All Times EDT
Keywords: Clinical Study Report, predictive model, machine learning, advanced analytics
Lack of insight to proactively predict, in a data-driven and objective manner, which Clinical Study Reports (CSR) are at risk of experiencing potential downstream amendment or errata is a problem in clinical development. To solve this problem, we create a predictive model and use its output, probabilities of having a CSR amendment or errata over the course of the study, to serve as a leading indicator. In this poster, we aim to introduce the CSR amendment predictive model and the specific machine learning methodology used. The model uses study characteristics, data attributes, and measures associated with the number of subjects to predict which studies are more likely to have a downstream amendment or errata. By using this predictive model, business not only can understand the related risk factors and minimize these risks but also can better allocate the resource to the studies that have the anticipated CSR amendment or errata.