Population differences affect the interpretation of fetal nonstress test results

OBJECTIVE: The object of the study was to determine whether population differences exist with respect to outcomes of women with reactive and nonreactive nonstress test results. STUDY DESIGN: An epidemiologic evaluation was conducted on 2579 women who underwent nonstress tests in the Fetal Assessment Center of the Johns Hopkins Hospital within a week of delivery. Risk factors such as hypertension, diabetes, and postterm pregnancy were used in a logistic regression model to evaluate the ability of the nonstress test to predict outcomes including proxies of fetal distress and fetal and neonatal death. The sensitivities, specificities, and predictive values of the nonstress test for predicting these outcomes in cohorts of black and white women were also determined. RESULTS: The nonstress test was consistently more sensitive for black women than for white women in predicting several perinatal outcomes, but specificity and negative predictive value were consistently lower for black women. The positive predictive value for fetal and neonatal death was higher for white women than for black women. Although the nonreactive nonstress test result seemed to be predictive of certain perinatal events, the odds ratio for predicting perinatal mortality in any study population was no greater than when the nonstress test result was reassuring. CONCLUSIONS: Epidemiologic characteristics affecting test results, such as disease prevalence and population differences, may lead to clinically significant differences in outcome prediction when these tests' results are used. These differences should be considered in the implementation of antepartum fetal testing programs.