MIRERC 074/2025: MODEL THE TIME TO HOSPITAL READMISSION OF TYPE 2 DIABETIC PATIENTS IN MERU COUNTY, USING THE COX PROPORTIONAL HAZARD MODEL WITH FRAILTY

Abstract

Hospital readmissions among diabetic patients are a growing healthcare concern worldwide, and Kenya is experiencing the burden rise. This study aims to identify and model the patient-level and clinical risk factors that influence the time to hospital readmission among type 2 diabetics in Meru County, Kenya. This study will adopt a retrospective cohort design and data will be collected from the Meru Teaching and Referral Hospital(MTRH) in Meru County. Data analysis will be performed using R software. Data exploration will be done to inform factor analysis and plots such as Kaplan-Meier will be used to inform the both the Cox and the extended Cox models. The study will adopt both the traditional CPHM and CPHM with frailty to determine the key predictors of time to readmission based on the hazard ratios. The study will fit two models; the traditional Cox proportional risk model and the Cox proportional risk model (CPHM) induced with frailty. A discriminative performance comparison will be established between the traditional CPHM and the extended CPHM with frailty using the Concordance Index. The results of this study will offer important knowledge of clinical, social and economic predictors of hospital re-admission, and provide a basis for designing effective interventions to reduce this burden. To the body of knowledge, this study will bridge clinical and statistical knowledge through the integration of statistical techniques with the clinical data in illustrating how the complex health phenomena can be quantified using statistical tools.