter containing a brief orientation of the opportunities to return to adjusted jobs while keeping sickness benefits, including cooperation on modified work measures between worker and employer.

Taking this evidence of recent Dutch and Norwegian trials in consideration, it seems hard to prove a superior effect on RTW of counseling by OPs and GPs using CBT-techniques among workers with common mental health problems. However, taking the results of this thesis into account, productivity loss can be avoided by relatively minimal RTW interventions in an occupational health care setting, especially for adjustment or ‘stress-related’ disorders (Waddell & Burton, 2004; Seymore & Grove, 2005; The Stress Impact consortium, 2006; Corbiere & Shen, 2006; Briand et al., 2007).

Promising interventions for workers with common mental health problems seem to be the ones that focus on an activating and combined intervention on individual and work aspects, based on CBT principles, in an occupational health care setting (de Boer, 2004; Taimela, 2007; Vahtera & Kivimaki, 2008; van Oostrom et al., 2008; Vlasveld et al., 2008). This has shown to be effective on RTW for workers with depression as well, since it was found that outpatient psychiatric care with adjuvant occupational therapy was more effective than outpatient care alone (de Vries et al. 2002; Schene et al. 2007, Nieuwenhuijsen et al., 2008). METHODOLOGICAL CONSIDERATIONS Strengths and limitations This is the first randomised controlled trial that evaluated the (cost-)effectiveness of guideline-based care (GBC) for mental health problems with RTW as a primary outcome. The pragmatic study design with broad inclusion criteria allowed variation in context, diagnosis and treatment. Since this was reflective of clinical practice in the occupational health care setting instead of ideal circumstances, external validity of our results was enhanced. Although the guideline has been evaluated in a pragmatic setting, many requirements for a high quality trial were met. A representative source population and intervention setting, in which patients were recruited over the same period of time and from the same source population, guaranteed external and internal validity. Losses to follow-up and principal confounders were taken into account, and the study had sufficient power to detect a clinically important effect. A main strength of our trial is that the results

GENERAL DISCUSSION

147

give a pragmatic evaluation of guideline-based care, compared to an intensive usual care. Additionally, objective data provided by the employer and the insurance company, could be used in the analyses on RTW and health care costs. Another strength of this study is that a new set of performance indicators (PIs) was developed and was used for rating purposes in a new study population, by different researchers. This set of 20 PIs was based on a validated set of 10 PIs.

This study represents a further step in the evaluation of guideline-based care for workers with common mental health problems. Results of this study contribute to treatment options in occupational health practice for workers on sick leave due to mental health problems. Additionally, evidence-based recommendations are proposed for new and better-suited guidelines. External validity Main comments on our study regarding external validity may be related to our study design, including the content of UC, and the lack of contrast between GBC and UC. Elaborations on these issues, and regarding our study population, are presented below. Content of usual care The high referral percentage to a psychologist reflects some of the occupational health care practices in recent years in The Netherlands. However, this UC does not seem to reflect UC in most other (western) countries. Occupational health care, especially on mental health problems, has a long history in The Netherlands due to our social legislation system and related societal impact, and has been more developed and evidence-based than in most other countries. UC, with easy access to commercial multidisciplinary rehabilitation psychologists paid by employers, may be regarded as unrealistic to those living in other countries. However, work incapacity due to mental health problems has been recognized more and more worldwide and resulted in guidelines in countries such as the USA and the United Kingdom (AGOEM, 2004; Mackay et al., 2004; Cousins et al., 2004). Referrals to secondary mental health specialists of workers suffering from mental health problems, is not only common in The Netherlands, but is getting more common in Western countries such as the USA as well (Wang et al., 2007; Taimela et al., 2007; Nystuen et al., 2006; Salmela-Aro et al., 2004; Perski & Grossi, 2004; Bower & Rowland, 2006; Gilbody et al., 2006). Therefore, we expect that the external validity of these results will grow, as mental health and occupational health care will integrate increasingly in the near future in a broader international context.

Lack of contrast between guideline-based and usual care Since randomisation was done at the individual level, OPs who were trained in the guideline treated all participants. Obviously this situation created a risk of treatment contamination (lack of contrast) between the groups. The trained OP was initiated to provide GBC to a worker in the intervention group. The same OP was initiated to treat a worker in the UC group with minimal involvement and if applicable, direct referral to a psychologist. A cross-over learning effect may have happened in UC, as the OPs appeared to have the same level of guideline adherence in UC as in the GBC group. The other way around, the OP could refer a worker in the GBC group to a psychologist as well. The guideline promotes this in case of stagnation in recovery or in case of severe mental health problems of the worker (NVAB, 2000; van der Klink, 2003). Although only 17% in the GBC group were immediately referred to a psychologist, still 46% of the workers in the GBC group received psychological treatment of a psychologist and/or psychiatrist during the follow-up period of 1 year.

GENERAL DISCUSSION

148

However, we tried to maximize the contrast by creating a situation in which referral to the psychologist in UC was pre-authorized and always granted by the health insurance company (DGVP). In UC 82 % of the participants were immediately referred to a ‘funded’ psychologist, while 85% in UC received treatment of a psychologist and/or psychiatrist. Additionally, results on related PIs indicate that this created a difference between the treatment groups in referral patterns and ‘referral delay’; i.e. the duration between inclusion and referral as well as the duration between referral by the OP and first contact with a psychologist. The percentage of referrals was significantly higher in UC compared to GBC, and referral delay appeared to be significantly higher in the GBC group compared to UC. However, the treatment by the psychologist in UC regularly started after the moment of full RTW and thus the impact of the referral should be doubted with regards to RTW of the worker (table 4). In UC, OPs referred more often to company social workers, who may have taken over the counseling role of the OP.

The intended and expected minimal involvement by the OP in UC did not seem to happen, because no differences in guideline adherence were found between the groups. This may have been due to the introduction of the ‘Gatekeeper Improvement Act’ as well, as OPs were obliged to have regular contact with worker and employer. Our design and these circumstances may have resulted in contamination of treatment in both groups, which may have contributed to an absence of effects of GBC, when compared to UC. Other explanations are the possible inapplicability of the trained skills, or an insufficient training in guideline-based care. However, Smits et al. (2003) showed that a comparable training of the guideline for OPs-in training resulted in an improved guideline-adherence.

Our results may have been biased also by the circumstance that the largest participating police department changed its occupational health service (OHS) during our study. As the OPs in this new OHS could not start immediately and worker, employer and OP had to adapt to this new context, this situation may have had a negative influence on the treatment in the GBC group. In UC these circumstances may have had less impact since workers had more chance to have a consistent therapy by a psychologist, besides the changes in their occupational health care.

Study population The study population, Dutch police workers, showed a higher risk of getting into stressful situations than other workers (Houtman et al., 2005; Slottje et al., 2008). This has been caused by a relatively high workload and by emotional pressure. To a certain extent this reflects that police workers have other occupational risks than the general working population. As the study population was not fully representative of the general working population, external validity of the results may be limited and caution has to be taken in generalising the results. However, this disadvantage regarding the ability to generalise our study results does not outweigh the advantages of this study population. The police are an organization with a relatively high incidence of common mental health problems and are therefore an interesting target population. Another advantage is the uniform sick leave registration of the police. As the police had connection to only one health insurance company, it also had a well-defined ‘usual care’. By applying broad inclusion criteria we tried to produce results with the highest possible external validity. Internal validity Although we faced problems in lack of contrast between the treatment groups, the design of randomised controlled trials positively affects the internal validity of the study. By randomisation, the possibility of bias from unknown confounders is minimised. However, we adjusted our regression models for potential confounders. There appeared to be a

GENERAL DISCUSSION

149

substantial impact of the different participating OPs that delivered GBC or UC, and the severity of the disorder, in confounding the results on RTW. Therefore, we adjusted the models for these (and other less substantial) confounders. We need to mention that there were only 5 participating OPs in the study, and that internal validity could have been better with more participating OPs. The best (contrasting) option would have been a cluster-randomized design, in which trained OPs would have provided GBC and in which OPs without a training provided UC. However, this appeared impossible in the occupational health care setting in which we conducted our study. In ethical sense it was unacceptable to have a UC group with no GBC or no easy access to psychologist, as each worker deserves the best care possible. Other elaborations regarding the internal validity are presented below. Assessment of guideline adherence When assessing guideline adherence by a physician, one can distinguish between direct and indirect methods of quality assessment (Schaafsma, 2007). While direct methods observe clinical performance, indirect methods use information from either medical files or self-report measures of physicians. In this thesis, guideline adherence was assessed by means of an audit of medical files. An audit has the advantage that physicians are unaware that their notes will be used for research purposes (Jamtvedt et al., 2006). However, disadvantage of this approach is that negative findings and routine activities may not have been recorded. The use of self-report forms counters this problem, and is still much less time-consuming than direct observations. However, it has been shown that self-report measures tend to lead to overestimation of guideline adherence (Schaafsma, 2007).

One of the strengths of the CO-OP study is that guideline adherence was audited by a set of performance indicators (PIs), as a process evaluation of the trial. A validated set of 10 PIs was used, which was extended to a 20 PI-set, based on recent literature (Nieuwenhuijsen et al., 2005; van der Klink & Terluin, 2005). Use of the initial 10 PI-set generated some problems, as this set was validated using registration forms, which requires the use of more stringent criteria. An overall score on guideline adherence was based on this initial 10 PI-set. The mean performance rates of our study resembled the performance rates reported elsewhere (Anema et al., 2006; Rebergen et al., 2006; Nieuwenhuijsen et al., 2003 & 2005; Smits et al., 2003; Hulshof et al., 2002). However, no relation was found between optimal guideline adherence conform the initial 10 PI-set and earlier RTW, while Nieuwenhuijsen et al. (2005) did find such a relation. In the 20 PI-set, a relation between optimal guideline adherence and faster RTW was found. This association was mostly explained by the PIs on regular contacts between OPs and workers and/or employers (PI14 and PI17). However, regular evaluation was at the time of the study ‘core business’ of the OP and was further enforced by new legislation. Therefore, we do not interpret this aspect of guideline adherence of additional value to UC, compared to other new elements. Per protocol-analyses confirmed this interpretation, since no difference was found in the relation between GBC in the subgroup characterized by optimal guideline adherence and RTW, compared to the relation between UC and RTW. To our opinion it was necessary to restructure and extend the initial 10 PI-set with an additional 10 PIs to assess contrast in guideline adherence between the two groups and to incorporate recent findings. Although a new set of 20 PIs was developed carefully, validity remains insecure. An instrument in validating and improving the quality of these indicators could be the recently developed AIRE-instrument (de Koning et al., 2007).

GENERAL DISCUSSION

150

Outcome measures The primary outcome measures of this trial were RTW and treatment satisfaction of the main occupational health care stakeholders in the RTW process. Mental health symptoms were only measured at baseline, though this is a frequently used outcome measure in mental health research. Our main focus was recovery of work functioning, as this has mostly been neglected in the field of mental health. As a result, we could not compare the outcome measures of RTW to results at symptom level, or compare results at symptom level to those found in other studies. This may cause some resistance in the acceptance of our results to the clinical field of mental health. However, our trial is one of the few in the field of mental health that focused on facilitating RTW, which is the main focus in occupational health care and received more attention of other disciplines in recent years. The idea that interventions which effectively reduce mental symptoms are effective for improving work-related outcomes has to be doubted, as RTW seems to follow a different pathway in recovery than reduction of symptoms. Recent studies (and results of the in this thesis presented Cochrane review) have shown that RTW does not negatively influence the recovery of symptoms, which is feared by many mental health clinicians (Blonk et al., 2006; Nieuwenhuijsen et al., 2005). Interventions regarding RTW may even enhance recovery of symptoms, although results on this subject remain unclear.

Results on treatment satisfaction confirmed earlier findings by Nieuwenhuijsen et al. (2006) and de Vente et al. (2008) that treatment satisfaction of workers diminishes, if treatment focuses on RTW. An activating treatment facilitating RTW confronts workers with an ‘unfavourable’ situation at forehand, as they feel ‘forced’ to resume work and feel vulnerable in showing their illness symptoms and functioning on the work floor. Surprisingly, however, employer satisfaction did not differ between both treatment groups, while GBC focused more on facilitating RTW and UC was more expensive due to referrals to the psychologist. Another interesting finding is that OPs felt more satisfied with delivering GBC, probably because they experienced more behaviour control and engagement (Rebergen et al., 2006). However, results on outcomes of treatment satisfaction by employers and OPs are difficult to interpret as these outcomes have been hardly measured in other studies.

RTW was defined as the duration of sick leave, which was used as a proxy measure of productivity loss in our cost-effectiveness analysis. By using sick leave days as proxy for productivity loss, we did not take into account work presenteeism (Lerner & Henke, 2008). Mental health problems can influence work presenteeism, as work performance may be suboptimal before and after periods of sick leave (Uegaki et al., 2007; Wang et al., 2007). As this may have biased our findings on cost benefits by reducing productivity loss, future research should focus on these productivity measures as well. However, trials on sick leave data as proxy of productivity loss regarding the mental health field have hardly been published (Brouwers et al., 2006; Schene et al., 2007; Uegaki et al., submitted). This study thus delivers a major contribution to this area, taking objective RTW-data into account.

In conclusion, our intention was to study the effects of the entire current guideline in practice, and it is fair to stat that our results give relevant insight in the overall effect of the guideline for a relevant study sample.

Applications of the ICF model Although no clear answers were found on issues regarding the most effective treatment, this thesis offers new applications of the ICF model (WHO, 2001; Heerkens et al., 2004). In the ICF model (figure 1), introduced in chapter 1, the studied elements of the guideline that proved to have higher adherence than UC and/or seemed promising in enhancing

GENERAL DISCUSSION

151

RTW, are underlined. Process management (regular evaluations with worker and employer) is underlined as it was associated with a faster RTW.

The ICF model offers three opportunities for interventions by OPs (Verbeek, 2006): a) regaining health by means of improvement of mental functioning; b) restoring activities by means of recovery of work performance; c) improving participation by partial or full RTW. Application 1 represents the guideline in that it introduced a simplified diagnostic classification of mental health problems, and a focus on work impairments/limitations/RTW (1). Optimal guideline adherence by the OPs on diagnosis, assessment of work impairments/limitations for RTW and advice on RTW to the worker was not associated with faster RTW. No differences were found on guideline adherence between GBC and UC, except for a significantly higher use in the GBC group of the questionnaire 4DSQ as a diagnostic instrument. Since use of the 4DSQ (or the DASS) may be helpful in differentiating between ‘minor’ and ‘more severe’ mental health problems and seems promising in speeding up RTW, use of the 4DSQ may be promoted (Terluin et al., 2006). In the GBC group, significantly more patients had a stepwise RTW. Apparently, OPs seem successful in incorporating RTW in their treatment. Stepwise RTW may work for the more context-oriented adjustment disorders, while more severe disorders initially need more person-oriented treatment.

The second application (2) advocates process management by a time contingent evaluation regarding work and curative care to prevent disability and facilitate RTW. Guideline adherence by the OP on process management did not show differences between the groups. Optimal guideline adherence on regular evaluations with the worker and the employer predicted faster RTW and, therefore, should be promoted. In the GBC group, the OP contacted the GP more often, although communication between OP and GP was rare.

In the third application (3), the OP may operate as counsellor by improving skills or changing recovery expectations regarding (return to) work or by counseling on personal characteristics, e.g. coping style (van Rhenen et al., 2007). In the GBC group the OP operated significantly more as counsellor (21 versus 6%), although numbers were small. The counseling role of the OP was the focus and most renewing aspect of the new guideline and related training, but uptake of this counseling role does not seem realistic after a 3-day training course.

In summary, potential improvement in mental functioning, recovery of work performance and participation by RTW, seem possible and necessary regarding more accurate detection of ‘severe’ symptoms by OPs using the 4DSQ. Continuous process management regarding worker and employer has already been put into effect by new legislation and has proven to be effective. The role of the OP as counsellor seems only relevant regarding facilitating (gradual) RTW. This is in line with the recent literature, which does not support an increase in the counseling position of the GP (Leone et al., 2006; Bower & Rowland, 2006; Bakker et al., 2007).

Additionally, this evidence underlines the importance of an occupational health care setting in facilitating and speeding up RTW, without negatively influencing recovery of symptoms (Blonk et al., 2006; Taimela et al., 2008). A referral to a psychologist seems beneficial if ‘more severe’ symptoms are detected or stagnation of recovery occurs, and counseling on the individual level becomes more relevant.

GENERAL DISCUSSION

152

IN-DEPTH ELABORATIONS ON THE CO-OP STUDY Our study design focused on effectiveness, because we wanted to know in the first place whether GBC was effective in daily practice for the receivers (workers). GBC appeared to be cost-effective from the societal and company perspective. For the worker, GBC did not seem effective in speeding up RTW and enhancing treatment satisfaction. However, effect modifications on RTW were found regarding the severity of disorder. This means that it is important that the OP is able to accurately diagnose sick-listed workers with mental health problems. Therefore, a more in-depth discussion on the accuracy of diagnosis by the OP is given in the next paragraph. Furthermore, the ‘counseling’ role of OPs for stress- and work-related problems is discussed. As a ‘bridge’ to the conclusions and implications, a brief summary is presented of the opportunities and challenges of evidence-based occupational health care.

Figure 1 New applications of the ICF model (Rebergen, 2008) Accuracy of the diagnosis by the OP Table 1 shows a summary of the main elements of the guideline, and shows the connection between type of diagnosis and the related treatment (role).

Health Condition

Mental health problems

a) Improvement in mental functioning

e.g. reduction fatigue, depressive symptoms

b) Restoring of activities

e.g. recovery of work performance

c) Improvement in Participation

e.g. return to work

Personal Factors e.g. age, gender, coping style

or recovery expectations

Environmental factors e.g. supervisor behaviour or

work characteristics

1. Problem orientation Diagnosis (4DSQ)

Work focused treatment

2. Process manager Evaluation

patient/work/GP Adapt work environment

3. Individual interventions

Counseling stress/work Support employment

GENERAL DISCUSSION

153

Table 1 Main elements of the guideline for OPs on mental health problems Diagnosis Treatment Adjustment (‘stress-related’) disorder: Distress, nervous breakdown, burnout

OP: case (‘process’) and care manager (‘counsellor’) In case of stagnation referral to GP/psychologist

Other mental health problems: Depression, anxiety, psychiatry

OP: case (‘process’) manager Communicates/refers to GP/psychologist

The OP may act as care manager by counseling workers with work-related problems for both diagnoses, but only operates as care manager of stress-related problems if an adjustment or stress-related disorder is diagnosed. If not, the OP needs to communicate with the GP and may refer, if necessary and appropriate, to a psychologist. This is a clear distinction with the UC in our study, in which an immediate referral to a psychologist was instigated for each subject. As this distinction appeared to be important for further treatment, an accurate diagnosis by the OP was deemed necessary. Therefore, we checked the accuracy of the diagnosis by the OP in different ways. Chapter 6 briefly described the results of different PIs on problem orientation and diagnosis. As medical files may be insufficient to reveal the real process of detection of symptoms, we made a comparison between the diagnosis of the OP with the diagnosis based on the DASS-scores and the diagnosis of the psychologist.

In this discussion the Depression Anxiety Stress Scales (DASS; Lovibond & Lovibond, 1993) is used as a ‘golden standard’ to evaluate the diagnosis made by the OP, according to cut-off scores used by Nieuwenhuijsen et al. (2003) in a comparable population. Table 2 shows that OPs diagnosed more often stress disorders in comparison with the DASS-scores. This was caused mainly by the finding that OPs diagnosed ‘only’ stress symptoms for participants who had depression, anxiety or remaining psychiatry according to the DASS. The detection of the diagnosis adjustment or stress-related disorder by the OP appeared to have a high sensitivity (82%) and a low specificity (24%). The results indicate that OPs have problems with recognizing depression and/or anxiety at an early stage. The same holds for comparison of diagnoses made by the OP and those available made by the psychologist (table 3), although the diagnosis made by the psychologist was assessed later.

Table 2 Diagnoses DASS and OP compared (n=208) Diagnosis DASS stress DASS other Total OP stress 115 (55%) 52 (25%) 167 (80%) OP other 25 (12%) 16 (8%) 41 (20%) Total 140 (67%) 68 (33%) 208 (100%)

Table 3 Diagnoses Psychologist and OP compared (n=68) Diagnosis Psychologist stress Psychologist other Total OP stress 44 (65%) 13(19%) 57 (84%) OP other 5 (7%) 6 (9%) 19 (16%) Total 49 (72%) 11 (28%) 68 (100%)

Problems among OPs in the making of an accurate diagnosis were noticed also in other studies (Anema et al., 2006; Nieuwenhuijsen et al., 2003 & 2005; Hughes et al., 2008). An explanation can be that the guideline is still not well implemented and/or that a 3-day training course is not sufficient to learn to work according to the guideline. Another reason can be that OPs do not have sufficient time in their consultations to come to an

GENERAL DISCUSSION

154

accurate diagnosis. Although the guideline recommends a first consultation of at least 30 minutes, the contracts of the commercially operating Dutch occupational health services often do not provide such time. A possible solution to this could be the use of mental health questionnaires by the OP. Therefore, OPs can be advised to use the DASS (or the 4DSQ) in the diagnostic process to alert the OP to all possible cases of anxiety disorder and depression (van Rhenen et al., 2008). Another option is to incorporate the expertise of a psychiatrist or a psychologist in the diagnostic process, although this may prolong the outcome. Is counseling by OPs the most effective solution? As mentioned in table 1, diagnosis was related to treatment (role). The treatment role regarding the connection between diagnosis and interventions contained two elements. The first element was the role of the OP as case or process manager, and the second element was the role of the OP as care manager, in which the OP may operate as a counsellor. Mean performance by the OP for the role as process manager was weak to moderate, did not show differences between the groups, but was associated with faster RTW. In the GBC group the OP operated significantly more often as counsellor and contacted the GP significantly more often. However, both ‘counseling’ and ‘contact with the GP’ were rarely performed by the OP.

The guideline stimulates the OP to counsel workers on work-related issues and workers with adjustment disorders on stress-related issues. Results of our study confirm the positive (cost-)effects on RTW of GBC, especially for this large subgroup of ‘minor’ mental disorders. However, counseling by OPs does not seem the cause of these positive (cost-) effects for this subgroup, when compared to UC.

For the group of workers with more severe mental health problems, the guideline is more consensus- than evidence-based. When ‘more severe’ mental health problems are diagnosed or recovery stagnates, the guideline prescribes referral to the GP and to a specialist in secondary care. For each worker with mental health problems, contact with the GP is instigated. As contact with and referrals to GPs by OPs rarely occurred in our population, results of other studies on this subject were confirmed (Houtman et al., 2002; Anema et al., 2006; Buijs et al., 2007). From this perspective, and based on the finding that contact between OP and GP showed promising effects on RTW for the subgroup, further communication between GP and OP should be encouraged.

In primary care, workers on sick leave due to mental health problems are not frequently referred to psychologists or other specially trained professionals (20%) (Houtman et al., 2002; Kovess-Masfety et al., 2007). In occupational health care, referrals to psychologists have been more common practice, mainly because employers are growingly keen to invest in these services to avoid productivity loss and related financial consequences. The last decade, commercial psychotherapeutic intervention centres have been developed, in which psychologists treat workers with common mental health problems. These workers are often referred by an OP. These psychologists often use stress-management counseling interventions based on CBT-principles (Blonk et al., 2006). As no conclusive evidence existed about the effectiveness of psychological interventions (e.g. counseling) carried out by physicians, the question remains if this work should be done by other occupational health professionals (e.g. occupational social workers) and/or psychological specialists (van der Klink, 2002).

As mentioned, counseling by OPs was hardly performed, although significantly more often in GBC. This leaves room for improvement, but for now the group of workers with adjustment disorders, counselled by the OP or not, does not seem to benefit from a referral to a psychologist. More important seems, as the results of our review indicate,

GENERAL DISCUSSION

155

that RTW is activated and facilitated by professionals in occupational health care. Although referrals to psychologists in UC were pre-authorized, still ‘referral delay’ seems to happen as can be seen in table 4.

For 68 workers referred to a psychologist of the commercially operating multidisciplinary centre, data were retraceable, 60 in UC and 8 in GBC. For the 8 workers in GBC a referral to the psychologist was done by the OP in a later stage, as was intended by our study protocol. Referral to a psychologist in UC was stimulated by pre-authorization and full funding by the insurance company, while in GBC half of the treatment costs were funded. The duration between inclusion by the OP in UC to an intake by a psychologist was on average one month, while duration between inclusion by the OP to a first session by a psychologist was almost two months. As median partial RTW for both groups was 50 days, OPs facilitated RTW already before psychologists were able to counsel participants.

Table 4 Referral delay OP-psychologists in calendar days. The cumulative results show the number of calendar days added to the ‘state’ before. Duration (days)

Sick-leave registration

Inclusion OP

Referral OP/OHS*

Intake Psy**

1st

session psy**

Duration treatment

N sessions (mean)

N

GBC Cumulative

0 0

15 15

63 78

15 93

15 108

165 273

9.6

8

UC Cumulative

0 0

23 23

18 41

12 53

26 79

133 212

7.6 60

* OHS = Occupational Health Service ** Psy = Psychologist of commercial provider, fully funded by the insurance company in UC

A different situation seems to occur if the diagnosis is ‘more severe’ and/or

recovery stagnates. Then, the beneficial influence of a psychological treatment may occur, since psychologists take over the care management. However, we should mention that psychologists in UC were working for a multidisciplinary rehabilitation centre. Therefore, these psychologists may have been more focussed on RTW in their treatment, than the average psychologist in primary and secondary care. Still, accurate detection of ‘more severe’ symptoms and immediate referral to a psychologist seem important for this subgroup. However, a counseling and process management role by the OP regarding the work situation and work-related stress remains necessary, as psychologists may not incorporate this in their treatment. Communication and cooperation between the OP and the psychologist is essential, both to divide treatment role and responsibility, and to evaluate results of recovery of symptoms and RTW. Improvement of guideline-based care in occupational health care In the field of occupational evidence-based medicine, guideline-based care has been embraced by prominent OPs and researchers, and it is now starting to reach daily occupational health practice. However, if we compare our results on guideline adherence for OPs on mental health problems to the studies by Smits (2002), Nieuwenhuijsen (2004) and Anema et al. (2006), no major differences were found. Results show equally weak performance rates on diagnosis, interventions and continuity of care. Equally high performance rates were found for work-related issues as evaluation of work causes and limitations in return to work. Although efforts need to be done on the measurement of guideline adherence in the field of occupational health care, implementation of the existing evidence-based guidelines may need adaptations to this field. More specific for occupational health care is the strong influence from legislation, from management of

GENERAL DISCUSSION

156

occupational health services and the presence of different stakeholders, i.e. workers, employers and sometimes insurance companies.

The study described in chapter 7 of this thesis showed that guideline adherence by Dutch OPs lagged behind its acceptance. Results of a recent thesis by Schaafsma (2007) confirmed these results in the broader perspective of evidence-based medicine (EBM). Most OPs were having a positive attitude towards EBM, but were aware of only a small portion of their actual information needs. This positive attitude and small awareness did not result in actual EBM practice. The most common way was to consult colleagues, while scientific databases were rarely consulted. Results of chapter 7 in this thesis revealed that self-reported guideline adherence correlated significantly with perceived behaviour control. Correspondingly, ‘no time available’ was an important reason not to search for literature in Schaafsma’s study, but to ask a colleague or an expert for an answer.

Schaafsma found that enhancement of EBM competence did not prove to be a reliable predictor for the quality of actual EBM performance or practice. Besides a stimulating environment, the fact that OPs were obliged to practice EBM made that the quality improved substantially. However, this effect decreased after four months, possibly indicating that the initial enthusiasm would need repeated stimulation or courses to continuously enhance quality. These results correspond with the significant correlation we found (chapter 7) between normative behaviour of important people around the OP and guideline adherence.

Recently, Hugenholtz (2008) concluded in her thesis that continuous medical education consisting of an EBM course combined with recurrent peer group session is effective in enhancing performance of OPs. However, many OPs lack the time for such an intensive intervention. Therefore, an introductory e-learning course on EBM is proposed, which appeared to be as effective as lecture-based learning, but may have a wider range of implementation. More research in this area is needed, to examine if e-learning relates to the implementation of guideline-based care. CONCLUSIONS AND RECOMMENDATIONS Case description (recommended continuation of General Introduction) The worker consults his occupational physician (OP) 3 weeks after he has been sick-listed. The OP diagnoses an adjustment disorder, with indications for an underlying depression. He agrees with worker, employer, and company social worker to continue work parttime with work accommodations, and daily structuring activities (e.g. cycling). During the second consultation 3 weeks later the worker is only focused on his deteriorated mental health symptoms. In consultation with general practitioner (GP) the OP notices stagnation of recovery, and refers the worker to a psychiatrist. OP and GP agree that OP wil be the process (case) manager. The OP informs the employer and advises a structuring work rhythm without productivity norm. During the third consultation 3 weeks later it appears that the worker has consulted the psychiatrist, who has diagnosed a depressive disorder. He has started medical treatment, combined with therapeutic sessions. The OP advises worker and employer to build up easy work activities and informs them about potential side-effects regarding work productivity. One month later the worker has regained control on activities and social interactions. The worker feels better now and is able to conduct self-instigated problem-solving activities. The treatment with the psychiater works well, and he has started running again. In the fifth consultation with the OP, five months after the start of the sick leave, the worker appears to have major improvements in his mental health state and work functioning. The OP advices full RTW and elaborates on ways to

GENERAL DISCUSSION

157

prevent relapses. One month later the worker is free of symptoms, has fully returned to work and is back at his initial level of functioning. After agreement by the worker, the OP informs GP and psychiatrist about the steps taken. GP and psychiatrist continue this treatment by reinforcing essential elements as structure, (social) activities, and experiments with problem-solving behaviour. Final conclusions In this thesis it was shown that guideline-based care by OPs, reflected by an activating approach facilitating RTW of the worker with mental health problems, is cost-effective in occupational health, if compared to usual care with easy access to a psychologist. These results support the overall evidence that a minimal intervention in occupational health care is (cost-)effective, especially regarding the substantial group of workers suffering from ‘minor’ mental health problems. Considering workers with ‘more severe’ mental disorders, such as depression and/or anxiety, early detection and more extensive combined interventions by occupational experts and/or specialised secondary care interventions by psychologists may be more effective. In a process evaluation, no contrast in guideline adherence by the OP was found between GBC versus UC. The per-protocol analysis showed that high guideline adherence by the OP in the GBC group did not result in faster RTW of the worker compared to UC. However, it remains unclear if these results are due to limitations of our study, the content of the guideline, or the implementation of the guideline in our intervention. We conclude that the guideline needs revision into a less ambitious version with more attention for detection and treatment of the more severe mental health problems. Guideline-based care may be improved by more rigorous implementation efforts, than the 3-days training course in the intervention of this study. Continuous training and supportive legislation by the relevant stakeholders may enforce the perceived behaviour control by OPs, which appeared to be associated with higher guideline adherence.

Revision of the guideline Recently, the guideline for OPs has been revised (NVAB, 2007). A critical reappraisal of the guideline appeared to be necessary since it was more consensus-based than evidence-based. Another reason was that the guideline seemed to be too ambitious in changing the OP’s performance in the current setting. Using the AGREE-criteria (Hulshof & Hoenen, 2007), revision of the guideline resulted in a less complex and less ambitious edition, to maximize applicability. There is a less prominent place of the counseling element. The new guideline promotes some elements that were measured by the extended 20 PI-set we used in our process evaluation. These are: more use of the 4DSQ, continuous evaluation with worker and employer to facilitate RTW, more attention for depression and anxiety, evaluation of and intervention on stagnation, consultation of the GP after 2 months, and (relapse) prevention of mental health problems. Recommendations for future research Although a recent review of evidence on the effectiveness of occupational health interventions showed that high quality evaluation studies have been conducted in all areas of occupational health care, the number of evaluation studies is still small compared to the number of aetiological studies (Ruotsalainen et al., 2006; Verbeek, 2007). As a result, there remains a challenge to conduct more trials on effectiveness, in particular within occupational health care. More research on the effectiveness of mental health

GENERAL DISCUSSION

158

interventions within occupational health care is obviously necessary, preferably from other countries than Norway and The Netherlands.

Furthermore, our studies show the importance of the use and evaluation of occupational health care interventions, e.g. evidence-based practice guidelines and systematic reviews. A good initiative that is currently in preparation is the publication of a list of available and downloadable evidence-based guidelines for occupational health care on the website of the ICOH committee on health services research and evaluation in occupational health (www.icohweb.org). We suggest stimulating research on impairments caused by mental health problems and on RTW-issues in various health care settings, including insurance medicine (Franche et al., 2005).

In this thesis, emphasis was put on the evaluation of the potential added value of guideline-based care.

Results showed that guideline-based care did not result in better

guideline adherence, faster RTW and higher treatment satisfaction, but led to more cost-effective care, compared to usual care. This leaves us with the question of what is the additional added value of training in the guideline. Will a revision of the guideline result in an increase of the effectiveness, if guideline adherence remains this poor? Future research should therefore investigate how guideline adherence can be improved in non-clinical settings, such as occupational health. Answering the question of why guideline adherence is low and acting upon this answer could be of greater influence on the effectiveness of the guideline than changing its contents alone. We expect that future research on behaviour change and (multidisciplinary) guideline implementation may help to answer these questions.

Finally, evidence-based occupational health care practice needs further investigation of output and outcome measures. Prevention of occupational and work-related diseases plays an important role in most countries, instead of sickness absence advice. Ongoing Dutch studies may confirm the need for more work-related interventions of mental health problems, i.e. by work adaptations, collaborative occupational health care or work-related guidelines for psychologists (van Oostrom et al., 2008; Vlasveld et al., 2008; Lagerveld et al., in press). Additionally, studies should investigate the potential effect of evidence-based practice on preventive actions and advices in occupational health care. Recent examples are the studies by Duits et al. (2007 & 2008) and van Rhenen (2007 & 2008), who found promising results of detective screening and preventive interventions on sick leave for workers with mental health problems. There may be potential for internet-based guided self-help and internet-administered cognitive behaviour therapy as well (Cuijpers et al., 2007 & 2008). Recommendations for occupational health care practice

Cognitive behavioural therapy interventions should be used for workers with adjustment disorders, preferably in an occupational health care setting.

Application of an activating guideline-based care by OPs should be promoted, as this is more cost-effective than usual care with easy access to a psychologist.

Workers with depression or anxiety symptoms should be detected in an early stage, and should be referred to pre-authorized secondary mental health care.

The guideline should be revised in a more simple and applicable version, with more attention for workers with severe mental health problems.

Implementation of guideline-based care should focus on higher applicability of the innovative elements of the guideline by OPs, and should be facilitated by continuous training and supportive legislation by relevant stakeholders.

GENERAL DISCUSSION

159

REFERENCES American College of Occupational and Environmental Medicine (ACOEM), Glass LS (editor). Occupational medicine practice guidelines: evaluation and management of common health problems and functional recovery of workers (2nd ed.) OEM Press, Beverly Farms, MA (2004). Anema JR, Jettinghoff K, Houtman ILD, et al. Medical care of employees long-term sick listed due to mental health problems: A cohort study to describe and compare the care of the occupational physician and the general practitioner. Journal of Occupational Rehabilitation 2006;16:41-52. Bakker IM, Terluin B, van Marwijk HW, van der Windt DA, Rijmen F, van Mechelen W, Stalman WA. A cluster-randomised trial evaluating an intervention for patients with stress-related mental disorders and sick leave in primary care. PLoS Clinical Trials 2007;2(6):e26. Blonk RWB, Brenninkmeijer V, Lagerveld SE, Houtman ILD. Treatment of Work-Related Psychological Complaints: a Randomized Field Experiment among Self-employed. Work & Stress 2006;20:129-144. Bower P, Rowland N. Effectiveness and cost effectiveness of counseling in primary care. Cochrane Database of Systematic Reviews 2006, Issue 3. Art. No.: CD001025. Briand C, Durand MJ, St-Arnaud L, Corbière M. Work and mental health: learning from return-to-work rehabilitation programs designed for workers with musculoskeletal disorders. Int J Law Psychiatry. 2007;30(4-5):444-57. Brouwers EPM, Tiemens BG, Terluin B, Verhaak PFM. Effectiveness of an intervention to reduce sickness absence in patients with emotional distress or minor mental disorders: a randomized controlled effectiveness trial. General Hospital Psychiatry 2006;28:223-229. Brouwers EP, de Bruijne MC, Terluin B, et al. Cost-effectiveness of an activating intervention by social workers for patients with minor mental disorders on sick leave: a randomized controlled trial. Eur J Public Health 2006;17(2):214–220. Bruinvels DJ, Rebergen DS, Verbeek JHAM, et al.. Return to work interventions for adjustment disorders. Submitted to Cochrane Database of Systematic Reviews, Occupational Health Field. Buijs PC, van Dijk FJ, Evers M, et al. Managing work-related psychological complaints by general practitioners, in coordination with occupational physicians: pilot study. Ind Health 2007;45(1):37-43. Cates JR, Young DN, Bowerman DS, Porter RC. An Independent AGREE Evaluation of the Occupational Medicine Practice Guidelines. The Spine Journal 2006;6(1):72-77. Cieza A, Chatterji S, Andersen C, Cantista P, Herceg M, Melvin J, Stucki G, de Bie R. ICF Core Sets for depression. J Rehabil Med 2004;(44 Suppl):128-34. Commissie Psychische Arbeidsongeschiktheid (CPA). Guideline approach on work disability due to mental health problems. (In Dutch). The Hague, 2001. Corbiere M, Shen J. A systematic Review of psychological Return-to-work interventions for people with mental health problems and/or physical injuries. CanJ com mental health 2006;25(2):261-288. Cousins R, Mackay CJ, Clarke SD, Kelly C, Kelly PJ, McCaig RH. ‘Management Standards’ and work-related stress in the UK: Practical development. Work & Stress 2004;18(2):113-136. Cuijpers P, Riper H. Internet-guided self-help for common mental disorders: State of the art and challanges for the future. In Sorbi, M.J.; Ruddel, H.; Buhring, M.E.F.; (Ed.), Frontiers in stepped e-care. (pp. 122-131). Utrecht: University of Utrecht, 2007. Cuijpers P, van Straten A, Andersson G. Internet-administered cognitive behavior therapy for health problems: a systematic review. J Behav Med 2008;00:1-9. de Boer AG, van Beek JC, Durinck J, Verbeek JH, van Dijk FJ. An occupational health intervention programme for workers at risk for early retirement; a randomised controlled trial. Occup Environ Med. 2004;61(11):924-9. de Koning, Smulders, Klazinga N. Appraisal of Indicators through Research and Evaluation 2007. de Vente W, Kamphuis JH, Emmelkamp PM, Blonk RW. Individual and group cognitive-behavioral treatment for work-related stress complaints and sickness absence: a randomized controlled trial. J Occup Health Psychol. 2008;13(3):214-31.

GENERAL DISCUSSION

160

de Vet HC, Heymans MW, Dunn KM, et al. Episodes of low back pain: a proposal for uniform definitions to be used in research. Spine 2002;27(21):2409-16. de Vries G, Kikkert MJ, Schene AH et al. Helpt arbeidshulpverlening bij patiënten met een depressie? Nederlands Tijdschrift voor Ergotherapie 2003;31:103-8. Duijts SFA. Prediction and early intervention in employees at risk for sickness absence due to psychological health complaints. PhD thesis. Maastricht University, 2007. Duijts SF, Kant I, van den Brandt PA, Swaen GM. Effectiveness of a preventive coaching intervention for employees at risk for sickness absence due to psychosocial health complaints: results of a randomised controlled trial. J Occup Environ Med. 2008;50(7):765-76. Fleten N, Johnsen R. Reducing sick leave by minimal postal intervention: a randomised, controlled intervention study. Occup Environ Med 2006;63(10):676-82 Franche RL, Baril R, Shaw W, et al.. Workplace-based return-to-work interventions: optimizing the role of stakeholders in implementation and research. J Occup Rehabil. 2005;15(4):525-42. Gilbody S, Bower P, Whitty P. Costs and consequences of enhanced primary care for depression. Systematic review of randomised economic evaluations. British J Psychiatry 2006;189:297-308. Harris JS, Sinnott PL, Holland JP, et al. Methodology to Update the Practice Recommendations in the American College of Occupational and Environmental Medicine's Occupational Medicine Practice Guidelines, Second Edition. J Occup Environ Med 2008;50(3):282-295. Heerkens Y, Engels J, Kuiper J, Van Der Gulden J, Oostendorp R. The use of the ICF to describe work related factors influencing the health of employees. Disability& Rehabilitation 2004;26(17):1060-6. Houtman I, Jettinghoff K, Brenninkmeijer V, van den Berg R. Work stress in the police force five years later: the effect of sectored agreements on (stress) management. [in Dutch]. TNO, 2005. Houtman ILD, Schoemaker CG, Blatter BM, de Vroome EMM, van den Berg R, Bijl RV. Psychological complaints, interventions and rehabilitation to work; the prognostic study of INVENT. (in Dutch). Hoofddorp: TNO Work & Employment, 2002. Hugenholtz NIR. Managing knowledge in occupational health care. PhD thesis. Amsterdam, 2008. Hughes E, Wanigaratne S, Gournay K, et al. Training in dual diagnosis interventions (the COMO Study): randomised controlled trial. BMC Psychiatry. 2008;8:12. Hulshof CTJ, Broersen JPJ, de Haan S. Primary care cooperation guideline on mental complaints and work (In Dutch: Samenwerkingsrichtlijn 1e lijns handelen bij psychische klachten en arbeid. Evaluatie fase 2: praktijktest) PARAG/SKB Utrecht/Amsterdam, 2002. Hulshof C, Hoenen J. Evidence-based practice guidelines in OHS: are they agree-able? Ind Health. 2007;45(1):26-31. Jamtvedt G, Young JM, Kristoffersen DT, O'Brien MA, Oxman AD. Audit and feedback: effects on professional practice and health care outcomes. Cochrane Database Syst Rev 2006;19(2):CD000259. Update of: Cochrane Database Syst Rev 2003;(3):CD000259. Kendrick T, Simons L, Mynors-Wallis L, et al. Cost-effectiveness of referral for generic care or problem-solving treatment from community mental health nurses, compared with usual general practitioner care for common mental disorders: Randomized controlled trial. Br J Psychiatry 2006;189:50-9. Kovess-Masféty V, Saragoussi D, Sevilla-Dedieu C, Gilbert F, Suchocka A, Arveiller N, Gasquet I, Younes N, Hardy-Bayle MC. What makes people decide who to turn to when faced with a mental health problem? Results from a French survey. BMC Public Health 2007;7(147):188. Lagerveld SE, Blonk RWB, Brenninkmeijer V, Schaufeli W. The role of self-efficacy on return to work among employees on sick leave with common mental disorders. In press. Leone SS, Huibers MJ, Kant I, van Amelsvoort LG, van Schayck CP, Bleijenberg G, Knottnerus JA. Long-term efficacy of cognitive-behavioral therapy by general practitioners for fatigue: a 4-year follow-up study. Journal of Psychosomatic Research 2006;61(5):601-7. Lerner D, Henke RM. What Does Research Tell Us About Depression, Job Performance, and Work Productivity? J Occup Environ Med. 2008;50(4):401–410.

GENERAL DISCUSSION

161

Lovibond SH, Lovibond PF (1993). Manual for the Depression Anxiety Stress Scales (DASS). University of New South Wales. Mackay CJ, Cousins R, Kelly PJ, Lee S, McCaig RH. ‘Management Standards’ and work-related stress in the UK: Policy background and science. Work & Stress 2004;18(2):91-112. Moreira C, May J, Mason M, Eccles A. New method of analysis enabled a better understanding of clinical practice guideline development processes. J Clin Epidem 2006;59(11):1199 –1206. Nieuwenhuijsen K, Verbeek JH, de Boer AG, Blonk RW, Dijk FJ van. The Depression Anxiety Stress Scales (DASS): detecting anxiety disorder and depression in employees absent from work because of mental health problems. Occupational and Environmental Medicine 2003;60:i77. Nieuwenhuijsen K, Verbeek JHAM, Siemerink JCMJ, Tummers-Nijsen D. Quality of rehabilitation among workers with adjustment disorders according to practice guidelines; a retrospective cohort study. Occup Environ Med 2003;60:21-25. Nieuwenhuijsen K, Verbeek JH, de Boer AG, et al. Validation of performance indicators for rehabilitation of workers with mental health problems. Med Care 2005;43(10):1034-42. Nieuwenhuijsen K, Bültmann U, Neumeyer-Gromen A, Verhoeven AC, Verbeek JH, van der Feltz-Cornelis CM. Interventions to improve occupational health in depressed people. Cochrane Database Syst Rev. 2008 Apr 16;(2):CD006237. Review. Noordik E, Klingen E, Nieuwenhuijsen K, van der Klink JJL, van Dijk F. Exposure Nystuen P, Hagen KB, Herrin J. Solution-focused intervention for sick listed employees with psychological problems or muscle skeletal pain: a randomised controlled trial Scandinavian BMC Public Health 2006;6:69. NVAB. Van der Klink J, ed. Guideline for Mental Health Problems. [in Dutch] Eindhoven: NVAB (Netherlands Society of Occupational Medicine), 2000. NVAB. Van der Klink J, ed. Guideline for Mental Health Problems. Revision 1

st edition 2000 [in Dutch]

Utrecht: NVAB (Netherlands Society of Occupational Medicine), 2007. Perski A, Grossi G. [Treatment of patients on long-term sick leave because of stress-related problems. Results from an intervention study] Lakartidningen. 2004;101(14):1295-8. Swedish. Rebergen DS, Hoenen JAHJ, Heinemans AMEC, Bruinvels DJ, Bakker AB, van Mechelen W. Adherence to a national guideline on mental health problems by Dutch occupational physicians. Occupational Medicine 2006;56:461-468. Ruotsalainen JH, Verbeek JH, Salmi JA, Jauhiainen M, Laamanen I, Pasternack I, Husman K. Evidence on the effectiveness of occupational health interventions. Am J Ind Med. 2006;49(10):865-72. Salmela-Aro K, Näätänen P, Nurmi JE. The role of work-related personal projects during two burnout interventions: a longitudinal study. Work & Stress 2004;18(3):208-230. Schaafsma FG. Evidence based medicine in occupational health care. Thesis. 2007 Schene AH, Koeter MWJ, Kikkert MJ, Swinkels JA, McCrone P. Adjuvant occupational therapy for work-related major depression works: randomized trial including economic evaluation. Psychological Medicine 2007;37(3):351-62. Seymour L, Grove B. British Occupational Health Research Foundation (BOHRF) Workplace interventions for people with common mental health problems: Evidence review and recommendations. United Kingdom. London, 2005. Slottje P, Smidt N, Twisk JW, Huizink AC, Witteveen AB, van Mechelen W, Smid T. Use of health care and drugs by police officers 8.5 years after the air disaster in Amsterdam. Eur J Public Health. 2008;18(1):92-4. Smits PB, de Buisonjé CD, Verbeek JH et al. Problem-based learning versus lecture-based learning in postgraduate medical education. Scand J Work Environ Health 2003;29(4):280-287. Staal JB, Hlobil H, van Tulder MW, Waddell G, Burton AK, Koes BW, van Mechelen W. Occupational health guidelines for the management of low back pain: an international comparison. Occup Environ Med 2003;60(9):618-26. Review.

GENERAL DISCUSSION

162

Taimela S, Justén S, Aronen P, Sintonen H, Läärä E, Malmivaara A, Tiekso J, Aro T. An occupational health intervention programme for workers at high risk for sickness absence. Cost effectiveness analysis based on a randomised controlled trial. Occup Environ Med 2008;65:242-248. Terluin B, van Marwijk H.W., Adèr H.J., et al. The Four-Dimensional Symptom Questionnaire (4DSQ): a validation study of a multidimensional self-report questionnaire to assess distress, depression, anxiety and somatization. BioMed Central 2006;6(34). The Stress Impact Consortium. Integrated report of Stress Impact: On the impact of changing social structures on stress and quality of life: Individual and social perspectives HPSE-CT-2002-00110 Work Package 8 Integrated report. 2006. Uegaki K, de Bruijne MC, Anema JR, et al. Consensus-based findings and recommendations for estimating health-related productivity loss from a company's perspective. Scand J Work Environ Health 2007;33:122-130. Uegaki K, Bakker IM, De Bruijne M, et al. Cost-effectiveness of a minimal intervention strategy for stress-related sick leave in general practice: results of an economic evaluation alongside a pragmatic randomized controlled trial. In press. Vahtera J, Kivimäki M. Reducing sickness absence in occupational settings. Occup Environ Med. 2008;65(4):219-20. van der Klink JJL, Blonk RWB. Schene AH, van Dijk FJH. Reducing long term sickness absence by an activating intervention in adjustment disorders: a cluster randomised controlled design. Occup Environ Med 2003;60:429-37. van der Klink JJL, van Dijk FJ. Dutch practice guidelines for managing adjustment disorders in occupational and primary health care. Scand J Work Environ Health 2003;29:478–487. van der Klink JJL, Terluin B. Mental health problems and work: guide for an activating management by general practitioner and occupational physician. (in Dutch:, Houten: The Netherlands, 2005. van Oostrom SH, Anema JR, Terluin B, de Vet HC, Knol DL, van Mechelen W. Cost-effectiveness of a workplace intervention for sick-listed employees with common mental disorders: design of a randomized controlled trial. BMC Public Health. 2008;8:12. van Rhenen W. From stress to engagement. Thesis. University of Amsterdam, 2008. van Rhenen W, Blonk RW, Schaufeli WB, van Dijk FJ. Can sickness absence be reduced by stress reduction programs: on the effectiveness of two approaches. Int Arch Occup Environ Health. 2007;80(6):505-15. van Rhenen W, van Dijk FJ, Schaufeli WB, Blonk RW. Distress or no distress, that's the question: A cutoff point for distress in a working population. J Occup Med Toxicol. 2008;3:3. Verbeek JH. How can doctors help their patients to return to work? PLoS Med 2006;3(3):e88. Verbeek JH. The occupational health field in the cochrane collaboration. Ind Health 2007;45(1):8-12. Vlasveld MC, Anema JR, Beekman AT, van Mechelen W, Hoedeman R, van Marwijk HW, Rutten FF, Hakkaart-van Roijen L, van der Feltz-Cornelis CM. Multidisciplinary collaborative care for depressive disorder in the occupational health setting: design of a randomised controlled trial and cost-effectiveness study. BMC Health Serv Res 2008:5;8:99. Waddell G, Burton AK. Concepts of rehabilitation for the management of common health problems. The Stationery Office, 2004. ISBN 0117033944. Wang PS, Simon GE, Avorn J, et al. Telephone screening, outreach, and care management for depressed workers and impact on clinical and work productivity outcomes: a randomized controlled trial. JAMA 2007 26;298(12):1401-11. World Health Organization (WHO). International Classification of Functioning, Disability and Health (ICF). WHO, 2001. Young AE, Roessler RT, Wasiak R, McPherson KM, van Poppel MNM, Anema JR. Measuring return to work. Jf OccupRehabil 2005;15(4):557-68 . Young AE, Roessler RT, Wasiak R, et al. A developmental conceptualization of Return to Work. J Occup Rehabil 2007;17:766–781.

SUMMARY

SUMMARY

164

Introduction This thesis focused on occupational health care of workers with common mental health problems, such as adjustment disorders, depression and anxiety. Common mental health problems in workers increasingly affect functioning to such an extent that they may lead to all kinds of productivity loss. Primary and occupational health care of workers with mental health problems usually focuses on recovery of symptoms instead of on return to work (RTW). In the Netherlands, workers have to visit an occupational physician (OP) for RTW purposes if they are on sick leave. Therefore, OPs may play a central role in the treatment of workers with common mental health problems. The usual approach towards these patients was the advice to take rest and not RTW before all complaints had disappeared. When complaints persist, workers could be referred to psychologists from primary and specialized secondary care. As an alternative to this minimal role by the OP in usual care, the Netherlands Society of Occupational Medicine (NVAB) published a new guideline in 2000. The guideline ‘The management by OPs of workers with mental health problems’ promotes a more active role of the OP, as counsellor facilitating RTW. The guideline received positive reactions from workers, employers and Dutch OPs. The main aim of this thesis is to contribute to quality improvement of occupational health care for workers with mental health problems, by evaluating the effectiveness of guideline-based care. Return to work for adjustment disorders Chapter 2 describes a Cochrane review on the efficacy of interventions aimed at RTW for workers with adjustment disorders. This was done in collaboration with the Cochrane Occupational Health Field, the Cochrane Depression, Anxiety and Neurosis Review Group and the Dutch Cochrane Centre. A literature search was done using 3 databases (Medline (PubMed), EMBASE, PsycINFO) and 3 filters (Adjustment disorders (ICD-10), Occupational health interventions, Controlled trials). Six randomized controlled trials (RCTs) on RTW interventions for adjustment disorders were identified. All dated from 2003 to 2006, four were from The Netherlands, two from Norway. Van der Klink was the first to evaluate an intervention on RTW of workers with mental health problems, by means of a randomised controlled trial. A decade ago, van der Klink developed a brief activating intervention for OPs, which was based on cognitive behavioural therapy (CBT) principles and included graded activity. As this intervention proved to be successful, it was reflected in the guideline that has been evaluated in this thesis. Most studies in the review used an intervention based on principles of cognitive behavioural therapy. This review found evidence that CBT, a commonly used type of psychotherapy, may facilitate RTW of workers with stress-related mental health problems. On average, workers who are offered CBT will start two weeks earlier with partial and full RTW compared to workers who received care as usual. A second finding of this review is that CBT improves the mental health of workers with adjustment disorders. This finding actually supports the hypothesis that early RTW may be associated with improved mental health. More studies are needed, preferably from other countries than Norway and The Netherlands. Design of the CO-OP study Chapter 3 describes the design of the central study of this thesis, the CO-OP study. This study aims to assess the effects of the NVAB-guideline, compared to usual care. The most renewing element of the guideline was defined as ‘counseling’, as a new role for OPs was introduced as counsellor of stress- and work-related problems. Therefore, main aim of the CO-OP study was to evaluate the (cost-)effectiveness of Counseling by OPs according to the NVAB-guideline. In a randomized controlled trial (RCT), subjects in the intervention

SUMMARY

165

group received guideline-based care, after a 3-days training course. The control group received usual care, with minimal involvement of the OP and easy access to a psychologist. Subjects were recruited from two Dutch police departments. The primary outcomes of the study were RTW and treatment satisfaction by the worker, employer, and OP. A secondary outcome was cost-effectiveness of the guideline-based care, compared to usual care. Furthermore, several prognostic measures were taken into account as potential confounders. The follow-up period was one year. A process evaluation would be done by means of performance indicators, based on the guideline. Effectiveness of guideline-based care The objective of chapter 4 was to evaluate the effectiveness of counseling by OPs compared to usual care on the primary outcomes RTW and treatment satisfaction. RTW was measured with and without including partial RTW, and with and without recurrent events during follow-up. Data analysis was done, based on the intention to treat principle, using Kaplan-Meier curves and Cox proportional hazard regression. From 2002 to 2005, 240 police workers on sick leave due to common mental health problems were recruited. Counseling by OPs did not result in earlier RTW, compared to usual care. Subgroup analysis showed a small effect on RTW in favour of counseling for workers with administrative functions and for workers with ‘minor’ stress-related symptoms. Treatment satisfaction between the two groups did not differ. Treatment satisfaction rated by the employee diminished significantly during treatment in both groups. Treatment satisfaction rated by the OP was significantly higher in the intervention group, when compared to usual care. Thus, workers with ‘minor’ stress-related disorders may benefit from counseling according to the guideline. OPs are more satisfied with their guideline-based care. Cost-effectiveness of guideline-based care In chapter 5 the cost-effectiveness of guideline-based care was evaluated and compared to usual care. An economic evaluation from both a societal and company perspective was conducted alongside the randomised controlled trial. Over a three-years period police workers on sick leave due to mental health problems (n=240) were included in the study. Duration of sick leave during one-year follow-up was the main outcome measure. Sick leave data and health care costs were gathered from computerized records of the police departments, the occupational health services, and the health insurance agency of the Dutch police. Analyses were based on intention to treat principles. Bootstrap techniques were used to estimate the 95%-confidence interval around the difference in mean costs and effects between the two groups. Cost-effectiveness planes and acceptability curves were calculated. Health care utilization costs (€574,532 in total) in the one-year follow-up period were significantly lower in the intervention group (mean difference -€520; 95% CI: -€980, -€59), while there were no significant differences in days of sick leave and productivity loss costs. These results suggest that from both society and company perspective, guideline-based care could be cost-effective, as lower direct costs lead to equal treatment outcomes of workers with common mental disorders. Process evaluation of the CO-OP study Chapter 6 describes a process evaluation of the CO-OP study. The aim was to examine guideline adherence by Dutch OPs, compared to usual care, as part of a process evaluation of a trial on the effectiveness of guideline-based care. Guideline adherence was assessed by means of an audit of medical files. A new set of performance indicators (PIs) was developed and independently rated by three researchers, resulting in a dichotomised

SUMMARY

166

score on each PI for each worker (optimal versus sub optimal). Performance rates on guideline adherence were related to RTW by means of the Cox proportional hazards model. Mean performance rates in guideline adherence by the OP were 50% and did not significantly differ between GBC and UC. PPA was not able to identify significant determinants on RTW. These results may be explained by the fact that the guideline was partially practice-based and contained many elements of UC. A possible explanation for the lack of results in the PPA may be that even in selected medical records with high performance rates, essential elements of the guideline were not applied. Guideline adherence by Dutch occupational physicians Chapter 7 describes a cross-sectional study that examined predictors of adherence by OPs to the national guideline for the management of workers with mental health problems. This study was conducted in 2001, one year after the guideline was introduced and disseminated to the Dutch OPs. Using the Theory of Planned Behaviour, a questionnaire was developed about self-reported guideline adherence of OPs and possible predictors of this behaviour. A total of 165 OPs were approached to complete the questionnaire and registration forms of first consultations of workers with mental health problems. Performance indicators based on the guideline were developed to calculate performance rates of guideline adherence by OPs. Eighty of 165 (48%) OPs approached completed the questionnaire. Fifty-six OPs returned one or more registration forms, totalling 344 consultations. On a five-point Likert scale, ranging from never (1) to always (5), the mean score on self-reported guideline adherence was 2.35, compared to a mean score of 4.06 on the intention to comply with the guideline. The mean performance rate of OPs ranging from 0 to 2 was 1.27 on diagnosis and 0.60 on guidance. No relation was found between self-reported guideline adherence and performance rates. Self-reported guideline adherence correlated significantly with perceived behaviour control (r = 0.48, P < 0.05), subjective norms (r = 0.33, P < 0.05) and positive job stress (r = 0.35, P < 0.05). These results show that guideline adherence by Dutch OPs lags behind its acceptance. Future implementation efforts need to focus on diminishing barriers and enhancing social norms of OPs to work according to the guideline. General discussion In the general discussion in chapter 8 some of the major findings described in the previous chapters of the thesis are reflected. The CO-OP study is the first randomised controlled trial to evaluate the effects of guideline-based care in a primary or occupational health care setting for mental health problems with RTW as a primary outcome. In this trial effectiveness, instead of efficacy was studied, as there were no ideal circumstances and the two research conditions were not highly contrasting. According to the ‘triad’ of effectiveness (or efficacy), i.e. a conceptual framework for the interpretation of trials, the lack of effectiveness in the trial can be explained by a number of findings. Firstly, participating OPs were not able to deliver guideline-based care appropriately. Secondly, guideline-based care as delivered in our intervention, was not effective in the treatment of workers with mental health problems. Thirdly, RTW of patients in the intervention group could not be attributed to guideline-based care. These findings should be interpreted in relation to our choice of design. An advantage of our design is that we were able to deliver realistic treatment outcomes, including an economic evaluation. A disadvantage of effectiveness studies is that they are not informative if the results are negative.

SUMMARY

167

Conclusions and implications In this thesis it was shown that guideline-based care by OP, reflected by an activating approach facilitating RTW of the worker with mental health problems, could be cost-effective in occupational health, if compared to an extensive usual care with easy access to a psychologist. These results support the overall evidence that a minimal intervention, CBT counseling, in occupational health care seems (cost-) effective, especially regarding the substantial group of workers suffering from ‘minor’ mental health problems. Considering workers with ‘severe’ mental disorders, such as depression and/or anxiety, early detection and a more extensive combined interventions by occupational experts and/or specialised secondary care interventions by psychologists may be more effective. Furthermore, the guideline needs revision with more attention for evaluation of stagnation of recovery and regular evaluation by the OP with worker and employer. Recently, the guideline for OPs has been revised (NVAB, 2007). Guideline-based care may be improved by facilitating higher applicability of the innovative elements of the guideline. This may be realized by continuous training and supportive legislation by the relevant stakeholders, which may enhance perceived behaviour control, normative behaviour by OPs, and work engagement of OPs. Opportunities and challenges regarding the development and implementation of evidence-based occupational health care for workers with mental health problems should be further analysed, especially in an international context. Recommendations for occupational health care practice

Cognitive behavioural therapy interventions should be used for workers with adjustment disorders, preferably in an occupational health care setting.

Application of an activating guideline-based care by OPs should be promoted, as this is more cost-effective than usual care with easy access to a psychologist.

Workers with depression or anxiety symptoms should be detected in an early stage, and should be referred to pre-authorized secondary mental health care.

The guideline should be revised in a more simple and applicable version, with more attention for workers with severe mental health problems.

Implementation of guideline-based care should focus on higher applicability of the innovative elements of the guideline by OPs, and should be facilitated by continuous training and supportive legislation by relevant stakeholders.

SAMENVATTING

SAMENVATTING

170

Casus Een werknemer loopt tegen problemen aan in zijn relatie en werk. Hij begint aan alles te twijfelen, voelt zich somber en ervaart de grip op zijn leven te verliezen. Ondanks een bezoek aan de huisarts nemen deze gevoelens toe. Als zijn relatie eindigt en hij weer tegenslag op zijn werk ervaart, meldt hij zich ziek. Na 3 weken bezoekt hij zijn bedrijfsarts. De vraag vanuit zowel werknemer, werkgever, bedrijfsarts, verzekeraar, overheid, en dit proefschrift, is: wat is de meest effectieve vorm van begeleiding om deze werknemer met psychische klachten succesvol het werk te laten hervatten? Introductie Context De afgelopen decennia is de visie op de relatie tussen gezondheid, werk en maatschappij veranderd. Gezondheid wordt niet meer alleen bepaald door de aan- of afwezigheid van klachten, maar ook door of we er mee kunnen functioneren. Wat betreft werk is de aandacht verschoven van fysieke naar mentale belasting, en werkbeleving. De vraag aan (bedrijfs)artsen is veranderd van bescherming tegen ziekmakend werk naar ondersteuning in een effectieve omgangsvorm met werk, dat zo bijdraagt aan de gezondheid en energie geeft. Ook in maatschappelijk opzicht zijn de veranderingen groot geweest. Het sociale zekerheidsstelsel is in korte tijd omgebouwd van een stelsel gericht op compensatie naar een stelsel gericht op participatie. De verantwoordelijkheid voor participatie is bij werknemer en werkgever gelegd, wat onder andere tot uiting komt in de Wet Verbetering Poortwachter (WVP) en de Wet Werk en Inkomen naar Arbeidsvermogen (WIA). Psychische bedrijfsgezondheidszorg in Nederland Dit proefschrift richt zich op de bedrijfsgezondheidszorg voor werknemers met vaak voorkomende psychische klachten, zoals aanpassings- of stressgerelateerde aandoeningen (spanningsklachten, overspanning, burnout), angst en depressie. Deze komen voor bij 10-18 procent van de Westerse beroepsbevolking en leiden tot persoonlijk leed, verlies aan arbeidsproductiviteit en aanzienlijke maatschappelijke kosten voor de samenleving. Voor de Nederlandse samenleving worden deze kosten geschat op 7.5 miljard Euro per jaar. In Nederland is eenderde van de mensen die langdurig arbeidsongeschikt zijn, uitgevallen met psychische problematiek. Binnen de bedrijfsgezondheidszorg speelt in Nederland de bedrijfsarts een cruciale rol, aangezien werknemers bij verzuim van werk een bedrijfsarts consulteren. Een bedrijfsarts is meestal werkzaam bij een arbodienst, waarmee werkgevers contracten afsluiten voor de te leveren bedrijfsgezondheidszorg. Eind vorige eeuw was de aandacht van hulpverleners nog vooral gericht op klachtherstel in plaats van functieherstel. Werknemers met psychische klachten kregen van huisartsen en bedrijfsartsen vaak het advies eerst thuis uit te rusten en de ‘accu op te laden’, alvorens weer aan het werk te gaan. Daarnaast werden deze werknemers wegens gebrek aan tijd en kennis regelmatig doorverwezen naar gespecialiseerde psychische zorg (psychiaters, psychologen en andere psychosociale hulpverleners). Echter, bij deze zorg stond klachtherstel nog meer centraal, met de angst dat functieherstel in de vorm van werkhervatting dit in de weg zou staan. NVAB-richtlijn psychische klachten (2000) Rond de eeuwwisseling kwam hierin verandering, mede door de ontwikkeling van nieuwe richtlijnen door de beroepsvereniging voor bedrijfs- en arbeidsgeneeskundigen (NVAB). In 2000 werd de geautoriseerde NVAB-richtlijn ‘Handelen van bedrijfsartsen bij werknemers

SAMENVATTING

171

met psychische klachten’ geïntroduceerd. De richtlijn beschrijft het proces van controleverlies van de cliënt op zijn eigen situatie, tot het moment waarop de cliënt weer zelf de regie in handen heeft en zijn rollen kan invullen. Dit proces werd verdeeld in 3 kenmerkende fasen waarbij per fase duidelijke taken zijn te verrichten door de cliënt en zijn omgeving. De richtlijn beschrijft daarnaast de volgende nieuwe uitgangspunten voor begeleiding door de bedrijfsarts:

een brede probleemoriëntatie en versimpelde diagnostiek een activerende aanpak gericht op controleherstel en werkhervatting een rol als procesbegeleider volgens een tijdschema inhoudelijke begeleiding ('counseling') bij stress- en werkgerelateerde klachten.

Bij dit laatste uitgangspunt kan de bedrijfsarts zelf cognitief-gedragsmatige interventies inzetten en is een optimale communicatie tussen bedrijfsarts en huisarts cruciaal. In een studie van Van der Klink e.a. (2003) bleek deze aanpak effectief voor stressgerelateerde aandoeningen. Echter, de richtlijn was voornamelijk gebaseerd op consensus, een stevig wetenschappelijk fundament ontbrak. De richtlijn werd verstuurd aan alle circa 2000 bedrijfsartsen in Nederland, waarvan een deel via hun arbodienst de mogelijkheid kreeg om een driedaagse training over de richtlijn te volgen. Hoewel de meeste bedrijfsartsen kennis hadden genomen van de richtlijn, bestond het risico dat een belangrijk deel van de bedrijfsartsen de richtlijnen niet zou toepassen in de praktijk, door bijvoorbeeld kennisgebrek of de hoge werkdruk. Bedrijfsartsen zouden derhalve een beroep kunnen doen op gespecialiseerde psychologische zorg. Probleemstelling Hoofddoel van dit proefschrift is bij te dragen aan kwaliteitsverbetering van de psychische bedrijfsgezondheidszorg, door de effectiviteit van de NVAB-richtlijn psychische klachten te evalueren. De vraagstelling bij het centrale onderzoek van dit proefschrift is of de NVAB-richtlijn effectief is op het bevorderen van werkhervatting en tevredenheid over de geboden zorg, ten opzichte van minimale begeleiding door de bedrijfsarts met laagdrempelige verwijzing naar een psycholoog. Als theoretisch raamwerk is in dit proefschrift het ICF-model (‘International Classification of Functioning, Disability, and Health’) gehanteerd. Dit biopsychosociale model geeft aangrijpingspunten voor verandering van het functioneren van iemand op basis van medische symptomen, persoonskenmerken, en de omgeving. De NVAB-richtlijn sluit met haar benadering goed aan op het ICF-model. Wetenschappelijke evidentie Om een overzicht te krijgen over de reeds bestaande wetenschappelijke literatuur met betrekking tot onze centrale vraagstelling, hebben we een systematische literatuurstudie (Cochrane review) uitgevoerd. Hoofdstuk 2 beschrijft de resultaten van de literatuurstudie naar de effectiviteit van interventies voor werknemers met een aanpassingsstoornis op werkhervatting. Een aanpassingsstoornis wordt ook wel een stressgerelateerde stoornis (acuut: overspanning; chronisch: burnout) genoemd en kenmerkt zich volgens het diagnostisch classificatiesysteem DSM-IV als een emotionele en/of gedragsmatige reactie op een aanwijsbare stressfactor. Exclusiecriteria zijn rouw en psychiatrische problematiek. Via een systematische zoekstrategie onderzochten we de elektronische databases van relevante Cochrane registers, MEDLINE (Pubmed), EMBASE, en PsycINFO op relevante gerandomiseerde, gecontrolleerde trials (RCTs). Deze zoekstrategie leverde 3789 publicaties op. Na een uitgebreide selectie-procedure bleven zes artikelen over. Het betrof vier Nederlandse en twee Noorse studies. Vier studies beschreven de effecten van

SAMENVATTING

172

cognitieve gedragstherapie op werkhervatting. De oudste studie is in 2003 gepubliceerd (Van der Klink e.a., 2003) en betreft de studie waarop de NVAB-richtlijn psychische klachten is gebaseerd. Een meta-analyse toonde aan dat behandeling met cognitieve gedragstherapie, een veelgebruikte methode binnen psychologische therapieën, werkhervatting gemiddeld met twee weken versnelt. Cognitieve gedragstherapie bleek ook psychische klachten op het gebied van stress, depressie en angst, significant te verminderen, ten opzichte van de groep met standaardbegeleiding. Deze resultaten ondersteunen de hypothese dat snelle werkhervatting samengaat met klachtvermindering bij stressgerelateerde problematiek. Onderzoeksopzet Project Reïntegratie Politie De achtergrond en opzet van het onderzoek dat centraal staat in dit proefschrift; 'Project Reïntegratie Politie' ofwel de 'CO-OP study' (Counseling by Occupational Physicians), staat beschreven in hoofdstuk 3. Aanleiding voor de politie om mee te werken aan dit onderzoek waren de hoge kosten van psychische arbeidsongeschiktheid, en structurele onderbezetting bij een toenemende maatschappelijke vraag om meer ‘blauw op straat’. Het onderzoek werd uitgevoerd bij twee politiekorpsen (Hollands Midden en Zaanstreek-Waterland), met totaal circa 2500 politiewerknemers. Standaard konden politiewerknemers via de bedrijfsarts van hun arbodienst (Commit, later ook KLM Health Services) relatief snel en gemakkelijk naar tweedelijns psychologische zorg worden verwezen. De 'gebruikelijke zorg' bestond zo uit minimale begeleiding door de bedrijfsarts, die laagdrempelig behandeling door de psycholoog in kon zetten (Psychopol genaamd). Laagdrempelig betekent dat deze vooraf geauthoriseerd en gefinancierd werd door de werkgever, het ministerie van Binnenlandse Zaken en Koninkrijks-aangelegenheden, via de zorg- en inkomensverzekeraar Dienst Geneeskundige Verzorging Politie. Door de introductie van de NVAB-richtlijn psychische klachten, was er een maatschappelijke en wetenschappelijke vraag of invoering van deze richtlijn een meer effectieve en efficiëntere zorg zou opleveren. Er was sprake van een gerandomiseerde, gecontroleerde onderzoeksopzet (RCT). Deelnemende bedrijfsartsen ontvingen voorafgaand aan het onderzoek als interventie een driedaagse training in de richtlijn. Politiewerknemers die zich ziek hadden gemeld met psychische klachten ontvingen informatie over het onderzoek. Als de bedrijfsarts tijdens het eerste consult van de werknemer schriftelijke toestemming kreeg voor deelname aan het onderzoek,opende de bedrijfsarts een gesloten envelop. De inhoud daarvan was vooraf door de onderzoekers op basis van toeval verdeeld over de interventiegroep (begeleiding conform richtlijn) en de controlegroep (gebruikelijke zorg). In de envelop stond dus welke zorg de betreffende werknemer zou gaan ontvangen van de bedrijfsarts. Op deze manier waren werknemer, werkgever en bedrijfsarts op de hoogte van de zorg die de werknemer ontving en in welke groep deze ingedeeld was. De onderzoekers wisten dit niet. Deelnemers werden een jaar lang gevolgd op het aantal dagen verzuim tot aan eerste en volledige werkhervatting op het niveau van voor de ziekmelding, eventuele nieuwe verzuimperiodes, tevredenheid over de begeleiding, en zorgconsumptie. Effectiviteit richtlijn De onderzoeksresultaten van de hierboven beschreven studie staan beschreven in hoofdstuk 4-6. De deelnemers werden tussen januari 2002 en januari 2005 geworven en een jaar lang gevolgd. De uiteindelijke twee groepen bleken bij aanvang van het onderzoek vergelijkbaar wat betreft persoonskenmerken. De onderzoeksgroep was

SAMENVATTING

173

gemiddeld 40 jaar, bestond vooral uit mannen (60%) en werknemers met uitvoerende functie 'op straat' (60%). Bedrijfsartsen oordeelden dat ruim 80% van de deelnemers een stressgerelateerde aandoening had, ofwel een (chronische) overspanning, en circa 20% ‘zwaardere’ problematiek als depressie/angst. De bij aanvang van het onderzoek afgenomen vragenlijsten (DASS, HADS) gaven echter aan dat zeker 35% depressie en/of angstproblematiek had. Dit duidt op een ondersignalering van deze problematiek door bedrijfsartsen. De hypothese was dat begeleiding conform de NVAB-richtlijn door bedrijfsartsen aan verzuimende werknemers met psychische problemen zal leiden tot snellere terugkeer in werk, en meer tevredenheid over de begeleiding bij werknemers, werkgevers en bedrijfsartsen dan in de groep met gebruikelijke zorg. Op basis van Cox-regressie analyse bleek geen verschil in gemiddelde duur tot werkhervatting (ruim 100 dagen) tussen de groepen. Wel bleek begeleiding conform de richtlijn tot snellere werkhervatting te leiden bij werknemers met een stressgerelateerde aandoening en/of een administratieve functie. De gebruikelijke zorg leek daarentegen beter te werken bij werknemers met depressie/angst (volgens zowel oordeel bedrijfsarts als vragenlijst) en/of uitvoerende functie ‘op straat’. Tevredenheid over de begeleiding bleek niet verschillend tussen de groepen voor werknemer en leidinggevende. De werknemerstevredenheid daalde significant gedurende de begeleiding (van gemiddeld 7,0 naar 6,0 op schaal van 1-10). De tevredenheid van de leidinggevenden bleef stabiel in beide groepen (6,6). De effectiviteit van de begeleiding bleek volgens de bedrijfsarts in de interventiegroep significant hoger (7,6) dan bij gebruikelijke zorg (7,0). Kosten-effectiviteit richtlijn Binnen dezelfde studie werd de hypothese getoetst dat begeleiding conform de NVAB-richtlijn kosten-effectief is, in de verwachting dat dit leidt tot een reductie in zorgconsumptie en relatief minder productiviteitsverlies ten opzichte van de gebruikelijke zorg. Data met betrekking tot zorgkosten werden verzameld bij de zorgverzekeraar DGVP, waarbij elke politiewerknemer verzekerd was. Data over productiviteitsverlies werden berekend vanuit de verzuimdata met één jaar follow-up, die door de werkgever geanonimiseerd werden aangeleverd. In de analyse werden voor de gegevens van de 240 deelnemers via bootstrap-technieken ‘cost-effectiveness planes’ en ‘acceptability curves’ berekend. De totale zorgkosten bedroegen €574.532,- voor alle 240 een jaar lang gevolgde deelnemers. De gemiddelde zorgkosten bleken in de interventiegroep gemiddeld €2.145,- per deelnemer. Dit was gemiddeld €520,- lager per deelnemer dan in de controle groep. Beide groepen kwamen tot gelijke resultaten wat betreft productiviteitsverlies (totale kosten -€3.397.555,-; gemiddelde kosten per deelnemer €14.156,-). De conclusie is dat vanuit financieel oogpunt inzet van richtlijnbegeleiding door de bedrijfsarts kosten-effectief en dus te prevaleren is. Dit wordt niet veroorzaakt door een effect op productiviteit (minder verzuim), maar door lagere zorgkosten als gevolg van relatief minder verwijzingen naar tweedelijns psychologische zorg. Procesevaluatie begeleiding volgens richtlijn Binnen de studie is aan de hand van een procesevaluatie bestudeerd in hoeverre de richtlijn in beide groepen werd toegepast door de deelnemende bedrijfsartsen. Tevens is bekeken of de mate van de richtlijntoepassing van invloed was op de werkhervatting van werknemers in beide groepen. De geanonimiseerde medische dossiers van de deelnemers werden aan een audit onderworpen met behulp van 20 op de richtlijn gebaseerde prestatie-indicatoren. De gemiddelde prestatiescores voor richtlijngebruik van de

SAMENVATTING

174

bedrijfsartsen zijn met een Cox-regressie analyse gekoppeld aan de duur tot volledige werkhervatting van de werknemers. Gemiddeld bleek de bedrijfsarts per patiëntendossier 50% van de richtlijn toe te passen. Er was geen verschil in mate van toepassing van de richtlijn tussen beide groepen. Mogelijke verklaringen voor deze resultaten zijn dat de richtlijn deels gebaseerd is op consensus en elementen uit de bestaande praktijk bevat, en indicatoren dus onvoldoende onderscheidend zijn geweest. Tevens begeleidde de bedrijfsarts in ons onderzoek na training in de richtlijn deelnemers uit beidde groepen, met waarschijnlijk een relatief intensieve gebruikelijke zorg door de bedrijfarts als gevolg. Daarnaast werd de bedrijfsarts door tijdens ons onderzoek ingevoerde regelgeving (zoals WVP) gedwongen meer procesmatig te werken, wat het contrast tussen beide groepen mede lijkt te hebben verkleind. Vervolgens hebben we een per–protocol analyse uitgevoerd om te onderzoeken of begeleiding conform de richtlijn in de interventiegroep bijdraagt aan het versnellen van werkhervatting van de werknemer. Echter, er bleek geen verschil te zijn in de duur tot werkhervatting tussen de dossiers met bovengemiddelde richtlijnprestatiescores in de interventiegroep, met de duur tot werkhervatting in de controlegroep. Deze resultaten kunnen mogelijk worden verklaard doordat zelfs in de groep met hoge prestatiescores mogelijk essentiële elementen uit de richtlijn niet zijn toegepast. Zo werd de richtlijn significant vaker toegepast in de interventiegroep op gebruik van de 4DKL, overleg met huisarts en de inzet van inhoudelijke begeleiding ('counseling’), maar betrof dit slechts 10-20% van de dossiers. Toepassing richtlijn in de praktijk Hoofdstuk 7 beschrijft een cross-sectionele studie (met één meetmoment) waarin de mate van toepassing van de NVAB-richtlijn psychische klachten is onderzocht, en eventuele daarmee samenhangende factoren. Het onderzoek is uitgevoerd in 2001, één jaar nadat de NVAB-richtlijn was uitgebracht en verspreid over de circa 2000 bedrijfsartsen in Nederland. Op basis van de Theorie van Planmatig handelen (Ajzen) werd in samenwerking met het Bureau Richtlijnen NVAB een vragenlijst ontwikkeld om gedragsdeterminanten met betrekking tot richtlijngebruik in kaart te brengen. Op basis van de richtlijn werden vier prestatie-indicatoren om richtlijngebruik van de bedrijfsartsen met behulp van registratieformulieren te toetsen. Van de 165 benaderde bedrijfsartsen retourneerden 80 bedrijfsartsen (48%) de vragenlijsten. 56 stuurden eveneens één of meer ingevulde registratieformulieren terug van in totaal 344 consulten. Het gemiddelde zelfgerapporteerde richtlijngebruik bleek 2,35 op een schaal van 1-5 (nooit-altijd), tegenover een score van 4,06 op intentie tot gebruik van de richtlijn. De gemiddelde prestatiescore op richtlijngebruik op basis van de registratieformulieren bleek even matig, er werd echter geen relatie gevonden tussen zelfgerapporteerd richtlijngebruik en de prestatiescores. Deze resultaten geven aan dat richtlijngebruik lager bleek dan men op basis van de intentiescores mocht verwachten. Zelfgerapporteerd richtlijngebruik hing samen met ervaren gedragscontrole, normatieve invloeden en een positieve werkbeleving. De aanbeveling is dat verbetering van de implementatie van de richtlijn zich dient te richten op vermindering van praktische barrières, en verhoging van werkbeleving en sociale normen onder bedrijfsartsen ten opzichte van werken volgens de richtlijn. Voorbeelden zijn het verhogen van regelmogelijkheden van bedrijfsartsen, door elementen uit de richtlijn mee te nemen in arbocontracten met werkgevers en/of meer aandacht te geven in intervisie/nascholing.

SAMENVATTING

175

Conclusies en aanbevelingen Conclusies Het onderzoek binnen 'Project Reïntegratie Politie' ('CO-OP study') is voor zover wij weten de eerste gerandomiseerde, gecontroleerde studie (RCT) die de effecten van een richtlijn voor psychische zorg evalueert op werkgerelateerde uitkomsten. In dit proefschrift is aangetoond dat training van bedrijfsartsen in een deels wetenschappelijk onderbouwde richtlijn binnen de psychische bedrijfsgezondheidszorg kosteneffectief is, in vergelijking met de voor deze specifieke populatie gebruikelijke zorg met laagdrempelige toegang tot een psycholoog. Het betrof een training van bedrijfsartsen waarin zowel activering van hun eigen rol als die van de werknemer centraal stond. Bedrijfsartsen werden tevens getraind in het toepassen van cognitief-gedragsmatige interventies, om het probleemoplossend vermogen van werknemers te versterken. Deze resultaten ondersteunen de algemene wetenschappelijke evidentie dat cognitief-gedragsmatige elementen in behandeling, met name binnen de werksetting, werkhervatting versnellen van werknemers met stressgerelateerde problematiek. Voor werknemers met ‘ernstigere’ psychische problematiek als depressie en angst, is snelle herkenning en doorverwijzing naar gespecialiseerde psychische zorg meer effectief. Ook hier blijft het echter van belang de individuele aanpak te combineren met werkgerichte interventies. De onderzoeksresultaten ondersteunen een gecombineerde aanpak binnen de psychische bedrijfsgezondheidszorg, wat aansluit bij het theoretische ICF-model, en zo handvaten biedt voor de ontwikkeling en evaluatie van nieuwe interventies op dit terrein. Het ontbreken van significante uitkomsten binnen de RCT kan volgens de ‘effectiviteitstriade’ verklaard worden door drie bevindingen. Allereerst bleken deelnemende bedrijfsartsen na een driedaagse training in de richtlijn niet in staat in de interventiegroep meer begeleiding conform de richtlijn te bieden, ten opzichte van de gebruikelijke zorg. Ten tweede bleek training van bedrijfsartsen in de richtlijn niet te leiden tot betere uitkomsten voor werknemers, ten opzichte van gebruikelijke zorg. Ten derde, snellere werkhervatting van werknemers bleek niet beïnvloed te worden door de mate van toepassen van de richtlijn. Deze uitkomsten zijn beïnvloed geweest door de keuze voor het type onderzoeksdesign. Het betrof een pragmatische RCT, waarbij maximaal gecontroleerde omstandigheden, en dus voldoende contrasterende onderzoeksgroepen ontbraken. Een nadeel van dit design is dat als uitkomsten ontbreken, het moeilijk te analyseren is waardoor dit komt. Een voordeel is dat de studie aansluit bij de praktijk, met relevante uitkomstmaten, waaronder een economische evaluatie. Hieruit bleek dat training van bedrijfsartsen in de richtlijn leidt tot minder zorgkosten, ofwel verwijzingen naar een psycholoog, bij gelijkblijvende uitkomsten op productiviteitsverlies (verzuim). Kanttekeningen Er zijn diverse kanttekeningen te plaatsen bij dit onderzoek. De belangrijkste is dat de deelnemende bedrijfsartsen na training in de richtlijn zowel deelnemers uit de interventiegroep als de controlegroep begeleidde, met mogelijk een relatief intensieve gebruikelijke zorg door de bedrijfarts als gevolg. Het betrof verder een specifieke onderzoeksgroep, politiewerknemers, die specifieke stressgerelateerde risico’s kent als gevolg van het type werk. Deze onderzoeksgroep had daarnaast een uitgebreide en intensieve ‘gebruikelijke zorg’, die zowel in binnen- als buitenland niet standaard is. Wel lijkt de toegang tot psychologische zorg eerder meer dan minder toegankelijk te worden.

SAMENVATTING

176

Aanbevelingen Professionals binnen de bedrijfsgezondheidszorg wordt aanbevolen bij werknemers met stressgerelateerde problematiek, interventies met cognitief gedragsmatige elementen toe te passen, liefst in de werksetting. Daarnaast dient bij deze patiënten handelen volgens wetenschappelijk onderbouwde richtlijnen ondersteund te worden, aangezien dit tot aanzienlijke kostenbesparingen kan leiden, bij gelijkblijvende effecten. Bij werknemers met zwaardere psychische problematiek (depressie, angst) is snelle herkenning en verwijzing naar gespecialiseerde psychische zorg aan te bevelen. Daarnaast versnelt regelmatig contact van de professional met de werknemer en de werkgever de duur tot werkhervatting. Bovenstaande aanbevelingen uit dit onderzoek zijn opgenomen in de in 2007 herziene versie van de NVAB-richtlijn ‘handelen van de bedrijfsarts bij psychische problemen’. De herziene versie van de richtlijn is gebruiksvriendelijker en toegankelijker, wat de kans op naleving verhoogt. Het gebruik van diverse implementatietechnieken wordt aanbevolen om aanpassing van de praktijk aan de richtlijn te bevorderen. Gedacht moet worden aan mogelijkheden die zich richten op het vergroten van regelmogelijkheden voor de betreffende professional. Voorbeelden zijn aandacht voor richtlijnen binnen contracten met arbodienst/zorgverzekeraar, nascholing en intercollegiale toetsing, audits van dossiers, integratie in software en reeds bestaande zorgmodellen (zoals 'collaborative care'). Aanvullend onderzoek is nodig om het effect van richtlijngebruik op de kwaliteit van de begeleiding aan te tonen, en implementatie van de herziene richtlijn te evalueren. Dit met name in de psychische bedrijfsgezondheidszorg met haar specifieke setting. Daarnaast is effectiviteits- en doelmatigheidsonderzoek in een bredere internationale context noodzakelijk. Casus (vervolg, aanbevolen volgens de resultaten van dit proefschrift) De bedrijfsarts diagnosticeert een aanpassingsstoornis met mogelijke onderliggende depressie. Bedrijfsarts en werknemer besluiten in overleg dat met enkele werkaanpassingen en in afstemming met bedrijfsmaatschappelijk werker en werkgever het verrichten van werk deels mogelijk lijkt. Tijdens het 2

e consult, 6 weken na

verzuimmelding, blijkt de werknemer volledig gericht op zijn verergerde klachten. De bedrijfsarts consulteert de huisarts en gezamenlijk is hun conclusie dat deze werknemer stagneert in zijn herstelproces en dat curatieve hulp geïndiceerd is, waarna een verwijzing volgt naar een psychiater. Ze spreken af dat de bedrijfsarts procesbegeleider blijft en contact onderhoudt met werknemer, werkgever en curatieve behandelaar. De bedrijfsarts informeert de werkgever en adviseert een structurerend werkritme zonder productienorm. Tijdens het 3e consult blijkt de werknemer bij de psychiater na diagnose van depressie gestart te zijn met antidepressiva behandeling, en psychotherapeutische gespreksvoering. De bedrijfsarts adviseert werknemer en werkgever over mogelijke bijwerkingen op functioneren en werkzaamheden worden deels uitgebreid. Een maand later ervaart de werknemer duidelijk meer grip en controle op zijn activiteiten en de sociale interactie. Hij besluit met bedrijfsarts 5 maanden na ziekmelding weer volledig het werk te hervatten. De bedrijfsarts bespreekt mogelijke terugvalscenario’s en brengt werkgever, huisarts en behandelaar op de hoogte van de ontwikkelingen en aanpak.

DANKWOORD

DANKWOORD

178

In het voorwoord van mijn stage-leerverslag, de laatste horde voor het behalen van mijn opleiding Psychologie van Arbeid, Gezondheid & Organisatie, koos ik voor het volgende citaat: ‘Als iets ten einde loopt, moeten we durven denken dat er ook iets aan het beginnen is’. Deze waarheid als een koe wordt uitgemolken als ik hem hier weer gebruik. Er zijn inmiddels ook nieuwe dingen begonnen, alleen ontbrak het nog aan de laatste punt van dit proefschrift. In de wetenschap (en bijna acceptatie) dat een geschreven tekst nooit af is, ga ik deze nu dan echt zetten. Maar niet voordat ik iedereen die mij in de afgelopen roerige jaren heeft geholpen dit proefschrift te realiseren, mijn dank heb betuigd. Deelnemers onderzoek De onderzoeksresultaten die beschreven staan in dit proefschrift waren nooit mogelijk geweest zonder de deelnemers zelf. Alle politiewerknemers die hebben deelgenomen wil ik dan ook hartelijk danken voor hun bereidheid ondanks hun problemen de diverse vragenlijsten in te vullen en terug te sturen. Hetzelfde geldt voor de leidinggevenden van de deelnemers aan Project Reïntegratie Politie, die ik regelmatig heb lastig gevallen met telefoontjes en vragenlijsten tijdens hun werkzaamheden. Hoe boeiend en stressvol het politiewerk kan zijn, heb ik mogen ervaren door diensten mee te lopen met onder andere Ronald van Riessen (Hollands Midden) en Gerrit Kaper (Zaanstreek-Waterland). Dank voor deze onvergetelijke avonturen, goede gesprekken, en een ander perspectief op het uitgaansleven. Politiewerk is mensenwerk en vakmanschap, diep respect!!! Projectteam Project Reïntegratie Politie Voor een succesvolle afronding van dit project is de samenwerking en afstemming binnen het projectteam onmisbaar gebleken, zeker toen de voortgang van het project halverwege aan een zijden draadje hing. Allereerst is dit project financieel mogelijk gemaakt door het Ministerie van Binnenlandse Zaken, waarbij ik met name Fons van Gessel wil bedanken voor zijn betrokkenheid. Organisatorisch viel het projectmanagement onder regie van de Dienst Geneeskundige Verzorging Politie (DGVP). Voor de realisatie daarvan, wil ik Bas de Beer, Ad Wennekes, en met name regisseur Theo Gallee mijn dank betuigen. Inmiddels is de DGVP veranderd in de PolitieZorgPolis (PZP), bestaande uit VTS Politie Nederland en de politiebonden ACP, NPB en VMHP. Hen wil ik danken voor de financiele steun bij het realiseren van dit proefschrift. Zonder de medewerking van de deelnemende politiekorpsen in de personen van Annemieke Madderom (Politie Zaanstreek-Waterland) en Hilda de Bruin (Politie Hollands Midden), was de uitvoering van dit project echt onmogelijk gebleken. Hilda, dank voor al je informatie en meelevendheid met het onderzoek, het was voor mij erg prettig te ervaren hoe wetenschap en praktijk zo goed samen kunnen gaan. En erg leuk dat we samen recent nog samen de resultaten hebben kunnen presenteren op een symposium. Ook de betrokkenheid van de arbodiensten is van onschatbare waarde geweest. De inzet en betrokkenheid van Hynek Hlobil, later bij het projectteam gekomen door de medewerking van KLM Health Services, heb ik bijzonder gewaardeerd. Hynek, je hebt me op een bijzonder prettige en menselijke manier geholpen bij de uitvoering van dit project. Mooi dat we binnenkort ook de afronding van jouw proefschrift gaan vieren! Mede dankzij de medewerking en cofinanciering van arbodienst Achmea Vitale (voorheen Commit), vertegenwoordigd door Arie Koster, is dit project succesvol afgerond. Ook wil ik Achmea Vitale danken voor hun financiële bijdrage aan de promotiekosten, evenals Ausems en Kerkvliet, medisch arbeidsadviseurs, en Lifeguard BV. Dit onderzoek is alleen mogelijk geweest door de inzet van de professionals uit de praktijk zelf. Daarom wil ik bij deze de deelnemende bedrijfsartsen en andere betrokken medewerkers bedanken voor hun medewerking bij de instroom van de deelnemers en het invullen van de vragenlijsten.

DANKWOORD

179

We hebben het jullie niet altijd makkelijk gemaakt, maar jullie zijn ons altijd van dienst geweest. Zonder jullie geen onderzoek, dus Freek Broekman, Tim Natte, Astrid Schaaper, Liane de Vries, Nancy Gruis, Bibi de Vries, Nel Sluijk (allen Commit), Tomas Rejda, Sandra Borgia, Jetty Csanky, Masha Linovitska, en Angelique Kater (allen KLM Health services), enorm bedankt! Ook wil ik de psychologen van De Gezonde Zaak bedanken, en met name Peter Westdijk, die ons van dienst is geweest bij het realiseren van dit onderzoek en de data-verzameling. Promoteam Mijn (co-)promotoren heb ik het niet altijd even makkelijk gemaakt. En hoewel alles niet altijd even soepel verliep, ben ik ze ontzettend dankbaar voor hun begeleiding. En tja, geweldig dat jullie mij als (co-) promoteren hebben getoond hoe bewegings-wetenschappen, sport- en bedrijfsgeneeskunde samengaan in de gezamenlijke hardlooplunchrondjes die ik met jullie allen heb gehad de afgelopen jaren! Tijdens een bijeenkomst over performance indicatoren van het Kwaliteitsbureau van de NVAB, waar ik de resultaten van mijn afstudeeronderzoek presenteerde, hoorde ik voor het eerst dat er een promovendus gezocht werd voor het Project Reïntegratie Politie. Nadat Karen Nieuwenhuijsen me er op attent maakte of dat niet wat voor mij zou zijn, besloot ik David Bruinvels te bellen. De volgende dag zat ik met mijn latere copromotor David Bruinvels in een kroeg in Utrecht, waar hij me met zijn bekende overtuigingskracht enthousiast maakte voor dit onderzoek. David, allereerst nog mijn complimenten voor hoe je dit project zelf hebt opgestart en een jaar draaiende hebt weten te houden. Ik heb je ervaren als een sprinter (om in hardlooptermen te blijven), die vol daadkracht soms wel eens kort door de bocht kan gaan. Door onze nog al verschillende karakters (weet je nog, de varkens van het filmpje op het PVA-congres in Noordwijkerhout..?), wist je me telkens weer te prikkelen om nieuwe stappen te zetten in het onderzoek en meer uit mezelf en het onderzoek te halen. Ook je originele blik en praktijkervaring leidden continu tot nieuwe inzichten, net als je 'Usual Suspects filmfilterbetoog' bij het schrijven. En dan heb ik het nog niet eens over je high-tech hobbies die je er op na houdt, en die je soms in buitenlandse steden plotseling in bosjes doen verdwijnen... Dank voor het geduld dat je soms toch echt op hebt moeten brengen voor mij, en de vele leerzame en grappige momenten. Allard, ook jou ontmoette ik voor het eerst op dezelfde bijeenkomst als David. Jou heb ik ervaren als een snel startende duurloper, die snel zicht heeft op de situatie, er even op reflecteert, en dan met een stevig eindschot komt. Er was al snel sprake van enige overeenkomst aangezien we allebei een link hadden met sportdocenten in de gemeente Ede, in Amsterdam-Oost woonden, graag over voetbal praten (hoewel je de verkeerde club aanhangt), en de marathon wilden lopen. Mooi dat dat laatste is gelukt (jou zelfs 2 keer), ook al had ik je natuurlijk graag ingehaald op het eind in Rotterdam… Dank voor je opbouwend kritische begeleiding en de persoonlijke aanpak daarbij. Willem, ik heb jou ervaren als de middenafstandsloper. Constant, betrouwbaar, gevat, de boel bij elkaar houdend, met af en toe een elleboogje hier en daar bij het positioneren voor de eindsprint. Hoewel je terecht al tijdens het sollicitatiegesprek bang was dat ik teveel dingen leuk vond, heb je me toch altijd je vertrouwen gegeven. Daarvoor ben ik je dankbaar, evenals voor je 'open deur' tijdens de begeleiding, en je verfrissende en humoristische stijl van leiding geven. Recent heb ik nog erg om en met je kunnen lachen in Zuid-Afrika, al was het maar om je onverwachte zangkwaliteiten...

DANKWOORD

180

Leescommissie Leden van de leescommissie; Jaq, Roland, Han, Aartjan, Aart Schene en Marc Koopmanschap, bedankt voor het kritisch lezen van het manuscript, en het opponeren bij de verdediging. Beste Jaq, ik ben blij dat ik met dit onderzoek mag voortborduren op jouw proefschrift, wat toch de basis is geweest voor de richtlijn. Regelmatig heb je me tijdens het onderzoek geinspireerd met nieuwe inzichten en/of kritische reflectie. Roland, je bent de meeste psychologen ver vooruit geweest met je pleidooi voor werk als onderdeel van het herstelproces. Je onderzoek en presentaties hebben me geholpen in mijn werkzaamheden, evenals je adviezen in een kroeg in Stockholm… Han, ik leerde je kennen als buurman op het EMGO die al een stuk verder was met zijn promotiewerkzaamheden, en me dan ook regelmatig van goede adviezen voorzag. Inmiddels ben je al wat stappen verder, en bedien je me nog steeds met goede raad binnen de huidige samenwerking en het Work Disability Program in Canada. Leuk dat je nu in mijn leescommissie zit. Collega’s Mijn collega's bij het EMGO-instituut wil ik bedanken voor de gezellige tijd en het mogen delen van de soms frustrerende onderzoekservaringen. In het bijzonder heb ik geluk gehad met mijn B-555 kamergenoten van het eerste uur Hidde, Stefan, Lando en Ivan, en later ook Maarten, Jolanda en Robin. De 'anti-RSI-breaks' ofwel pingpongsessies hebben zeker bijgedragen in mijn werkplezier, dus dank Harm-Jan voor het vormen van een onverslaanbaar koppel, en Sander, Ellen, Matthijs, Jurriaan, Roelof, Alwin, Maurice, David S en andere lamme hazen, dank voor het bieden van enige tegenstand. Sander, hopelijk doe je nou wat langer met een auto, aangezien je die nu met iemand anders deelt… Sandra en Ludeke, dank voor veel leuke collegiale momenten op het EMGO en in Canada. Amika, dank voor het delen van lief en leed, en veel dierbare herinneringen. Alle andere EMGO-collega’s bedankt voor inhoudelijke inspiratie, leuke afdelings- en EMGO-uitjes, en natuurlijk niet te vergeten de zaalvoetbaltoernooien, het Pak EM&GO-team, en een onvergetelijke ervaring met de EMGO-band. Inge, Patrick, en Brahim…bedankt voor de ondersteuning die een chaotisch persoon als ik af en toe nodig heeft. Milou van Paridon, Marcel van der Meer en Chris Bos, het was een plezier jullie als stagiaires te begeleiden. Dank voor jullie bijdragen aan het doen slagen van dit project! Na mijn tijd bij het EMGO-instituut heb ik tijdelijk met veel plezier gewerkt bij het Coronel-instituut voor Arbeid en Gezondheid. Monique Frings, Frank van Dijk, Judith Sluiter, Teake Pal, dank voor jullie begeleiding tijdens en bijdrage aan dit leerzame project, dat me zeker van dienst is geweest bij het afronden van mijn proefschrift. Ook alle andere ex-collega’s bij het Coronel, bedankt voor de leuke tijd daar! Ik werd toen ook lid van de redactie van www.psychischenwerk.nl de website op dit terrein voor professionals. Monique, Frank, Karen, Nathalie, Moniek, en Susan, dank voor vele leerzame en gezellige ontmoetingen. Dear Work Disability Prevention (WDP) program colleagues, thanks for inspiring me in the last phase of my thesis. See you in Montreal! Inmiddels ben ik al enige tijd werkzaam bij het Trimbos-instituut binnen het programma Diagnostiek en Behandeling. Christina van der Feltz wil ik bedanken voor de mogelijkheid hier te werken op het gebied van arbeid en psyche, en je plaats in de oppositie. Ook wil ik met name Jasper Nuijen en Moniek Vlasveld bedanken voor de leuke samenwerking. En natuurlijk mijn overige D&B collega’s, jullie zijn (over de) top..! En ook bij het Trimbos hebben het zaalvoetbal-, volleybal- en fietsteam al voor broodnodige ontspanning gezorgd tijdens de laatste loodjes van mijn proefschrift, dank daarvoor.

http://www.psychischenwerk.nl/

DANKWOORD

181

Vrienden De afgelopen jaren was de steun van goede vriendschap onmisbaar. Jeroen, wat een geluk voor mij dat je Inez bent tegengekomen. Dank voor het delen van een appartement, gitaarlessen, fietslol, een emmer vol met SAUS, en je hulp bij de vormgeving van dit boekwerk. Berry the Blues, bedankt voor de vrijgezellige tijden in Amsterdam, steun in moeilijke momenten, en vele mooie concerten en gesprekken. Gert en Mart, de bonobo’s are still alive! en onze eeuwige vriendschap van onschatbare waarde. Erik, Vicky en Bart, jullie houden me scherp met altijd weer bijzondere ontmoetingen. Het herendiner was en is een inspirerend en gastronomisch festijn, want 'heersen doen de zeven wederom'. Voetbal is altijd een prachtige uitvlucht geweest van de kantoorwerkzaamheden, en al was het Sporting '70, UVV of VSC, veld of zaal, kampioen worden we toch niet… De kleedkamerlol blijft echter onmisbaar. De sportieve hoogtepunten met het Fartmedia Cycling Team zijn nog immer een genot, met dank aan Saus, Garate, Valveevers, Schlager, Schlock en Sapman, en de gezonde competitie met Anticlimax en DiscoAnal, die ongetwijfeld dit jaar in de GdA weer tot een nieuw hoogtepunt zal komen. Familie Lieve ouders, broers en schoonzussen, onze ontmoetingen zijn me ontzettend dierbaar, ben ik de afgelopen jaren nog meer achter gekomen. Het is heerlijk met jullie, en jullie kids, soms weer even kind te zijn. Er en Clas, dank voor jullie steun in barre tijden. Lieve pa en ma, fijn dat jullie er altijd voor me zijn, en de betekenis van onvoorwaardelijke liefde doen voelen. Dat ik nog maar heel lang van jullie, en de andere naaste familie, mag genieten. Paranimfen Eef, mooi dat je me tijdens de promotie tot steun wil zijn als broer, vriend, politieman, en paranimf. Je broederlijke adviezen en humor gaan me vast goed van pas komen op 2 juli. Stefan, precies een jaar na jouw promotie ben ik nu ook aan de beurt. Bedankt voor het goede voorbeeld en je hulp. Je energie, eigenwijsheden, grappen en grollen zullen me er ongetwijfeld doorheen slepen. Ik ben blij dat jullie mijn paranimf willen zijn na zo'n intens jaar. Melissa Melief, nu komt er zelfs voor mij een punt in plaats van een komma. Wat een geluk dat ik jou ben tegengekomen. Dank voor je liefde en steun de afgelopen jaren. Ik kijk er naar uit om met je te gaan genieten van wat komen gaat, want ik geloof, ik geloof, ik geloof, in jou en mij.

183

Curriculum Vitae 1977 Born in Ede 1990-1995 Christelijk Streeklyceum, Ede 1995-1996 Voluntary work, France & Israel 1996-2002 Masters degree Psychology Work, Health & Organization Faculty of Social Sciences, University of Utrecht 2003-2007 Masters degree Epidemiology, Faculty of Medicine, VU University Amsterdam 2003-2008 PhD-student Department of Public and Occupational health EMGO Institute, VU University Medical Center Amsterdam 2008-2008 Senior researcher Coronel Institute of Work and Health, Academic Medical Center 2008-present Student Work Disability Prevention Program University of Sherbrooke, Montreal, Canada 2008-present Research Associate

Trimbos Institute / Netherlands Institute of Mental Health and Addiction

185

List of publications Rebergen D, Hoenen J, Heinemans A, Bruinvels D, Bakker A, van Mechelen W. Adherence to mental health guidelines by Dutch occupational physicians. Occup Med (Lond). 2006;56(7):461-8. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W. [In Dutch] Project Reïntegratie Politie. Wetenschappelijk onderzoek naar de effectiviteit van interventies bij werknemers met psychische klachten. In opdracht van BZK/DGVP. EMGO+ Institute for Health and Care research. VU Medical Center. Amsterdam, 2006. Rebergen DS, Bruinvels DJ, van der Beek AJ, van Mechelen W. Design of a randomized controlled trial on the effects of counseling of mental health problems by occupational physicians on return to work: the CO-OP-study. BMC Public Health. 2007;7:183. Rebergen DS, Bruinvels DJ. The Management of Employees with common Mental Health Problems: Occupational Physician or Psychologist? In “Psychosocial Resources in Human Service Work”, Organizational Psychology and Health care, Vol.5. Edited by Peiro JM, Schaufeli W. Rainer Hampp Verlag, Munchen. 2007. Rebergen DS, Ashgari E, Sluiter JK, Pal TM, van Dijk FJH, Frings-Dresen MHW. [In Dutch] De invloed van arbeidsrisico's op gezondheid en arbeidsparticipatie. In opdracht van SZW/VWS/BZK. Coronel Institute of Work and Health & Netherlands Center of Occupational Diseases. Academic Medical Center. Amsterdam, 2008. Rebergen DS, Bruinvels DJ, Bezemer PD, van der Beek AJ, van Mechelen W. Guideline-based care of common mental disorders by occupational physicians (CO-OP study): a randomized controlled trial. J Occup Environ Med. 2009;51(3):305-12. Rebergen DS, Bruinvels DJ, van Tulder MW, van der Beek AJ, van Mechelen W. Cost-effectiveness of guideline-based care for workers with mental health problems. J Occup Environ Med. 2009;51(3):313-22. Rebergen DS e.a. Guideline adherence in treatment of workers with mental health problems by Occupational Physicians (CO-OP study): process evaluation of a randomized controlled trial on return to work. Submitted. Bruinvels DJ, Rebergen DS, Verbeek JH, Nieuwenhuijsen K, Madan I, Neumeyer-Gromen A. Return to work interventions for adjustment disorders. (Protocol) Cochrane Database of Systematic Reviews 2007, Issue 1. Submitted to Cochrane Occupational Health Field December 2008.

chapters of thesis - politieacademie › kennisenonderzoek › k... · nederlandse titel...

Documents