Reliability and validity of the Canadian Occupational Performance Measure in stroke patients

Clinical Rehabilitation 2003; 17: 402 409 Reliability and validity of the Canadian Occupational Performance Measure in stroke patients EHC Cup Department of Occupational Therapy, University Medical Centre Nijmegen, WJM Scholte op Reimer Department of Health Policy and Management, Department of Cardiology, Erasmus Medical Centre Rotterdam, MCE Thijssen and MAH van Kuyk-Minis Department of Occupational Therapy, University Medical Centre Nijmegen, The Netherlands Received 15th December 2001; returned for revisions 12th February 2002; revised manuscript accepted 5th May 2002. Objective: To research test retest reliability and discriminant validity of the Canadian Occupational Performance Measure (COPM), a client-centred outcome measure, in stroke patients. Design: The COPM was administered twice with a mean interval of eight days (SD 2.5, range 5 16). On both occasions the patient identied a maximum of ve problems in daily activities. The problems of both interviews were compared. The problems identied during the rst COPM were rated by the patient on a performance and satisfaction rating scale on both occasions. The individually identied items with use of the client-centred COPM were compared with the xed items of standardized measures (Barthel Index, Frenchay Activities Index, Stroke Adapted Sickness Impact Prole-30, Euroqol 5D and Rankin Scale). Setting: Patients were interviewed at their place of residence. Subjects: Twenty-six stroke patients participated, 11 men and 15 women, aged from 26 to 83 years (mean 68, SD 15). Twenty-four patients were six months, two patients were two months post stroke. Results: Of the 115 problems identied during the rst COPM, 64 (56%) were also identied the second time. Correlation coefcients for the scores were 0.89 (p < 0.001) for performance and 0.88 (p < 0.001) for satisfaction. Of the individual problems identied with the COPM, 25% or less were present in the standardized measures. Correlations between the scores on the COPM and the standardized measures were low and nonsignicant, while all standardized measures correlated signicantly with each other. Conclusions: Test retest reliability of the COPM was moderate for the item pool but was good for the performance and satisfaction scores. Discriminant validity was conrmed. Many patient-unique problems identied with the COPM were not evaluated by standardized measures. Address for correspondence: EHC Cup. Department of Occupational Therapy 304, University Medical Centre Nijmegen, PO Box 9101, 6500 HB Nijmegen, The Netherlands. e-mail: E.Cup@ergo.umcn.nl Arnold 2003 10.1191/0269215503cr635oa

Canadian Occupational Performance Measure in stroke patients 403 Introduction Occupational performance refers to the ability to choose, organize and satisfactorily perform meaningful daily activities that are culturally dened and age appropriate for looking after one s self (self-care), enjoying life (leisure), and contributing to the social and economic fabric of a community (productivity). 1 Occupational performance is unique to the person. In rehabilitation it should therefore be assessed by individualized measures, which are sensitive to varying needs and situations. 2 However, most outcome measures used in rehabilitation are society-perceived and focus on independence in xed activities, not on the needs or problem-experience of individual patients. 3,4 This is also the case for frequently used functional outcome measures in stroke such as the Barthel Index, 5 Frenchay Activities Index, 6 Stroke Adapted Sickness Impact Prole-30 7 and Rankin Scale. 8 In clientcentred rehabilitation, patients and therapists work together to dene problems in occupational performance and to decide on the focus of and need for intervention and the preferred outcomes. 1 Studies have found that motivation, participation and functional recovery are enhanced when patients choice and self-evaluation are incorporated in the assessment and treatment process. 9,10 An example of a client-centred outcome measure is the Canadian Occupational Performance Measure (COPM). The COPM is designed for use by occupational therapists to detect change in patients self-perception of their occupational performance over time. 11 The COPM is based on the occupational therapy guidelines for clientcentred practice and the Canadian Model of Occupational Performance. 1,11,12 The COPM is a generic measure, meaning that it is not diagnosis-specic. With a semi-structured interview the patient is encouraged to identify problems in selfcare, productivity or leisure activities. It concerns those activities the patient wants, needs or is expected to do, but cannot do, or those in which the patient is not satised with current performance. Then the patient rates importance of the problems on a 10-point scale from not important at all (score 1) to extremely important (score 10). Rating of importance helps to prioritize the problems. For the ve most pressing or important problems, the patient is asked to rate current performance of each of these activities on a 10-point scale from not able to do it to able to do it very well. The patient is also asked to rate satisfaction with performance on a 10-point scale from not satised at all to extremely satised. These scores range from 0 to 10, higher scores reect better performance and satisfaction with performance as perceived by the patient. The performance and satisfaction can be reassessed following a period of treatment. The average change scores of performance and satisfaction can be used as outcome measures in rehabilitation. Although the COPM is widely known and often used in rehabilitation research, 13 18 evidence for its reliability and validity is limited. Regarding the reliability, the manual 11 refers to three unpublished studies that showed acceptable test retest reliability scores. In another study with seven adults with traumatic brain injury, Spearman rank order correlations were moderate for both performance and satisfaction. 19 Until now, only the reproducibility of the scores for performance and satisfaction has been researched. No studies have researched whether patients identify the same problems and priorities when interviewed twice. Regarding validity, convergent validity of the COPM has shown to be moderate or low. 20,21 This is not surprising because COPM differs from other health care measures in that the item pool is not xed, but dened by the patient. Evaluation of the discriminant validity, which provides evidence that instruments measure different constructs, 22 would therefore be more appropriate. Responsiveness of the COPM has been researched by Wressle et al., who found that 79 of 108 patients had change scores of two or more on the COPM before and after the intervention. 23 There was no criterion measure used to conrm this improvement. In another study among patients with hand trauma, the change scores of the COPM exceeded change scores in other measures (Action Research Arm test and Sequential Occupational Dexterity Assessment), supporting the responsiveness of the COPM. 24 In the present study test retest reliability of the COPM is researched for

404 EHC Cup et al. the item pool, dened as the type of problems identied by the patient, and the scores for performance and satisfaction with performance. The discriminant validity of the COPM is evaluated in two ways: by comparing the individual problems identi- ed with the COPM with the xed items on four standardized functional measures and by correlating the COPM performance scores with the performance scores on the standardized functional measures. To demonstrate discriminant validity, we expect that the item pool of the COPM differs from the item pool of the standardized functional measures and that the COPM performance scores will be unrelated to scores on the standardized functional measures. Methods Patients were recruited from a multicentre study to evaluate stroke services in the Netherlands. In this study patients were included during their hospital stay and followed through the stroke service. Their health status was measured two and six months following stroke. From February 2000 until June 2000, 26 patients were randomly selected and asked to participate in this additional COPM study. Inclusion criteria were: they had residual impairments and disabilities (Rankin score 2), communication (understanding and producing language) and general condition were suf- cient to participate in two additional interviews (judged by the research occupational therapist) and they gave informed consent for the extra visits by the occupational therapist. Patients were interviewed at their place of residence. The COPM interviews were administered by two trained occupational therapists. The medical ethics committees of the participating approved this study. Instruments In this study the client-centred COPM is compared with measures that were used in the study to evaluate stroke services in the Netherlands. These are all frequently used standardized functional measures in stroke research: The Barthel Index (BI) measures disability in basic self-care activities and mobility. 5 Total scores range from 0 to 20; higher scores denote greater independence. The Frenchay Activities Index (FAI) measures disability in instrumental activities and some aspects of handicap. 6 In this study the total scores (range 0 45) are divided by the number of questions that are applicable (maximum 15 questions). Therefore FAI scores range from 0 to 3, higher scores reect greater independence. The Stroke Adapted Sickness Impact Prole- 30 (SA-SIP30) assesses quality of life after stroke. 7 Scores are weighted percentile scores and range from 0 to 100; higher scores mean lower independence. The Euroqol 5D (EQ-5D) is a nondiseasespecic health-related quality of life instrument. 25 Scores are weighted and range from 1 to 1; higher scores indicate better quality of life. Rankin Scale can be viewed as a global functional health index with a strong accent on physical disability. 8 Scores range from 0 to 5, a higher score reects a more severe handicap. Test retest reliability Each patient was interviewed twice by the same occupational therapist with a mean interval of eight days (SD 2.5, range 5 16). The rst COPM interview consisted of problem identication up to and including scoring performance and satisfaction of the ve most pressing problems. During the second COPM interview, again the patient identied the ve most pressing problems. First, to research the test retest reliability of the item pool, the problems identied during the rst interview were compared with the problems identied the second time. Secondly, to research the test retest reliability of the COPM performance and satisfaction scores, only the problems that were identied during the rst

Canadian Occupational Performance Measure in stroke patients 405 COPM interview were scored again the second time. Discriminant validity Discriminant validity was established by comparing the COPM with ve standardized functional measures: the BI, the FAI, the SA-SIP30, the EQ-5D and the Rankin Scale. First, the item pool was compared. We checked whether the activities that were identied as problems with the COPM were present as specic items within the standardized functional measures. If the activities were not present as specic items, we checked whether the item tted into one of the domains of the scale. For instance cutting one s nails is not a specic item in any of the standardized functional measures, but it ts within the question grooming of the BI and the domain body care and movement of the SA-SIP30. Also many hobbies, like playing volleyball or collecting stamps or attending the bingo, are not specic items in the standardized measures, but they all belong to the FAI question actively pursuing hobby. Secondly correlations between the performance scores on the COPM and the performance scores on each of the standardized functional measures were calculated. Statistics Descriptive statistics, scatterplots and Spearman s rho correlation coefcients were used to show the results of the test retest reliability of the performance scores and satisfaction scores. Spearman s rho correlation coefcients were also calculated for the outcome on the COPM performance scale and standardized functional measures. Data were analysed using the SPSS/PC+ Statistics version 10.0 (SPSS Inc. Illinios, USA). Results The study group consisted of 26 stroke patients, 11 men and 15 women with ages varying from 26 to 83 years (mean 68 years, SD 15). Twenty-four patients were six months and two patients were two months post stroke. Fourteen patients were living at home (54%), nine patients (35%) stayed in a nursing home on a rehabilitation ward and three patients (11%) in a rehabilitation centre. The mean scores were for the BI 14.7 (SD 4.6, range 6 20), for the FAI 0.5 (SD 0.6, range 0 1.9), for the SA-SIP30 20.7 (SD 5.9, range 4.7 30.8), for the EQ-5D 0.34 (SD 0.35, range 0.36 0.85) and for the Rankin 3.1 (SD 0.7, range 2 4). Test retest reliability The initial COPM interview revealed a total of 115 problems. During the second COPM interview 112 problems were identied. Of the 115 problems identied the rst time, 64 problems (56%) were also identied the second time. The mode was that three of ve problems were identied on both occasions. The rst time, the mean score for performance was 3.5 (SD 1.8, range 1.0 7.0) and for satisfaction 3.3 (SD 1.9, range 1.0 7.5). At the second time, the mean score for performance was 3.7 (SD 1.9, range 1.0 6.8) and for satisfaction 3.5 (SD 2.1, range 1.0 7.4). Figures 1 and 2 provide scatter-plots of all individual mean scores on both occasions. The Spearman s rho correlation coefcient for the test retest performance scores was 0.89 (p < 0.001) and for the test retest satisfaction scores 0.88 (p < 0.001). Discriminant validity Of the 115 problems identied with the initial COPM, 45% of the problems belonged to the self-care domain, 17% to the productivity domain and 38% to the leisure domain. Table 1 provides an overview of the number (percentage) of problems identied with the COPM, that were present as a specic item in the standardized functional measures or belonged to a domain of that measure. The number of problems that were actually present as specic items in one of the measures varied from 17% (EQ-5D) to 25% (BI). The additional number of items that were not specically addressed but tted into a domain of a scale varied from 4% (BI) to 77% (EQ-5D). The total number of problems that were addressed either as a specic item or tting within a domain varied from 30% (BI) to 94% (EQ-5D). Correlations of the performance scores of the COPM with the performance scores on the standardized functional measures are presented in Table 2. None of the standardized functional

406 EHC Cup et al. Figure 1 Mean COPM performance scores on two occasions. Figure 2 Mean COPM satisfaction scores on two occasions.

Canadian Occupational Performance Measure in stroke patients 407 Table 1 (n = 26) Presence of problems identied with the client-centred COPM in four standardized functional measures Number (percentage) of problems identied with client-centred COPM in: BI FAI SA-SIP30 EQ-5D Present as item 29 (25%) 26 (23%) 22 (19%) 20 (17%) Fitting within a domain 5 (4%) 39 (34%) 56 (49%) 88 (77%) Total 34 (30%) 65 (57%) 78 (68%) 108 (94%) BI, Barthel Index; FAI, Frenchay Activities Index; SA-SIP30, Stroke Adapted Sickness Impact Prole-30; EQ-5D, Euroqol 5D. Table 2 Spearman s rho correlation coefcients of the COPM performance scores with the scores on ve standardized functional measures (n = 26) COPM BI FAI SA-SIP30 EQ-5D BI 0.225 FAI 0.115 0.794** SA-SIP30 0.102 0.517** 0.426* Euroqol 0.143 0.701** 0.648** 0.483* Rankin 0.209 0.806** 0.797** 0.468* 0.675** *p 0.05; **p 0.01. measures signicantly correlated with the COPM, but they all signicantly correlated with each other. Discussion Frequently used standardized functional measures in stroke like the BI, FAI and SA-SIP30 evaluate daily performance from a society perspective, meaning that it is focused on independence and not on the needs and problems of the individual patient. To measure individual prob- Clinical messages The Canadian Occupational Performance Measure (COPM) evaluates patient-unique problems that are not evaluated by frequently used standardized functional measures. Although the stability of the COPM item pool is moderate, the test retest reliability of the COPM scores is good. lems in daily activities perceived by patients, the COPM has been developed. The COPM is very helpful in identifying goals and in making a treatment plan together with the patient even without the rating part. The big advantage of the allocation of scores is that the effect of the treatment becomes measurable. This idea of measuring goals is not new. In 1990 Ottenbacher and Cusick described the Goal Attainment Scaling (GAS), a method which has been used to evaluate programmes in psychotherapy and mental health since 1974. 26 The GAS provides a framework for the development of goals that are measurable, attainable, desired by the patient and the patient system, socially, functionally and contextually relevant. The main difference with the COPM is the scoring procedure. Although both the COPM and the GAS aim at measuring client-centred treatment goals, in the COPM the patient is the rater and with the GAS the therapist allocates the scores. In a true client-centred approach, the patient should be the rater. In case patients have cognitive problems as a result of acquired brain damage such as stroke or brain injuries, they may have difculties judging and rating their own performance. Then a combination of the COPM and the GAS, as done by Trombly et al. 19 seems a

408 EHC Cup et al. good alternative. In their study with adults suffering from traumatic brain injury, the COPM was used to set treatment goals with the patient and the GAS was used to make these goals measurable. 19 The current study showed that the test retest reliability or stability of the item pool (the activities identied by the patient) is subject to change when the COPM is administered twice. In this study generally three of ve problems were the same when patients were interviewed twice. The semi-structured character of the COPM may result in a somewhat different interview on different occasions. On every single day a patient may experience different problems. It is therefore not surprising that the item pool is not completely stable. One can argue whether this is acceptable or not. In clinical practice this will not be a problem. New problems can be added if they become a priority, even when they were not mentioned before. However, this is not easily done when the COPM is used as an outcome measure in clinical trials. Then the item pool needs to be valid and reliable. In this study the interviewer was the same person on both occasions. Because every therapist has his or her own style of interviewing, there may be even more variation in the outcome when different therapists interview the patient. This needs further evaluation. Regarding the validity, it was expected that the problems identied with the COPM would not be covered by the frequently used standardized functional measures. This would demonstrate the additional value of the COPM. The COPM covers a broad spectrum of daily activities in the areas of self-care, productivity and leisure. Patients are allowed to nominate any activity that is of importance to them. The BI only investigates basic personal care activities and the FAI addresses productivity and leisure. Although the SA-SIP30 evaluates problems in all areas, many patient-unique problems are not specically addressed with the SA-SIP30. Examples of problems identied with the COPM, which were not assessed in any of the other measures, include reading the newspaper, going to the hairdressers, using the video-recorder or planning the day. Although many problems t into the questions of the EQ-5D, this measure is so general that it will not be very useful for client-centred goal-setting and individual treatment planning. When comparing the COPM performance scores with frequently used standardized functional measures, there are no signicant correlations between the COPM and any of the standardized health measurement scales, but the standardized scales all signicantly correlate with each other, conrming discriminant validity of the COPM. In conclusion, the COPM is a valuable tool to assist in client-centred rehabilitation. In contrast to standardized functional measures it truly focuses on the problems and needs of the individual patients. For research purposes, the reliability of the COPM item pool is doubtful. However, the test retest reliability of the performance and satisfaction scores is good. Acknowledgements This study was supported by the Occupational Therapy Department of the University Medical Centre Nijmegen and the Department of Health Policy and Management of the Erasmus Medical Centre Rotterdam and ZorgOnderzoek Nederland (ZON). References 1 Townsend E ed. Enabling occupation. Ottawa: CAOT Publications ACE, 1997. 2 Pollock N. Client-centred assessment. Am J Occup Ther 1993; 47: 298 301. 3 Pollock N, Baptiste S, Law M, McColl M, Opzoomer A, Polatajko H. Occupational performance measures: a review based on the guidelines for the client-centred practice of occupational therapy. Can J Occup Ther 1990; 57: 77 81. 4 Cardol M, Brandsma JW, Groot de IJM, Bos van den GAM, Haan de RJ, Jong de BA. Handicap questionnaires: what do they assess? Disabil Rehabil 1999; 21: 97 105. 5 de Haan R, Limburg M, Schuling J, Broeshart J, Jonkers L, van Zuylen P. Klinimetrische evaluatie van de Barthel Index; een meetinstrument voor beperkingen in activiteiten vn het dagelijks leven. Ned Tijdschr Geneesk 1993; 137: 917 21. 6 Schuling J, de Haan R, Limburg M, Groenier KH. The Frenchay activities Index; assessment of functional health status in stroke patients. Stroke 1993; 24: 1173 77. 7 van Straten A, de Haan RJ, Limburg M, Schuling J, Bossuyt PM, van den Bos GAM. A Stroke Adapted

Canadian Occupational Performance Measure in stroke patients 409 30-item version of the Sickness Impact Prole to assess quality of life (SA-SIP30). Stroke 1997; 28: 2155 61. 8 de Haan R, Limburg M, Bossuyt P, van der Meulen J, Aaronson N. The clinical meaning of Rankin handicap grades. Stroke 1995; 26: 2027 30. 9 Dickerson AE. Should Choice Be a Component in Occupational Therapy Assessments? Occup Ther Health Care 1996; 10: 23 32. 10 Gibson W, Schkade JK. Occupational Adaptation Intervention with patients with cerebrovascular accident: a clinical study. Am J Occup Ther 1997; 51: 523 29. 11 Law M, Baptiste S, Carswell A, McColl M, Polatajko H, Pollock N. Canadian Occupational Performance Measure, third edition. 1998, Toronto: CAOT Publications ACE, 1998. 12 CAOT. Occupational therapy guidelines for client-centred practice. Toronto: Canadian Association of Occupational Therapists, 1991. 13 Davis SE, Mulcahey MJ, Smith BT, Betz RR. Outcome of functional electrical stimulation in the rehabilitation of a child with C-5 tetraplegia. J Spinal Cord Med 1999; 22: 107 13. 14 Bailey A, Starr L, Alderson M, Moreland J. A comparative evaluation of a bromyalgia rehabilitation programme. Arthr Care Res 1999; 12: 336 40. 15 Law M, Russell D, Pollock N, Rosenbaum P, Walter S, King G. A comparison of intensive neurodevelopmental therapy plus casting and a regular occupational therapy programme for children with cerebral palsy. Dev Med Child Neurol 1997; 39: 664 70. 16 Bodiam C. The use of the Canadian Occupational Performance Measure for the assessment of outcome on a neurorehabilitation unit. Br J Occup Ther 1999; 62: 123 26. 17 Carpenter L, Baker GA, Tyldesley B. The use of the Canadian Occupational Performance Measure as an outcome of a pain management programme. Can J Occup Ther 2001; 68: 16 22. 18 Gilbertson L, Langhorne P. Home-based occupational therapy: stroke patients satisfaction with occupational performance and service provision. Br J Occup Ther 2000; 63: 464 68. 19 Trombly CA, Radomski MV, Davis ES. Achievement of self-identi ed goals by adults with traumatic brain injury: phase I. Am J Occup Ther 1998; 52: 810 18. 20 Chan CCH, Lee TMC. Validity of the Canadian Occupational Performance Measure. Occup Ther Int 1997; 4: 229 47. 21 McColl MA, Paterson M, Davies D, Doubt L, Law M. Validity and community utility of the Canadian Occupational Performance Measure. Can J Occup Ther 2000; 67: 22 30. 22 Streiner DL, Norman GR. Health measurement scales. A practical guide to their development and use. New York: Oxford Medical Publications, 1998. 23 Wressle E, Samuelsson K, Henriksson C. Responsiveness of the Swedish version of the Canadian Occupational Performance Measure. Scand J Occup Ther 1999; 6: 84 89. 24 Efng T, Dirx M, Sleegers E, van Meeteren N. De responsiviteit van de COPM, ARA en SODA voor veranderingen in handvaardigheid bij plastisch-chirurgische handpatiënten. Ned Tijdschr Ergother 1999; 27: 109 16. 25 Group TE. Euroqol: A new facility for the measurement of of health related quality of life. Health Policy 1990; 19: 199 208. 26 Ottenbacher KJ, Cusick A. Goal attainment scaling as a method of clinical service evaluation. Am J Occup Ther 1990; 44: 519 25.