Longitudinal trajectories and persistence of frequent emergency department use: A two-year individual-level cohort study

INTRODUCTION: Whether frequent emergency department (ED) use (≥4 ED visits per calendar year) reflects a persistent patient phenotype or predominantly transient episodes remains unknown in France. This study characterised longitudinal trajectories and predictors of persistence, lack of primary care affiliation and very frequent use (VFU). METHODS: Two-year retrospective cohort study at a French tertiary-care university hospital (2023-2024). Individual patient-level linkage classified patients as persistent, incident, or remitting frequent users. Forced-entry multivariable logistic regression identified predictors of persistence, lack of primary care affiliation, and VFU (≥6 visits/year). RESULTS: Linkage identified 196 persistent users (16.7% of 2023 frequent ED users), 982 incident, and 976 remitting. Persistent users were younger (49.3 vs 55.4 years; p = 0.001) and more often in the ≥6 visits/year stratum (51.5%vs 24.9%; p < 0.001). Persistence increased steeply with index-year visit intensity (adjusted OR 8.78 for ≥10 vs 4 visits/year). Repeat attendance was mostly for differing reasons (identical chief complaint at every visit: 1.9%). Mental-health, haematological, and addiction-related presentations were over-represented among VFU. Lack of primary care affiliation was predicted by male sex (aOR = 2.25) and younger age (aOR = 0.97/year), and was higher among out-of-catchment patients (28.1%vs 13.5%; p = 0.002). CONCLUSIONS: Frequent ED use was predominantly transient, with only 16.7% of patients remaining frequent users across consecutive years; persistent users were younger and concentrated in the highest-use stratum. Lack of primary care affiliation reflected sociodemographic factors (male sex, younger age). These findings identify persistent users as a priority target for future coordinated interventions.