Modeling Time-to-Good Control of Hypertension using Cox Proportional Hazard and Frailty Models: the case of Bahir-Dar Felege Hiwet Referral Hospital

Abstract

Hypertension is a chronic disease that has a major health problem over the centuries due to its significant contribution to the global health burden. According to World Health Organization, hypertension is the seventh leading cause of death in Ethiopia and it ranks as one of the first causes of cardiovascular related mortality. Objective:The main objective of this study is to explore possible modeling approaches of time-togood control of hypertension using Cox proportionalhazard and frailty models, using data from Bahir-Dar Felege Hiwet Referral Hospital. Methods: An institutional-based retrospective cohort studywas conducted in June 2014. The study population consists of all hypertensive patients measured repeatedly at least three times with a three month follow up between1 st January, 2009 to last December, 2013. Five hundred patients were selected using simple random sampling. The data were collected by trained data collectors using check list. SPSS version 16 and R software were used for data entry and processing of the data, respectively. First, single covariate analysis was done using Cox proportional hazard and univariate frailty models. Then all variables that are significant were included in the multi-variable analysis. Results: The median survival time of hypertensive patientsto attain good control is 48 months and the mean survival time is 43.6 months. Age and systolic blood pressure of patients have a negative relationship with outcome variable. However, fasting blood sugar has positive relationship with the outcome the variable. Moreover, the result showed that, the progression of outcome depends on patient’s baseline socio-demographical characteristic such as age. Conclusion: Cox proportional hazard based analysis reveled thatthe major factors that affect good control of hypertensive patients are age, systolic blood pressure, fasting blood sugar and creatinine. The result of univariate frailty analysis showed that there is unobserved heterogeneity between individuals in the study set-up, which indicates, there are unmeasured covariates.the clinicians should have give an attention to the younger age and lower systolic blood pressuregroup to attain good control of hypertension earlier like that of older age and higher systolic blood pressuregroups.