Table 2: Proportional Hazard Assumption This suggests that PH assumption is met at o = 5% significant level. The Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) for each model is presented in Table 5. In order to compare the performance of the conventional cox proportional hazard model with cox proportional hazard model with random effect. The AIC and BIC values were used. The result as presented in Table 3 revealed that the cox proportional hazard model with random effect parameter has least value of AIC and BIC as compared to the conventional cox proportional hazard model (without random effect parameter). This suggest that the cox proportiona hazard model with random effect parameter outperformance the conventional cox proportional hazard model. Hence, the researcher focused on the interpretation of he result based on the cox proportional hazard model with random effect parameter. In addition, the performance of AFT model with random effect and the conventional AFT model were also asses using the AIC and BIC values. The result of the analysis as presented in Table 3 indicated that the Weibull AFT model with Gamma random effect distribution has the least AIC and BIC. This suggest that Weibull AFT model with Gamma random effect outperformance the exponential, log-logistics and lognorma models. Hence, it was considered as the bes AFT model for predicting the survival time for Hypertension patient. This implies tha Weibull Gamma random effect AFT mode it is more efficient model to describe the determinant factors of time-to-event of hypertension patient. Hence, the study focused on the interpreting the result of Weibull Gamma random effect AFT model.
