Replicate Engineered Virtual Patient Populations as Surrogates for Real Patient-Level Data

ObjectivesTo demonstrate a new method for generating virtual, individual-level data by testing it on a known clinical trial population. DesignVirtualization of aggregate data from a clinical trial. SettingVirtual Participants936,100 virtual patients InterventionsNone Main Outcomes MeasuresOdds ratios for adverse outcomes in virtual patient populations compared to clinical trial participants. MethodsThe replicate engineered virtual patient populations (RE-ViPPs) method, based on aggregate cross-tabulated categorical population data, does not require access to individual-level data. Using sequential regression combined with randomization, it generates virtual individual patients to comprise populations that, on average, closely resemble the real population in question. The method is validated by applying it to aggregated data from the seminal SPRINT trial, which compared intensive versus standard blood pressure treatment goals on major adverse cardiovascular events. ResultsThe method yields virtual populations, each with 9361 patients, faithfully mimicking the real SPRINT participants. Multiple logistic regression on 100 such populations shows that factors with the highest odds ratios for the primary event are, in descending order, past clinical cardiovascular disease, age [≥] 75, chronic kidney disease, high non-HDL, and smoking history. Intensive blood pressure treatment, the trials intervention, had an odds ratio of 0.74 [0.63-0.87]. On all these measures, the 100 RE-ViPPs mirrored the real SPRINT participants, including the intensive therapy result (actual SPRINT odds ratio: 0.74 [0.62-0.88]). ConclusionsClinical data dissemination has limitations. The most coveted data is descriptive at the individual level but comes with significant cost, effort, and time. There is potential for privacy breaches, and the open-data movement has progressed slowly due to data-ownership concerns. RE-ViPPs closely matched the true SPRINT population. Applied to trials, registries, and databases, RE-ViPPs could reduce open-data burdens by encouraging dissemination of aggregate cross-tabulated real data that allow investigators to generate and measure virtual patients.