A Bayesian Meta-Analysis of the Effect of Alcohol Use on HCV-Treatment Outcomes with a Comparison of Resampling Methods to Assess Uncertainty in Parameter Estimates

• Activity Number: 358
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Katherine Cauthen* and Gregory Lambert and Patrick Finley and David Ross and Maggie Chartier and Victoria J. Davey
• Companies: Sandia National Laboratories and Sandia National Laboratories and Sandia National Laboratories and U.S. Department of Veterans Affairs and U.S. Department of Veterans Affairs and U.S. Department of Veterans Affairs
• Keywords: Bayesian meta-analysis ; resampling ; hepatitis C ; alcohol ; uncertainty quantification

Abstract

Heavy alcohol use is associated with lower HCV treatment response rates in interferon-based therapies, and although health care providers recommend reducing or abstaining from alcohol use prior to treatment, many patients are not successful. This meta-analysis systematically summarizes the English-language literature up through June 30, 2014 on the relationship between alcohol use and HCV treatment outcomes, among patients who were not required to abstain from alcohol use in order to receive treatment. Seven articles studying 1,751 HCV-infected patients were identified. Log-ORs of HCV treatment response for heavy versus light alcohol use were calculated. We employed a Bayesian model to accommodate the small sample size. The summary estimate for the log-OR of HCV treatment response was -0.775 (CI -1.397, -0.236). We compare various resampling methods to assess the uncertainty in parameter estimates. Heterogeneity accounted for 60.28% of the variation among log-ORs, but meta-regression was unable to capture its sources. This meta-analysis confirms that heavy alcohol use is associated with decreased HCV treatment response compared to lighter levels of alcohol use.