A geographically determined cohort of all infants of less than 1,500 g born in 1980 and 1981 were clinically followed up to age four to determine their disabilities which were classified as mild, moderate or severe. A quality adjustment coefficient, determined by the severity of the disability, was used to calculate the number of quality adjusted lives produced. The total cost of care for these children was assessed as the sum of hospital costs to age four (which were specifically determined) and an estimate of the life-time costs of care of disabled children from information provided by the Education and Social Service departments. A very poor predictive power of birthweight with cost was obtained with the ungrouped birthweight data, whereas clinical factors explained up to 60% of the variance of the initial hospital costs for survivors and up to 30% of the variance of life-time costs and the cost of quality adjusted lives produced. Rules for the allocation of resources based on discrimination by birthweight are flawed but the application of clinical discretion is important.