Bayesian Shared Frailty Models for Time to First Birth of Married Women in Ethiopia: Using EDHS 2016

Abstract

Introduction. The primary effect of the fertility process is the birth of the first child. The ages at which women establish marital union and give their first birth depend on and result in varying demographic features. This research demonstrates how to examine the effect of numerous factors on married women’s delay to first birth in Ethiopia using Bayesian parametric models with gamma shared frailty distribution. Methods. This study analyzed data from the 2016 EDHS on factors related to the time of married women to first birth. A sample of 8810 married women from all parts of Ethiopia participated in the study. The Akaike information criterion (AIC) and Bayesian information criterion (BIC) were used to compare several parametric models with gamma shared frailty distributions to find the best model (BIC). Finally, when the prior data was taken into account, the chosen model was proven to be accurate (Bayesian approach). Results. The median survival time for the first birth after marriage is 24 years (95% CI; 23.4, 25.3). The result shows that the place of residence, the access to media, the level of education of the mother, the education level of the husband, the use of the head of the contraceptives, and the sex of the household are statistically associated with the time to first birth of married women. The Weibull-gamma shared frailty model under the Bayesian approach was found to be the best model that fit the time to first birth data in this study. The result also showed that there is heterogeneity between regions of married women. Conclusion. To slow the increase in the Ethiopian population, families must be taught how to use contraception, and rural populations must be educated on the necessity of increasing the length of the first birth gap rather than encouraging early marriage. In general, attempts to reduce fertility by raising the age of the first marriage must consider the social and cultural settings in which marriage takes place. On the other hand, the campaign against early marriage should focus on the sociocultural, physiological, and psychological effects, as well as the reduction of reproduction.

Time to first birth and its predictors among reproductive-age women in Ethiopia: multilevel analysis using shared frailty model

“…The median time to give first birth in this study is slightly lower than the findings from EDHS, 13 which was 18 years and 24 years. This result is comparable with the finding by Negash et al in SSA countries where there is a high fertility rate.…”

A Multivariate Proportional Odds Frailty Model with Weibull Hazard under Bayesian Mechanism

“…From the study it is observed that the newly https://www.indjst.org/ Proposed Multivariate Proportional Odds Frailty Model using Weibull Hazard perform better as compared to the traditional proportional odds model without the frailty term. Like the earlier works done by (1,7,8,19) etc. this study also recommended the application of frailty models instead of traditional survival model to analyse survival data.…”

“…From the Table 3 it is observed that the Multivariate Proportional Odds Frailty Model using Weibull Hazard has a larger LMPL and Smaller DIC as compared to the Multivariate Proportional Odds (PO) Model without the frailty term. So in case of the above analysis it is observed that the Proposed Multivariate Proportional Odds Frailty Model using Weibull Hazard is better than the model without frailty term (1,8,19) also developed some new proportional hazards and proportional odds frailty models and compared its performance with other traditional survival models using AIC, BIC, DIC and many other criteria and observed that fit of the frailty models are better than the traditional survival models without the frailty term.…”

“…(7) proposed a Semi-Parametric Cure Rate Proportional Odds Models with Spatial Frailties for Interval-Censored Data. (8) developed a Bayesian shared Proportional odds gamma frailty models for time to first birth of married women in Ethiopia. (9) in their study used a Bayesian hierarchical survival model in the setting of competing risks for the spatially clustered HIV/AIDS data.…”

“…Form the literature review it is observed that various researchers work in the field of Bayesian analysis of Multivariate Proportional hazards frailty model but so far our knowledge goes, it is observed that a very less amount of work have been done yet in case of Bayesian analysis of Proportional Odds frailty models. However, some researcher work in the fields of Bayesian analysis of Proportional Odds models under spatial frailty approach (7,8) . Still, some distributions exist, which are very simple but till now not used for modeling the baseline hazard function in case of Bayesian analysis of Multivariate Proportional odds frailty models.…”