Judging who needs emergency department care: a prerequisite for policy-making.

The debate regarding risks and benefits of triaging nonurgent patients away from emergency departments (EDs) stems from widely varying estimates as to what proportion of ED visits are inappropriate. A study was undertaken based on the hypothesis that these discrepant estimates might be due to differences in how "appropriateness" is defined. This cross-sectional study included 596 ED patients. Seven different indicators of "Inappropriate" ED visits were used. Two could be determined by the patient; two were based on the triage nurse's assessment; three were determined retrospectively, by chart review. All 21 possible pairs of indicators were compared for agreement using the kappa statistic. The proportion of ED visits classified as inappropriate by the different indicators ranged from 10% to 90%. Kappa values for agreement between indicators ranged from -0.04 to 0.31, indicating poor agreement beyond that expected due to chance alone. Decisions as to which ED visits are appropriate depend heavily on the criteria used. Limiting patients' access to EDs without the aid of a valid and reliable standard for what constitutes an appropriate ED visit could create harmful barriers to care.