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.