A Data-Driven Approach to Risk-Based Source Data Verification.

Abstract

Source data verification (SDV) is the process of confirming that reliable, accurate information collected from participants during a clinical trial has been reported successfully to the trial's sponsor by investigators conducting the study. Over the past 15 years or so, there has been considerable discussion in the literature of alternate (reduced and risk-based) approaches to the traditional 100% SDV approach, but these discussions have been theoretical rather than data driven. This research therefore employed data from studies conducted by the authors' company to answer the following research question: Can historical data and simulation methodology be employed to understand the risks (unidentified problems) and benefits (cost reductions) of specific reduced SDV scenarios? The methodological approach was based upon a 2010 paper published in the Drug Information Journal that proposed 4 hypothetical risk-based monitoring approaches. The paper's authors proposed well-thought-out and defined scenarios that were readily replicated in simulation algorithms. These scenarios therefore facilitated the exploration of whether real data could be used to simulate reduced SDV scenarios. These data came from 30 trials that had utilized electronic data capture and were completed between 2005 and 2010. Findings revealed that real study data can successfully be used to simulate reduced SDV scenarios, bringing a data-driven analytical approach to the determination of efficient and effective approaches to reduced SDV, hence translating our theoretical understanding to data-driven methodology.