Prospective evaluation of a risk scoring system for predicting preterm delivery in black inner city women.

Nine percent of United States births are preterm. The ability to efficiently identify women destined to deliver before term would enable obstetricians to initiate early intervention. The Papiernik-Creasy risk scoring system is being extensively applied for this purpose, without prospective validation in this country. We evaluated prospectively its ability to predict high risk women in a black inner city population, when applied once early in pregnancy. The system, based on social, past medical, and current pregnancy problems, failed to identify at-risk patients. Adjusting the score cutoff and reweighting factors with discriminate analysis did not improve the predictive value. Analysis of an additional 60 social, demographic, and medical factors failed to produce a useful alternative risk assessment tool. We suggest that risk scoring systems such as these may not be helpful in socioeconomically at-risk populations.