Supporting Data for Modeling of randomized hepatitis C vaccine trials: bridging the gap between controlled human infection models and real-word testing
Description
Supporting Data for "Modeling of randomized hepatitis C vaccine trials: bridging the gap between controlled human infection models and real-word testing" to appear in PNAS Nexus. Abstract: Global elimination of chronic hepatitis C (CHC) remains difficult without an effective vaccine. Since injection drug use is the leading cause of hepatitis C virus (HCV) transmission in Western Europe and North America, people who inject drugs (PWID) are an important population for testing HCV vaccine effectiveness in randomized-clinical trials (RCTs). However, RCTs in PWID are inherently challenging. To accelerate vaccine development, controlled human infection (CHI) models have been suggested as a means to identify effective vaccines. To bridge the gap between CHI models and real-world testing, we developed an agent-based model simulating a two-dose vaccine to prevent CHC in PWID, representing 32,000 PWID in metropolitan Chicago and accounting for networks and HCV infections. We ran 500 trial simulations under 50% and 75% assumed-vaccine efficacy (aVE) and sampled HCV infection status of recruited in silico PWID. The mean estimated vaccine efficacy (eVE) for 50% and 75% aVE was 48% (standard deviation (SD ±12) and 72% (SD±11), respectively. For both conditions, the majority of trials (~71%) resulted in eVEs within 1SD of the mean, demonstrating a robust trial design. Trials that resulted in eVEs >1SD from the mean (lowest eVEs of 3% and 35% for 50% and 75% aVE, respectively), were associated with imbalances in acute infection rates across trial arms. Modeling indicates robust trail design and high success rates of finding vaccines to be effective in real-life trials in PWID. However, with less effective vaccines (aVEs~50%) there remains a higher risk of concluding poor vaccine efficacy due to post-randomization imbalances. The three folders contain data and Stata code for the 3 simulation conditions: 1) “monthly 50” : monthly testing, 50% assumed vaccine efficacy 2) “monthly 75” : monthly testing, 75% assumed vaccine efficacy 3) “biweekly 50” : biweekly testing, 50% assumed vaccine efficacy Each folder contains: - A subfolder holding data from 500 simulations in the form of 500 .csv files (events_history_1.csv, etc.), 500 files containing summary data computed in-simulation, (summarytable_1.csv, etc.), and an .xml file with the batch parameters - Three Stata command files: 1. A program that imports the raw data, computes outcomes, and posts the results 2. A program that imports and consolidates the summary data 3. A program to analyze the outcome data created by #1 - Two Stata data files 1. The outcome data created by program #1 2. The consolidated summary data - A Stata log file showing the results of the meta-analysis The “monthly 50” folder contains a fourth Stata program to compare the outcomes from monthly and biweekly testing simulations.