Statistical Techniques for the Analysis of Repeated Measures: An Application to Diabetes Mellitus Data

Abstract

The focus of the research is to model some risk factors called covariates that influence the blood sugar level of diabetes patients using Generalized Linear Mixed Models (GLMM) and Generalized Estimating Equations (GEE). Both primary and secondary data was obtained from the Diabetes Centre of Maamobi Polyclinic in Accra. Data was obtained on 155 patients for four consecutive times they visit the hospital for review. Analysis of the data was done using SPSS version 20 and R. The findings revealed that the covariates (Age, Weight, Systolic blood pressure, Family history) were significant in both the GLMM and GEE models. A conclusion was drawn to the effect that since factors influencing the rise and fall of the sugar content in the blood of diabetes have been established, health providers should note these and lay emphases on them in controlling the epidemic (diabetes). Health providers must also be more vigilant when controlling high or low systolic and diastolic blood pressure of patients. Patients must be educated regularly on these factors so they can control their sugar levels.