Patient demographic and health factors associated with frequent use of emergency medical services in a midsized city

Amy Knowlton, Brian W. Weir, Brenna S. Hughes, R. H.J.Hunter Southerland, Cody W. Schultz, Ravi Sarpatwari, Lawrence Wissow, Jonathan Links, Julie Fields, Junette McWilliams, Wade Gaasch

Research output: Contribution to journalArticlepeer-review


Objectives: To optimize health care utilization, health outcomes, and costs, research is needed to improve an understanding of frequent users of emergency health services. Frequent use of emergency services is associated with high costs of health care and may be indicative of challenges accessing, or poor outcomes of, health care. Patient demographics and health factors related to frequent use of the emergency medical services (EMS) system of a midsized city were identified. Study findings will aid in the development of targeted interventions to improve population health. Methods: The authors reviewed 9-1-1 call dispatch data and Baltimore City Fire Department (BCFD) EMS records from 2008 through 2010. Frequent use was defined as six or more EMS incidents in the 23-month period. Analyses used census data to compare demographics of EMS users to their population distribution and examined differences in demographics and health problems of frequent EMS users compared to nonfrequent users. Results: Frequent EMS users (n = 1,969) had a range of six to 199 EMS incidents (mean = 11.2) during the observation period, and although they accounted for only 1.5% of EMS users, they were involved in 12.0% of incidents. Frequent users, compared to nonfrequent users and to the population, were more likely to be male, African American, and 45 years of age or older. Of frequent users, the modal age group was 45 to 54 years, accounting for 29.7% of frequent users, which represented twice this age group's population distribution. Furthermore, this age group had the greatest overrepresentation of males (63.0% of frequent users) and was the peak age group for incidents related to substance abuse (28.0% of frequent users' incidents in this age group). Frequent users, compared to nonfrequent users, had lower levels of incidents related to trauma (5.1% vs. 16.7%) and higher levels of medical incidents (94.8% vs. 82.9%). As proportions of EMS incidents among frequent versus nonfrequent users, respiratory, mental health, and seizure-related incidents were highest in the youngest age groups; substance abuse-related incidents were highest in those middle-aged (35 to 44 and 45 to 54 years). Of health problems, behavioral health (mental health or substance use) contributed most to frequent EMS use (23.4% of frequent users' incidents). Across all incidents, 65.8% of frequent users had indications of behavioral health problems, representing 6.6-fold higher odds than nonfrequent users (22.5%). Frequent compared to nonfrequent users also had higher levels of select chronic conditions (diabetes, 39.9% vs. 14.6%; asthma, 40.9% vs. 13.4%; and HIV, 9.1% vs. 2.4%), with unadjusted odds almost four to seven times higher. Conclusions: The study findings revealed the major role of chronic somatic and behavioral health problems in frequent EMS use and that rates of frequent use were highest among those middle-aged, African American, and male. These results suggest the need for coordination of EMS with community-based, integrated medical and behavioral health services to improve access and use of preventive services, with implications for health outcomes and costs. This study demonstrates the value of EMS patient data in identifying at-risk populations and informing novel, targeted approaches to public health interventions.

Original languageEnglish (US)
Pages (from-to)1101-1111
Number of pages11
JournalAcademic Emergency Medicine
Issue number11
StatePublished - Nov 2013

ASJC Scopus subject areas

  • Emergency Medicine

Fingerprint Dive into the research topics of 'Patient demographic and health factors associated with frequent use of emergency medical services in a midsized city'. Together they form a unique fingerprint.

Cite this