Novel Algorithm to Accurately Identify Bleeding in Electronic Health Records
Researchers from Bayer AG, Integrated Evidence Generation (IEG), have developed a novel algorithm to identify two widely used clinical trial definitions of bleeding events in electronic health records, allowing for the comparison of clinical trials and observational studies results.
The team, led by IEG Data Scientist, Alexander Hartenstein, selected bleeding definitions by the International Society for Thrombosis and Haemostasis (ISTH) that contain detailed text descriptions and lists of criteria that allow clinical trial physicians to identify and categorise bleeding events they observe. Currently, these ISTH definitions of bleeding are not available for usage in electronic health records. The proposed algorithm for ISTH bleeding definitions in electronic health record data will bridge this gap and enable meaningful comparison of clinical trial and observational study results.
With the rise of digitisation in the healthcare sector, huge amounts of “real-world” patient clinical data are collected during routine medical care including electronic health records. Observational studies performed using these real-world data have the potential to greatly expand our knowledge of public health by observing millions of patients at a time. These observational studies complement (randomized controlled) clinical trials that represent an experiment performed under controlled conditions with prospective data collection. For this reason, it is of critical importance to combine the knowledge gained from clinical trials and observational studies.
In order to compare and merge evidence generated in clinical trials and observational studies, it is necessary to have a common ‘language’ describing medical events. For example, if a clinical trial reports that 2 out of 100 patients experienced major bleeding within one year, an observational study that hopes to replicate this finding on a larger or different population must know exactly how to define ‘major bleeding’ as done in the clinical trial.
Lead author Alexander Hartenstein emphasises the importance of common definitions saying, "Accurately defining clinical events in real-world data is of the highest priority to increase the value of evidence generated from observational studies"
Alexander Hartenstein, MD, IEG Data Scientist, Bayer AG
Hartenstein, A., Abdelgawwad, K., Kleinjung, F., Privitera, S., Viethen, T. and Vaitsiakhovich, T. (2023) “Identification of International Society on Thrombosis and Haemostasis major and clinically relevant non-major bleed events from electronic health records: a novel algorithm to enhance data utilisation from real-world sources”, International Journal of Population Data Science, 8(1). doi: 10.23889/ijpds.v8i1.2144.