# Formative Assessment and Cluster Membership of Homeless Persons’ Emergency Department Visits

## TL;DR

This study explores how homeless individuals' emergency department visits compare to others, aiming to improve healthcare for them through targeted interventions.

## Contribution

The paper introduces a cluster analysis framework to identify patterns in ED visits among homeless individuals and other groups.

## Key findings

- Homeless individuals with comorbidities and substance abuse had higher rates of recurrent ED visits.
- Cluster analysis predicted residence type and 72-hour ED recurrence based on health and social factors.
- Substance abuse and depression were significant predictors of ED revisit patterns.

## Abstract

Background

Evidence is limited on whether homeless individuals who visit emergency departments (EDs) share similar clinical characteristics as individuals from private households who, upon examination, require emergency health services beyond preventative healthcare. While the literature is rich with studies on homeless persons in the ED, a comprehensive assessment of similarities and differences with other social groups is lacking. Consequently, there is a gap in knowledge as it pertains to the appropriate approaches that will further support the development of targeted healthcare and emergency health services for the homeless person.

Aim

The goal is to provide a framework for targeted interventions that hospitals can develop based on behavioral, health, and social characteristics and, in the process, better meet the healthcare needs of homeless patients.

Methods

Data were collected from the records of 85,350 patients in the National Hospital Ambulatory Medical Care Survey (NHAMCS) database who visited the ED between the years of 2013 and 2016; each patient was classified as having either a private residence, living in a nursing home, or homeless (n = 83,446, n = 1,459, and n = 925, respectively).

Results

Cluster analysis of the data confirmed that individuals with multiple comorbidities, as well as substance abuse or dependency, regardless of residence type, were more likely to have recurrent ED visits within 72 hours. Nominal regression analyses revealed that cluster membership generated from ED data could predict patient residence and suggest that substance abuse and depression can predict 72-hour ED visit recurrence.

Conclusion

Cluster analyses have the potential to reveal social health and group characteristics and can support targeted solutions respective to group individualities in the ED.