Individual and Area-Level Predictors of Direct Acting Antiviral Uptake for Hepatitis C Virus in a New England Medicaid Population

Abstract

Research Objective: Direct acting antivirals (DAAs) offer significant improvement over previous options for hepatitis C virus (HCV) treatment. Although uptake has increased since their introduction in 2013, many with HCV remain untreated. While some individual- and area-level disparities in prescribing have been documented, including age, disease stage, substance use disorder, and rural residence, research focusing in the Medicaid population is sparse. We examined individual- and area-level predictors of DAA use in a multi-state Medicaid population.

Study Design: This retrospective cohort study used enrollment, medical, and pharmacy claims from Medicaid enrollees in three New England states, linked to databases with area-level information about socioeconomic indicators and healthcare providers. Predictors of DAA uptake (1+ pharmacy claims for a DAA) were examined overall, by individual-level characteristics, including demographic (age, gender, homeless status) and clinical (disease stage, HIV, psychiatric diagnoses, substance use disorder, other comorbidity) characteristics, and area-level characteristics (rural-urban designation, education and unemployment levels in zip code, number of gastroenterologists and primary care providers (PCPs)/100,000 population in county). Generalized estimating equations with binomial distribution and log link identified independent predictors of uptake. Adjusted risk ratios (aRR) are presented.

Population Studied: The population included Medicaid members ages 18-64 years with HCV (2+ claims with ICD-9/10 code for HCV diagnosis or 1+ claim for chronic HCV) from 12/2013–6/2018. Uptake was identified through 6/2018.

Principal Findings: Overall, 55,207 members with HCV comprised the sample and 18.2% received a DAA. The strongest individual-level predictors of uptake included older age (aRR = 1.25, 95% CI 1.20-1.31 and 1.25, 95% CI 1.19-1.31 for ages 35-49 and 50-64 years, respectively, compared with ages 18-34 years), HIV diagnosis (aRR = 1.44, 95% CI 1.36-1.54), and advanced liver disease (aRR = 1.42, 95% CI 1.35-1.50). Individuals with alcohol use disorder (aRR = 0.74, 95% CI 0.70-0.78) and opioid use disorder (aRR = 0.78, 95% CI 0.74–0.72) were less likely to receive treatment. Female gender, other substance use disorders, some psychiatric diagnoses, tobacco use, and documented homelessness were also associated with lower uptake. The strongest area-level predictor of uptake was living in a county with few gastroenterologists (aRR = 1.31, 95% CI 1.11-1.53 for 0-1 gastroenterologists /100,000 vs. 6-20/100,000 population). Uptake was also higher among members living in a zip code with the lowest area-level education attainment and lower in counties both with the lowest and highest number of PCPs/1000 population relative to counties with an average density of PCPs.

Conclusions: Among the high percentage of Medicaid members with HCV who remain untreated, women, younger adults, individuals at early disease stage, homeless, and with substance use disorders appear at higher risk of not being treated. Medicaid members living in areas with higher socioeconomic indicators and highest density of specialists and PCPs are also at higher risk.

Implications for Policy or Practice: These findings identify groups in the Medicaid population that may benefit from targeted interventions designed to increase HCV treatment. The unexpected finding of lower uptake among individuals living in geographical areas with higher socioeconomic indicators and high density of healthcare providers merits further exploration into healthcare access of Medicaid members living in these areas.