Primary Care Physician Visit Continuity

Abstract

To the Editor:—For more than 40 years there has been interest in answering these questions: Is provider continuity beneficial and how do we measure it? From many correlative studies based on patient report, provider continuity seems beneficial.1 If we seek stronger evidence than correlative studies we have the results of one controlled, double-blind trial.2 When 776 patients were randomly placed into two groups (but each group had the same clinicians) provider continuity resulted in the following: a) higher patient assessments both of continuity and of clinician knowledge, thoroughness, and ability to provide excellent education; and b) fewer emergent admissions and shorter average length of stay. It is noteworthy that continuity per se influenced several patient perceptions of care, (such as clinician knowledge and ability to provide excellent education), even when the clinicians in the continuity and discontinuity groups were the same. Regarding measurement, Rodriguez et al.1 in their correlative study recapitulate the fact that patient-reported measures of provider continuity and “administrative measures” give different results. The one double-blind, controlled trial documented decades ago that “patient responses were not sensitive to low continuity of care indices: when the measured continuity of care index was less than 0.25, 51% of patients claimed continuity as compared with 83% of the patients who claimed continuity when the index was greater than 0.60.”2 From a statistical perspective both the correlative and controlled trial results confirm evidence of bias in patient report about continuity of care. This is old news and consistent with the fact that humans generalize from their good and bad experiences. Yet, based on their confirmation of old news, Rodriguez et al. now assert that “studies and quality initiatives seeking to evaluate visit continuity should rely on administrative measures whenever possible.” I hope that the elimination of patient report is not the end result the authors intend. Imagine a future in which we are presented with only statistically different administrative indices (using Rodriguez et al.’s method) of 0.53 from one setting and 0.48 from another. We might grasp the statistical implications of the observed difference between the settings. But how would we understand let alone “rely on” these particular indices when such indices: a) are subject to considerable bias,3,4 and b) tell us nothing about how the patients interpret the continuity of care they are receiving?