A Case Control Study with the Statistical Predictive Modeling of Child Burn in Pakistan

Abstract

In any country children are the precious asset and burns are targeting them usually in low income country regions so there is pressing need for some kind of predictive modeling to prevent the children from burning. This study aims to find out about the Risk Estimates of each factor, fitting of Conditional Logistic model along with the predicted probabilities for different choices of each Risk Factor. Simple random samples of 100 cases and 100 age and gender matched controls were selected from the Burn Unit and Pediatric OPD of Mayo Hospital, Lahore. It is found that males are at high risk of getting a burn injury 1.291 times more than females; children of last or high birth order are at risk of 1.222 times more than others; a child with ≤4 number of siblings has risk of burn 1.545 times more; if the child is cared by other persons, then it increase the risk by 2.042 times more; illiterate mother (1.556 times more), children that do not go to school (1.390 times more), low income occupation of father (1.430 times) rural residence (1.35 times), rented house (1.364 times), kitchen burden (1.176 times), joint family system (1.128 times), poor living standard (3.026 times), mother pregnancy (2.345 times), no past burn experience (6.192 times) will increase the risk of burning of any child with high Odds Ratios. Conditional Logistic model reported 7 variables as significant and gave best fit to it; residence (rural), house status (rented), availability of kitchen (no), living standard (poor), use of heating device (no use), availability of conveyance (no), past burn (no), all these socioeconomic factors have p-value <0.10 and they have strong significant association with child burning.