Patient characteristics and neighborhood attributes associated with hepatitis C screening and positivity in Philadelphia

Dong Heun Lee; Edgar Y Chou; Kari Moore; Steven Melly; Yuzhe Zhao; Hal Chen; James W Buehler

doi:10.1016/j.pmedr.2022.102011

Patient characteristics and neighborhood attributes associated with hepatitis C screening and positivity in Philadelphia

Prev Med Rep. 2022 Oct 3:30:102011. doi: 10.1016/j.pmedr.2022.102011. eCollection 2022 Dec.

Authors

Dong Heun Lee¹, Edgar Y Chou¹, Kari Moore², Steven Melly², Yuzhe Zhao², Hal Chen², James W Buehler^{2

3}

Affiliations

¹ Drexel University College of Medicine and Drexel University Physicians Practice Plan, Philadelphia, PA, United States.
² Urban Health Collaborative, Dornsife School of Public Health, Drexel University, Philadelphia, PA, United States.
³ Department of Health Management and Policy, Dornsife School of Public Health, Drexel University, Philadelphia, PA, United States.

Abstract

Among patients of an urban primary care network in Philadelphia with a universal hepatitis C virus (HCV) screening policy for patients born during 1945-1965, we examined whether being unscreened and HCV positivity were associated with attributes of the census tracts where patients resided, which we considered as proxies for social health determinants. For patients with at least one clinic visit between 2014 and mid-2017, we linked demographic and HCV screening information from electronic health records with metrics that described the census tracts where patients resided. We used generalized estimating equations to estimate adjusted relative risk ratios (aRRs) for being unscreened and HCV positive. Overall, 28% of 6,906 patients were unscreened. Black race, male gender, and residence in census tracts with relatively high levels of violent crime, low levels of educational attainment and household incomes, and evidence of residential segregation by Hispanic ethnicity were associated with lower aRRs for being unscreened. Among screened patients, 9% were HCV positive. Factors associated with lower risks of being unscreened were, in general, associated with higher HCV positivity. Attributes of census tracts where patients reside are probably less apparent to clinicians than patients' gender or race but might reflect unmeasured patient characteristics that affected screening practices, along with preconceptions regarding the likelihood of HCV infection based on prior screening observations or implicit biases. Approaching complete detection of HCV-infected people would be hastened by focusing on residents of census tracts with attributes associated with higher infection levels or, if known, higher infection levels directly.

Keywords: CDC, Centers for Disease Control and Prevention; CI, confidence interval; Census tracts; EHR, electronic health record; GEE, generalized estimating equation; Geographic analysis; HCV, hepatitis C virus; Hepatitis C screening; Preventive health care services; Primary care; Record linkages; Social determinants of health; aRR, adjusted relative risk.