1053. Beyond Racial Concordance: Patients Perceptions of Relational Similarity
Paul Haidet, MD, MPH, Houston VAMC, K O'Malley, Houston VAMC

Objectives: Cross-cultural relationship research has traditionally relied on self-reports of race to determine concordance. This process relegates race alone to be the parameter that determines the similarity of patient-physician dyads, and places decisions regarding similarity in the hands of the researcher, rather than the research subjects.

Methods: We developed a scale that measures relational similarity from the patients’ perspective. The scale has ten items (in 2 dimensions of ‘culture’ and ‘communication’) that ask patients about how similar they think they and their doctor are. Scores on the items are summed into a ‘similarity score’ than can range from 10 (low similarity) to 60 (high similarity). We recruited 114 patients and their physicians to complete a survey that measured demographics and patients’ and physicians’ race using US Census 2000 items. We asked patients to complete our similarities scale along with measures of trust and perceptions of physicians’ participatory decision-making style (PDM). We examined associations between outcomes (trust, PDM) and both racial concordance and perceptions of similarity, respectively.

Results: Patients’ and physicians’ mean ages were 55 (SD 16) and 40 (SD 7), respectively. Race/ethnicity among patients included black (47%), Hispanic (16%), and white (37%); among physicians included Asian (38%), black (26%), and white (36%). Cronbach Alphas were 0.9 and 0.8 for the culture and communication dimensions of the similarity scale. There was a wide range of similarity scores both within racially concordant and discordant patient-physician dyads. In linear regression models, patient perceptions of higher similarity were associated with higher trust (p=0.007) and greater participatory decision-making (p<0.001). Racial concordance was not significantly associated with either trust or PDM.

Conclusions: The similarity scale has good psychometric properties and allows the subjects (rather than the researcher) to determine similarity. The similarity construct is more powerful in predicting outcomes than racial concordance. Future work will explore relationships between similarity and patient-physician communication