August 2, 2011 | Pilot testing DecisionView’s StudyOptimizer provided Merck & Co. with ample evidence that the predictive analytics platform significantly improves the odds of clinical trials getting done on time and within budget. The clinical enrollment optimization and decision support tool has since become the standard for large phase II and III studies across legacy Merck, and will soon be the standard for big-investment studies across legacy Schering-Plough as well, says Christopher Heider, director of information technology at Merck. The two former rivals merged in late 2009.
Merck expects to receive the same return on investment across the entire portfolio of trials that it observed in the pilot studies, says Heider. These include a reduction in overall cycle time variance by approximately two to eight weeks, reduction in the time trial managers spend aggregating study data by roughly 50%, improvement in timelines and accuracy of reporting study data to management by 50%, and reduction in the time trial managers spend identifying recruitment and data cleanup issues by 20%.
The insights and productivity gains Merck has achieved using the enrollment modeling capabilities of StudyOptimizer were recognized in April with a Bio•IT World Best Practices Award in the category of clinical and health-IT research.
StudyOptimizer is a “leading-edge,” cloud-based application for planning, tracking, and optimizing clinical trial enrollment performance, says Heider. Eight of the ten top global pharmaceutical companies are now customers. Traditionally, patient enrollment projections and course correction strategies were based on the experience and intuition of study managers with inconsistent and often costly consequences. Merck previously used a custom solution that looked only at first patient enrolled/last patient enrolled and the number of sites. It had no way to model additional variables such as the impact of additional sites, vendor tactics, and screen failure ratio differences between geographies.
The tool automates the business process of enrollment through a collaborative platform, allowing headquarters and regional trial management teams to work together to create realistic enrollment plans, validate plan assumptions, test multiple scenarios, and approve a baseline against which performance will be monitored, says Linda Drumright, president and CEO of DecisionView. Underperforming sites can be quickly pinpointed and closed, and rescue sites identified to keep studies on track.
Importantly, StudyOptimizer captures current enrollment plans and historical enrollment metrics in a single database, updated from the organization’s clinical trial management system nightly. Study managers no longer need to aggregate data into Excel files to see overall study enrollment progress. As part of the project with DecisionView, Merck centralized data loads from disparate sources—interactive voice recognition, electronic data capture, and central lab systems—enabling creation of this single “source of truth,” Heider says.
The “real value” of StudyOptimizer comes from underlying algorithms that produce usable charts and graphs, allowing the impact of various strategies to be visualized, Heider says. When several targeted countries dropped out of one diabetes study during enrollment with only seven months remaining, for example, Merck’s Global Trial Optimization (GTO) group was able to use StudyOptimizer to develop three recovery strategies using validated assumptions about new and existing countries: the number of additional sites that could be brought on board, site-ready ramp-up time, and fluctuations in screening rates during winter holidays. Seven months later, the diabetes trial finished enrollment within three weeks of projections.
Feedback from Merck’s GTO group, the primary users of StudyOptimizer, has been extremely positive, says Heider. StudyOptimizer gets a daily data feed from Merck, which is loaded into the application to update projections, estimate completion dates, display alerts, recalibrate the forecasting model, and execute many of the tasks once performed manually. Trial managers can thus concentrate their efforts on analyzing trends and spotting potential problems. •