In conjunction with a Patient Origin Study which surveyed all patients discharged from hospitals throughout most of California, a separate feasibility study was conducted in northeastern California to evaluate the effectiveness of utilizing provisional discharge diagnoses recorded by Medical Record Librarians at the time of discharge as an added factor to incorporate into future traffic pattern analyses. This paper reports on a validation study of the diagnosis portion of the study. A 5 per cent sample (156 cases) of the 3,308 patients discharged from the pilot area was selected for chart review. Approximately 10 per cent of the discharge diagnoses showed discrepancies between the original survey results and the validation review, including 3 per cent of moderate or severe differences in interpretation and 1 per cent numerical transposition. On the basis of these figures, it is felt that the use of provisional discharge diagnoses yields useful information of acceptable accuracy.