The value of Diagnosis Related Groups, Refinement Diagnosis Related Groups and Pediatric Modified Diagnosis Related Groups in measuring the output of specialist children's hospitals was evaluated by the examination of discharge data for 1987-88 from three major Australian children's hospitals and four district hospitals. The study included all patients aged 0-18 years but excluded those with specific neonatal diagnoses. Findings indicated that Refinement Diagnosis Related Groups seem to give a better measure of the output of specialist children's hospitals than Version 5 Diagnosis Related Groups in that they explained a higher proportion of variation of length of stay. Pediatric Modified Diagnosis Related Groups developed in the United States of America for specialist children's hospitals overall did not seem to have major advantages over Refinement Diagnosis Related Groups but there were some specific Pediatric Groups that appeared beneficial. Further modification of the Refinement Diagnosis Related Groups could allow these advantages to be incorporated. Overall it seems preferable for there to be a similar system measuring output of both children's hospitals and general hospitals and this could be achieved by some minor changes to the Refinement Diagnosis Related Groups.