Development of machine learning models for the prediction of complications after colorectal and small intestine surgery in psychiatric and non-psychiatric patient collectives (P-Study)

Abstract

Psychiatric and psychosomatic diseases are an increasingly cumbersome burden for the medical system. Indeed, hospital costs associated with mental health conditions have been constantly on the rise in recent years. Moreover, psychiatric conditions are likely to have a negative effect on the treatment of other medical conditions and surgical outcomes, in addition to their direct effects on the overall quality of life. Our study aims to investigate the impact of preoperative Risk factor, psychiatric and psychosomatic diseases on the outcomes of colorectal surgery and length of hospital stay via predictive modeling techniques. Method: Patient data will be collected from several participating national and international surgical centers The machine learning models will be calculated and coded, but also published according to the TRIPOD guidelines (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis). Result: It is conceivable to arrive at generalizable models predicting the abovementioned endpoints through large amounts of data from several centers. The models will be subsequently deployed as a free-to-use web-based prediction tool using the Shiny environment [47]. The cost for the hosting server and digital infrastructure will be covered by Dr. Anas Taha. The data, which has been gathered, will be saved for ten years by the sponsor.