Information bias in the Canadian occupational performance measure: A qualitative study

Abstract

Purpose

The Canadian Occupational Performance Measure (COPM) is designed to measure outcomes from client-centred occupational therapy. We explored clients’ views and experiences of participating in an initial COPM assessment in order to examine the potential information bias which may influence COPM scores.

Material and methods

We used qualitative methods to analyse semi-structured interviews (qualitative thematic analysis) on clients at a typical Japanese rehabilitation hospital, to examine the potential information bias affecting their scores in their initial COPM assessment.

Results

19 of 20 clients (13 men; 7 women, aged 19–84 years) demonstrated potential sources of information bias in their COPM scores. We identified 15 sources of information bias, grouped into three domains: (1) bias during the selection of occupational areas (Misunderstanding client-centeredness, Misunderstanding meaningfulness, Misunderstanding occupation, and Composite occupations), (2) bias during the scoring of performance and satisfaction (Imaginary scores, Confusing scores of performance or satisfaction with importance, Ambiguous scores, Hopeful scores, Target scores, Emotional scores, Considerate scores and Humbleness scores) and (3) bias interfering future scoring (Changes in selected occupation, Forgotten scores, Ceiling effects).

Conclusion

This study identifies potential sources of bias affecting COPM scores, and taking account of this result would facilitate better collaboration with clients through COPM.

Introduction

The Canadian Occupational Performance Measure (COPM) is considered to be a gold standard, client-centred outcome measure in occupational therapy (OT). Canadian occupational performance measure is used in over 40 countries and has been translated into more than 36 languages (Law et al., 2014). It allows clients to score their own performance and satisfaction in individually selected areas of occupational performance on a scale of 1–10. Canadian occupational performance measure has been shown to have good reliability and validity and has been used in many randomized controlled trials as the primary outcome measure (Arman et al., 2019).

However, information about the validity of COPM has been largely based on criterion validity and predictive validity (Ohno et al., 2021). Criterion validity involves comparing COPM scores with scores on other outcome measures that are thought to be related. Predictive validity involves testing hypotheses about how well the COPM can predict scores on other outcome measures over time. To our knowledge, there are no previous studies that have examined how clients make choices about their preferred areas of occupational performance when being interviewed for the COPM, or how they make decisions around the scores they give for performance and satisfaction. Errors in the selection of areas of occupational performance or the scoring of COPM could contribute to bias in the measure.

Bias in outcome measurement can arise from both random and systematic errors (Tripepi et al., 2010). Random errors are caused by chance and can be minimized as a source of bias in clinical studies by increasing sample sizes, thereby decreasing the overall variation in measurements. Systematic errors are a result of selection bias or information bias and cannot be minimized by increasing sample sizes. Selection bias occurs when data collection is influenced by the methods used to recruit participants or by factors affecting their participation (Tripepi et al., 2010). For example, in one hospital in Japan, we found that the COPM was used with 37% of clients, with OTs being statistically more likely to use the measure when clients had a higher cognitive score (Sawada et al., 2020). Information bias occurs when there are systematic distortions in the collection of data (Tripepi et al., 2010). For example, past studies have demonstrated that clients change their preferred area of occupational performance when generating COPM scores 27–34% of the time when being interviewed a second time (Eyssen et al., 2005; Verkerk et al., 2006). Furthermore, a study of the application of COPM in Denmark identified that clients were much more likely to select basic tasks and actions (e.g. self-care tasks) over more complex, meaningful occupational performance issues during the COPM interview (Larsen et al., 2020). The authors of this study suggested that this may be because the OTs conducting these interviews steer clients towards types of occupational performance issues that they viewed as easier for clients to score or because these were problems that well-meaning OTs had a preference to solve. The authors also noted that there was a need for more research to be conducted on this important aspect of the COPM process – on how occupational performance issues are selected during the interview process and the influence of these on the validity of the COPM.

Examining the potential for information bias to influence scores generated by the COPM is crucial for OT. Therefore, the aim of this study was to explore clients’ views and experiences of selecting their preferred areas of occupational performance and scoring their performance and satisfaction in these areas when using the COPM.

Materials and methods

Design and setting

We used a general qualitative descriptive method for this study. This study was approved by the Ethics Committee of the Tokyo University of Technology (No. E18HS-004, 2018) and written informed consent was obtained from all participants in the study. The protocol for this study has been reported according to the Consolidated Criteria for Reporting of Qualitative Research (COREQ). We have provided information about the characteristics of our research team in the supplementary file.

We conducted this study in one subacute rehabilitation hospital in Japan. This hospital promotes an interdisciplinary team approach whereby patients and their families are provided with a comprehensive monthly rehabilitation plan, including information about planned goals, achieved goals, rehabilitative approaches, discharge planning and the social resources necessary for home discharge.

Participant recruitment

Study participants needed to be hospitalized, receiving OT services, score more than 24 on the Mini-Mental State Exam (MMSE) and have participated in a COPM assessment delivered by their primary OT during their initial assessment. Participants were recruited by OTs who were not otherwise involved in their care and were informed about this study in detail before providing signed consent.

COPM assessment

At the hospital where this study was conducted, COPM is routinely administered every month to monitor the progress of rehabilitation. The OTs in this study implemented the COPM in accordance with the COPM manual (1). Occupational performance issues definition, (2). Rating importance, (3). Selecting problems for scoring and (4). Scoring performance and satisfaction) (Law et al., 2014). All OTs receive comprehensive training in use of COPM, following the process outlined in the COPM guideline (Law et al., 2014), and in occupational interview when they are first employed at the hospital. Furthermore, their interview skills are regularly checked even after joining the hospital (Table 1).

Table 1.

Training program about COPM in this hospital.

Training module for new employee	Education time/frequency
Client-centred occupational therapy model (CMOP-E, MOHO)	3 h
Presentation of an experts OTs case study (including COPM)	1 h
Methodology of COPM	1 h
Group working about COPM	1 h
Introduction of interview reasoning sheet (hospital original)	1 h
Sharing client’s context (communication skills)	2 h
Simulation of using COPM to understand the client’s context	10 h
Observation of COPM conducted by experts	5–10 times
Objective Structured Clinical examination of COPM	1 time
Supervised application of COPM with client in clinical practice	5 times
Case presentation and report using COPM	1 time
For staff
Regularly COPM skill check	1 time/year
Special interest Group about COPM (interested)	1 h per month

CMOP-E: Canadian Model of Occupational Performance and Engagement; MOHO: Model of Human Occupation; COPM: Canadian Occupational Performance Measure.

Data collection

One OT researcher (TS, male, PhD), experienced in qualitative research, conducted semi-structured interviews in Japanese with each client within 24 h of completing their first COPM assessment (Figure 1). Although the researcher did not attend the COPM assessment in order to minimize the risk of social desirability bias (Althubaiti, 2016), he read each participant’s COPM report prior to conducting their interview. The interviewer conducted all interviews in a private, comfortable room, and had no previous interactions with the participants other than to view the results of the initial COPM assessment. All participants were assured that their responses would not be disclosed to their treating OTs. During the interview, each participant was asked about how they chose their preferred areas of occupational performance, how they arrived at their performance and satisfaction scores, and what they thought of the COPM process. When a participant provided confusing or contradictory responses (for example, the interview process produced information that appeared to conflict with the COPM assessment scores), the researcher probed for more information to understand the participants’ perspectives and experiences. All interviews were audio-recorded and transcribed verbatim. All data were collected between 31 July 2018, and 22 June 2019.

Figure 1.

Tendency of performance score of Canadian Occupational Performance Measure (COPM).

Data analysis

Researchers read each interview transcript several times for accuracy. NVivo software (QSR International) was used for managing the interview data and data coding. We analysed the interviews following the principles of qualitative thematic analysis (Smith et al., 1995) with a particular focus on identifying potential sources and mechanisms of information bias. Interview data was read several times to enable the researchers to become immersed in the data before any coding was applied. Sections of text were then coded in NVivo for areas of potential information bias. Concepts from this initial coding were then grouped into categories, clustering similar types of bias together. During this analysis, we took the participants opinions and experiences at face value. For instance, we did not discount apparent examples of information bias in instances where we thought the treating OT might have not conducted the COPM correctly. As noted above, the OTs in the study had received full training in the COPM as part of their hospital employment, so bias arising from misapplication of the COPM still amounted to information bias.

To strengthen the credibility and trustworthiness of the analysis, the initial coding was first undertaken independently by three researchers (TS, KT and KO). These three researchers then compared and debated their findings in Japanese, along with personal reflection on the data collection and study process, before translating the findings into English to share and discuss with other members of the research team. Discussion at this point focused on interpreting and explaining the findings, including alternative interpretations of the data (Smith et al., 1995 & Holliday et al., 2009). Regarding coding, each researcher summarized similar codes they had developed to categorize types of information bias, then decided on broader domains of bias through open discussion. Where the opinions of researchers were divided, we consulted with other researchers (KT, YS, WL) to resolve disagreement in the coding.

Results

Participant characteristics

Twenty clients (13 men and 7 women; mean age 75.9 years, standard deviation (SD) 19.2 years; mean MMSE 28.0, SD 2.4) participated in the study (see Appendix). No clients declined to participate in this study. 10 participants were in hospital for orthopaedic problems (e.g. fractures), seven for cerebrovascular problems (e.g. stroke) and three for epilepsy. Each interview lasted from 3 to 39 min.

Overview of findings

19 of the 20 interviews produced examples of potential sources of information bias in the scoring of the COPM. We identified 15 sources of information bias across all interviews, which we grouped into three overall domains: (1) bias during the selection of occupational areas (4 types of bias); (2) bias during the scoring of performance and satisfaction (8 types of bias) and (3) bias interfering the future scoring (3 types of bias) (Table 2). Data saturation was reached with the 19th participant.

Table 2.

Domain and type of information bias for COPM.

Domain	Type
Bias during the selection of occupational areas	Misunderstanding client-centeredness
	Misunderstanding meaningfulness
	Misunderstanding occupation
	Composite occupations
Bias during the scoring of performance and satisfaction	Imagined scores
	Confusing scores of performance or satisfaction with importance
	Ambiguous scores
	Hopeful scores
	Target scores
	Emotional scores
	Considerate scores
	Humbleness scores
Bias interfering the future scoring	Changes in selected occupation
	Forgotten scores
	Ceiling effects

1. Bias during the selection of occupational areas

These biases occurred when participants gave responses that indicated they had selected areas of occupational performance that did not fit with the COPM concept of client-centred occupational therapy. These problems arose from the participants misunderstanding what information the OT was asking them to provide during the COPM assessment – areas of activity that were client-centred and related to a specific occupational activity that was meaningful to them.

i) Misunderstanding client-centredness

Despite the OTs being trained in the correct application of the COPM, some participants had not understood that the selection of occupational performance issues should reflect their opinions and preferences, and instead thought that the OT tended to decide what occupations they should pursue.

“I think the [COPM] questions are phrased in a way that makes it difficult to include personal opinions. Therapists should consider the intentions of the patients a little more.” (ID4)

“On the other hand, the rehabilitation staff here should assign the score from a third-party perspective. This would better ensure uniformity … After all, patients are compared to other patients, are not they? Even if [a client] thinks he or she is no good, they may be better than another person.” (ID4)

ii) Misunderstanding meaningfulness

Some participants had identified an occupational performance issue in the COPM interview that they later said to the field researcher was not actually important to them, and therefore was not something that they actually wanted to address in therapy.

“[Regarding cooking simple meals] I do not have difficulty [with that occupation] so I do not think I need to do rehabilitation.” (ID7)

“[Regarding doing carpentry or gardening on the weekends, which is mentioned in the COPM score] This is something I wanted to bring up at the end so I am bringing it up … I do not want to do things like that and if I had the time I might do those things … But it is not necessary.” (ID16)

iii) Misunderstanding occupation

Some participants had identified a movement (e.g. getting up, standing up, sitting) rather than an occupation.

“I have scored five points for being able to get up, stand up, or sit, because I have difficulty with those tasks.” (ID4)

“[The reason I scored six points for walking on my own without anxiety is that] I can walk without a cane or other aid, but when I walk I become unsteady sometimes. If I fall because I am unsteady on my feet, then that would be a problem. So, the most important thing is that I want to be able to walk with steady steps.” (ID15)

iv) Composite occupations

Some participants identified a cluster of occupations as a single occupation, which made scoring of performance and satisfaction more difficult and less meaningful for them.

“And, I would like the items [i.e. the occupations] to be divided more specifically … I think using the toilet, grooming, and bathing can be categorized separately. It is difficult to assign a score since these tasks are not separated into different items.” (ID4)

“I think going to the toilet, dressing, beauty, and bathing should be categorized as different items. It was difficult to provide a score since they were not separated into different items.” (ID4)

“In terms of moving around, people differ as to the distance they are able to move. I think the items should be more specifically divided up so that they ask whether one is able to move around on one’s own or with the help of some mechanical support. Shopping and moving should be separated, and I think spending time with friends should also be separated.” (ID4)

2. Bias during COPM scoring

These biases occurred when participants gave responses that indicated they had misunderstood the COPM scoring system in some way such that their COPM scores did not reflect the intended measurement construct of occupational performance issues.

i) Imagined scores

Some participants had no current experience of performing the occupation they had identified in the COPM assessment, so provided scores based on their imagined level of performance and satisfaction rather than based on any actual lived experience.

“I will not know until I try. I do not wear Kimono right now. But even if I were, I do not think it would be much different from before [my disability].” (ID1)

“I imagine that if I did it, I would be able to do around this well, so I gave myself an overall score of 8.” (ID13)

ii) Confusing scores of performance or satisfaction with importance

Some participants did not understand what information they needed to provide when asked to score performance or satisfaction, and instead provided a score to indicate how important the area of occupational performance was for them.

“This [performance score of] eight is intended to reflect how important it is to take a bath every day.” (ID20)

“The [performance score of] five does not mean I think I can do that well, it just means that I scored in order of 9, 8, 7, 6, 5. It is more an indication of importance than it is a score.” (ID14)

iii) Ambiguous scores

Some clients provided ambiguous scores without reference to any apparent standard. These clients could not understand the scoring method but felt obliged to provide a number because it had been requested by the OT.

“When asked how well I can do it, I gave myself a score of 4, but this number does not really mean much. I was asked how well I can do it, and it was the first impulse.” (ID10)

“It was sort of a careless way of scoring. There’s really no way to know how to determine a score, so I am sorry but I guess I was a bit careless in my scores.” (ID11)

iv) Hopeful scores

Occasionally, clients used to the COPM scores to indicate what they hoped they would be able to achieve, with higher score indicating a higher level of hope, rather than using the score to reflect their current situation as intended by the measure.

“I gave a satisfaction score of eight because if I keep trying to do it, I suppose I will become able to do it in the future. I guess the hospital [staff] will help me become able to do it here [in the hospital].” (ID13)

“I gave a score of eight because that’s about how satisfied I am right now. I think that if I keep trying to do it, I will become able to do it. I cannot do it right now, but I hope I will be able to do it in the future.” (ID20)

v) Target scores

A few clients used the COPM score to set themselves a target to achieve in the future, rather than to score their performance and satisfaction in the present. These clients intended to use these scores as a target to work towards.

“I scored myself 8 [on the accomplishment level for driving a car]) not because I am only able to drive at a level of eight but because I’d like to be able to drive that well. In other words, that score is my “passing score.” (ID2)

“If we say that before I could do it at level 10 and then use that as the ideal level, then now I guess I am around a 7. That’s my ideal or my hope.” (ID12)

vi) Emotional scores

Clients also used the performance and satisfaction scores to indicate their emotional attachment to a particular outcome. These scores were not based on how well they thought they could perform the occupation, but to indicate their enthusiasm for it.

“This score [of five for the performance of going to a day service] means that since I went in the past, now, if I go again my friends will be there, and I feel nostalgic, and also that if I go, I will make efforts on my own to recover my physical condition. That’s why I scored this item at 5. [The performance score] is less about what I am able to do than it is about the fact that I’d like to keep trying. I have the will to keep trying.” (ID10)

“Yes, I scored eight to take a bath every day. I think I can take a bath by myself, but it shows the degree of my emotion.” (ID20)

vii) Considerate scores

This situation occurred when clients were grateful for the OT they had received, and did not want to be rude by giving a low COPM score. Although the client believed their actual performance or satisfaction was lower than the score they had given, they provided higher scores in consideration of others.

“[Regarding the performance score of five for getting up and standing up] Well, though I can get up on my own, it takes quite a time. So I guess from the perspective of someone who is a 10, I would be a 3 or 4. The fact that I am frustrated with myself is also reflected in my score. So, I guess it is around a 3. I said “watch” and stood up. I was praised that I did well [by the OT], so my score is half [i.e. Five points]. I think it should really have been about three points. Am I allowed to say this?” (ID4)

“[Regarding playing Pokemon GO] How should I put this? When I think about the praise I received from the rehabilitation doctor I suppose the score should be around a 5. I mean, when I consider the fact that I went out and paid attention [to the game]. While I certainly think that Pokemon GO is a great game, it is just like using a smartphone while walking. You look at the screen, look at the scenery, and play inside the [world of the] game. Cars appear and people appear…] But that probably is not so good from the perspective of the rehabilitation doctor, is it? That [level of] attention. (The employee in charge of rehabilitation) hasn’t said it is no good, but still. That’s what he would say, I suppose. So, I considered those things and decided on a score of 5.” (ID17)

viii) Humbleness scores

In contrast, some clients did not want to appear boastful about how well they thought they were doing. These clients avoided expressing their real opinions regarding their COPM scores because they wanted to be modest or humble.

“[Regarding the performance score for driving a car] This score of eight reflects my experience driving throughout my life. I am trying not to be too greedy. Overall, you see? Normally, 10 would mean you can do it just normally, but well, I suppose if I were able to do it at a level of about eight then I could do that without much problem. [Also] I have to think about my age. It is not that I cannot [drive], it is more like I’d like to be able to do it at about that level.” (ID2)

“It is not like this is a reason, but I should not overestimate myself. So, by being modest and not prideful – this really is not a reason, but – I did not [want to] score myself too high.” (ID10)

3.Bias interfering the future scoring

These biases included errors with the COPM process that could potentially result in problems with using future score to measure change in performance or satisfaction over time. Examples of these errors related to problems with changes in the clients’ interpretation of the target occupational performance issue for scoring, clients being unable to remember why they had ever provided a particular score, and problems with ceiling effects in the COPM scores.

i) Changes in selected occupation

These errors occurred when the clients described providing a score for a different area of occupational performance than the one listed as the select of occupational performance issue for the COPM scores. This indicates that either the original occupational performance issue identified in the COPM interview was recorded incorrectly or the clients had changed what occupational performance issue they were thinking of when they provided a rationale for their score later in the day.

“[For the occupation of going to Akabane station by train] You see, I enjoy buying something I like and returning. So, I provided a score of 7 points because I am still unable to make calculations [for my shopping].” (ID3)

“[Regarding going to Disneyland with his grandchild] What I mean to say is that even if I cannot go to Disneyland with my grandchild, I can still play with him, like playing house or something like that. So, it is not like going to Disneyland is the only way to have fun. When I think about “play” in general, then right now I am a 6.” (ID15)

ii) Forgotten scores

Some clients could not remember why they selected the scores that were reported in their notes, or even complained that they could not remember providing such a score.

“[Regarding a recorded performance score of two] I suppose I gave myself a score of two over the course of the conversation. What is “2” and how did I arrive at that score? I cannot really remember why it is that I gave myself a score of 2.” (ID5)

“[Regarding the score on the task of “becoming able to run with your grandchild”] I do not really know why I gave myself that score.” (ID20)

iii) Ceiling effects

A few clients gave themselves the top score of 9 or 10 out of 10 points for their performance and satisfaction right at the initial assessment. This is problematic because it means that repeat scoring would be unable to demonstrate any clinical significant improvement (a change of two points on the COPM scores).

“About taking a bath by myself, well, probably I thought that I could do it now, unlike before. That’s why I scored myself a 10.” (ID8)

“[Regarding the satisfaction score for going shopping] That score of nine is because I do not forget what it was that I wanted to buy and I almost never make mistakes when I go shopping.” (ID9)

Discussion

Although the COPM has established evidence for its criterion validity and predictive validity, this study demonstrates the potential for problems with information bias in the production of COPM scores. To our knowledge, this is the first study that has explored how clients make decisions during a COPM assessment, either in their selection of occupational performance issues or their interpretation of the COPM scoring system.

Arguably, some of these errors could have been noticed and corrected by the treating OTs before COPM scores were finalized, that is, these could be considered errors in the application of COPM by the OTs. Examples of these errors include those related to misunderstanding the concept of occupation or in the selection of composite occupations, which are apparent in the documented area of occupational performance. Other sources of bias however may not be apparent to the treating OT even if the OT correctly follows COPM protocols. For instance, a treating OT may not know if a patient has selected an area of occupational performance that is actually meaningful to them. They may also not be able to tell if a client is providing a score that reflects how important they consider an area of occupational performance rather than their views on their performance or satisfaction in that area, nor may they be able to tell if a client is modifying their scores to be considerate, hopeful or humble. These sources of information bias may be hidden from the treating OT even if they prompt the client to provide more of a rationale for their scores. Sources of information biases like those identified in this study can mean that patients might not actually provide a score for the intended measurement construct but for something else.

Bias during the selection of occupational performance areas

Some sources of information bias occurred during the selection of occupational performance areas. There are a variety of reasons why clients may have problems with the selection of occupational performance areas. One of these reasons may be cognitive impairment (Plant et al., 2009). However, the patients in our study did not have any serious cognitive problems, so this is unlikely to have been a cause of this problem.

Prior studies have identified that OTs, like all health professions, struggle to develop a common understanding of treatment goals with their clients (Holliday et al., 2005; Maitra and Erway, 2006; Saito et al., 2021). For instance, Saito et al. (2021) analysed perspectives on goals from 100 clients and 79 OTs from seven subacute rehabilitation wards in Japan and found that while both the OTs and clients perceived that they engaged in goal setting together, only 21% of reported goals actually matched between pairs of OTs and clients. These difficulties in developing a shared understanding of person-centred goal setting could be similar to the difficulty that clients have in selecting the kinds of areas of occupational performance that treating OTs expect when applying the COPM assessment. To overcome this difficulty, attention needs to be paid to how OT introduces the concept of person-centred goal setting and the purpose of COPM to their clients. Further training tool and communication strategy may need to be developed and tested for effectiveness.

Bias during COPM scoring

Problems with clients providing COPM scores may have related to their misunderstanding of what was being requested of them, social barriers to clients being candid with their treating OT, or simply clients having a different opinion about what information is important to measure. One common type of bias that arose in COPM scores related to clients trying to score an area of occupation that they had no current experience of performing (i.e. imagined scores). Wæhrens et al. (2012) reported that self-reported measures of activity of daily living ability (conducted either on questionnaire or interview) do not correlate well with measure of performance of activities of daily living based on observation, perhaps for a similar reason – self-reported score may reflect imagined performance rather than actual performance. This source of information bias is likely to result in increased measurement error if clients gain more experience of a target area of occupational performance between time points when they provide COPM scores.

Other types of information bias related to the scoring of COPM may suggest a lack of importance attribute to the measurement exercise by clients and therefore a lack of engagement in it by them. An example of this are ambiguous scores, where clients did not appear to have invested sufficient attention to the COPM process to be able to recall the reason for their scores.

Bias interfering the future scoring

In addition to the problems described above, this study also identified sources of information bias that related more to problems with scoring change in occupational performance over time. A ceiling effect in the initial scoring of COPM is a problem we have noted in an earlier study: In a study of 232 COPM assessments in an inpatient rehabilitation hospital in Japan, we found that 5% of all initial COPM assessment involved scores of 9 or 10 (Sawada et al., 2020). Baseline scores of 9 or 10 are problematic because they cannot be used to detect the clinically meaningful difference of >2 points on the COPM scale at reassessment (Law et al., 2014). Therefore, these scores should indicate to the occupational therapist that there is a problem with the initial selection of occupational performance issues.

These types information biases can also arise from response shift errors in COPM scores. Response shift errors occur when a person’s evaluation of a measurement construct (such as performance of a particular occupation) alters as a result of changes in his or her internal standard of measurement, personal values, or understanding of what is being measured (Ring et al., 2005). If a person’s understanding of a named occupation changes over time, perhaps due to a misunderstanding of the purpose of COPM, then changes in their COPM scores for this occupational performance issue may not be related to any actual change in occupational performance.

As highlighted by other authors (Larsen and Carlsson, 2012), the validity and reliability of COPM depends in a large part on the quality of training and ongoing support for its implementation. Although COPM education was conducted at the hospital in this study, an education program which takes into account the bias which revealed in this study may be useful for more accurate use of COPM. We encourage other OT services in other countries to actively evaluate possible instance of information bias in their application of COPM to practice in order to test their own assumptions about the quality of their COPM training and delivery.

Limitations and future research

It is possible that some of the biases identified in this study are sociocultural in origin and particular to Japanese people. For example, the instances of considerate scores and humbleness scores may have been particularly pronounced in this study because of cultural values in Japan, which often emphasize politeness ahead of self-promotion. Japanese culture also makes a distinction between tatemae, behaviour adopted by a person in public for the good of the collective, and honne, a person’s true (internal) opinions, thoughts and desires (Iwama, 2006). The cultural need to express tatemae may impact on clients being able to give candid opinions regarding their COPM scores. It may also have made it difficult for the treating OT to correct a client who provided a response to the COPM interview or a COPM score that indicated misapplication of the measure. Indeed, cultural differences such as these led Iwama (2006) to propose that Western concept of ‘client centred’ occupational therapy theory and practice may not be possible to fully implement in the Japanese social context. Our study ought to be replicated in rehabilitation wards in Western and other Asian countries in order to examine these assumptions. However, even if cultural differences do exist, it is important for Western OTs to be aware of these differences when working with Asian people in their own countries. Furthermore, it is noteworthy that the COPM has been translated into more than 36 languages (Law et al., 2014). Cultural assumptions in the use of the COPM ought to be tested whenever it is used in another language.

As this is a qualitative study, it should not be used to generalize about the frequency of type of biases or errors with COPM scores. Instead, this study highlights the potential for such errors and provides a way to categorize them in future studies. A larger quantitative study would be needed to examine the frequency of these types of problems in general rehabilitation populations. Additionally, although previous studies recommend the training on COPM usage (Colquhoun et al., 2012; Tuntland et al., 2016), COPM does not require certification for its use. Therefore, our results may reflect real world application of COPM in practice rather than its idealized application in more controlled research environments. In this study, all OTs had received internationally recommended COPM training, but it remains unclear what type of training and ongoing support is required to maximize the validity of the measure. Future studies will need to clarify how to minimize the biases we revealed and how to use COPM more effectively.

Conclusions

This study demonstrates that COPM scores can be associated with information bias. This is the first study to our knowledge to highlight the potential for these errors contributing to biases when implementing the COPM. When collecting or interpreting COPM data, OTs need to be aware of the potential for these types of bias. They should also introduce the purpose of the assessment to clients more clearly and question clients about their rationale for providing certain scores. Japanese OTs may also need to develop strategies to address sociocultural barriers to the application of the measure, although further international research is required to test whether these issues are particular to Japan. By taking these points into consideration, OTs should be able to establish a better collaborative relationship with their clients, which should in turn result in better health outcome.

Key findings

• • We identified 15 sources of information bias, grouped into three domains

• • Three domains of information bias indicated ‘bias during the selection of occupational areas’, ‘bias during the scoring of performance and satisfaction’, and ‘bias interfering future scoring’.

What the study has added?

This result contributes that OTs need to be aware of the potential for these types of bias when collecting or interpreting COPM data and cultural assumptions in the use of the COPM ought to be tested whenever it is used in another language.

Appendix A. Participant’s demographic data and COPM results.

ID	Age	Sex	Diagnosis	MMSE	Interview time	Occupational performance areas	Performance score	Satisfaction Score
1	78	Female	Meningioma	29	12:23	① Walking inside and outdoors wearing tabi (Japanese socks)	5	1
						② Performing Bon-odori (a traditional Japanese dance)	5	1
						③ Rowing bicycle (tricycle)	1	1
						④ Watching my grandchild play baseball	1	1
2	75	Male	Traumatic subarachnoid hemorrhage	29	11:09	① Driving a car	8 (10)	8
						② Returning to work	7	9
						③ Walking	8	8
						④ Playing with my grandchild	6	9
3	74	Female	Traumatic subarachnoid hemorrhage	24	6:02	① Shopping	7	5
						② Cleaning the balcony	6	5
						③ Take the train to Akabane station	7	5
4	84	Male	After total hip joint replacement (both side)	30	20:22	① Getting up from the floor	5	6
						② Toileting, changing clothes, dressing, bathing	3	6
						③ Walking inside and outside	3	1
						④ Going out to go shopping	1	1
						⑤ Participating in activities at the community centre	1	1
5	51	Male	Putaminal hemorrhage	30	4:16	① Driving the car	6	2
						② Cooking a meal for my family (4 person)	7	8
						③ Returning to work as a site foreman	3	5
						④ Returning to work providing education for younger employees	5	7
						⑤ Going shopping on foot	7	8
						⑥ Going shopping by car	6	2
6	66	Female	Destructive coxopathy	28	12:50	① Putting on socks	6	5
						② Walking in my home with an aid	7	7
						③ Showering at home	3	6
						④ Getting into and out of my husbands car	8	5
						⑤ Cooking simple meals	10	10
7	87	Female	Proximal femoral fracture	28	5:01	① Removing my trousers	5	1
						② Toileting safely	6	4
						③ Walking with an aid to go shopping in the neighbourhood	1	1
						④ Preparing a meal and clearing up afterwards each morning, noon and evening	1	1
						⑤ Hanging out the washing and folding the laundry	1	1
8	53	Male	Epilepsy	29	14:20	① Taking a bath independently	10	9
						② Catching a train or bus	7	4
						③ Going to a concert	3	1
						④ Going on a trip	3	1
						⑤ Working as a care worker	2	1
9	54	Male	Epilepsy	30	38:28	① Going out by myself using public transportation	5	4
						② Cooking easily smoothly	4	3
						③ Going shopping safely	7	9
						④ Using a personal computer	6	6
						⑤ Returning to work	7	6
10	93	Female	Proximal femoral fracture	23	11:34	① Toileting independently	4	3
						② Changing into pajamas independently	5	5
						③ Going up and down the stairs in my house independently	3	2
						④ Attending a day service (Japanese day care facility)	5	4
						⑤ Cooking with my daughter	4	2
11	76	Male	Both calcaneus fracture Spinal compression fracture	27	9:41	① Taking a bath	8	8
						② Going to the toilet and managing my underwear easily	8	8
						③ Going o shopping by foot	8	8
12	93	Male	Proximal femoral fracture	24	9:20	① Returning to work	7	1
						② Driving a car	8	1
						③ Playing golf	5	1
13	67	Female	Proximal femoral fracture	28	6:48	① Taking a bath	3	8
						② Toileting	7	5
						③ Going shopping	5	2
						④ Cooking	6	2
14	34	Male	Subarachnoid hemorrhage	30	2:58	① Returning to work	5	5
						② Using transportation independently	9	9
						③ Cleaning up rooms	8	8
						④ Washing the laundry	7	7
						⑤ Cooking a meal	6	6
15	79	Female	Proximal femoral fracture	28	9:55	① Walking without anxiety	6	4
						② Going to Disneyland with my grandchild	5	6
						③ Going shopping, going out to a friends house	6	6
						④ Completing housework such as cooking and cleaning	6	6
						⑤ Growing plants in the garden	5	4
16	64	Male	Calcaneus fracture	30	12:49	① Taking a bath independently	4	2
						② Drive a car, go shopping, participating as a volunteer, and attending a reunion	3	1
						③ Meeting with friends using public transportation	1	1
						④ Completing household chores (taking out the trash, cooking a meal, cleaning rooms)	5	5
						⑤ Gardening and DIY as a pastime activity.	2	2
17	47	Male	Cerebral infarction	28	9:58	① Cleaning my ears and cutting my nails	2	2
						② Performing housework (shopping, cooking, laundry, cleaning) when living alone	7	7
						③ Return to work (telephone technical support)	5	5
						④ Travelling far to play Pokemon Go	5	5
						⑤ Using a computer for my job	5	5
18	58	Male	Epilepsy	27	16:07	① Changing my clothes	5	5
						② Toileting independently without aids	3	4
						③ Taking a bath independently	2	3
						④ Eating meal with chopsticks	5	6
						⑤ Walking to the nearby convenience store by stairs	2	2
19	19	Male	Cervical cord injury, brachial plexus palsy	30	2:31	Returning to work as a tobi (Japanese scaffolder)	1	1
20	65	Male	Cerebral infarction	30	10:48	① Running to be able to play with my grandchild	4	1
						② Returning to work	8	8
						③ Driving a car	5	5
						④ Taking a bath (even if the hospital)	8	8

MMSE: Mini-mental state examination.

ORCID iDs

Tatsunori Sawada https://orcid.org/0000-0002-6467-8537

William Levack https://orcid.org/0000-0001-6631-908X