Abstract:

COVID-19 pandemic showed how fast new drugs can be introduced and the need for reliable drug efficacy and safety evaluation methods. Lack of past data that could be used for evaluation and validation of such methods limits possibility to work on them before real data is available. Using a common electronic health record data type as well as possibility to generatesynthetic data could allow early validations and end to end testing of such methods. To address these limitations, a system was proposed system that allows for end-to-end patient population generation, data transformation and method evaluation. The system utilizes existing and well-known patient generation framework Synthea, as well as OMOP CDM data model along with patient level prediction framework. Exisitng solutions were extended for the purposes of the system including possibility of modelling adverse drug reactions in Synthea and orchestrated for seamless execution.