Online Program

Return to main conference page
Saturday, May 19
Applications
Public Health Applications
Sat, May 19, 1:15 PM - 2:45 PM
Lake Fairfax A
 

Application of Algorithms With Serial Hepatitis C RNA Tests to Predict Treatment and Sustained Virologic Response Among Patients Infected With Chronic Hepatitis C. (304439)

Presentation

Aaron M Harris, Centers for Disease Control and Prevention 
Lom Jennifer, Emory University School of Medicine,Division of General Medicine and Geriatrics 
Lesley Miller, Emory University School of Medicine,Division of General Medicine and Geriatrics 
Alexander J Millman, Centers for Disease Control and Prevention 
*Ademola B Osinubi, Centers for Disease Control and Prevention 
Claudia Vellozzi, Centers for Disease Control and Prevention  

Keywords: Hepatitis C virus, DAA treatment, medical records, HCV RNA tests, Sensitivity ,Specificity.

Background: Data from electronic medical records (EMR) present opportunities to monitor population-level trends in hepatitis C virus (HCV) treatment uptake and cure, but missing data and lack of automated data extraction processes are barriers to using these systems. Methods: Using a registry of chronically-infected HCV patients created from an automated EMR data extraction process, we developed multiple algorithms using serial HCV RNA test results as proxy measures for direct acting antiviral (DAA) treatment initiation and sustained virologic response 12 weeks after treatment completion (SVR). All algorithms started with a detectable HCV RNA test. The algorithm definitions varied based on the number of additional RNA tests (i.e. one to three) or by time requirements between tests (i.e. tests performed > 6 weeks apart). We calculated the sensitivity, specificity, positive predictive and negative predictive values by comparing the algorithms and using DAA treatment initiation and SVR results from the registry as the gold standard. Results: Of 4224 with chronic HCV infection in the registry from 2004–2016, 698 patients initiated non-interferon based DAA treatment after 1 December 2013 based on medical record abstraction; 256 were excluded from the SVR analysis because of lack of both treatment start and end dates. Of the 442 patients with treatment start and/or end dates, 314 had documented SVR. For the five algorithms to identify DAA initiation (n=698), the range for sensitivities was 74–96%, specificities 57–95%, positive predictive values 58–91%, and negative predictive values 85–95% compared to the registry. For the five algorithms to identify SVR (n=314), the range for sensitivities was 79–83%, specificities 64–85%, positive predictive values 85–93%, and negative predictive values 61–62%. Conclusions: Algorithms defined by serial quantitative HCV RNA results can be used as proxy measures for evaluating population-level DAA treatment initiation and SVR outcomes