Complications following surgical procedures strike upwards of 15% of the 234 million patients operated annually worldwide. Upwards of 4% of patients die as a direct result of the complication, and survivors are often severely affected by these events. Furthermore, complications increase the cost of treatment by upwards of 130%, and is thus a significant economic burden on society.
Accurate risk prediction prior to surgery have the potential to reduce morbidity, as well as provide substantial health-care related savings. Current prediction models are, however, not able to take into account precision medicine effects such at personal genomic profile
With this project, we seek to use artificial intelligence algorithms to assess the how personal genomics affects the risk of postoperative complications. We primarily seek to train algorithms on data from the UK biobank in order to teach the model how genetics can influence disease. Subsequently, we seek to transfer these trained models to a the surgical setting, to investigate whether they can assist in more accurate risk prediction.