Available online at www.sciencedirect.com

ScienceDirectBehavior Therapy 45 (2014) 199–211

www.elsevier.com/locate/bt

The Gap Between Science and Practice: How Therapists MakeTheir Clinical Decisions

Alex GyaniRoz ShafranPamela Myles

University of Reading

Suzanna RoseBerkshire Healthcare NHS Foundation Trust

Recent surveys have found thatmany patients are not receivingempirically supported treatments and that therapists may notupdate their knowledge of research. Studies have found thattherapists prefer to use their clinical experience rather thanresearch findings to improve their practice, although cognitivebehavioral (CB) practitioners have been found to use researchmore frequently than therapists of other theoretical orienta-tions. The organization in which therapists work has beenshown to impact attitudes toward working practices, butstudies have not examinedwhetherworkplace requirements touse research affect therapists’ practice. Studies to date havemainly been conducted inNorth America. These findings maynot be generalizable to the United Kingdom where there is aNational Health Service (NHS), which requires the use ofempirically supported treatments. The first part of this studyaimed to investigate which factors were influential intherapists’ choice of theoretical orientation and to see whetherCB practitioners differed from other therapists in the factorsthat influenced their choice of theoretical orientation. Thesecond part tested whether therapists’ theoretical orientation

Alex Gyani is now Policy Advisor at the Behavioural InsightsTeam, Cabinet Office, London, UK. Roz Shafran is now Professorat the Behavioural and Brain Sciences Unit. University CollegeLondon, London, UK.

We would like to thank the reviewers for their helpful commentsand the Medical Research Council and the Economic and SocialResearch Council for their funding (Ref: G0900201-1/1).

Address correspondence to Alex Gyani, School of Psychology andClinical Language Sciences, University of Reading, Reading, Berkshire,RG6 6AL, UK; e-mail: alex.gyani@cabinet-office.gsi.gov.uk.0005-7894/45/199-211/$1.00/0© 2013 Association for Behavioral and Cognitive Therapies. Published byElsevier Ltd. All rights reserved.

or their workplace influenced the frequency with which theyused research in their clinical decision-making. The final partinvestigatedwhether being aCBpractitioner orworking in theNHS was associated with having a favorable attitude towardresearch. An online survey was sent to 4,144 psychologicaltherapists in England; 736 therapists responded (18.5%).Therapists reported that research had little influence over theirchoice of theoretical orientation and clinical decision-makingcompared to other factors, specifically clinical experience andsupervision. CB practitioners and NHS therapists, regardlessof their orientation, were significantly more likely to useresearch than other therapists and were more likely to have apositive attitude toward research.

Keywords: evidence-based treatments; research; sources; cognitivebehavioral therapy; dissemination

THE EVIDENCE-BASED MEDICINE MOVEMENT argues thatthe treatments offered to patients should be informedby the most robust research evidence available(Sackett, Rosenberg, Gray, Haynes, & Richardson,1996). Research evidence refers to data gatheredsystematically from a range of methodologies, includ-ing qualitative studies, case studies, and randomizedcontrolled trials. This movement has led to variousorganizations, including the American PsychologicalAssociation (APA), recommending the use of "clearlyspecified psychological treatments shown to beefficacious in controlled research with a delineatedpopulation" (Chambless&Hollon, 1998, p. 7). Thesetreatments are referred to as empirically supportedtreatments (ESTs).

200 gyan i et al .

Cognitive behavioral therapy (CBT) is the theo-retical orientation with the most empirical supportfor its efficacy in treating a range of disorders. It hasbeen consistently found to be efficacious in treatingnumerous psychological disorders in both adults(Hollon & Ponniah, 2010; Öst, 2008; Rector &Beck, 2001) and children (Chorpita et al., 2011).Some treatments from other theoretical orientationsare backed by empirical support for the treatment ofsome disorders, but not others. For example,interpersonal psychotherapy and brief psychody-namic therapy are efficacious in the treatment ofdepression and are recommended in the U.K. by theNational Institute for Health and Care Excellence(NICE, 2009; Shedler, 2010), but psychodynamictherapy has been shown to be harmful in treatingschizophrenia (Mueser & Berenbaum, 1990).Recent reviews of care have found that patients are

not receiving ESTs for psychological disorders(Baker,McFall,&Shoham, 2008;Dobson&Beshai,2013; Insel, 2010; Shafran et al., 2009). Even iftherapists practice ESTs, it does not mean they useresearch to update their practice. For example, thereis evidence that some CB practitioners rely on theirclinical experience and omit key components oftherapy for obsessive-compulsive disorder that arerequired for their practice to be consideredminimallyeffective, as demonstrated in controlled trials (Stobie,Taylor, Quigley, Ewing,& Salkovskis, 2007). This isoften referred to as “therapist drift” (Waller, 2009).There are significant problems with therapists’

reliance on their personal experience tomake clinicaldecisions, as this type of experience has been found tobe subject to cognitive biases (Dawes, Faust, &Meehl, 1989; Grove et al., 2000). For example,Hannan and colleagues (2005) compared the judg-ment of clinicians in a college counseling center witha research-derived algorithm designed to predictwhichpatientswere likely to deteriorate in treatment.Therapists predicted deterioration correctly in 1 outof 550 cases, even though it occurred in 40 patients,compared to the algorithm, which identified 100%of cases. To change therapist behavior, it is necessaryto examine the sources of information therapists useto inform their practice and use this knowledge tofocus efforts to change behavior.One of the earliest investigations into the sources

therapists use to influence their clinical decision-making found that only 10% of clinical psycholo-gists used psychotherapy research as the main sourceof informationonwhich to base their clinical practice(Morrow-Bradley & Elliott, 1986). A further study,which surveyed psychological therapists of varioustheoretical orientations as a sample, found that only35% of respondents read peer-reviewed journals(Beutler,Williams,&Wakefield, 1993). The authors

found that clinicians tended to have positive orneutral attitudes toward research and stated thatthey did use science to increase their knowledge.Unfortunately, clinicians searched for knowledgein places where peer-reviewed research articles wereunlikely to be found. For example, 76% respondedthat they readprofessional newspapers, 58%accessedpractice-oriented journals, and 51% mentionedpopular books.Recent studies have had a greater focus on con-

trasting the use of research between therapists ofdiffering theoretical orientations (Ogilvie, Abreu, &Safran, 2009; Safran, Abreu, Ogilvie, & DeMaria,2011; Stewart&Chambless, 2007; Stewart, Stirman,& Chambless, 2012). By understanding whichinformation sources therapists from different theo-retical orientations prefer, efforts to disseminate ESTscan be tailored to therapists to maximize impact(Noar, Benac, & Harris, 2007). These studies havefound substantial differences between therapists’ useof research dependingon their theoretical orientation.Stewart and Chambless (2007) completed a largesurvey of U.S. therapists in private practice and foundthat after controlling for research emphasis ingraduate school, CB therapists maintained signifi-cantly more positive views of research than psycho-dynamic practitioners and eclectic practitioners.These findings are perhaps not surprising given thatCBT is supported by a large body of researchdemonstrating its efficacy in treating a range ofdisorders, whereas there is only supportive empiricalevidence for the use of psychodynamic psychotherapyin the treatment of certain disorders (Mueser &Berenbaum, 1990; NICE, 2009; Shedler, 2010).Studies have also found that there are differences

between therapists in what drives them to choosetheir theoretical orientation. Pope and Tabachnick(1994) surveyed 800 therapists and found thatpsychodynamic therapists were more likely than CBpractitioners to believe that receiving therapy shouldbe a requirement for practicing therapy. Similarly,Orlinsky, Botermans, Rønnestad, and the Society forPsychotherapy Research Collaborative ResearchNetwork (2001) surveyed nearly 4,000 therapists inGermany, theU.S.,Norway, Portugal, and Switzerlandto see what they believed had influenced theirprofessional development. In this study, non-CBpractitioners reported that an important factor inguiding their choice of theoretical orientation wasreceiving their own therapy. CB practitioners did notview it as being as important.In addition to understanding whether results

generalize across theoretical orientations, it is impor-tant to consider whether findings generalize acrosscultures and health systems. A recent study under-taken by Safran and colleagues (2011) used a sample

201how therap i s t s make cl in i cal dec i s ions

of therapists from around the world who were allmembers of the SPR. A large proportion of thissample (82%) had conducted research themselves.The results showed that therapists believed thatresearch publications were the least helpful sourceof information for their clinical practice compared toongoing clinical experience, whichwas rated to be themost helpful source of information. Other sourcesconsidered helpful by therapists included supervision/consultation with others, conducting psychotherapyresearch, theoretical publication and/or presenta-tions, and the experience of being a client. Theauthors also compared respondents’ ratings of thehelpfulness of quantitative versus qualitative researchand foundno significant differences. Like Stewart andChambless’s study (2007), differences between theo-retical orientations were found. Psychodynamictherapists rated the importance of psychotherapyresearch significantly lower than therapists of othertheoretical orientations and ongoing experience withclients significantly higher than therapists from othertherapeutic orientations.The sample in the study conducted by Safran and

colleagues (2011) was largely North American, with52% of the sample hailing from the U.S., 7% fromCanada, and very few from the U.K. (4%). The U.K.has a unique health care system, the National HealthService (NHS), which offers treatments that are freeat the point of access for all patients and primarilyfunded through general taxation rather than requir-ing insurance payments from the government orprivate companies.In the U.K. therapists can choose to work in the

NHSandbe paid a flat rate salary by the organization,choose to solely work in private practice, or workacross sectors. All NHS workers have line managersand typically work in specific buildings. Unlike otherenvironments, the NHS working environment man-dates the delivery of evidence-based medicine inaccordance with national guidance (Department ofHealth, 2011). EachNHS service is required to submitreports about their adherence to ESTs and is regulatedby two independent government organizations: theCare Quality Commission and NHS Monitor (De-partment of Health, 2011). As well as a generalemphasis on the use of evidence-based treatments,separate services have been created to increase the useof evidence-based treatments for common mentalhealth disorders. In October 2008 the Department ofHealth invested £173million in training psychologicaltherapists in ESTs and delivering them to thepopulation as part of the Improving Access toPsychological Therapies (IAPT) initiative (Clark,2011). A further £260 million was invested in 2010to ensure that the development of the services wassustained until 2015. Investigations into these services

has found that most patients in IAPT services receiveESTs and the minority that do not have a reducedlikelihood of recovery (Gyani, Shafran, Layard, &Clark, 2013). It is important to explore a system witha unique environment as the commitment andmandate to use ESTs may result in a change ofattitude towards research.There is evidence that the organization in which

therapists work can have an impact on their attitudestoward working practices. Safran and colleagues(2011) found that therapists in their sample, whowere mainly university based, were more likely tofavor research than therapists in Morrow-Bradleyand Elliott’s (1986) sample, who were mainly basedin private practice. Hatfield and Ogles (2007) foundthat therapistsworking in an institution that requiredthem to use outcome measures were more likely touse outcome measures than those in private practice.Conversely, studies by Aarons and colleagues

have found that therapists who had less favorableattitudes towards the use of ESTs stated that theywere unlikely to adopt them, even if they wererequired to do so by the organization in which theyworked (Aarons, 2004; Aarons, Cafri, Lugo, &Sawitzy, 2012; Aarons & Sawitzy, 2006). Aarons(2004) also found that therapists who worked in aless bureaucratic organization were likely to havemore favorable attitudes toward research, indicat-ing that the type of institution is important. Surveysin the NHS have found that family physicians viewthe emphasis on using evidence-based medicine as abureaucratic exercise (Harrison & Dowswell,2002) and has been described as “the latest movein an audit revolution” (Charlton & Miles, 1998,p. 372).No previous study has been conducted exclusively

in the U.K., with therapists working in a range ofworkplace settings from a variety of orientations(including “creative arts”) and with major Govern-ment investment in the delivery of ESTs and majorrequirements to use ESTs in practice. The primaryaim of the study was to better understand the role ofresearch in (a) therapists’ theoretical orientation,(b) clinical decision-making, and (c) workplacesetting.Specifically, it was hypothesized that:

(1a) Clinical experience will be more influentialthan research when therapists choose theirtheoretical orientation.

(1b) Research will be more influential in CBpractitioners’ choice of theoretical orienta-tion than other therapists.

(2a) Clinical experience will be used more fre-quently than any other measured factor intherapists’ routine clinical decision-making.

202 gyan i et al .

(2b) CB practitioners will use research more fre-quently than other therapists when makingclinical decisions.

(2c) NHS therapists will use research morefrequently than therapists who do not workin the NHS when making clinical decisions.

(3) Both CB practitioners and therapists workingin theNHS, regardless of orientation,will havemore favorable attitudes toward research thanother therapists.

Methodparticipants

Participants were psychological therapists based inEngland. They were recruited through the websitesof the British Association for Counselling &Psychotherapies (n = 2,988), British Associationfor Behavioural and Cognitive Psychotherapies(n = 602), British Association of Art Therapists(n = 115), British Psychological Society (n = 208),and the British Association of Drama Therapists(n = 231).Seven hundred and thirty six therapists responded

(18.5%). The majority of respondents were female(n = 569, 78.5%)and theirmeanagewas54.11 years(SD = 9.18). The mean number of years of therapists’clinical experience was 12.28 (SD = 7.57) and themean number of hours per week therapists practicedwas 13.64 (SD = 7.76). Seventy-three participantsstated that their highest professional qualificationwasan undergraduate degree, 252 participants statedthat they had amaster’s degree, 325 stated they had apostgraduate diploma (a qualification obtainedafter graduating from university but without therequirement of an independent research project),38 stated they had a doctorate (8 of which weremedical doctorates), and 30 stated that they eitherdid not have one of the qualifications listed, or thattheir highest qualification was not listed. This is inline with the sample in Aarons’s (2004) study, inwhich 10% had doctorates, and Aarons andcolleagues’ (2012) study, in which 7% had doctor-ates. However, this is different from the sample inStewart and Chambless’s (2007) study, in which95% had a doctorate, and from the sample in Safranand colleagues’ (2011) study in which 68% hadPh.D.-level qualifications.The theoretical orientations that respondents

identified with can be seen in Figure 1. Eighty-seventherapists used the open text box to state that theywould consider themselves eclectic, but preferred theterm “integrative therapists.” The categories “inte-grative” and “eclectic” were combined. As very fewtherapists stated that they were cognitive analyticaltheory therapists (n = 3), family-based interventionstherapists (n = 4), or eye movement desensitization

and reprocessing therapists (n = 10), these categorieswere collapsed into the “other” category. A largeproportion of therapists stated that they worked inboth the NHS and in private practice. Theseparticipants were analyzed as being in the NHS toisolate the effect of the NHS and to ensure thatanalyses were suitably powered. Respondents whoidentified with the “other” category when statingtheir workplace tended to respond that they workedin the charitable sector, an education setting or for anEmployment Assistance Program. A logistic regres-sion indicated that CB practitioners were nearly fourtimes more likely to work in the NHS (eitherexclusively in the NHS or by working in both theNHS and private practice) than therapists of anyother theoretical orientation: Wald Statistic (1) =42.22, Odds ratio = 3.92, p b .001, 95% LowerCI = 2.60, 95% Upper CI = 5.93.

materials

A questionnaire was developed for this study. Itcomprised the following four sections: (a) questionsabout respondents’ demography, (b) questionsabout the influences of therapists’ choice of thera-peutic orientation, (c) questions about the regularitywithwhich therapists used certain sources to enhancetheir clinical practice, and (d) questions abouttherapists’ attitudes toward research. Sections b, c,and d each took the form of a single question towhich participants responded on a 7-point Likertscale. The four parts of the questionnaire aredescribed in detail below.

DemographyThedemographic variables collectedwere: theoreticalorientation, workplace context, age, gender, years ofclinical experience, respondents’ highest qualificationand number of hours a week they spent treatingpatients. The categories that therapists could choosefrom were as follows: humanistic/experiential,cognitive/behavioral, family-based interventions,art, drama or music (referred to as creative artstherapists), eye movement desensitization and repro-cessing, psychodynamic, cognitive analyticaltheory, and eclectic. Family-based interventionsrefer to Maudsley-based family therapy devised byChristopher Dare and colleagues for eating disorders(Russell, Szmukler, Dare, & Eisler, 1987). Thisorientation was included in the Stewart andChambless study (2007). An open text box wasincluded for participants who felt that they werenot represented in these groups. Therapists pickedtheir one primary orientation. To assess whererespondents worked, they were asked: “Do youwork in” and were presented with four tick boxes:NHS, Private Practice, both, or other. Respondents

Fre

qu

ency

FIGURE 1 Chart showing the theoretical orientation and workplace setting of the study’s participants.

203how therap i s t s make cl in i cal dec i s ions

who entered “other” were given an open text boxto respond.

Influences on Theoretical OrientationThis section assessed how certain sources of infor-mation influenced therapists’ choice of therapeuticorientation. The following question was posed:“How influential was each of the following indeciding your favoured theoretical orientation?”Participants were then given a list of sources, whichwere taken from the Stewart and Chambless (2007)study. The list of sources included: research evidence,personal experience with clients, clinical training, thetreatment’s appeal to clients, intuition, and pastexperience with personal therapy. Participants wererequired to give a response for each source ofinformation. A value of 1 was recorded if partici-pants responded “Not at all influential,” a score of 4was recorded if participants stated “Sometimes” anda value of 7 was recorded if participants respondedwith the answer “Extremely influential.”

Investigating Factors Influencing ClinicalDecision-MakingThis section assessed the frequency with whichtherapists used certain sources to enhance theirclinical practice. The following question was posed:“Which sources do you currently use to increase yourskills and effectiveness as a clinician?” Participantswere then given a list of sources, which were adaptedfrom the Stewart and Chambless (2007) study. Thelist of sources included: personal experience withclients, empirical research from controlled trials,supervision, empirical research from case studies,peer discussion, popular books, clinical case obser-vations, outcomemeasures, and clinical guidelines. Avalue of 1 was recorded if participants responded“Never,” a score of 4 was recorded if participants

stated “Sometimes,” and a value of 7 was recorded ifparticipants responded with the answer “Always”.Participants were required to give a response for eachsource of information.

Research Attitudes IndexThe section used to investigate therapists’ attitudestoward research was based on the Esteem for andUtilization of Research Scale (Stewart & Chambless,2007) and the Evidence-Based PracticeAttitude Scale(EBPAS; Aarons, 2004). It was necessary to adaptthese measures as the EBPAS specifically focuseson the use of manualized treatments rather thanresearch in general and the Esteem for andUtilizationof Research Scale only focuses on controlled trials.The Research Attitudes Index was developed by AG,RS, and SR to include beliefs about the value ofresearch above clinical experience. All the questionsasked in the index can be seen in Table 1. Thequestionswere adapted to suit theU.K. context of thestudy. A value of 1 was recorded if participantsresponded “Strongly disagree,” a score of 4 wasrecorded if participants stated“Neutral,” anda valueof 7 was recorded if participants responded with theanswer “Strongly agree.” Cronbach’s alpha was .78and the Kaiser–Meyer–Olkin Measure of SamplingAdequacywas .79 on the basis of the responses of the450 participants who gave responses to all items. Asingle factor solution accounted for 19.4% of thevariance.

procedure

An email was sent to all respondents to ask whetherthey would like to take part in a study investigatingtheir attitudes toward research and the sourcesthey use to make clinical decisions (n = 4,144). Theemail contained a link to the questionnaire, hosted

Table 1Questions Asked for the Development of the Research Attitudes Index

Forward Coded Reverse Coded

I am willing to use new and different types oftherapy developed by researchers

My clients are more complex and diverse than those in clinical trials

It is my professional duty to keep up with newdevelopments in treatment research

My clients prefer other treatments than evidence based treatments

My training emphasized research findings Therapy cannot be manualisedMy supervisors require me to use evidence basedtreatments*

The diagnoses used in treatment trials are too simple

Evidence based treatments are cost effective The treatments I favour have not been tested in a randomizedcontrolled trial

I may attract new clients by learning an evidencebased treatment

I have an individual treatment approach for each client

It is important to incorporate scientific findings intomy everyday practice

I do not have the time to learn evidence based treatments

I would try a new therapy even if it was very differentfrom what I am used to doing

Training in evidence based treatments would cost me too muchmoney personally

I am interested in learning evidence based treatments I do not know which treatments are evidence basedThe treatments I use with my clients have an empiricalbasis

My clinical training did not provide sufficient information aboutevidence based treatmentsThe therapeutic alliance is more important than learning how to do aspecific form of psychotherapyMost therapies are equally effectiveMy employer does not have the funds for training in evidence basedtreatments*Clinical experience is more important as a guide to treatment thanresearch evidence

Note. An asterisk indicates that the item was removed due to missing data.

204 gyan i et al .

by Surveymonkey.com. This was sent in September2011. After the initial email, a reminder was sent3 weeks later. Of the 4,144 emails sent, 156 emailsdid not reach their intended recipient. The responserate was calculated from the 3,988 emails that weresent and did not bounce. The contact details oftherapists were obtained through the websites ofthe organizations listed above. Therapists were toldthat for every person taking part in the survey, 10pwould be donated to the Samaritans or MIND, twoU.K.-based mental health charities.

analytical approach

All analyses were conducted using SPSS 19 and Excel2011. Repeated measures ANOVAs were used tocompare the sources used to influence the respon-dents’ choice of theoretical orientation and thefrequency with which respondents stated they usedsources of information to make clinical decisions.Planned comparisons were used to understandwhether research was used more or less often thanother sources. Where the assumption of normalitywas not met, Friedman’s test andWilcoxon sign testswere used as a replacement for repeated measuresANOVAs and planned comparisons, respectively. Amultiple linear regression was used to investigate

whether being a CB practitioner and whether atherapist who worked in the NHS was associatedwith attitudes toward research. The multiple linearregression allowed the simultaneous investigation ofwhether or not a therapist worked in theNHS and/orwas a CB practitioner, age, gender, years of clinicalexperience, and number of hours worked as atherapist. Power analyses were modelled on theresponse rates from the previous literature (Stewart& Chambless, 2007). The use of a simultaneousentry method was important, as CB practitionerswere more likely to work in the NHS than othertherapists. Cronbach’s alpha was used to assess theinternal consistency of the Research Attitudes Index.A Principal Component Analysis was used tounderstand how much variance was explained by asingle factor.MANOVAs were conducted to investigate the

effect of theoretical orientations and organizationson the sources they used to influence their choice oftheoretical orientation or clinical decision-making. Tosimplify the analyses, all non-CB practitioners werecollapsed into one group and compared to CBpractitioners. Box’s M test and Levene’s test wereused to test for violations in the assumptions ofequality of covariance matrices and homoscedasticity.

205how therap i s t s make cl in i cal dec i s ions

Rank transformations were used when the assump-tions of normality were violated (Zwick, 1985). Ifthese measures failed, a Mann-Whitney U test wasused. Post hoc planned comparisonsweremadewith aBonferroni correction. Partial eta-squared, r, and betaare reported as effect sizes for MANOVAs, nonpara-metric tests, and linear regression analyses, respec-tively. Sensitivity analyses were undertaken to testwhether IAPT therapists were more likely to useevidence from controlled studies than other NHStherapists and to test whether therapists whoworkedexclusively in the NHS were more likely to useevidence from controlled studies than therapists whoworked in both the NHS and private practice.

Resultshypothesis 1a: clinical experiencewill be more influential thanresearch when therapists choosetheir theoretical orientation

All dependent variables violated the assumptions ofnormality according to Kolmogorov-Smirnov tests;therefore, a nonparametric Friedman test was used. Asignificant difference between the importance of thesources on therapists’ theoretical orientation wasobserved, χ2(5) = 726.50, p b .001; see Table 2.Therapists’ choice of theoretical orientationwasmostfrequently based on intuition, personal experience,and clinical training. Use of research was reported tobe the least influential of all the options provided. Posthoc planned comparisons found that research studieswere used significantly less than personal experiencewith clients (z = 18.15, p b .001, r = .71), pastexperiences with personal therapy (z = 12.03,p b .001, r = .51), clinical training (z = 17.15,p b .001, r = .67), intuition (z = 18.52, p b .001,

Table 2Mean Responses to the Question, “How influential were each of th

Theoretical Orientation(n)

ResearchEvidence (SD)

Personal Experiencewith Clients (SD)

PasPer

Humanistic orExperiential (173)

3.29 (1.93) 5.95 (1.39) 5.3

CBT (62) 5.31 (1.68) 5.89 (1.16) 4.0Music Drama or Artstherapist (35)

3.74 (1.77) 6.20 (1.16) 5.5

Psychodynamic (87) 3.03 (1.67) 5.67 (1.65) 5.7Integrative orEclectic (103)

4.10 (1.61) 6.19 (0.89) 5.0

Other (64) 3.73 (1.98) 5.98 (1.25) 5.3Total (524) 3.73 (1.91) 5.97 (1.30) 5.2

Note. A value of 1 was recorded if participants responded “Not at all influeand a value of 7 was recorded if participants responded with the answe

r = .72), and whether therapists felt their choiceof theoretical orientation would appeal to clients(z =14.90, p b .001, r = .59).

hypothesis 1b: research will be moreinfluential in cb practitioners’choice of theoretical orientationthan other therapists

A MANOVA was used to investigate whether CBpractitionersweremore likely than other therapists toreport that research was influential in choosing theirtheoretical orientation. A rank transformation wasused. Tests of equality of variance and equality ofcovariances were nonsignificant, indicating that theanalyses met the required assumptions. Identifying asa CB practitioner was found to have an impact on thesources therapists felt were important when firstchoosing their theoretical orientation, Hotelling’sTrace = .162, F(6, 515) = 13.95, p b .001, ηp

2 =.140. CB practitioners were more likely to useresearch evidence to inform their choice of theoreticalorientation than other therapists, F(1, 520) = 54.49,p b .001, ηp

2 = .095. CB practitioners were less likelyto report that their past experience with personaltherapy, F(1, 520) = 29.00, p b .001, ηp

2 = .053, andintuition, F(1, 520) = 6.40, p = .012, ηp

2 =.012,influenced their choice of theoretical orientation.There were no significant differences between CBpractitioners and other therapists in the perceivedinfluence of the following factors on choice oftheoretical orientation: personal experience withclients, F(5, 520) = 1.68, p = .196, ηp

2 = .003),clinical training, F(1, 520) = 0.27, p = .605, ηp

2 =.001, and whether therapists felt their choice oftheoretical orientation would appeal to clients,F(1, 520) = 0.02, p = .904, ηp

2 b .001.

e following in deciding your favoured theoretical orientation?”

t Experiences Withsonal Therapy (SD)

ClinicalTraining (SD)

Intuition(SD)

Appeal ToClients (SD)

8 (1.61) 5.90 (1.36) 6.43 (0.94) 5.62 (1.42)

2 (1.98) 5.85 (1.19) 5.89 (1.31) 5.45 (1.43)4 (1.72) 6.37 (0.81) 6.23 (1.19) 5.89 (1.21)

4 (1.50) 5.91 (1.27) 6.13 (1.05) 4.92 (1.56)9 (1.44) 5.76 (1.06) 6.05 (0.99) 5.38 (1.39)

0 (1.72) 5.97 (1.19) 6.41 (1.05) 5.73 (1.29)2 (1.69) 5.91 (1.22) 6.23 (1.06) 5.47 (1.43)

ntial,” a score of 4 was recorded if participants stated “Sometimes,”r “Extremely Influential.”

Tab

le3

Mea

nRes

pons

esto

theQue

stion,

“Which

source

sdo

youcu

rren

tlyus

eto

increa

seyo

ursk

illsan

deffectiven

essas

aclinician?

”

The

oretical

Orie

ntation(n)

Perso

nalE

xperienc

ewith

Clients(SD)

Sup

ervision

(SD)

Empiric

alEvide

nce

from

Con

trolledTria

ls(SD)

Empiric

alEvide

nce

from

Cas

eStudies

(SD)

Pee

rDiscu

ssion

(SD)

Pop

ular

Boo

ks(SD)

Clinical

Cas

eObs

erva

tions

(SD)

Outco

me

Mea

sures

(SD)

Clinical

Guide

lines

(SD)

Human

isticor

Exp

eriential

(209

)6.04

(1.16)

6.23

(0.94)

3.68

(1.31)

3.92

(1.25)

4.81

(1.14)

3.11

(1.22)

4.53

(1.28)

4.15

(1.39)

4.05

(2.19)

CBT(112

)5.83

(1.21)

5.90

(1.00)

4.97

(1.21)

4.68

(1.13)

4.84

(1.20)

3.61

(1.29)

5.15

(1.13)

5.20

(1.18)

4.53

(1.92)

MusicDramaor

ArtsTherap

ist(46)

6.33

(0.94)

6.33

(0.87)

3.80

(1.56)

4.78

(1.50)

5.28

(1.15)

2.83

(1.22)

4.87

(1.50)

4.57

(1.60)

4.98

(1.82)

Psy

chodyn

amic

(111

)6.06

(0.97)

6.28

(0.89)

3.71

(1.38)

4.16

(1.19)

4.87

(1.03)

2.86

(1.25)

4.64

(1.19)

4.11

(1.32)

4.19

(2.01)

Integrative

or

Eclec

tic(129

)5.86

(1.07)

6.05

(1.04)

3.78

(1.35)

3.95

(1.25)

4.63

(1.15)

3.20

(1.24)

4.55

(1.27)

4.16

(1.39)

4.01

(1.93)

Other

(85)

6.15

(1.07)

6.20

(0.99)

3.92

(1.62)

3.93

(1.49)

4.84

(1.35)

2.98

(1.40)

4.51

(1.52)

4.26

(1.60)

3.52

(2.06)

Total(69

2)6.00

(1.11)

6.15

(0.97)

3.95

(1.44)

4.15

(1.31)

4.83

(1.17)

3.13

(1.28)

4.67

(1.31)

4.36

(1.44)

4.14

(2.05)

Note.

Ava

lueof

1was

reco

rded

ifpa

rticipan

tsresp

onde

d“N

ever,”asc

oreof

4was

reco

rded

ifpa

rticipan

tsstated

“Som

etim

es,”an

dava

lueof

7was

reco

rded

ifpa

rticipan

tsresp

onde

dwith

the

answ

er“Alway

s.”

206 gyan i et al .

hypothesis 2a: clinical experiencewill be used more frequently thanany other measured factor in therapists’routine clinical decision-making

A Friedman’s ANOVA showed that therapistsreported using some sources of information signifi-cantly more often than others, χ2(8) = 2296.39,p b .001. The most used source was supervision,followed by personal experience with clients, clinicalguidelines, peer discussion, clinical case observations,outcome measures, case studies, controlled trials,followed by popular books. Table 3 shows the meanresponses to questions about clinical decision-making.Planned comparisons found that controlled studieswere used significantly less than personal experiencewith clients (z = 19.74,p b .001, r = .59), supervision(z = 20.89, p b .001, r = .79), empirical researchfrom case studies (z = 4.86, p b .001, r = .19), peerdiscussion (z = 12.51, p b .001, r = .48), clinical caseobservations (z = 10.63, p b .001, r = .40), and out-come measures (z = 7.38, p b .001, r = .28). Therewas no significant difference between the frequency ofuse of clinical guidelines and controlled trials after theBonferroni correction was applied (z = 2.42, p =.008, r = .19). Popular books were used significantlyless than controlled studies (z =11.02, p b .001, r =.42). As previous studies have found that personalexperience was used more often than all othersources, a Wilcoxon test was used to clarify whethertherapists’ preference for supervision to past experi-enceswith clientswas significant, and this was foundto be the case (z = 4.03, p b .001, r = .15).

hypothesis 2b: cb practitioners willuse research more frequently thanother therapists when makingclinical decisions

A MANOVA found that therapists’ theoreticalorientation was associated with the frequency withwhich they reported using certain sources to makeroutine clinical decisions, Hotelling’s Trace = .191,F(9, 679) = 14.38, p b .001, ηp

2 = .160. The analysiswas undertaken with a rank transformation due toinequality of covariance matrices. CB practitionersreported using the following sources of informationmore frequently thanwas reportedbyother therapists:research evidence from controlled trials, F(1, 687) =76.56, p b .001, ηp

2 = .100, empirical research fromcase studies, F(1, 687) = 24.19, p b .001,ηp

2 = .034,popular books, F(1, 687) = 19.69, p b .001, ηp

2 =.028, clinical case observations, F(1, 687) = 19.94,p b .001, ηp

2 = .028, outcome measures, F(1, 687) =50.48, p b .001, ηp

2 = .068, and clinical guidelines,F(1, 687) = 4.48, p = .035, ηp

2 = .006. CB practi-tioners were also likely to use supervision lessfrequently than other therapists, F(1, 687) = 9.89,

207how therap i s t s make cl in i cal dec i s ions

p = .002, ηp2 = .014. There was no association

between theoretical orientation and the frequencywith which therapists stated that they usedtheir personal experience with clients, F(1, 687) =2.71, p = .100, ηp

2 = .004, or peer discussion, F(1,687) =0.085, p = .770, ηp

2 b .001, to inform theirpractice.

hypothesis 2c: nhs therapists will useresearch more frequently thantherapists who do not work in thenhs when making clinical decisions

A third MANOVA was used to test whethertherapists working in the NHS were more likely touse research to make clinical decisions than thera-pists working outside the NHS. A significant impactof working in the NHS was found, Hotelling’sTrace = .049, F (8, 680) = 4.17, p b .001, ηp

2 =.047. Therapists working in the NHS were likely touse the following sources more frequently than othertherapists: empirical evidence from controlled trials,F(1, 687) = 29.41, p b .001, ηp

2 = .041, case studies,F(1, 687) = 12.74, p b .001, ηp

2 = .018, and out-come measures, F(1, 687) = 10.03, p = .002, ηp

2 =.014. There were no significant differences betweenNHS therapists and other therapists in the frequencywithwhich they stated theywould use: their personalexperience with clients, F(1, 687) = 0.53, p = .465,ηp

2 = .001, supervision, F(1, 687) = 0.03, p = .866,ηp2 b .001, peer discussion, F(1, 687) = 3.14, p =

.077, ηp2 = .005, popular books, F(1, 687) = 1.51,

p = .219, ηp2 = .002, and clinical case observations,

F(1, 687) = 1.15, p = .284, ηp2 = .002. Whether or

notNHS therapistsweremore likely to use guidelineswas assessed using a Mann-Whitney U test as thisvariable failed to meet the assumption of homo-scedascity and was removed from the MANOVAdescribed above. The results showed that NHStherapists were significantly more likely to use clini-cal guidelines than other therapists, Mann-WhitneyU = 36865, Z = 3.35, p = .001, r = .13.Fifty-five of the therapists who stated they worked

in the NHS reported that they worked in an IAPTservice (30.2%). To test whether NHS therapistswere only more likely to use research because IAPTtherapists were using research, a Mann-Whitney testwas used to investigate the extent to which IAPTtherapists used research from controlled studiescompared to other NHS therapists. No significantdifference was found, Mann-Whitney U = 2921.5,Z = 0.394, p = .694, r = .03.Similarly, only a small number of therapists

stated that they worked exclusively in the NHS(n = 37). The majority of those who worked in theNHS also worked in private practice (n = 145). Totest whether these groups differed in their use of

research, a Mann-Whitney test was used to comparethe extent to which therapists who only worked inthe NHS used research from controlled studiescompared to therapists who worked in both theNHS and private practice. No significant differencewas found, Mann-Whitney U = 1923.5, Z = 0.956,p = .339, r = .07.

hypothesis 3: both cb practitionersand therapists working in the nhswill have more favorable attitudestoward research than other therapists

Two variables had significant missing data (47%and 74% of responses) and were not entered intothe regression. These missing data are indicatedwith an asterisk in Table 1. Respondents’ answersto the questions were totaled to create the ResearchAttitudes Index, which was found to be normallydistributed: Kolmogorov-Smirnov (450) = .035,p = .200.A multiple linear regression was used to investigate

if a therapist’s theoretical orientation, and whether ornot they work in the NHS, influenced their attitudestoward research. This model was found to explain35.0% of the variance in Research Attitudes Indexscores (adjusted R2 = .339) and was significantlybetter at predicting scores than a model that justcontained the constant, F(6, 343) = 30.782, p b .001.The analysis shows that when demography wasconsidered, CB practitioners (B = 15.91, LowerCI = 13.02, Upper CI = 18.80, p b .001) and NHStherapists were still more likely to have favorableresearch attitudes (B = 2.59,LowerCI = 0.21,UpperCI = 4.98, p = .033). Older therapists were likely tohave less favorable research attitudes (B = -0.18,Lower CI = -0.32, Upper CI = -0.05, p = .007).Gender (B = 0.08, Lower CI = -0.39, UpperCI = 4.60, p = .097), years of clinical experience(B = -0.02, Lower CI = -0.18, Upper CI = 0.13,p = .732) and hours of therapy practiced per week(B = -0.04, Lower CI = -0.20, Upper CI = 0.08,p = .400) had no significant effect on researchattitudes. Casewise diagnostics indicated that therewere no residuals larger than one would expect.The tolerance and Variance Inflation Factor (VIF)statistics indicated that multicollinearity was notan issue with these data (1.075 b all VIFs b 1.445;.692 b all tolerances b .930).

DiscussionThe findings of this study are consistentwith previousdata demonstrating that research only plays a minorrole in influencing therapists’ choice of theoreticalorientation and clinical decision-making. Whenchoosing their theoretical orientation, therapistsreported that they were more likely to rely on their

208 gyan i et al .

intuition, clinical training, and personal experiencewith therapy (either received or given) than research.This study also found that self-identified CB practi-tioners were more likely to use research evidence tochoose their theoretical orientation. These results fitwith Stewart and Chambless’s (2007) findings thatCB practitioners are more likely to use research toinform their choice of theoretical orientation thantherapists of other orientations.The finding that, generally, clinicians prefer to rely

on clinical experience rather than research to informtreatment decisions replicates previous studies (Beutleret al., 1993; Beutler, Williams, Wakefield, &Entwistle, 1995; Morrow-Bradley & Elliott, 1986;Stewart & Chambless, 2007). Unlike previoussurveys, this study found that therapists consideredsupervision to be significantly more important thanclinical experience. The studies undertaken by Beutlerand colleagues (1993; 1995) and Stewart andChambless (2007) did not investigate whether thera-pists considered supervision to be an important factorin making clinical decisions. These factors were con-sidered by Morrow-Bradley and Elliott (1986) andSafran and colleagues (2011), neither of whom foundsupervision to be as important as clinical experience.Interestingly, supervision was reported to be used

more frequently by non-CB practitioners. Thisfinding might also explain why the current studyfound that supervisionwas themost important factorin clinical decision-making compared to Safran andcolleagues’ (2011) and Stewart and Chambless’s(2007) studies. These studies had a greater propor-tion of CB practitioners than the current study. Inthis study 16% of respondents stated that theywere CB practitioners compared to 20% in Safranand colleagues’ study and 45% in Stewart andChambless’s study. This result indicates that super-visionmay be a way to reach therapists whomay notbe currently practicing ESTs. The way in whichtrainers and supervisors could be encouraged tovalue and discuss research evidence with trainees andsupervisees, and whether this will increase itslikelihood of impact on practice, is not clear.Research on evidence-based supervision is sparse(Beidas&Kendall, 2010), but there is some evidencedemonstrating that supervision based on acceptanceand commitment training can increase the likelihoodof therapists implementing ESTs (Luoma et al., 2007;Varra, Hayes, Roget & Fisher, 2008). In the IAPTprogram, supervisors are given additional university-based training according to a national curriculumthat emphasizes the importance of ESTs and researchmore generally. Training supervisors to use theseinterventions may increase the implementation ofresearch evidence into practice, but challenges interms of time, interest, and funding still remain.

Stewart and Chambless (2007) did not reportwhether there were any differences between theoret-ical orientations in terms of therapists’ use ofpersonal experience with clients when makingclinical decisions. This study found that CB practi-tioners were no more likely to use their personalexperience with their clients when making clinicaldecisions than other therapists. This finding suggeststhat CB practitioners do not feel they undervaluetheir own clinical experience when using researchand integrate both research and clinical experiencewhen making clinical decisions. Such integration isconsistent with original descriptions of evidence-based medicine (Sackett et al., 1996), rather thanseeing the choice between the use of research and theuse of clinical experience as mutually exclusive.Similar to Stewart and Chambless’s (2007) find-

ings, this study found that therapists reportedpopular books to be the least important sourcewhen making clinical decisions, unlike previouswork by Beutler and colleagues (1993, 1995). Thecurrent study found that CB practitioners were morelikely to report that they used popular books to aidtheir clinical decision-making than other therapists.This means that popular books may be a moreeffective way of keeping CB practitioners up to datewith research than it would be for other therapists.The finding that non-CB practitioners used outcomemeasures less often than CB practitioners replicatesHatfield andOgles’s work (Hatfield&Ogles, 2007).The benefits of using outcome measures in treatmenthave been demonstrated across many theoreticalorientations (Hannan et al., 2005; Lutz et al., 2006).Overall, self-identified CB practitioners were signif-icantly more likely to use external sources, such ascase studies, controlled studies, outcome measures,clinical case observations, and clinical guidelines inclinical decision-making; other therapists relied moreon their own or their supervisors’ experiences. Thesample as a whole reported that their past experienceas patients was a strong influence on their choice ofinitial theoretical orientation. This replicates previ-ouswork (Orlinsky et al., 2001; Pope&Tabachnick,1994). However, Macran and Shapiro (1998)reviewed nine studies that looked at the link betweenreceiving therapy and patient outcomes. They foundno conclusive link between having received psycho-logical treatment andhaving better patient outcomes.Work by Aarons (2004) has found that therapists

who state that they have less favorable attitudestowards the use of ESTs are less likely to adopt them,despite any requirements by the organization inwhich they work. In this study, clinicians working inthe NHS were more likely to use research whenmaking clinical decisions and had more favorableattitudes toward research than therapists working

209how therap i s t s make cl in i cal dec i s ions

outside the NHS. It is not clear whether research-focused clinicians are moving to the NHS or if theNHS is training therapists to become more researchfocused, as this was a cross-sectional study. None-theless, the results do suggest that the emphasis in theNHS on ESTs is associated with therapists with agreater inclination to use research compared to thoseworking outside the organization. IAPT therapists,who were trained specifically to deliver ESTs (Clark,2011), were no more or less likely to use researchthan otherNHS therapists. This finding suggests thatthere is a broader reason why therapists in the NHSare more likely to use research than purely becausethere has been a recent initiative to increase their useacross English mental health services.In this study older therapists were found to be less

favorable toward research than younger therapists,even once the number of years of experience thattherapists had was controlled for. This replicates afinding by Aarons and Sawitzky (2006)—that oldertherapistsweremore likely to perceive evidence-basedpractice as not clinically useful and less importantthan clinical experience.The present study has a number of limitations.

Although the absolute sample size in this studycompares favorably to previous work (Beutler et al.,1993; Safran et al., 2011; Stewart & Chambless,2007), the response rate for the study was 18.5%,which couldmean that therewas a selection bias in theresults. The response rate compares favorably toSafran and colleagues’ study (2011), which had aresponse rate of 12%, but unfavorably to previousstudies by Stewart and Chambless (2007), which hada response rate of 25%.Another limitation of this study is that personality

was not considered as a factor in therapists’ use ofresearch, or their choice of theoretical orientation. Ina recent study of 46 therapists, CB practitioners werereported as less angry and impulsive than psychody-namic or humanistic therapists, but less open totheir feelings and the values of others (Boswell,Castonguay, & Pincus, 2009). However, the choiceof orientation for psychodynamic psychotherapistsmay be more dependent on their training than theirpersonality (Buckman & Barker, 2010). Furtherwork could be undertaken to investigate the associ-ation between personality variables and theoreticalorientation, clinical decision-making and the influ-ence of the workplace.This study used questions based on those in Stewart

and Chambless’s study (2007). Therapists could havebeen asked about a broader range of research sources,which would have potentially led to further under-standing aboutwhat type of research therapists prefer.For example, therapists could be asked whether theyprefer studies that had no exclusion criteria versus

studies that had a high number of exclusion criteria, orstudies reporting treatments with unified protocols orindividualized treatment manuals, or more proces-s-oriented research. This would be an interestingavenue for further exploration.Likewise, the theoretical orientations that therapists

used to describe themselves could have been brokendown into further categories. This is a limitation fortwo reasons. First, therapists may feel that their choicemay not accurately reflect their practice, although thiswas mitigated by the inclusion of an open text box.Second, some types of psychodynamic treatments, likebrief psychodynamic therapy, have been shown to beefficacious in the treatment of depression (NICE,2009; Shedler, 2010), but other studies have foundthat some types of psychodynamic therapy may beharmful to patients with schizophrenia (Mueser &Berenbaum, 1990). If the study had asked psychody-namic therapistswhich specific types of treatment theyuse, it may have been possible to establish with moreaccuracy whether therapists who practice ESTs aremore likely to use research to inform their clinicaldecisions. Such a distinction would have the disad-vantage of leading to a large number of groups thatwould have reduced the power of the analyses.In this investigation, therapists were not asked

about their reasons for their limited use of research.A follow-up qualitative study has been conductedto investigate why therapists prefer certain sourcesof information, which offers some insights intoways that could improve dissemination of researchinto practice (Gyani, Shafran, Rose & Lee, 2013).This follow-up study also provided insight intowhat therapists considered useful research. Gyaniand colleagues found that therapists’ objections arebased on their beliefs about the translation ofresearch into practice, but they also object to thetools used to disseminate research.

conclusions

The current study aimed to investigate whetherresearch is used to guide therapists’ choice oftheoretical orientation and the extent to whichresearch is used to inform clinical decisions. Thera-pists reported that they used research less frequentlythan other sources of information when makingclinical decisions. CB practitioners were more likelyto use research than therapists of other theoreticalorientations and not feel that they compromised theiruse of clinical experience by using research to informtheir practice. Furthermore, therapists that workedin theNHSweremore likely to use research to informtheir practice than other therapists.The exclusive use of clinical experience could lead

therapists to use treatments that are not optimallyeffective (Stobie et al., 2007) and could ultimately lead

210 gyan i et al .

toworse patient outcomes (Gyani, Shafran, Layard&Clark, 2013). This study provided some positive newideas for encouraging therapists to use research intheir clinical practice. In particular, the data indicatedthe importance of supervision. Special supervisionsessions for non-CB practitioners could be used todiscuss the relevance of research to clinical decision-making. For self-identified CB practitioners, popularbookswere felt to be important and therefore could beused to keep therapists up to date with researchfindings to avoid therapist drift. The finding thatNHStherapists were more likely to use research to informtheir practice highlights that the organization inwhichone works can have an impact on the use of researchto inform practice and highlights that having anorganization that mandates the use of evidencemay increase the quality of care that patients receive.In conclusion, the study indicates that therapistscontinue to use their personal experience in clinicaldecision-making more than research data and thatmuch remains to be done to bridge the gap betweenscience and practice.

Conflict of Interest StatementThe authors declare that there are no conflicts of interest.

References

Aarons, G. A. (2004). Mental health provider attitudes towardadoption of evidence-based practice: The Evidence-BasedPractice Attitude Scale (EBPAS). Mental Health ServicesResearch, 6(2), 61–74.

Aarons, G. A., Cafri, G., Lugo, L., & Sawitzky, A. (2012).Expanding the domains of attitudes towards evidence-basedpractice: The Evidence based practice attitude scale-50.Administration and Policy in Mental Health and MentalHealth Services Research, 39(5), 331–340.

Aarons, G. A., & Sawitzky, A. C. (2006). Organizationalculture and climate and mental health provider attitudestoward evidence-based practice. Psychological Services,3(1), 61–72.

Baker, T. B.,McFall, R.M.,& Shoham, V. (2008). Current statusand future prospects of clinical psychology. PsychologicalScience in the Public Interest, 9(2), 67–103. http://dx.doi.org/10.1111/j.1539-6053.2009.01036.x

Beidas, R. S., & Kendall, P. C. (2010). Training therapists inevidence-based practice: A critical review of studies from asystems-contextual perspective.Clinical Psychology, 17, 1–30.http://dx.doi.org/10.1111/j.1468-2850.2009.01187.x

Beutler, L. E., Williams, R. E., & Wakefield, P. J. (1993).Obstacles to disseminating applied psychological science.Applied and Preventive Psychology, 2(2), 53–58.

Beutler, L. E., Williams, R. E., Wakefield, P. J., & Entwistle,S. R. (1995). Bridging scientist and practitioner perspec-tives in clinical psychology. American Psychologist, 50(12),984–994.

Boswell, J. F., Castonguay, L. G., & Pincus, A. L. (2009). Traineetheoretical orientation: Profiles and potential predictors.Journal of Psychotherapy Integration, 19(3), 291–312.

Buckman, J. R., & Barker, C. (2010). Therapeutic orientationpreferences in trainee clinical psychologists: Personality ortraining? Psychotherapy Research, 20(3), 247–258.

Chambless, D. L., & Hollon, S. D. (1998). Defining empiricallysupported therapies. Journal of Consulting and ClinicalPsychology, 66(1), 7–18.

Charlton, B. G., & Miles, A. (1998). The rise and fall of EBM.Quarterly Journal of Medicine, 91(5), 371–374.

Chorpita, B. F., Daleiden, E. L., Ebesutani, C., Young, J.,Becker, K. D., Nakamura, B. J., Phillips, L., Ward, A.,Lynch, R., & Trent, L. (2011). Evidence-based treatmentsfor children and adolescents: An updated review ofindicators of efficacy and effectiveness. Clinical Psychology:Science and Practice, 18(2), 154–172.

Clark, D. M. (2011). Implementing NICE guidelines for thepsychological treatment of depression and anxiety disorders:The IAPT experience. International Review of Psychiatry,23(4), 318–327. http://dx.doi.org/10.3109/09540261.2011.606803

Dawes, R. M., Faust, D., &Meehl, P. E. (1989). Clinical versusactuarial judgment. Science, 243(4899), 1668–1674.

Department of Health. (2011). Operational guidance to theNHS – extending patient choice of provider. London.Retrieved October 3, 2013, from https://www.gov.uk/government/publications/operational-guidance-to-the-nhs-extending-patient-choice-of-provider–2.

Dobson, K., & Beshai, S. (2013). The theory-practice gap incognitive behavioral therapy: Reflections and a modest pro-posal to bridge the gap. Behavior Therapy. http://dx.doi.org/10.1016/j.beth.2013.03.002

Grove, W.M., Zald, D. H., Lebow, B. S., Snitz, B. E., Nelson, C.,& Snitz, B. E. (2000). Clinical versus mechanical prediction:A meta-analysis. Psychological Assessment, 12(1), 19–30.

Gyani, A., Shafran, R., Layard, R., & Clark, D. M. (2013).Enhancing recovery rates: Lessons from year one ofIAPT. Behaviour Research and Therapy, 51(9), 597–606.http://dx.doi.org/10.1016/j.brat.2013.06.004

Gyani, A., Shafran, R., Rose, S. & Lee, M.J. (2013)A qualitativeinvestigation of therapists’ attitudes towards research: Horsesfor courses? Manuscript submitted for publication.

Hannan, C., Lambert, M. J., Harmon, C., Nielsen, S. L.,Smart, D. W., Shimokawa, K., & Sutton, S. W. (2005). Alab test and algorithms for identifying clients at risk fortreatment failure. Journal of Clinical Psychology, 61(2),155–163.

Harrison, S., & Dowswell, G. (2002). Autonomy and bureau-cratic accountability in primary care: What English generalpractitioners say. Sociology of Health & Illness, 24(2),208–226.

Hatfield, D. R., &Ogles, B. M. (2007). Why some clinicians useoutcome measures and others do not. Administration andPolicy in Mental Health and Mental Health ServicesResearch, 34(3), 283–291.

Hollon, S. D., & Ponniah, K. (2010). A review of empiricallysupported psychological therapies for mood disorders inadults. Depression and Anxiety, 27(10), 891–932.

Insel, T. R. (2010). Translating scientific opportunity intopublic health impact. Archives of General Psychiatry, 66(2),128–133.

Luoma, J., Hayes, S. C., Twohig, M., Roget, N., Fisher, G.,Padilla,M., Bissett, R., &Kohlenberg, B. (2007). Augmentingcontinuing educationwith psychologically focused groupconsultation: Effects on adoption of group drug counseling.Psychotherapy: Theory, Research, Practice, Training, 44,463–469.

Lutz, W., Lambert, M. J., Harmon, S. C., Tschitsaz, A., Schürch,E., & Stulz, N. (2006). The probability of treatment success,failure and duration—What can be learned from empiricaldata to support decision making in clinical practice? ClinicalPsychology & Psychotherapy, 13(4), 223–232.

211how therap i s t s make cl in i cal dec i s ions

Noar, S. M., Benac, C. N., & Harris, M. S. (2007). Doestailoring matter? Meta-analytic review of tailored printhealth behavior change interventions. Psychological Bulletin,133(4), 673–693.

Macran, S., & Shapiro, D. A. (1998). The role of personal therapyfor therapists: A review.British Journal ofMedical Psychology,71(1), 13–25. http://dx.doi.org/10.1111/j.2044-8341.1998.tb01364.x

Morrow-Bradley, C., & Elliott, R. (1986). Utilization ofpsychotherapy research by practicing psychotherapists.American Psychologist, 41(2), 188–197.

Mueser, K. T., & Berenbaum, H. (1990). Psychodynamictreatment of schizophrenia: Is there a future? PsychologicalMedicine. 20(2), 253–262.

National Institute for Health and Clinical Excellence. (2009).Depression: The treatment and management of depressionin adults (update). Retrieved from www.nice.org.uk/CG90.

Ogilvie, A. J., Abreu, I., & Safran, J. D. (2009). What findingsdo psychotherapy researchers use in their own practice? Asurvey of the Society for Psychotherapy Research. The NewSchool Psychology Bulletin, 3(2).

Orlinsky, D. E., Botermans, J., & Rønnestad, M. H., & the SPRCollaborative Research Network. (2001). Towards anempirically grounded model of psychotherapy training:Four thousand therapists rate influences on their development.Australian Psychologist, 36(2), 139–148.

Öst, L.-G. (2008). Cognitive behavior therapy for anxietydisorders: 40 years of progress.Nordic Journal of Psychiatry,62(S47), 5–10.

Pope, K. S., & Tabachnick, B. G. (1994). Therapists as patients:A national survey of psychologists' experiences, problems,and beliefs. Professional Psychology: Research and Practice,25(3), 247–258.

Rector, N. A., & Beck, A. T. (2001). Cognitive behavioraltherapy for schizophrenia: An empirical review. The Journalof Nervous and Mental Disease, 189(5), 278–287.

Russell, G. M., Szmukler, G. I., Dare, C., & Eisler, II. (1987).An evaluation of family therapy in anorexia nervosa andbulimia nervosa. Archives of General Psychiatry, 44(12),1047–1056.

Sackett, D. L., Rosenberg, W. M. C., Gray, J. A. M., Haynes, R.B., & Richardson, W. S. (1996). Evidence based medicine:

What it is and what it isn't. British Medical Journal,312(7023), 71–72.

Safran, J. D., Abreu, I., Ogilvie, J., & DeMaria, A. (2011). Doespsychotherapy research influence the clinical practice ofresearcher–clinicians? Clinical Psychology: Science andPractice, 18(4), 357–371.

Shafran, R., Clark, D. M., Fairburn, C. G., Arntz, A., Barlow,D. H., Ehlers, A., . . . Wilson, G. T. (2009). Mind the gap:Improving the dissemination of CBT. Behaviour Researchand Therapy, 47(11), 902–909. http://dx.doi.org/10.1016/j.brat.2009.07.003

Shedler, J. (2010). The efficacy of psychodynamic psychotherapy.American Psychologist, 65(2), 98–109.

Stewart, R. E., & Chambless, D. L. (2007). Does psychotherapyresearch inform treatment decisions in private practice?Journal of Clinical Psychology, 63(3), 267–281.

Stewart, R. E., Stirman, S. W., & Chambless, D. L. (2012). Aqualitative investigation of practicing psychologists' attitudestoward research-informed practice: Implications for dissem-ination strategies. Professional Psychology: Research andPractice, 43(2), 100–109.

Stobie, B., Taylor, T., Quigley, A., Ewing, S., & Salkovskis, P.M.(2007). “Contents May Vary”: A pilot study of treatmenthistories of OCD patients. Behavioural and CognitivePsychotherapy, 35(03), 273–282. http://dx.doi.org/10.1017/s135246580700358x

Varra, A. A., Hayes, S. C., Roget, N., & Fisher, G. (2008). Arandomized control trial examining the effect of acceptanceand commitment training on clinician willingness to useevidence-based pharmacotherapy. Journal of Consultingand Clinical Psychology, 76(3), 449–558.

Waller, G. (2009). Evidence-based treatment and therapist drift.Behaviour Research and Therapy, 47(2), 119–127. http://dx.doi.org/10.1016/j.brat.2008.10.018

Zwick, R. (1985). Nonparametric one-way multivariate analysisof variance: A computational approach based on thePillai-Bartlett trace. Psychological Bulletin, 97(1), 148–152.

RECEIVED: April 29, 2013ACCEPTED: October 17, 2013Available online 28 October 2013