Between 1979 and 1987, there was an estimated 361% increase in the number of drug-affected newborns discharged from the 6000 non-federal, short-stay hospitals in the United States. Per 10 000 newborns, the rate increased 339%, mostly occurring after 1983. The estimated number of drug affected newborns in 1987 was about 13 000 (95% confidence interval: 10 000-15 000). Recognizing that underreporting could have occurred, multiplicative correction factors derived from the literature were employed to produce 'adjusted' estimates. While somewhat arbitrary, the number of drug-affected newborns, adjusted for underreporting, was about 38 000 (95% confidence interval: 30 000-45 000). Both adjusted and unadjusted estimates of the numbers of newborns identified as drug-affected by the present study is much smaller than most of the estimates reported in the literature. Possible reasons for this finding are discussed.