'More shades of grey in my answers': an interview study revisiting attitude erosion during clerkships

Small declines in patient-centred attitudes during medical education have caused great concern. Although some of the self-report scales applied have solid psychometric foundations, validity evidence for the interpretation of attitude erosion during clerkships remains weak.

We sought to address this gap in a qualitative study of the relationships between scores on four commonly used attitude scales and participants' experiences and reflections. Our aim was to gain a better understanding of the score changes from the participants' perspectives.

We conducted semi-structured interviews with 15 junior doctors from a cohort (n = 37) that had previously shown a small decline in patient-centred attitudes during clerkships, measured on four self-report scales. In the interviews, we explored interviewees' experiences of their development of patient-centredness and subsequently discussed their scale scores, particularly for those items that contributed to a rise or decline in scores. We analysed the data using a process of constant comparison among personal experiences, scale scores and participants' explanations of score changes, applying the coding techniques of grounded theory.

The analysis revealed important response distortions that might be responsible for small declines in scores during clerkships separately from changes in attitudes. The drastic alterations to the participants' frame of reference, attributable to the transition to clinical practice, represented the most prominent cause of distortion. More nuanced, context-specific, patient-centred reasoning resulted in more neutral responses after clerkships, paradoxically causing a decline in scores. In addition to response distortions, the interviews revealed shortcomings in content validity such as an 'extreme' construct of patient-centredness.

This study calls into question the validity of the interpretation of attitude erosion during clerkships. The findings suggest that small declines in scores on self-report attitude scales are related to a recalibration of trainees' understandings of patient-centredness as they grow more clinically experienced. The evolved construct of patient-centredness and the way attitudes are measured require special attention in the development of future instruments.