Pitfalls in the statistical examination and interpretation of the correspondence between physician and patient satisfaction ratings and their relevance for shared decision making research

Abstract

BACKGROUND: The correspondence of satisfaction ratings between physicians and patients can be assessed on different dimensions. One may examine whether they differ between the two groups or focus on measures of association or agreement. The aim of our study was to evaluate methodological difficulties in calculating the correspondence between patient and physician satisfaction ratings and to show the relevance for shared decision making research. METHODS: We utilised a structured tool for cardiovascular prevention (arriba™) in a pragmatic cluster-randomised controlled trial. Correspondence between patient and physician satisfaction ratings after individual primary care consultations was assessed using the Patient Participation Scale (PPS). We used the Wilcoxon signed-rank test, the marginal homogeneity test, Kendall's tau-b, weighted kappa, percentage of agreement, and the Bland-Altman method to measure differences, associations, and agreement between physicians and patients. RESULTS: Statistical measures signal large differences between patient and physician satisfaction ratings with more favourable ratings provided by patients and a low correspondence regardless of group allocation. Closer examination of the raw data revealed a high ceiling effect of satisfaction ratings and only slight disagreement regarding the distributions of differences between physicians' and patients' ratings. CONCLUSIONS: Traditional statistical measures of association and agreement are not able to capture a clinically relevant appreciation of the physician-patient relationship by both parties in skewed satisfaction ratings. Only the Bland-Altman method for assessing agreement augmented by bar charts of differences was able to indicate this.