comparing nurses' and patients' pain evaluations: a study of hospitalized patients in...

Sm. Sci. Med. Vol. 36, No. 5, pp. 683492, 1993 Printed in Great Britain. All rights reserved

0277-9536/93 $6.00 + 0.00 Copyright 0 1993 Pergamon Press Ltd

COMPARING NURSES’ AND PATIENTS’ PAIN EVALUATIONS: A STUDY OF HOSPITALIZED PATIENTS

IN KUWAIT

ANN HARRISON*

Faculty of Medicine, University of the United Arab Emirates, P.O. Box 17666, Al Ain, United Arab Emirates and Faculty of Medicine, University of Kuwait, Kuwait

Abstract-All eligible patients hospitalized on the general medical, surgical and pediatric wards of a district hospital in Kuwait during the first 2 weeks of April 1990 (N = 199) were interviewed about their pain and tde medical care provided. Patients rated thei; current pain using a Cl0 visual analogue scale (VAS) on which 0 was labelled ‘no pain’ and 10 ‘unbearable pain’, and also the least and worst levels of pain which they had experienced during the previous 24 hr. Pediatric patients rated their mood at these times using a cartoon faces scale ([I]: McGrath P. A., DeVeber L. L. and Hearn M. T. Aduunces in Pain Research and Therapy, pp. 387-393. Raven Press, New York, 1985). Patients were asked to describe the most painful procedure which they had experienced whilst hospitalized, to detail their reactions to any uncontrolled pain, and recount the outcome of any requests made for additional analgesics. Whichever nurse on duty had had prime responsibility for caring for the patient was asked to furnish comparable VAS ratings of the patient’s pain, and to categorize the patient as providing an accurate picture of, exaggerating, or understating his/her pain. Overall, nurses returned significantly lower pain estimates than their patients. Pediatric patients and patients described by their nurses as ‘understating their pain’, however, were exceptions. The vast majority of patients approved of the medical care provided and the pain management, and considered that nurses are well able to judge patients’ pain and are concerned about what pain patients experience. And yet the VAS reports from patients indicated that uncontrolled pain was common. Also, when specifically questioned, a substantial proportion of patients stated that pain had had a negative impact on their mood and activity. Various explanations for these findings are discussed, and their implications for effective pain management. Some of the special problems facing nurses in Kuwait are considered.

Key words-pain, evaluation, hospital, cross-cultural

INTRODUCTION

Accurate pain assessment is essential for good medi-

cal care, for judging the status and progress of patients, the impact and efficacy of treatments and sometimes for reaching a proper diagnosis [2]. Nurses play a pivotal role in monitoring pain, and yet often doubt their ability to do so adequately [3,4]. Despite the availability of a range of appropriate methods, quantitative pain assessment is not yet a regular component of nursing care [2,5,6]. Even the pain information which nurses currently record has been shown to be incomplete and inaccurate, and to describe location more often than severity [7-lo]. In one study, only half of the pain information provided by patients was noted, and over a third of the information recorded did not accord with patients’ reports [lo].

An early technique used to investigate nurses’ pain assessments involved providing them with written descriptions of patients and asking them to estimate how much pain such patients were experiencing [ll-161. This approach enabled researchers to vary the information provided and to investigate the im-

*Address correspondence to: Dr Ann Harrison, Faculty of Medicine, University of the United Arab Emirates, P.O. Box 17666 Al Ain, United Arab Emirates.

pact of different variables (such as the patient’s diagnosis, prognosis, signs, symptoms, age, sex, social status, observed reactions and verbal comments) in a systematic and controlled fashion. Marginal factors, such as the social status of the patient, were found to influence nurses’ pain assessments, and nurses were shown to rate pain consistently differently from pro- fessionals not involved in health care. A major draw- back of the vignette method is that it does not simulate the usual clinical situation. A vignette pro- vides the nurse with a preselection of information; whereas nurses normally choose which pain cues to attend to, and can use further questioning and obser- vation to pursue issues until they are confident of their evaluations. Under normal clinical conditions, the nurse also has access to subtle cues from the patient’s voice, facial expression, body tone and posture. Studies in which nurses have rated video recordings of patients indicate that such cues do influence nurses’ pain evaluations [ 171. Video recordings share with vignettes, however, the limitation that normal interaction between the patient and the nurse is precluded.

More recent studies have compared the pain evaluations returned by actual patients and their nurses. The technique often employed was to ask both the nurse and the patient to rate the severity of the

683

684 ANN HARRISON

patient’s pain using a visual analogue scale (VAS) [18]. A study of 36 chronic pain and 33 acute pain patients found that both groups reported higher pain levels than their nurses, and that the difference was significant for the chronic group [19]. A study of 47 patients in the early phase of acute myocardial infarc- tion showed that 40% of the pain ratings provided by patients and nurses did not match [20]. In a study involving 42 severe burn patients, the amounts of pain experienced by patients both ‘at rest’ and during painful therapeutic procedures, such as hydrotherapy and wound dressing, were considered [2]. Disagree- ments were common: 70% of the therapy evaluations returned by patients and their nurses, and 51% of the ‘at rest’ evaluations, were discordant. Nurses who had worked longer in burn units agreed more fre- quently with their patients about how painful therapeutic procedures are. Longer experience, however, was also associated with a growing tendency by nurses to return lower pain estimates than their patients, as other researchers have reported [l5, 171. Eighty-live percent of the discordant ratings returned by nurses with more than 2 years experience of burn nursing were underestimates, contrasted with 22% for nurses who had worked on such units less than 6 months.

The above studies involved clinical conditions where high levels of pain are encountered. The aim of the present study was to investigate the agreement shown by nurses and patients from general medical, surgical and pediatric wards of a district hospital in Kuwait. where a broader range of pain levels should pertain. Patients on burn and intensive care units were specifically excluded. Again, VAS rating was employed. VAS scaling has proved a reliable and valid tool for measuring clinical pain [IB]. The form selected (Fig. 1) has been used for a number of previous hospital studies in Kuwait, and has proved effective for securing pain information from illiterate and poorly educated patients, and young children [21l23]. Patients were asked to rate their current pain, and also the worst and least pain they had experienced during the previous 24 hr. Comparable data were then collected from whichever nurse on the ward had had major care contact with that patient.

Previous research has shown that nurses disagree regularly with the pain ratings provided by their patients [24]. These studies. however, did not explore

the issue of whether nurses misjudge their patients responses or simply disagree with them. Nurses may be able to predict very accurately what pain evaluations will be returned by their patients, but select discordant ratings because they believe these consti- tute a more appropriate evaluation [25]. In the current study, nurses were asked to characterize their patients as ‘exaggerating’, ‘understanding’ or ‘providing an accurate picture of’ their pain, so that the frequency of disagreements for these groups could be compared. Nurses and patients were asked for their opinions about how accurately nurses perceive patients’ pain. Patients were asked to evaluate the care they had received, to comment on any difficulties that they had experienced regarding pain management and to report on how pain had affected their mood and activities.

METHOD

One hundred and ninety-nine patients (85 medical, 89 surgical and 25 pediatric) were interviewed. The age profiles and hospitalization histories of the patient groups are shown in Table 1. Two of the children were 4 years of age, 5 were less than 7 years. Five adult patients refused to participate blaming their current discomfort; but all of the parents, children and nurses who were approached agreed.

Ethical clearance was received from the Ministry of Health, State of Kuwait. The protocol set was to interview all eligible patients hospitalized on pediatric, surgical and medical wards in the Mubarak Al-Kabir hospital, Kuwait, during a 2 week period. The following patients were excluded: patients who had been hospitalized for less than 24 hr, children under 3 years of age, psychiatric patients and patients judged by the medical staff to be too ill to participate. The nursing director of the hospital informed staff about the general aims of the study and appealed for their co-operation, it was stressed that nurses’ answers would not be used to evaluate individual staff. Pretesting was used to identify ambiguities and inadequacies in the interview schedules, and the final versions contained very few open-ended items. Allied Health majors in the second year of a 4 year degree

Table I. The subject groups

Medical Surgical Pediatric

Fig. I. Visual analogue pain scale

Number 85 89 25

Age (years) Mean (SD) 48.4 (15.8) 38.1 (15.3) 8 I (2.4) Range 12-92 13-73 4412

Sex % Males 46% 64% 72%

% Females 54% 36% 28% Patients who had been hospitalked precious/)

Number 53 49 25

% Group 63% 55% 44%

Previous hospitalizations Mean (SD) 4.3 (8.1) 2.0 (1.4) 6.4 (14.7)

Range I-50 l-6 l-48

Hospitalized for same condition 72% 33% 82%

Nurses’ and patients’ pain evaluations in Kuwait 685

program acted as interviewers. They were trained in general interviewing skills, and in how to administer and score the questionnaire in a standard fashion. Potential subjects (nurses and patients) were informed about the purpose of the study, and their co-operation was sought. It was made clear that participation was voluntary and could be curtailed at any time, and that all replies would be treated with strict confidentiality. In the case of pediatric patients, initial informed consent was provided by a parent, and then children were given the choice of whether or not to participate. Interviewers worked in pairs. One interviewed the patient, and later, the other interviewed whichever nurse on duty had had prime responsibility for caring for that patient. The nurse and second interviewer were never aware of a patient’s replies.

The patient interview began by gathering information about age and hospitalization history. The patient rated his/her current pain on a t&10 VAS on which 0 was labelled ‘no pain’ and 10 was labelled ‘unbearable pain’ (Fig. 1). Arabs read from right to left, hence the format selected. The patient then rated the worst pain and least pain he/she had experienced during the preceding 24 hr. Patients were asked whether they had received sufficient medication to control their pain, and about the outcome of any requests for additional drugs. Patients were asked whether, and how, pain had interfered with their activities and mood, and what was the most painful procedure they had experienced while hospitalized. Patients were also asked for their views about nurses’ responses to patients’ pain. Interviews with pediatric patients followed the same format, except that each time children provided a VAS pain rating they were also asked to select which face (Fig. 2) best signified their feelings at that time. The nurse interview began

4 3

6

by asking for VAS ratings of the patient’s current pain, and worst pain and least pain experienced during the preceding 24 hr. The nurse was then asked to characterize the patient as exaggerating, understating, or providing an accurate picture of his/her pain. Finally, nurses were asked about how accurately nurses perceive patients’ pain.

RESULTS

Subjects displayed no difficulty or hesitation in providing VAS pain ratings: they responded quickly and did not change their minds. Children showed the same facility when asked to select a face from the mood scale (Fig. 2). Non-parametric tests which assume ordinal, but not interval, measurements were used to analyse the VAS and mood ratings. The worst pain estimates from 2 children were not recorded, and these missing data are indicated by an asterisk in the data tables.

Comparing patients’ and nurses’ pain evaluations

Patients chose consistently higher VAS levels than nurses (Fig. 3) when describing current pain (Kruskal-Wallis one way analysis of variance, x2 = 34.49, P < O.OOOS), worst pain (Kruskal-Wallis one way analysis of variance, x2 = 36.44, P < 0.0001) and least pain (Kruskal-Wallis one way analysis of variance, x2 = 21.96, P < 0.05). The pain ratings provided by a given patient and his/her nurse were also contrasted (Table 2). When the data from all patients were included, nurses were found to select reliably lower VAS levels than their patients (Wilcoxon matched-pair signed-ranks test, P < 0.05) for describing current pain, worst pain and least pain. Nurses’ and patients’ pain estimates correlated significantly (Spearman’s rank order coefficient of correlation,

Fig. 2. Mood scale (reproduced from McGarth P. A., DeVeber L. L. and Hearn M. T. Multidimensional pain assessment in children. In Advances in Pain Research and Therapy (Edited by Fields H. L., Dubner

R. and Cervero F.), pp. 387-393. Raven Press, New York, 1985).

686 ANN HARRISON

Table 2. ComDarine oatients’ and nurses’ oain estimates

N of patients with VAS WilCOXOIl Spearman’s rank matched pair order coefficient

N > Nurse = Nurse <Nurse signed-ranks test of correlation

Current pain Worst pain Least pain






AN patients

199 110 50 39 z = 5.95, P < 0.0001 197 126 40 31 ; = 7.72, P < 0.0001

199 99 55 45 I = 4.77, P < 0.0001

Medical parients 85 31 28 20 z = 2.83, P < 0.005 85 54 I8 I3 z = 5.04. P < 0.0001 85 39 28 I8 I = 3.09, P < 0.005

Surgicul putients

89 60 I4 I5 ; = 5.28, P < 0.0001 89 57 I9 I3 z = 5.62, P < 0.0001 89 46 22 21 2 = 3.30. P < 0.005

Pediatric patients

25 13 8 4 z = 1.59, P < 0.05

23’ 15 3 5 z = 1.83. P < 0.05

25 I4 5 6 z = 1.74, P < 0.05

Patients who are characterized by ntmes as exaggerating their porn

33 I9 8 6 z = 2.87, P < 0.005

32* 28 3 I -=4.11. P<0.0001 33 I6 IO 7 z = 2.43, P < 0.05

Patients who are characterized by nurses us understating their pain

II 5 4 2 i = 1.52, P < 0.05 II 6 3 2 T = 1.54. P < 0.05 II 2 3 6 z = 1.05, P < 0.05

rs = 0.39, P < 0.0001 IS = 0.37, P i O.OOOl rs = 0.27, P < 0.0001

rs = 0.39, P < 0.0001 i-s = 0.42, P < 0.001 rs = 0.35, P < 0.005

i-s = 0.35, P < 0.0001 rs = 0.33, P < 0.001 rs = 0.15, P c 0.05

I = 0.52, P < 0.005 rz = 0.33, P < 0.05 rs = 0.45, P < 0.05

r., = 0.48, P < 0.005 rs = 0.17, P < 0.05 rs = 0.25, P < 0.05

r.5 = 0.59, P i 0.05 r.v = 0.69, P < 0.01 rs = 0.17, P < 0.05

Patients who ore characterized by nurses as prooiding an accurate picture of their pain

Current pan 151 85 35 31 z = 4.92, P < 0.0001 rs = 0.35, P < 0.0001 Worst pain 150; 90 33 27 z = 6.57, P < 0.0001 r.7 = 0.40, P < 0.0001 Least pan 151 80 40 31 z = 4.49, P < 0.0001 r.~ = 0.29, P < 0.0001

Patients who believe that nurses haue a good idea about patients’ pain

Current pain 151 83 39 29 I = 5.42, P < 0.0001 r,, = 0.40, P < 0.0001 Worst pain I so* 92 32 26 : = 6.64, P < 0.0001 IS = 0.32, P < 0.0001 Least pain I51 76 41 34 : = 4.45, P < 0.0001 r.v = 0.32. P < 0.0001

Patients who believe that nurses do not haue a good idea about patients’ pain

Current pain 47 21 II 9 z = 2.93, P < 0.005 rs = 0.35, P < 0.01 Worst pain 46* 33 8 5 z = 3.95, P < 0.0001 rs = 0.53, P < 0.0001 Least pain 41 23 I3 11 z = 1.86, P <0.05 rs = 0.05, P < 0.05

Nurses who belieoe nurses have o good idea about patients puin

Current pain 160 89 42 29 7 = 5.62, P < 0.0001 I.7 = 0.34, P < 0.001 Worst pain 160 98 34 28 : = 6.43, P < 0.0001 rs = 0.36, P < 0.001 Least pain 160 71 49 34 z = 4.39, P < 0.0001 rs = 0.32, P < 0.0001

Nurses who believe nurses do not haue a good idea about patients pain

Current pain 34 19 8 7 z =2.15. P <0.05 i-s = 0.52, P < 0.001 Worst pain 33* 25 6 2 i = 4. IS, P < 0.0001 rs = 0.49, P < 0.005 Least pain 34 20 5 9 2 = 1.99, P < 0.05 rs=O.l2, P<O.O5

‘The worst pain estimates from 2 children were not recorded

P < 0.05). When the results for different groups of patients were analysed (Table 2) in most instances it was found that nurses returned lower pain estimates than their patients. Pediatric patients, and patients described by their nurses as understating their pain, were exceptions. For most groups, the current pain assessments returned by patients and their nurses correlated significantly and positively, as did their worst pain reports. The worst pain reports for pediatric and ‘exaggerating’ patients were exceptions. For only half of the groups considered did nurses’ and patients’ least pain reports correlate significantly.

Patient variables

The three patient groups (Tables 3 and 4) differed in their current pain reports (Kruskal-Wallis one way analysis of variance, x2 = 10.58, P < O.Ol), with pediatric patients selecting lower VAS levels and adult surgical patients higher pain ratings. The 3 patient groups did not differ significantly, however, in terms

of their worst pain (Kruskal-Wallis one way analysis of variance, 1’ = 1.05, P > 0.05) or least pain (Kruskal-Wallis one way analysis of variance, x2 = 2.65, P > 0.05) reports. When nurses’ VAS evaluations were analysed, the 3 patient groups did not differ significantly in their current pain (Kruskal-Wallis one way analysis of variance, x2 = 5.12, P > 0.05), worst pain (Kruskal-Wallis one way analysis of variance, x2 = 0.29, P > 0.05) or least pain (Kruskal-Wallis one way analysis of variance, x2 = 1.46, P > 0.05) reports. Male and female patients did not differ significantly in the levels of current pain (Kruskal-Wallis one way analysis of variance, x2 = 14.86, P > 0.05) worst pain (Kruskal-Wallis one way analysis of variance, x2 = 5.91, P > 0.05) or least pain (Kruskal-Wallis one way analysis of variance, x2 = 13.43, P < 0.05) which they reported. The worst pain reports of patients who had been hospitalized previously did not differ reliably from those of patients hospitalized for the first time (Kruskal-Wallis one way analysis of variance, x2 = 70.69, P > 0.05).


Worst pain

a 0 1 2 3 4 5 6 7 8 9 10

VAS level

2 60

4 8 50 Least pain

& 40 e

H Patients

2 30 q Nurse

& 20 I 2 s

10

5 0.

0 012345 6 7 8 9 10

VAS level

2 50 0 B 40 Current pain

Eb W Patients 2 30

It! q Nurse

CG 0 1 2345678 9 10

VAS level

Fig. 3. Nurses’ and patients’ pain ratings.

General ratings

When asked to evaluate the medical care provided, patients responses were overwhelmingly positive: 58% rated it as ‘very good’, 42% as ‘good’, 4% as ‘poor’ and 2% as ‘very poor’. Surgical, medical and pediatric patients did not differ significantly in their judgements (x2 test, x2 = 5.96, df = 6, P > 0.05). Fifty-two percent of patients stated that the pain care they had received was ‘very good’, 42% rated it as ‘good’, 4% as ‘poor’ and 2% as ‘very poor’. Surgical, medical and pediatric patients again did not differ significantly in their judgements (x2 test, x2 = 4.23, df = 6, P > 0.05).

Pain managemenl

Twelve percent of the adults and 32% of the pediatric patients reported that they had not received sufficient medication to control their pain. Just over half of these adults (57%), but no children, had requested additional medication (x2 test, x2 = 7.06, df = 2, P < 0.05). None had received additional medication, but only 3 were told that their request

could not be met. In the other cases, the request was either ignored or staff promised to bring more medication but did not do so. Of the patients who had not asked for additional drugs, 40% gave as the reason that they thought patients should not ask, 30% reported that they thought that they had to put up with the pain.

Patients were asked whether pain had interfered with designated activities. Sixty-four percent reported that pain had interfered with sleeping, 46% with eating, 38% with enjoying visitors and 43% with their ability to concentrate. Medical, surgical and pediatric patients did not differ significantly in their responses (x2 test, P > 0.05). A higher proportion of surgical (72%) than medical (54%) or pediatric (36%) patients, however, reported that pain had interfered with moving around (x2 test, x2 = 12.46, df = 2, P < 0.005). The groups did not differ significantly in what impact they reported pain had had on them emotionally (x2 test, P > 0.05): 52% reported that it had made them miserable, 65% anxious, 41% angry and 45% frightened. When asked to nominate the most painful procedure they had experienced whilst hospitalized, 52% of the children stated venipuncture and 28% injection. The most commonly mentioned procedure by surgical patients was the surgery itself (33%) 25% nominated venipuncture and 17% injection. Thirty-five percent of the medical patients selected venipuncture and 33% injection.

The majority of patients (76%) stated that nurses have an accurate picture of patients’ pain, and the surgical, medical and pediatric group did not differ in this regard (x2 test, x2 = 1.52, df = 2, P > 0.05). Eighty-five percent believed that nurses care enough about their patients’ pain, and again the patient groups did not differ (x2 test, x2 = 0.98, df = 2, P > 0.05). Half of the patients (51%) reported that nurses try techniques other than drugs in an attempt to relieve pain, among the approaches mentioned were conversation, repositioning, exercise and mas- sage.

Nurse interviews

Nurses were asked to characterize the patients interviewed as exaggerating, understating, or providing an accurate assessment of their pain. Most patients (77%) were classified as accurate reporters, 17% as exaggerating and 6% as understating; the three patient groups did not differ significantly (x2 test, x2 = 4.66, df = 4, P > 0.05) in their composition. The vast majority of nurses (83%) stated that nurses have an accurate picture of their patients’ pain. A greater proportion of nurses on pediatric wards (33%) stated that nurses do not, but the replies of nurses on the three types of ward did not differ significantly (x2 test, x2 = 4.77, df = 2, P > 0.05).

Correlating children ‘s pain and mood

Children selected a different subset of VAS levels and faces when describing their least and worst pain

688 ANN HARRISON

Table 3. Patients’ pain reports

VAS pain level selected (%)

0 I 2 3 4 5 6 7 8 9 IO N

Current pain Medical 32 4 8 6 2 21 2 5 5 4 II 85 Surgical 13 5 I7 I2 IO 21 6 5 3 I 7 89 Pediatric 28 20 32 8 4 0 8 0 0 0 0 25

Worst pain Medical I8 5 3 5 I II 2 I3 8 II 23 85 Surgical IO I 5 6 4 I6 3 7 12 IO 26 89 Pediatric 9 4 I3 9 4 4 4 9 17 9 I7 23*

Least pain Medical 41 13 I5 7 2 9 2 2 0 4 4 85 Surgical 32 9 I8 10 7 II 6 I I 2 3 89 Pediatric 12 44 24 I2 4 4 0 0 0 0 0 25

*The worst pain estimates from 2 children were not recorded.

(Fig. 4). Children’s mood and VAS ratings correlated significantly for current pain (Spearman’s rank order coefficient of correlation, rs = 0.57, N = 24, P < 0.005) and least pain (Spearman’s rank order coefficient of correlation, rs = 0.66, N = 23, P < O.OOl), but not worst pain (Spearman’s rank order coefficient of correlation, rs = 0.09, N = 22, P > 0.05).

Child’s pain

n Worst

p1 Least

c 0

e, 20 2 t

Y 10 ; a

0 0 I 2 3 4 5 6 7 8 9 10

VAS level

Child’s mood

40

1 n Worst pain

vl FaL east pain

Face

Fig. 4. Pediatric patients’ pain and mood ratings.

DISCUSSION

Comparing patients’ and nurses’ VA.9 ratings

A major finding of the present study was that the nurses returned consistently lower pain estimates than their patients. This was not because nurses thought that most patients exaggerated their pain. Indeed, over three quarters of the patients were looked upon by their nurses as providing accurate data, and the nurses provided consistently lower pain evaluations than these patients. The only groups where disagreement was not significant were patients who were thought by their nurses to ‘understate’ their pain and pediatric cases. Although the number of ‘understating’ patients was small, this finding sup- ports the interpretation that nurses were underestimating their patients’ pain. In the case of pediatric patients, the data do not necessarily imply that children evoked a more generous pain rating from nurses. The children reported consistently lower levels of pain than the adult patients interviewed, and so there was less scope for disagreement. Further research is needed to disentangle whether the lower pain reports of the children were due to a different representation of clinical conditions, differences in pain reporting or differences in pain management. Other studies have shown that hospitalized children sometimes underreport pain in order to avoid medical procedures, such as injections, which they find more threatening than the pain itself [7]. In the present study, more than a quarter of the pediatric patients stated that injection was the most painful procedure that they had experienced while hospitalized, over a half chose another very common procedure- venipuncture.

The pain ratings provided by patients and their nurses usually correlated significantly, indicating that while nurses were biased towards understating patients’ pain, they were nonetheless sensitive to gradations in patients’ pain. Similar correlations have been reported previously [2,20]. Further research is required to establish why nurses and children’s evaluations did not always correlate significantly. It may be linked to the lower pain levels of the pediatric group, the reticence of children to complain of pain,


Table 4. Summary of VAS selections


Medical Surgical Pediatric Medical Surgical Pediatric Medical Surgical Pediatric

By Patients Mean 3.15 3.92 1.72 5.85 6.39 5.14 2.19 2.63 1.64 Median 3 4 2 I I I I 2 I Mode 0 5 2 IO IO S/l0 0 0 0 Number 85 89 25 85 89 23’ 85 89 25

By Nurses Mean 2.64 2.03 1.12 3.59 3.82 3.88 1.19 1.56 1.12 Median I I 0 4 3 5 0 0 I Mode 0 0 0 0 0 0 0 0 0 Number 85 89 25 85

*The worst pain estimates from 2 children were not recorded.

89 25 85 89 25

or the difficulties which nurses face when trying to interpret pain information provided by pediatric patients because of the children’s linguistic and cognitive immaturity [6,21,23,25-271. More nurses working on pediatric wards than adult wards returned the judgement that nurses do not have a good idea of their patients’ pain.

Disagreement. Given the subjective nature of pain evaluations, it is not surprising that nurses and patients sometimes provide discordant ratings. Among other factors which jeopardize agreement are the ‘plastic’ nature of pain, which means that nurses cannot generalize from their experience with similar patients, and the fact that pain is a multi-dimensional experience with sensory, affective and evaluative facets [5,25]. What is interesting about this and previous studies is the preponderance of underestimates returned by nurses. Nurses are not simply inaccurate, they are systematically biased in their estimates of patients’ pain. Various explanations for this have been offered [2,24,25]. Pain is a common component of medical care, and so nurses may become inured to it. Nurses probably have a different frame of reference when evaluating pain than their patients, and this is important. Nurses will have had some contact with patients with exceptionally painful clinical conditions, and so may be less willing to employ high pain ratings than patients without this experience. It has been shown that women who have experienced significant pain before pregnancy rate childbirth as being less painful than women without this experience [28]. It has been speculated that student nurses who underestimate patients’ pain are more likely to complete their training and, after- wards, to select positions which bring them into direct contact with patients in high pain [2]. There are also grounds for believing that underestimations will arise when nurses face patients who are experiencing more pain than the nurse believes is acceptable, but the nurse lacks an effective way of reducing the pain (maybe the nurse does not have an effective way of securing additional drugs or maybe the nurse is wary of addiction risks). When faced with such a dilemma (cognitive dissonance), one solution is to devalue the patient’s pain. In Kuwait, pro re nata prescriptions were rarely encountered in the sample studied and so nurses’ direct control was indeed limited.

If the patient’s estimate is accepted as a valid index of the amount of pain suffered [29], then consistent underestimation by nurses is likely to result in inadequate pain control. Nurses may consciously opt not to administer analgesics requested by the patient because they believe that they have a more appropriate assessment of the patient’s needs. Alternatively, such a bias may lead nurses to misperceive pain cues or to be inattentive to them.

Pain management

The vast majority of patients commended both the general care provided by the hospital and its pain management, and believed that nurses have an accurate perception of how much pain patients are experiencing and care sufficiently about it.

The VAS ratings from patients indicate, however, that less than a third of the patients in any group were pain-free when interviewed, and less than a fifth had been free of pain throughout the preceding 24 hr. About half of the adult patients had moderate to severe pain when interviewed (VAS levels 4-lo), and 15% of medical patients and 8% of surgical patients reported being in extreme pain (VAS level 9 or 10). A third of the adult patients interviewed, and a quarter of the children, reported having experienced extreme pain (VAS level 9 or 10) during the preceding 24 hr; 8% of medical patients and 5% of surgical patients rated their least pain in this range. Over 40% of the children chose one of the three most miserable faces to describe how they had felt when pain was at its worst during the previous day. Over 10% of the adult patients and a third of the pediatric patients stated that they would have liked additional analgesics. When asked specifically whether pain had caused them to be miserable, anxious, angry or frightened, about half of the patients interviewed endorsed each one of these. More than a third of the patients stated that pain had interfered with their enjoying visitors, over 40% stated it had disrupted their concentration or eating, and almost two-thirds stated that pain had interfered with sleep.

There was a marked contrast, therefore, between patients’ general comments about pain care and their answers to specific probes. This is not unique to patients in Kuwait, but has been found when interviewing patients in western developed

690 ANN HARRISON

societies [ 11,301. Part of the explanation may be that patients do not expect to be kept pain-free, and so uncontrolled pain is not considered grounds for evaluating the pain management as inadequate. In- deed, this may be an expectation which patients acquire from the medical staff taking care of them [3l]. In the current study, when patients were asked why they had not requested more analgesics when they thought they needed them, almost a third stated that they thought that they had to put up with the pain, and almost half explained that they did not think patients should ask. Again, similar responses have been reported from hospitalized patients in developed countries [7]. Such data clearly point to a need for sophistication and caution when interpreting patients’ general appraisals of pain management.

Various concerns have been raised by pain experts about the prevalence of uncontrolled pain among hospitalized patients in western developed countries [8, 11, 30, 3 I], and about the knowledge, attitudes and skills which medical care staff bring to pain management. Pain not only causes distress to patients, it can also actively compromise recovery: and so encouraging stoicism in patients can be counterproductive [2,29, 311. Yet the express aim of most medical staff is nor to keep patients free from pain postoperatively [3l]. Medical staff have been shown not to be sufficiently cognizant of modern pain management options or knowledgeable about pharmokinetics; and an irrational fear of using opiates persists which results in unnecessary suffering for patients whose pain can only be controlled effectively with their use [32]. Specific suggestions for amending and updating the education of medical care staff working in clinical settings, such as burn units, where high levels of uncontrolled pain have been recorded, have been put forward because of the severity of the problem ex- posed [2]. In the current study, the fact that the majority of patients did not feel free to request additional medication, and that nurses generally failed to explain why those requests made could not be met, are causes for concern.

Various authors have pointed out the need to educate patients about what pain is expected and reasonable, and to encourage them to provide staff with accurate feedback about the pain management approach employed and whether it needs revising [2, 241. The ramifications of undercomplaining by patients and underestimation by nurses are numerous and interrelated. Both will lead to unnecessary, uncontrolled pain among patients. Undercomplaining will prevent nurses from developing an accurate appreciation of how much pain patients experience during routine medical procedures, and of what pain prevention and pain management needs exist [2]. This will limit how usefully nurses can prepare future patients, and may well result in unnecessary fear and pain for them [23]. Many of the patients interviewed listed routine procedures, such as injections and blood sampling, as being very painful. Yet there are

established behavioral and drug management techniques for making such procedures tolerable [23]. While the levels of uncontrolled pain reported are no worse than those found in hospitals in developed countries [e.g. 8, 11, 30. 311, this is certainly no reason for complacency. Pain management in western countries has been shown to be generally suboptimal, and ways of lessening uncontrolled pain are actively being sought. The negative impacts that uncontrolled pain can have on patients’ recovery and well-being, and on their future reactions to medical treatment, are beginning to be recognized, as too are the difficulties that patients’ excessive pain and suffering create for nurses [33]. This is not to suggest that all procedures should be pain-free; children and adults do benefit from learning that they can cope with reasonable stress and pain. Whenever possible, however, medical procedures should be tolerable to the patient. The challenges for the future are to develop effective tools for predicting how a given patient will respond to a given procedure, and a rich set of effective options for alleviating excessive stress, fear and pain.

VAS and mood ratings

An encouraging feature of the present study was the facility with which both adults and children provided VAS pain ratings. Patients had no difficulty interpreting the scales or making a selection, and they did not change their minds. There are encouraging indications that the data provided by patients were valid: the rank orderings of least and worst pain VAS evaluations were appropriate, patients’ and nurses’ VAS ratings generally correlated significantly, and the present findings accord with ones reported previously. The lack of difficulty shown by the youngest children [cf. 341 is consistent with previous studies in Kuwait [21,23]. This may reflect the fact that formal education in Kuwait begins earlier than in Europe and North America, most children are in school by 4 years of age. Also, the graphical VAS representation used is less abstract than the standard horizon- tal line, and so probably can be coped with adequately by younger children. Children between 3 and 7 years of age have been shown to provide valid pain reports using pictorial pain scales [6]. Children evidenced no difficulty in using the faces scale, and showed an appropriate change in their selection of faces as pain increased. Children seemed reticent, however, to select the extremely unhappy faces. This may reflect an unwillingness to select a crying face unless they had actually cried. Alternatively, the relatively low levels of pain reported by the children in the current study may have rendered such mood descriptions inappropriate.

Routine documentation of pain. Visual analog scaling has proved a robust pain assessment technique when used with patients varying in age, educational attainment and cultural background [21-23,34-361. VAS assessment is quick and simple, and the present


study confirms that both pediatric patients and adult patients with limited education had no difficulty in providing useful pain ratings with it. Regular VAS evaluations would provide nurses and physicians with information about the pain progress of patients, and with feedback concerning the effectiveness of the pain management methods employed. As other authors have pointed out, there is a great need for improving the documentation of pain in hospitals [2,9, lo]. VAS ratings could be charted in much the way that temperature is logged. This would not represent a total solution. The VAS measures primarily the sensory component of the pain experience, while pain behaviors provide an index of the affective aspects of the pain experience [22,37]. Supplementary pain assessment tools will, therefore, be needed to do justice to non-sensory components; and some clinical conditions may require dedicated assessment schedules to do justice to their idiosyncratic characteristics. VAS assessment, nonetheless, would provide a general screening tool, alerting care staff to pain complaints which require further exploration. Routine pain assessment would also provide patients with an oppor- tunity to mention their needs and concerns regarding pain management, and so could have the added benefits of encouraging more input from patients and demonstrating the concern of care staff [24].

Local conditions

The data were collected during the Islamic month of Ramadan, which is when Muslims fast throughout the hours of daylight. During this period, people generally prefer not to be hospitalized if they have the option, and occupancy rates are low. There are, therefore, sound grounds for assuming that the current sample is not representative of patients hospitalized on these wards at other times of the year. There are no grounds for believing, however, that the consistent underestimation of pain by nurses observed in the current study would not be replicated with a more typical patient group. This view is supported by the findings that different patient groups showed the same phenomenon, that patients who had been hospitalized previously did not differ consistently in their pain reports from ones hospitalized for the first time, and that underestimation has been reported in a variety of clinical contexts in other countries. However, the present study was not in- tended, and cannot be considered, to provide a representative profile of uncontrolled pain amongst general medical, surgical and pediatric patients in Kuwait [cf. 81. Further research is needed to characterize the pain experiences of such patients, and to explore the impact of different medical conditions, clinical settings, and nurse and patient characteristics on the agreement shown by nurses and patients.

There are some features of medical care in Kuwait which may hinder pain evaluation by nurses. Kuwait is a multiethnic community, and patients and nurses often do not share a first language. Nurses and

patients sometimes come from cultural backgrounds which differ in how pain is expressed verbally and behaviorally, and what amount and type of pain complaining is considered acceptable [22,37,38]. There are active attempts being made in Kuwait to improve the training and status of nurses; but currently nurses in Kuwait exercise less initiative in patient care, and have less independent control, than nurses in North America or Europe. The lack of pro re nata prescribing is probably a reflection of this. The result is that nurses in Kuwait have fewer resources than nurses in western countries for effectively relieving their patients’ pain, and so it is probable that they will be more prone to cognitive dissonance and to the underestimations which stem from it. Therefore, while the nurses studied in Kuwait behaved like nurses studied elsewhere in consistently underestimating their patients’ pain [24], some of the reasons for this may be peculiar to the social and clinical setting in which they are operating, and some factors may play a bigger role than is typical elsewhere.

Conclusions

Previous studies had demonstrated that nurses are often inaccurate, and sometimes biased, in their evaluations of patients’ pain [24]. The current study showed that nurses working on general medical, surgical and pediatric wards in Kuwait consistently underestimated their patients’ pain, even though the clinical settings were less painful than most investi- gated previously [2], and even though most patients were looked upon by their nurses as providing an accurate picture of their pain.

VAS rating is considered to provide an index primarily of the sensory component of the pain experience (nociception) [22,37], rather than of the emotional suffering triggered by being in pain, being ill and the diverse negative consequences these can have. The discussion, therefore, has focussed on trying to explain the observed disparity between nurses’ and patients’ pain evaluations in terms of the different frames of reference which they bring to judging nociception and on factors which may encourage underestimation by nurses, rather than on explanations dealing with the wider pain experience and how different this is for the sufferer than the observer. Further work is needed to explore which, if any, of the explanations offered played a role, and to establish whether pain management is compromised by nurses’ underestimations.

Regular VAS assessment appears to be a viable technique to use with both adult and pediatric patients to probe certain aspects of their pain experience, and could lead to a much-needed improvement in hospital pain documentation. VAS ratings should enable staff to gauge more accurately the effectiveness of pain management approaches adopted, and when changes are indicated. Regular pain evaluation could also have the benefits of encouraging a more fruitful

692 ANN HARRISON

dialogue between patients and staff about pain and 16.

pain management, enabling members of the medical team to acquire a better picture of the pain associated with different clinical conditions and medical pro- 17. cedures so that they are in a position to help future patients more effectively.

Acknowledgements-The permission of Raven Press and the 18.

authors to use Fig. 2 is gratefully acknowledged. The data were collected by nursing, physical therapy and medical information students from the Faculty of Allied Health 19.

Sciences and Nursing, Kuwait University, as part of the research component of course 211 FM ‘The Psychology of Medical Care’. Without the assistance of the directors, 20.

quality assurance section and nursing staff of the Mubarak Al-Kabir hospital the project would not have been possible. Mrs Fawzia Yousif and Mrs Hiclea Lima supervised the data collection and assisted with data preparation. The 21.

work was supported by Kuwait University Project MU073. 71

Lenburg C. B., Glass H. P. and Davitz L. J. Inferences of physical pain and psychological distress. II. In relation to the-stage of the patient’s illness and occupation of the perceiver. Nursing Res. 19. 392-398. 1970. von Baeyer C. L., Johnson M. E.. and McMillan M. J. Consequences of nonverbal expression of pain: patient distress and observer concern. Sot. Sci. Med. 19, 1319-1324, 1984. Heft M. W., Gracely R. H., Dubner R. and McGrath P. A. A validation model for verbal descriptor scaling of human clinical pain. Pain 9, 363-373, 1980. Teske K., Daut R. L. and Cleeland C. S. Relationships between nurses’ observations and patients’ self-reports of pain. Pain 16, 289-296, 1983. Bondestam E., Hovgren K., Gaston-Johansson F., Jern S., Herlitz J. and Holmbera S. Pain assessment by

Li.

REFERENCES 23.

I.

2.

3.

4.

5.

6.

7.

8

McGrath P. A., DeVeber L. L. and Hearn M. T. Multidimensional pain assessment in children. In 74 Advances in Pain kesearch and Therap), (Edited by - ” Fields H. L.. Dubner R. and Cervero F.). pp. 387-393. 25. Raven Press, New York, 1985. Choiniere M., Melzack R., Girard N., Rondeau J. and 26. Paquin M.-J. Comparisons between patients’ and nurses’ assessment of pain and medication efficacy in severe burn injuries. Pain 40, 143-152. 1990. Heidrich G. and Perry S. Helping the patient in pain. 27. Am. J. Nursing 82, 1828-1833, 1982. Watt-Watson J. H. Nurses’ knowledge of oain issues: a survey. J. Pain Symplom Managemen; 2,2b211, 1987. McGrath P. A. Pain in Children. Guilford, New York, 1990.

patients in the early phase of acute myocardial infard- tion. .I. adv. Nursing 12, 671- 682. 1987. Harrison A. Arabic children’s pain descriptions. Pediatric Emergency Care 7, 199-203, 1991. _ Harrison A. Childbirth in Kuwait: the exoeriences of three groups of Arab women. J. Pain. Symptom Management 6, I-10, 1991. Harrison A. Preparing children for venous blood sampling. fain 45, 299.-306, 1991. Lander J. Clinical judgments in pain management. Pain 42, 15m~22, 1990. Harrison A. Assessing patients’ pain: identifying reasons for error. J. adll. Nursing. 16, 1018.-1025, 1991. Lavigne J. V., Schulein M. J. and Hahn Y. S. Psycho- logical aspects of painful medical conditions in children. 1. -Developmental aspects and assessment. Pain 27, 133-146, 1986. Thompson K. L. and Varni J. W. A developmental cognitive-behavioral approach to pediatric pain assessment. Pain 25. 283 296. 1986.

9

10.

11.

12.

13.

14.

15.

28.

McGrath P. J. and Unruh A. Pain in Children and Adolescents, Vol. I. Elsevier, Amsterdam, 1987. Mather L. and Mackie J. The incidence of postoperative pain in children. Pain 15. 271-282, 1983. Donovan M., Dillon P. and McGuire L. Incidence and characteristics of pain in a sample of medical-surgical inpatients. Pain 30, 69-78, 1987. Camp L. D. and O’Sullivan P. S. Comparison of medical, surgical and oncology patients’ descriptions of pain and nurses’ documentation of pain assessments. J. adv. Nursing 12, 593-598, 1987. Camp L. D. A comparison of nurses’ recorded assessments of pain with perceptions of pain as described bv cancer patients. Cancer Nursing Ii, 237-243. 1988. . Cohen F. L. Postsurgical pain relief: patients’ status and nurses’ medication choices. Pain 9, 265-274, 1980. Davitz L. J. and Pendleton S. H. Nurses’ inferences of suffering. Nursing Res. 18, 100-107, 1969. Dudley S. R. and Helm K. Assessment of the pain experience in relation to selected nurse characteristics. Pain 18, 179-186. 1984. Baer E., Davitz L. J. and Lieb R. Inferences of physical pain and psychological distress. I. In relation to verbal and nonverbal pat&t communication. Nursing Res. 19, 388-392, 1970. Lenburg C. B., Burnside H. and Davitz L. J. Inferences of physical pain and psychological distress. III. In relation to the length of time in the nursing education program. Nursing Res. 19, 399-401, 1970.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Niven C. and Gijsbers K. The study of labour pain using the McGill pain questionnaire. Sot. Sci. Med. 19, 1347.-1351, 1984. McCaffery M. Understanding your patient’s pain. Nursing 10, 26-31, 1980. Marks R. M and Sacher E. J. Undertreatment of medical inpatients with narcotic analgesics. Ann. intern. Med. 78, 17lLl81, 1973. Weis 0. F., Sriwatanakul K., Alloza J. J., Weintraub M. and Lasagna L. Attitudes of patients, housestaff, and nurses towards postoperative analgesic care. Aneslh. Analg. 62, 70-74. 1983. Melzack R. The tragedy of needless pain. Sci. Am. 262, 27-33. 1990. Atchinson N., Guercio P. and Monaco C. Pain in the pediatric burn patient: nursing assessment and perception. Issues in Comprehensive Pediatric Nursing 9, 399-409, 1986. Abu-Saad H. The assessment of pain in children. Issues in Comprehensive Pediatric Nursing 5, 327-335, 1981. Jerrett M. D. Children and their pain experience. Children’s Health Care 14, 83-89, 1985. McGrath P. A. An assessment of children’s pain. A review of behavioral, physiological and direct scaling techniques. Pain 31, 147-176, 1987. Weisenberg M. and Caspi Z. Cultural and educational influences on pain of childbirth. J. Pain Symptom Management 4, 13-19, 1989. Streltzer J. and Wade T. C. The influence of cultural group on the undertreatment of postoperative pain. Psychosomat. Med. 43, 397-403, 1981.

comparing nurses' and patients' pain evaluations: a study of hospitalized patients in...

Documents