Homeless status documentation at a metropolitan hospital emergency department

Abstract

Objective: This study compared the prevalence of homelessness in consecutive patients presenting to a metropolitan hospital ED measured via a prospective housing screen with the prevalence of homelessness determined via retrospective audit of hospital data. Factors that altered the odds of patients being homeless and service outcomes that differed were examined for screened patients. Methods: All patients presenting to the ED during a 7 day period in 2017 were invited to complete a housing screen. A retrospective audit of all ED presentations during the same period also occurred. Demographic (e.g. age, gender), clinical (e.g. reason for presentation, ED presentation history) and arrival mode (e.g. time, how arrived) predictors of homeless status were examined alongside care outcomes (e.g. ED length of stay, admission and 28 day re-presentation). Results: Of 1208 presenting patients, 504 were prospectively screened and 7.9% were homeless. This compared with 0.8% of ED presentations coded as homeless in the Victorian Emergency Minimum Dataset and 2.3% of the 704 non-screened patients identified as homeless using Victorian Emergency Minimum Dataset Usual Accommodation alongside primary diagnosis and registration address. Within the screened sample, homeless patients were more likely to be male, arrive by emergency ambulance/with police, have a psychosocial diagnosis, and be frequent presenters. Representation within 28 days occurred for 43% of homeless and 15% of not-homeless patients. Conclusions: Hospital ED administrative data substantially under-recognises the prevalence of homelessness in presenting patients. Standardised use of brief housing screens could improve identification of and provision of support to this often highly vulnerable population.

Comparison of the Victorian Emergency Minimum Dataset to medical records for emergency presentations for acute cardiovascular conditions and unspecified chest pain

“…The item usual living arrangements was one the most discordant item, particularly for patients who reside in a private residence. A previous report also which found issues with the accuracy of the coding of this VEMD item, and found homelessness was underreported in the VEMD . Researchers and policy makers need to be aware that this VEMD item may not be accurate in its current form, but can be used to identify patients who reside at home with excellent accuracy by combining the two private residency codes.…”

“…Examples of recent published studies using the VEMD data include: exploring trends in ED presentations; examining the impact of major events and public awareness campaigns; and examining ED readmission rates . However, to date there is limited published research describing the accuracy of the VEMD data, and only one study was found that used the medical record as the ‘reference standard’ …”

“…5 Examples of recent published studies using the VEMD data include: exploring trends in ED presentations; [6][7][8][9] examining the impact of major events 10 and public awareness campaigns; 11 and examining ED readmission rates. 12,13 However, to date there is limited published research describing the accuracy of the VEMD data, [14][15][16] and only one study was found that used the medical record as the 'reference standard'. 14 Validation studies, comparing the accuracy of administrative datasets to the original source (ideally prospectively collected data at the time of the healthcare encounter), are necessary to verify that data are coded accurately and that observed trends are real.…”

“…12,13 However, to date there is limited published research describing the accuracy of the VEMD data, [14][15][16] and only one study was found that used the medical record as the 'reference standard'. 14 Validation studies, comparing the accuracy of administrative datasets to the original source (ideally prospectively collected data at the time of the healthcare encounter), are necessary to verify that data are coded accurately and that observed trends are real. For example, studies examining patients arriving to hospital by ambulance require accurate entry of the arrival mode variable in the VEMD.…”

COVID‐19 and emergency care for adults experiencing homelessness

“…Most existing healthcare systems are poor at identifying homeless patients and underestimate the prevalence of homelessness in the ED. 5 Underrecognition is exacerbated in emergency settings where homelessness is frequently perceived to be synonymous with 'rough sleeping', despite rough sleepers accounting for only 7% of Australia's homeless population. 1 In the midst of a global pandemic, the urgency to identify people who are homeless is magnified, both because homelessness poses inherent risks that exacerbate potential exposure to infection, but also because there are widespread implications for public health, including community transmission, contact tracing and the provision of isolation and quarantine services.…”