Medidata Rave Coder
Medidata, a Dassault Systèmes company
Vanessa Platt, Director of Product Marketing
vplatt@medidata.com
The evolving clinical trial landscape demands faster and cheaper clinical trial execution while improving quality. Achieving these seemingly conflicting objectives places tremendous pressure on clinical trial teams. Trial teams must innovate to perform more work with less staff and with increasing pressure to deliver within shorter timelines and with greater accuracy. This pressure is particularly felt by clinical coding teams, which have been subject to team reductions and outsourcing.
Coding is the process of standardizing concomitant medications, adverse events, or medical histories collected during a clinical trial to standardized terminology. However, this process is time-consuming and expensive. For this reason, Medidata, a Dassault Systèmes company, has begun applying machine learning methods to automate coding in clinical trials through Medidata Rave Coder. These coding suggestions are provided by a machine learning algorithm that was trained on over 20 million coding decisions across 2,000 studies.
After going through the algorithm, Rave Coder provides the top three recommended standard terms to the user, which helps to reduce the amount of time spent browsing the dictionary in search of the appropriate standard term. So far, Rave Coder has proven to present the coding decisions with 98% accuracy for adverse events and medical histories and 97.29% accuracy for concomitant medications.
Before now, coding has traditionally been performed manually by trained coders or automatically by using synonym lists. Both provide coding results but with extensive labor and cumbersome workflows. Rave Coder has helped coding teams reduce coding time by 50% on average for MedDRA. The reduction in manual coding time enables coding teams to deliver more accurate clinical coding faster, reducing the pressure to achieve trial milestones and improving team efficiency to meet organizational financial mandates.
