18 severely malnourished children (IM) who participated in a 3-year home-visiting program were compared with 2 other comparison groups comprising 17 severely malnourished (NIM) and 19 adequately nourished children (controls). On enrollment, all the groups were in the same hospital, and both malnourished groups had lower developmental levels than the controls. The IM group received intervention for 3 years after hospitalization, consisting of weekly or 2 weekly home visits with toy demonstrations. At 7, 8, 9, and 14 years after leaving the hospital, the 3 groups were compared on tests of school achievement and IQ. The NIM group showed no sign of reducing their deficits, and at the 14-year follow-up they had markedly lower scores on the WISC verbal and performance scales, the Wide Range Achievement Test (WRAT), and the Peabody Picture Vocabulary Test (PPVT), than the controls. Throughout the follow-up the IM group's scores were intermediate between the NIM and the controls in every test. At the 14-year follow-up, their scores were significantly higher than those of the NIM group in the WISC verbal scale, and the difference approached significance in the WRAT. We conclude that psychosocial intervention should be an integral part of treatment for severely malnourished children.