Identifying Heterogeneous Treatment Effects with Machine Learning

Medications affect individuals differently. For some, the medication may be effective and well tolerated, while others may experience a lack of effect or significant adverse events. Currently, when medications are tested in clinical trials, the existence of patient subgroups with different responses is rarely considered. Consequently, medications are recommended and used in our health care system based on average responses although there may be large underlying differences in the patient response to treatment. My project will implement and evaluate the use of machine learning on health register data to identify such subgroups with varying responses. This is pioneering research and if successful, will result in personalized and targeted recommendations on medication usage and better treatment outcomes for patients