Abstract / Summary
Abstract Background Emergency department (ED) utilization continues to rise globally, and a disproportionate share of visits is attributable to a small subgroup of patients. This subgroup is referred to throughout the present report by the single term high-frequency users (HFUs), defined as patients with four or more ED visits within a 12-month period. Characterizing the clinical profile, utilization patterns and cost structure of HFUs is relevant to emergency care planning and resource allocation. Methods This retrospective, single-center observational study analyzed all adult ED presentations at a high-volume urban tertiary hospital over a 12-month period. HFU status was defined at the patient level as four or more ED visits within the study period. The unit of analysis was specified for every comparison: age, sex, the ICD-10 diagnostic spectrum and all logistic regression models were analyzed at the patient level (one observation per unique patient), whereas triage category, timing of presentation, mode of arrival, resource utilization and per-visit costs were analyzed at the visit level. For the regression models, visit attributes and diagnostic categories were aggregated to the patient level as ever/history indicators. Groups were compared using the Mann–Whitney U and Pearson chi-square tests, with effect sizes reported alongside p values. Results Among 466,884 unique adult patients accounting for 726,696 ED presentations, 26,871 patients (5.8%) were HFUs. HFUs represented 5.8% of adult patients but accounted for 148,322 of 726,696 presentations (20.4%) and approximately one-fifth (21.0%) of cumulative laboratory, imaging and medication costs. Total cost per visit was similar between groups (median 6.39 USD in both; p = 0.647), and differences in individual cost components, although statistically significant, were small in practical terms (rank-biserial correlations +0.03 for laboratory, −0.05 for radiology and +0.08 for medication cost). A higher proportion of HFU than non-HFU presentations involved any laboratory test (29.3% vs 26.7%) and any medication order (57.8% vs 51.4%), whereas any radiologic imaging was less frequent among HFUs (40.5% vs 44.7%). HFUs were marginally younger, presented more often during the night shift and were less likely to arrive by ambulance (3.5% vs 5.0%). Yellow-triage presentations predominated among HFUs, and symptom-based and pain-related diagnoses were over-represented. In the patient-level multivariable model, male sex, a history of yellow- or red-triage presentation, a history of evening- or night-shift presentation and multiple symptom-based diagnostic groups were associated with HFU status, whereas a history of ambulance arrival was associated with lower odds. Conclusions HFUs constitute a small proportion of adult ED patients but account for a disproportionate share of presentations and of cumulative diagnostic and pharmacologic resource use. Because total cost per visit was similar between the groups and the differences in individual cost components were small in practical terms, the principal finding is the cumulative resource burden associated with repeated attendance rather than a higher cost of the individual encounter. This pattern is consistent with a contribution of out-of-hours access and symptom-driven presentation; however, the present observational data cannot establish these mechanisms or identify which interventions would reduce frequent ED use.