Identifying Risk Factors causing Kidney Transplantation Using Generalized Estimating Equations

Generalized Linear Models (GLMs)and Generalized Estimating Equations (GEEs) allow response variables to follow exponential distributions and accept correlated or clustered data in longitudinal or repeated measures research. GEE is ideal for modeling non-normal outcomes and delivers consistent parameter estimates even with a misspecified working correlation structure. GEE estimates robustly. GEE using the Poisson distribution and log link function are used to evaluate six critical predictors. Which represent risk factors for kidney transplant statues, our data collected from 150 patients with kidney transplant cases (done, preparing to transplant), and for our model of prediction, the researcher compares two types of correlation structures (Exchangeable and independent) to represent within-subject correlations to determine which model is better for our data. Future patients can benefit from this medical study. Exchangeable correlation structure was better and adequate, finding five risk factors more important than independent correlation. GEE was resolved repeatedly using robust variance estimation to estimate parameters. The test effect model found that renal artery stenosis, polycystic kidney disease, diabetes, congenital kidney disease, and uncontrolled hypertension predicted kidney transplantation results. Systemic lupus erythematosus did not significantly effect, and exchangeable model has lower QIC and QICC values which’re model selection criteria., making it a better model. Gradient Vector Hessian Model The Matrix converged estimators with stable solutions. GEE provides average effects across a population, which should help policymakers allocate resources and design preventive health programs that target the most important risk factors to delay or reduce kidney transplantation. GEE limitation cannot forecast the effects on an individual's outcome trajectory and sensitive to limited sample sizes or a small number of clusters.