Use of regression models to predict glomerular filtration rate in kidney transplanted patients

Abstract

Despite of the numerous progress of modern medicine, organ transplantation is never a free risk procedure. Kidney transplant, in particular, can bring to numerous short and/or long-term problems, like infection or diabetes. Since such problems can appear up to years, a constant monitoring of Kidney transplanted patients is necessary to try and avoid a bad prognosis. Glomerular Filtration rate (GFR) is an important marker to evaluate kidney transplanted patients. Since it is necessary to measure values of GFR over time, new predictive approaches can reveal useful to such scope. In this work we present a Multiple Linear Regression model and a Machine Learning method to correlate GFR with glycaemia (mg/dL) and the dosage of a calcineurin inhibitor. Results show how such model can be useful in a long term evaluation of kidney transplanted patients.