Background: Expectant parents of very preterm infants, physicians, and policy makers require estimates for chances of survival and survival without morbidity. Such estimates should derive from a large, reliable, and contemporary data base of easily available items known at birth.
Objective: To determine short-term outcome and risk factors in very-low-birth-weight preterm infants based on administrative data.
Methods: Anonymized routine data sets transmitted from hospital administrations to statutory health insurance companies were used to assess survival and survival free of major morbidities in a large cohort of preterm infants in Germany.
Results: After exclusion of infants with lethal malformations, there were 13,147 infants with a birth weight below 1,500 g admitted to neonatal care 2008-2012, of whom 1,432 infants (10.9%) died within 180 days. Estimated 180 days survival probabilities were 0.632 (95% confidence interval 0.583-0.677) for infants with 250-499 g birth weight, 0.817 (0.799-0.834) for 500-749 g, 0.931 (0.920-0.940) for 750-999 g, 0.973 (0.967-0.979) for 1,000-1,249 g, and 0.985 (0.981-0.988) for 1,250-1,499 g. Estimated probabilities for survival without major morbidity (surgically treated intraventricular hemorrhage, necrotizing enterocolitis, intestinal perforation, or retinopathy) were 0.433 (0.384-0.481) for 250-499 g, 0.622 (0.600-0.643) for 500-749 g, 0.836 (0.821-0.849) for 750-999 g, 0.938 (0.928-0.946) for 1,000-1,249 g, and 0.969 (0.964-0.974) for 1,250-1,499 g, respectively. Prediction of survival and survival without major morbidities was moderately improved by adding sex, small for gestational age, and severe or moderate congenital malformation, increasing receiver operating characteristic areas under the curve from 0.839 (0.827-0.850) to 0.862 (0.852-0.874) (survival) and from 0.827 (0.822-0.842) to 0.852 (0.846-0.863) (survival without major morbidities), respectively.
Conclusion: The present analysis encourages attempts to use administrative data to investigate the association between risk factors and outcome in preterm infants.
Keywords: morbidity; mortality; predictive power; preterm infant; risk factor.