[Corroboration of concurrent assessment of inappropriate hospitalization with retrospective evaluation based on...

Since the availability of a German adaptation of the "appropriateness Evaluation Protocol" (AEP) for hospital utilization review, physicians and health care managers were concerned about a potential overestimation of the extent of inappropriate hospital use in a retrospective assessment. The objective was to assess the agreement of concurrent and retrospective assessment of inappropriate hospital use.

The appropriateness of 42 admissions to surgical wards in a teaching hospital was tested concurrent to the admission. In 25 hospitalized patients the appropriateness of hospital days was assessed. Results were compared with the retrospective testing of the same patients after three months.

Agreement between concurrent and retrospective assessment within the same reviewer was 86% of admissions (95% CI = 75%-96%) and 96% of hospital days (88%-100%). Kappa showed values of 0.85 and 0.78, which can be described as excellent. A comparison of two different reviewers, one testing concurrently, the other retrospectively showed agreements of 86% and 90%. The proportion of inappropriate days was the same in both designs.

The results refute prior concerns that a retrospective assessment of the appropriateness of hospital use is impossible because in a retrospective design important clinical information would remain unconsidered or clinical information not yet available at that time would be used for the evaluation. The possibility of an incorrect assessment due to a retrospective design is within the variability inherent to the instrument. Despite the strong agreement, the small sample of surgical patient limits the generalizability of results.