Contextual analysis of determinants of late diagnosis of hepatitis C virus infection in medicare patients.

Abstract

Patient- and county-level characteristics associated with advanced liver disease (ALD) at hepatitis C virus (HCV) diagnosis were examined in three Medicare cohorts: (1) elderly born before 1945; (2) disabled born 1945-1965; and (3) disabled born after 1965. We used Medicare claims (2006-2009) linked to the Area Health Resource Files. ALD was measured over the period of 6 months before to 3 months after diagnosis. Using weighted multivariate modified Poisson regression to address generalizability of findings to all Medicare patients, we modeled the association between contextual characteristics and presence of ALD at HCV diagnosis. We identified 1,746, 3,351, and 592 patients with ALD prevalence of 28.0%, 23.0%, and 15.0% for birth cohorts 1, 2, and 3. Prevalence of drug abuse increased among younger birth cohorts (4.2%, 22.6%, and 35.6%, respectively). Human immunodeficiency virus coinfection (prevalence ratio [PR] = 0.63; 95% confidence interval [CI]: 0.50-0.80; P = 0.001), dual Medicare/Medicaid eligibility (PR = 0.89; 95% CI: 0.80-0.98; P = 0.017), residence in counties with higher median household income (PR = 0.82; 95% CI: 0.71-0.95; P = 0.008), higher density of primary care providers (PR = 0.84; 95% CI: 0.73-0.98; P = 0.022), and more rural health clinics (PR = 0.90; 0.81-1.01; P = 0.081) were associated with lower ALD risk. End-stage renal disease (PR = 1.41; 95% CI: 1.21-1.63; P = 0.001), alcohol abuse (PR = 2.57; 95% CI: 2.33-2.84; P = 0.001), hepatitis B virus (PR = 1.32; 95% CI: 1.09-1.59; P = 0.004), and Midwest residence (PR = 1.22; 95% CI: 1.05-1.41; P = 0.010) were associated with higher ALD risk. Living in rural counties with high screening capacity was protective in the elderly, but associated with higher ALD risk among the disabled born 1945-1965.