Analysis of Costs, Length of Stay, and Utilization of Emergency Department Services By Frequent Users : Implications for Health Policy

In efforts to decrease emergency department (ED) crowding and health care costs, frequent users of ED services have been targeted for interventions to decrease their utilization. Previous studies have had different definitions for frequent users and have considered all frequent users as a homogeneous group. To the authors' knowledge, no study has examined visit characteristics and resource utilization of different levels of frequent use. Objectives: 1) to determine the rates of ED utilization by five user groups defined by number of annual visits, 2) to examine variations in visit characteristics by frequency of ED use, and 3) to compare levels of resource utilization among frequent user groups. Methods: This was a retrospective, cross-sectional study of clinical and financial records for all ED visits to an urban, academic hospital in 2001. Multinomial logistic and linear regression models were used for analyses. Estimates were corrected for multiple comparisons (with Bonferroni corrections), where applicable, and adjusted for clustering within individuals (with Huber-White estimators). Outcome measures were triage acuity, diagnosis-related group (DRG) severity, disposition status, primary complaint, medical diagnosis, hospital inpatient length of stay, payment method, costs, and demographics. Results: Patients with three to 20 visits were more likely to be admitted to the hospital than patients visiting once or twice. Patients visiting more than 20 times were less likely to require hospital admission and more likely to present with non-urgent conditions, have lower severity scores, and elope or leave the ED without medical attention than patients visiting the ED once. The group had fewer inpatient days and lower average costs than patients visiting once. Patients with six to 20 visits had traditional Medicaid coverage more often than those with one or two visits. Virtually no patients visiting more than 20 times had Medicare or Medicaid managed care, a health maintenance organization, or a preferred provider organization. Conclusions: Frequent ED users are a heterogeneous group. Many patients previously thought to overutilize the ED for socioeconomic or insignificant medical problems are as sick as less-frequent ED users. There is a small subgroup with more than 20 visits who are less ill or injured but also incurred lower-than-average costs per visit