An obstetric scoring system: its development and application in obstetric management.

Abstract

To develop a statistically derived but clinically usable antenatal risk scoring system. Data from 20,985 pregnancies were statistically analyzed to identify significant risk factors. Logistic regression analysis was then used to produce a final scoring system, which was subsequently tested for validity on a separate population of 3120 pregnancies. Twenty-seven significant antenatal variables were included in the final scoring system. Application of the system in early pregnancy resulted in a predictive accuracy of 0.73; at the onset of labor, predictive accuracy was 0.91. At the time of labor, 87% of poor outcomes were accurately identified by allocation of only 16% of the women to the high-risk group. It was possible to develop a risk scoring system with a predictive accuracy higher than any previously reported statistically derived score. Summation of the logistic coefficients provides a score that by comparison with a chosen threshold identifies a high-risk pregnancy. In this way, despite the complexity of statistical analysis, all clinicians can quickly apply this scoring system.