We conducted a case--control study to examine the efficacy of non-stress testing in preventing fetal death in post-term pregnancy. The analysis was based on data from the 1988 National Maternal and Infant Health Survey, which was a nationally representative sample of live births, fetal deaths and infant deaths that occurred in 1988. Information on whether a woman had non-stress testing was obtained from a questionnaire sent to prenatal care providers and hospitals. Cases were post-term women (with 42 weeks or more gestation) who had fetal deaths. Three post-term controls, who had live births and who delivered at the same time or later than the cases, were randomly chosen and individually matched to each case by maternal race. The proportion of women who had one or more non-stress tests during pregnancy was compared between cases and controls. Non-stress testing was used in 30.9% of the 126 cases and in 28.5% of the 375 controls. The race-adjusted odds ratio for exposure to non-stress test was 1.12 [95% CI 0.72, 1.75]. After controlling for other important confounding variables the odds ratio was 1.05 [95% CI 0.57, 1.91]. These results do not support the efficacy of non-stress testing in post-term pregnancies. A more detailed evaluation of this widely used screening procedure is needed.