Prediction of Hospitalization During Infancy: Scoring the Risk of Admission

Abstract

A cohort of 5,243 live newborn infants in Haifa was followed to determine hospital admissions during the first two years of life. Parameters of the family and the neonate, routinely available at birth, were collected and studied in order to select those which would identify co-host infants with an increased risk of hospitalization during infancy. A total of 767 cohort children (14.6%) were admitted to hospital. Multivariate analysis showed that many group factors are interrelated and therefore inconsistent. Six single independent factors predisposing to hospital admission could be defined (in order of decreasing significance): increasing birth order, male sex, poor education (up to four years of schooling), birth weight less than 2,250 gm, Jewish mothers born in Asia or Africa, and maternal age up to 24 years. Numerical risk coefficients allocated to each of the above six single (multivariate) factors enabled the calculation of a predicted risk coefficient scoring the risk of admission to hospital. The use of a score predicted 43.3% of the total initial hospitalizations affecting 23.3% of the total cohort. The allocation of health resources might be improved by the use of such a method of scoring to select families in need of special services.