Decision support system minimizing the risk of post-operation complications after surgeries

Abstract

If colorectal cancer requires surgical intervention, it is performed in either the laparoscopic or open method. It is desirable to choose the best methodology for each patient. We first consider the POSSUM model to estimate the probability of potential complications. The model was not found accurate enough, so the logistic regression was used to find new parameters of the POSSUM model. This new model was not accurate enough as well. A decision was made to decompose the POSSUM model to its 18 core variables and try to find the ones which are statistically significant. Using the method of logistic regression we isolated the variables providing us with the most information and based a new model on them. The quality of this model is further going to be discussed and a simple formula is provided to decide on the optimal operation technique while minimizing the risk of post operation complications.