Assessing Preweaning Mortality in Goat Kids Using Logistic Regression and Cox Proportional Hazards Models

A study was aimed to assess risk factors for preweaning mortality (PWM) and to develop incidence and time-to-event models using logistic regression and Cox proportional hazards analysis in Malabari, Attapady Black and crossbred kids. Data on 3935 kids born between 2011 and 2022 at the University Goat and Sheep Farm, Mannuthy (Thrissur), Kerala, were screened and a complete case dataset of 769 kids (284 deaths, 485 censored) was used for modelling. Predictor variables included breed, sex, type of birth, birth weight, parity, age and weight of dam. Stepwise logistic regression identified breed and birth weight as the most important determinants of PWM, with the odds of PWM 3.98 and 6.08 times higher in Attapady Black and crossbreds than in Malabari kids, and a one kg increase in birth weight reduced the odds by about two-thirds (odds ratio 0.32). Cox proportional hazards models demonstrated that breed, type of birth and birth weight significantly (P<0.001) influenced time to PWM; lower birth weight was associated with an earlier occurrence of mortality. The selected models showed satisfactory discrimination and calibration for identifying at-risk kids. Llogistic regression and Cox proportional hazards models thus provide robust basis for risk stratification and early decision-making to mitigate PWM in kids.