Many aspects of medicine would be well served by a simple method to assess the outcome of care in specified groups of patients. This study examined charts of patients with stomach cancer on a surgical service. Two digits were added to the ICD-9 number in the routine discharge data; one for the nature and severity of case and the other for the outcome of care. The digits were designed for on-line registration at discharge. Information was also obtained on resource consumption in the various groups of patients. Most of the variables had to be evaluated implicitly as there were no explicit judgement criteria and few empiric data available for comparison. Implicit evaluation of the results was significant and prompted steps for improving care. With current systems, the information obtained from traditional hospital statistics is limited and partly misleading. By slight modification, however, hospital statistics may provide valuable information for assessing quality of care and resource allocation during hospitalization.