Background: The Canadian Occupational Performance Measure (COPM) is a commonly used outcome measure in rehabilitation. In this study it was adapted for very young children by deleting paid/unpaid work and household management categories and having parents act as proxies to rate child performance and their own satisfaction.
Purpose: To assess the internal consistency reliability, content and construct validity, responsiveness, and impact of half scores (20 not 10-point scale) of the adapted COPM.
Method: Parent proxies of subjects aged 2 - 8 (mean 3.9) years with spastic hemiplegic cerebral palsy (n = 41) participating in a clinical trial. There was a total of 214 occupational performance problems for analysis and an additional 56 which had used half score ratings. Internal consistency reliability and construct validity were evaluated using Cronbach alpha statistic. Proxy views explored content validity. Responsiveness was evaluated using pre-post intervention scores and a comparison with Goal Attainment Scaling scores which were assumed to be a suitable benchmark measure. The effect of half scores was assessed by two-sample t-tests.
Results: The COPM adaptations did not have a negative impact on internal consistency reliability as this was acceptable for performance (0.73) and satisfaction (0.83). The high Cronbach alpha scores indicated good construct validity. Content of occupations and rating approach was considered valid by proxies. Use of half scores did not result in significantly different performance ratings, but mean satisfaction ratings were significantly higher when half scores were used (p = 0.0001). This suggests that half scores may provide more precise proxy satisfaction ratings, but at the cost of rigour as internal consistency with satisfaction half scores was lower (0.63 vs. 0.82). Responsiveness to change in clinical status was demonstrated by significant pre-post scores and moderate correlations with goal attainment scores.
Conclusion: The adapted COPM is a psychometrically robust tool and the use of half scores is not recommended.