Sourik Beltrán1, Lanair A Lett2, Peter F Cronholm3. 1. Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; Department of Medical Ethics and Health Policy, University of Pennsylvania, Philadelphia, Pennsylvania. Electronic address: sourik.beltran@pennmedicine.upenn.edu. 2. Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania; Department of Biostatistics, Epidemiology, and Informatics, University of Pennsylvania, Philadelphia, Pennsylvania; Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, Pennsylvania. 3. Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, Pennsylvania; Center for Public Health Initiatives, University of Pennsylvania, Philadelphia, Pennsylvania; Department of Family Medicine and Community Health, University of Pennsylvania, Philadelphia, Pennsylvania.
Abstract
INTRODUCTION: Little is known about how provider bias can influence nonadherence labeling. Therefore, a retrospective cohort analysis was conducted to assess the risk of patients with Type 2 diabetes being labeled nonadherent by sociodemographic factors. METHODS: Patients with Type 2 diabetes were identified from 4 primary care sites of the University of Pennsylvania Health System. Demographics, HbA1c, and ICD-10 codes for Type 2 diabetes and nonadherence were extracted from the electronic health record and analyzed in October 2017. Log-binomial regression models were used to estimate patients' risk of nonadherence labeling by race, age, sex, BMI, and insurance payer while controlling for HbA1c as a proxy for medication use. RESULTS: This study included 3,768 adults aged 18-70 years with Type 2 diabetes who received care from 1 of 4 primary care sites at University of Pennsylvania from 2014 to 2017. An increased risk was found for black patients relative to white patients (RR=2.86, 95% CI=1.91, 4.27) and Medicaid (RR=1.8, 95% CI=1.45, 2.22) or Medicare (RR=1.69, 95% CI=1.36, 2.1) relative to private insurance to be labeled as nonadherent while adjusting for HbA1c. Though statistically insignificant, Hispanic patients also showed increased risk of nonadherence labeling. BMI, age, and sex showed no association. CONCLUSIONS: Black race and nonprivate insurance status were shown to be associated with increased risk of nonadherence labeling. The findings may indicate a concerning bias among providers in their perception of patient behavior by race and insurance.
INTRODUCTION: Little is known about how provider bias can influence nonadherence labeling. Therefore, a retrospective cohort analysis was conducted to assess the risk of patients with Type 2 diabetes being labeled nonadherent by sociodemographic factors. METHODS:Patients with Type 2 diabetes were identified from 4 primary care sites of the University of Pennsylvania Health System. Demographics, HbA1c, and ICD-10 codes for Type 2 diabetes and nonadherence were extracted from the electronic health record and analyzed in October 2017. Log-binomial regression models were used to estimate patients' risk of nonadherence labeling by race, age, sex, BMI, and insurance payer while controlling for HbA1c as a proxy for medication use. RESULTS: This study included 3,768 adults aged 18-70 years with Type 2 diabetes who received care from 1 of 4 primary care sites at University of Pennsylvania from 2014 to 2017. An increased risk was found for black patients relative to white patients (RR=2.86, 95% CI=1.91, 4.27) and Medicaid (RR=1.8, 95% CI=1.45, 2.22) or Medicare (RR=1.69, 95% CI=1.36, 2.1) relative to private insurance to be labeled as nonadherent while adjusting for HbA1c. Though statistically insignificant, Hispanic patients also showed increased risk of nonadherence labeling. BMI, age, and sex showed no association. CONCLUSIONS: Black race and nonprivate insurance status were shown to be associated with increased risk of nonadherence labeling. The findings may indicate a concerning bias among providers in their perception of patient behavior by race and insurance.
Authors: Sourik Beltrán; Daniel J Arenas; Itzel J López-Hinojosa; Elizabeth L Tung; Peter F Cronholm Journal: J Am Board Fam Med Date: 2021 Sep-Oct Impact factor: 2.395