Based on 9350 pregnant Japanese women who were screened by serum triple-marker determination, accuracy of predicted risk for Down syndrome was examined using 24 Down syndrome cases detected either prenatally or postnatally. The correlation is statistically very high (r = 0.98) between the predicted risks and the prevalence of Down syndrome cases. Here we emphasize that this could be accomplished only by an extensive follow-up study, implemented in our prospective intervention programme.