Analyzing Menstruating Interval to the First Conception: An Application of CPH and AFT Models

Survival analysis techniques analyze the time-to-event data. In survival analysis, there are two essential methods: Semiparametric and parametric methods. The semiparametric Cox proportional hazard (CPH) model is frequently used to analyze demographic survival data. However, the presumption of proportional hazard (PH) is not always met in real-life data. Meanwhile, one can use parametric Accelerated Failure Time (AFT) models as an alternative when the PH assumption does not hold. The main purpose of this article is to analyze and compare the performances of the CPH model and AFT models in identifying the significant covariates affecting the menstruating interval to first conception (MIFC). In this article, three parametric AFT distributions based on Exponential, Weibull, and Log-normal distributions are used to check the performance. We have shown the violation of having a proportional hazard assumption with the help of a graphical technique and statistical test. According to the Akaike Information Criteria (AIC) and Bayesian Information Criterion (BIC), we have found that the Weibull AFT model is best suited for the data. Further in this paper, we have identified the significant factors affecting the MIFC. Analysis reveals that covariates such as age at marriage, place of residence, wealth index, mother’s education, and Mass media exposure are significant factors affecting the menstruating interval to first conception in Uttar Pradesh.