Using information extraction to improve patient accrual

Abstract

The high complexity of clinical trial protocols impacts their executability and the overall performance of clinical studies. Reduction of unnecessary complexity would result in the timely execution of planned phases and prevent overspendings. Currently, there is no publicly available solution providing actionable insights that would support protocol designers in the definition of less complex protocol documents. Leveraging the value of the historical clinical trial data could support the development of less complex protocols for the upcoming studies and reduce the amount of protocol deviations. Before such a solution could be built, the significant data points need to be identified and adequately pre-processed, so the learning process can be conducted, and the value of the historical data uncovered. Clinical trial protocols often consist of unstructured data points, so they need to be firstly adequately interpreted and structured. Pharmaceutical companies are particularly interested in the evaluation of the information extraction methods in the biomedical domain, as they hold a significant value especially when it comes to informed decision making....