Do differences in patient-provider relationships explain racial differences in side effect burden among women with breast cancer in the pathways study?

Abstract

48 Background: Prior research suggests that black patients with cancer report a greater burden of symptoms compared to white patients. Differences in patient-provider relationships may be an underlying and modifiable factor in these observed disparities. We examined differences in side effect burden by race/ethnicity and tested the hypothesis that patient-provider communication and trust may partly explain differences in symptom burden among women with breast cancer (BC). Methods: We conducted a cross-sectional analysis of data collected from 4,505 women diagnosed with BC from 2005-2013 at Kaiser Permanente Northern California. Women were asked two months post-diagnosis how much they were bothered by treatment side effects (FACT-B). Using modified Poisson regression, we assessed the association of race/ethnicity with side effect bother, adjusting for clinical characteristics, socioeconomic status (SES), and patient-provider communication and patients’ trust in providers in a step-wise fashion. Provider communication and trust were measured using the Interpersonal Processes of Care survey and a one-item assessment of patients’ trust in providers, respectively. Results: Before adjustment, Black, Hispanic, and Asian patients were at least 75% more likely than white patients to report high (vs. low) side effect bother (p < 0.0001). Bivariate associations of communication and trust with side effect bother and race/ethnicity were statistically significant. For example, patients with high (vs. low) scores for shared decision-making were more likely to be white than black (46% vs. 39%) and less likely to experience high vs. low side effect bother (10% vs. 14%, p < 0.0001 for both comparisons). However, after adjusting for clinical and SES characteristics in a multivariable analysis, the addition of patient-provider communication and trust to the model did not substantially alter disparity estimates. Conclusions: In our sample, patient-provider communication did not modify racial disparities in side effect bother. Further research is needed to explore additional modifiable factors underlying potential disparities in side effect management.