reducing the risk of low or high birth weight for women with low or high body mass index under the care of high quality hospital by using instrumental variable

background and purpose: women with low (high) pre-pregnancy body mass index (bmi) recently delivered infants with approximately normal or close to normal birth weights under the high quality of prenatal care. this study estimated the effect of pre-pregnancy bmi when concerns about the effects of different quality levels of prenatal care and the health status of mothers and their infants existed. materials and methods: the sample consisted of the female patients who referred to one of the two hospitals with different quality levels of prenatal care in a city of iran. the logistic mixed effect model and chi-square test did not show any significant effect of low (high) bmi on the risk of low (high) birth weight. then, the two-stage residual inclusion instrumental variable (iv) method was used to estimate the effect of bmi in order to overcome the effects of the levels of quality care and the health status of the mothers and their infants. results: adjusted iv analysis revealed that women with a low bmi experienced an approximately 18% (rr=0.82; 95% ci (0.69, 0.97)) reduction in the risk of delivering a lbw infant and women with a high bmi experienced an approximately 26% (rr=0.74; 95% ci (0.57, 0.96)) reduction in the risk of delivering a hbw infant when they were under the care of a high quality hospital. conclusion: this study revealed that the effect of bmi is confounded by the effects of quality of care and the health status of the mothers and their infants. further, these results contributed to providing the conditions in improving the health status of mothers and their infants during pregnancy in local areas.