A multiple regression analysis of admission of infants to hospital. A report from the Jerusalem Perinatal Study

Abstract

A stepwise multiple regression technique was used to select the combination of variables which together give the best estimate of the number of hospital admissions in the first year of life for the 17,244 infants born in West Jerusalem in 1966-68 who survived the first 28 days of life. From the 60 demographic and health predictors tested, 30 variables together explained 13% of the variation in hospital admissions. Significant predictors were the presence of major and minor congenital malformations, parity, linear and quadratic functions of birth weight, social class, and parental ages, Cesarean section, season of birth, parental education, ethnic group, and area of residence. There was no significant effect of the mother's going out to work, nor of breech delivery, fetal distress or asphyxia, and a number of other obstetric complications. Admissions decreased with increases in birth weight, social class, parental ages, and education. There was a significant interaction between social class and paternal age, the effect of class increasing with increasing paternal age. There were more hospital admissions with increasing parity, delivery by Cesarean section, the presence of malformations and among infants of Arab fathers and Jewish mothers of Asian and North African origin. There was also a significant interaction between sex and ethnic group, males being more at risk if the father or paternal grandfather had immigrated from Asian or North African countries. The data support the hypothesis that sex differences in hospital admissions of Jerusalem infants arise from a status difference rather than from a biological difference between males and females.