Loss to follow-up is a major problem in indigent inner-city populations. We evaluated a large, well-described, inborn indigent population of high-risk infants (HRI) and control infants (CI) to assess possible selection biases in loss to follow-up at one year adjusted age. Serial clinic visits, phone calls, and letters and payment of $20.00 for attending at 1 year was used to minimize patient loss. Yet, the 1 year loss rate was high, and among HRI, greater for ventilator-treated infants greater than 1500 g birthweight (71/114; 62%) than for ventilator-treated very-low-birthweight (VLBW; less than 1500 g) infants (39/108; 36%) or non-ventilated VLBW infants (62/145; 43%) (P less than 0.05). Multivariate analyses indicated that those lost to follow-up were at no greater risk of a poor outcome on the basis of prenatal and perinatal medical and socioeconomic findings than were those in the same risk group (HRI or CI) or subgroup of HRI who were examined at 1 year. In a review of hospital records, similar rates of hospitalization and neurologic problems during infancy were identified for HRI examined and HRI lost to follow-up. The identification of such morbidity during infancy may be less complete for HRI lost to follow-up than for those examined. Thus, the high frequency of deficits observed in follow-up evaluation of indigent HRI is unlikely to result from loss of unaffected infants.