All Times EDT
Keywords: Data quality, central statistical monitoring, risk-based monitoring, statistics, fraud detection
Central statistical monitoring (CSM) aims at detecting data anomalies in clinical trials. A wide range of statistical tests are applied to detect atypical patterns in data from various sources including EDC, CTMS, ePRO and Lab data. CSM relies on the underlying assumption that all centers are broadly comparable save for the natural variations that can occur across sites (e.g. due to patient ethnicity if centers are located in different regions of the world).
Over the last years, we have developed and enhanced a cloud-based monitoring platform to enable centralized monitoring of clinical trials. Sponsors run a battery of statistical tests on their clinical and operational data and statistical results are generated using our statistical engine, SMART™. This platform has been used to analyze a large number of industry-sponsored trials for various indications and therapeutic areas. Statistical anomalies have been reviewed and documented by various study teams using our signal management system.
In this poster, we perform an analysis of all metadata collected across trials and sponsors. Metadata include data about statistical data inconsistency scores and signals defined to capture data anomalies. Our results show that central statistical monitoring is an effective way to detect anomalies that point towards actionable interventions to improve data quality in clinical trials.