Comparison of Disagreement and Error Rates for Three Types of Interdepartmental Consultations
American Journal of Clinical Pathology
A b s t r a c t Previous studies have documented a relatively high rate of disagreement for interdepartmental consultations, but follow-up is limited. We reviewed the results of 3 types of interdepartmental consultations in our hospital during a 2-year period, including 328 incoming, 928 pathologist-generated outgoing, and 227 patient-or clinician-generated outgoing consults. The disagreement rate was significantly higher for incoming consults (10.7%) than for outgoing pathologist-generated
... ogist-generated consults (5.9%) (P = .06). Disagreement rates for outgoing patient-or cliniciangenerated consults were not significantly different from either other type (7.9%). Additional consultation, biopsy, or testing follow-up was available for 19 (, but differences were not significant (P >.05 for each). Review of the individual errors revealed specific diagnostic areas in which improvement in performance might be made. Disagreement rates for interdepartmental consultation ranged from 5.9% to 10.7%, but only 33% to 79% represented errors. Additional consultation, tissue, and testing results can aid in distinguishing disagreements from errors. Diagnostic accuracy is crucial in anatomic pathology, including surgical pathology and cytopathology. Many studies concerning diagnostic accuracy have relied on diagnostic agreement, often from interdepartmental consultation, as a measure of accuracy. 1-11 These studies have shown a wide range of disagreement rates. As a result, many pathologists and pathology organizations recommend review of outside material before undertaking procedures within their own institutions. 6 Nevertheless, although some aspects of interdepartmental consultation have been well studied, others have not. For example, there are several different types of interdepartmental consultation, including patient-generated, pathologistgenerated, and incoming and outgoing consultations, and direct comparison of these different types of consultation from the same practice setting have not been performed. In addition, virtually all published data derive from the institution receiving the consultation; there are almost no data from institutions that send out these consultations. Whether the data from these institutions are similar to published data is not known, although it is always easier to find error in someone else's performance than in your own. Finally, diagnostic disagreement is not the same as error. Unfortunately, few of the previous studies have examined this issue, and most simply assume that the consulting opinion is correct, even when the reported clinical findings suggest otherwise. 6 Correlation of the results of consultation with additional information, including additional consultation, tissue sampling, and testing on cases with disagreement is limited. To help address these issues, we reviewed the results of interdepartmental consultation from our department for a 2-year period.