''Inappropriate'' emergency department use: A comparison of three methodologies for identification

Objective: To determine the level of agreement between the rates of ''inappropriate'' ED visits assigned to a cohort of ambulatory patients based on three methods of defining ED use appropriateness. Methods: Ambulatory adult patients seen at one urban, university-based teaching hospital ED between 8 AM and midnight during select days from April to June 1994 were assessed regarding the appropriateness of their ED visits. Patients triaged to acute resuscitation rooms in the ED were excluded. Eligible patients were asked to complete a 90-question survey including demographics and health service use (response rate 81%), The appropriateness of ED use was assessed for consenting respondents by 1) application of a list of 51 nonemergent complaints that have been used by managed care providers and previously published (TRIAGE), 2) use of ten explicit criteria (e.g., need for parenteral medication) from prior publications (EXPLICIT), and 3) the consensus of two emergency physicians (EPs) reviewing the records of ED patients (PHYS). All three methods were applied at the time of retrospective chart review. The agreement between methods was evaluated using kappa scores. Results: Of the 892 eligible respondents, 64% were white, 54% were employed, 50% were female, and 29% were uninsured. Of the respondents, 26% had no regular source of ambulatory care and 25% considered the ED their regular source of care. The assigned rates of ''inappropriate'' visits using the three definitions were TRIAGE, 58%; PHYS, 47%; and EXPLICIT, 42%. Of those deemed ''inappropriate'' by the EXPLICIT criteria, 81% also were judged as ''inappropriate'' by the TRIAGE criteria, and 72%, by the PHYS criteria. Of those patients deemed ''inappropriate'' by the TRIAGE criteria, 59% also were judged as ''inappropriate'' by the EXPLICIT criteria, and 66%, by the PHYS criteria. Levels of agreement (kappas) were TRIAGE/EXPLICIT, 0.39; TRIAGE/PHYS, 0.42; and EXPLICIT/PHYS, 0.42. Conclusion: There is only moderate agreement between different methods of determining appropriateness of ED use. Until further refinement is made in triage assessment, managed care organizations and EPs should remain cautious when implementing a protocol that defines and restricts ''inappropriate'' ED visits.