Increased long-term mortality in patients with repeated visits to the emergency department

Eur J Emerg Med. 2010 Oct;17(5):274-9. doi: 10.1097/MEJ.0b013e3283104106.

Abstract

Background: Revisits to the emergency department (ED) are common. It is not clear whether the number of revisits, the time between revisits or the reason for the revisits are associated with increased mortality.

Methods: During 1 year, the number of visits, the reason for the visit and the time between visits were recorded in 15 607 nonsurgical ED patients and related to 1-year and 5-year mortality.

Results: Five-year mortality was dependent on the number of revisits in an inverse U-shaped manner. When compared with one-time visitors, patients with three visits showed an increased 5-year mortality (hazard ratio 1.85, 95% confidence interval 1.58-2.16, P<0.0001), whereas in patients with four or five visits mortality decreased. Patients with six or more visits had a 5-year mortality not different from one-time visitors. The impact of the number of visits was, however, dependent on the presenting complaint (P<0.0001). Furthermore, the time between two adjacent visits influenced long-term mortality in an inverse U-shaped manner. In patients not admitted to the ward, a revisit after 2-3 days was associated with increased mortality (hazard ratio 1.89, 95% confidence interval 1.06-3.35, P=0.03). In patients revisiting the ED with the same adjacent presenting complaint, mortality differed depending on the complaint (P<0.0001).

Conclusion: In nonsurgical patients revisiting the ED, long-term mortality was dependent on both the number of revisits, as well as the time between two visits in an inverse U-shaped manner. This indicates a possibility of detecting the transition level between appropriate medical utilization and inappropriate frequent ED use.

MeSH terms

  • Analysis of Variance
  • Confidence Intervals
  • Emergency Service, Hospital / statistics & numerical data*
  • Health Services Needs and Demand / statistics & numerical data*
  • Humans
  • Kaplan-Meier Estimate
  • Mortality / trends*
  • Proportional Hazards Models
  • Retreatment / statistics & numerical data
  • Statistics, Nonparametric
  • Sweden
  • Time Factors