Article Title

Classification and Mapping of Emergency Department Use in Three San Francisco Bay Area Hospitals

Publication Date



emergency department overcrowding, inappropriate utilization


Background/Aims: Preventable or nonurgent encounters to the emergency department (ED) disrupt patient flow and contribute to the burden of overcrowding. Furthermore, overcrowding is expected to increase as insurance coverage expands. We sought to create two categorization schema for nonurgent and/or preventable encounters to the ED and to visualize the resulting patient segments geographically in order to identify gaps in health care delivery services.

Methods: From electronic health records we extracted demographics, payer type and diagnoses for all encounters from June 2013 to May 2014 at three Sutter Health EDs in Alameda County, California. Based on primary diagnosis, we developed an algorithm to classify each encounter as urgent/nonurgent, and preventable/nonpreventable, collectively defined as inappropriate. We geocoded patient addresses and the addresses of all federally qualified health centers (FQHCs) in the area.

Results: During the study period, 69,595 individuals made 111,535 encounters to the study EDs; 36% (n = 25,349) of individuals had at least one nonurgent encounter, whereas only 19% (n = 13,132) had one or more preventable encounters. Overall, approximately half of all patients had at least one type of inappropriate encounter. Demographics differed between those who used the ED inappropriately and those who did not. Nonurgent users were younger and those with preventable encounters were older compared to urgent and nonpreventable users, respectively. The distribution of minorities differed with more African-Americans having inappropriate encounters of both types. The proportion of uninsured did not vary by type of user (n = 57,290 [82%] of all patients had some type of insurance). We identified that 49% (23,535 of 48,156) of inappropriate encounters were made by individuals who lived within one mile of an FQHC.

Conclusion: Using the methods described, we found that individuals who used the ED for nonurgent and/or preventable conditions tended to have some type of insurance coverage and were disproportionately African-American. By mapping the encounters and FQHCs, we were able to define catchment areas for each hospital and estimated that half of all inappropriate encounters could have been diverted to nearby, more suitable resources. Further identification and mapping of alternative health care and community resources will likely show that even more of these patients have alternative options available to them.




June 30th, 2016


August 12th, 2016