A 'snap shot' of the health of homeless people in inner Sydney: St Vincent's Hospital

Abstract

Objectives. The poor health profile of people who are homeless results in a disproportionate use of health resources by these people. An in-hospital count of demographic and health data of homeless patients was conducted on two occasions at St Vincent's Hospital in Sydney as an indicator of health resource utilisation for the Sydney region.

Methods. Two in-hospital counts were conducted of homeless patients within the boundaries of St Vincent's Hospital to coincide with the inaugural City of Sydney homeless street counts in winter 2008 and summer 2009. Data collected included level of homelessness, principal diagnosis, triage category, bed occupancy and linkages to services post hospital discharge.

Results. Homeless patients at St Vincent's utilised over four times the number of acute ward beds when compared with the state average. This corresponds to a high burden of mental health, substance use and physical health comorbidities in homeless people. There was high utilisation of mental health and drug and alcohol services by homeless people, and high levels of linkages with these services post-discharge. There were relatively low rates of linkage with general practitioner and ambulatory care services.

Conclusion. Increasing knowledge of the health needs of the homeless community will assist in future planning and allocation of health services.

What is known about the topic? The poor health status of people who are homeless has been previously noted in the USA, Canada and Scotland.

What does this paper add? Homeless people living in Sydney also have a poor health profile and a disproportionate use of health resources when compared to people in the general population.

Whatare the implications for practitioners? Health services for homeless people should be equipped to deal with mental health, substance use and physical health comorbidities.

The numbers of homeless people in cities appear to be increasing, according to the Australian Bureau of Statistics (ABS) Census data.1 These data also reflect a higher density of homelessness in the City of Sydney local government area (LGA) than for New South Wales (NSW) in general.2 The City of Sydney Council has initiated biannual street counts to monitor the numbers of 'rough sleepers' and those people in crisis accommodation in the Sydney LGA. Similar initiatives are occurring internationally in cities such as Vancouver and London. Two street counts have been conducted in Sydney: the first winter count was in August 2008 and first summer count in February 2009.

Homelessness is commonly stratified into three levels: primary, secondary and tertiary. Primary homelessness refers to any individual without some form of shelter, and is also known as 'sleeping rough'. Secondary homelessness refers to those individuals who reside in temporary and unsecured accommodation, such as in hostels or at the goodwill of friends or relatives. Longterm residents of private boarding houses, without a secure rental lease, fall into the category of tertiary homelessness.3

The correlation between homelessness and poor health status has been previously noted in the USA, Canada and Scotland.4-8 Homeless people suffer from a diverse range of physical and mental health problems. Disease is more likely to be detected at an advanced stage rather than at a primary or preventable stage.9 Diabetes and hypertension may go undiagnosed and untreated for many years, by which stage secondary complications may have manifested themselves.10 Hwang found that amongst men accessing homeless refuges in Toronto, mortality rates were 8.3 times higher than the mean for 18-24-year-olds. Mortality rates amongst youth sleeping rough in Montreal were also unacceptably high, at nine times higher for males and 31 times higher for females. Musculoskeletal disorders, respiratory tract infections, skin infections, and poor oral health are prevalent within the homeless community.6 Mental health issues, increased rates of substance use, and injury due to violence and self-harm, are also well documented in this population.8

The Australian healthcare system has been reported to be in crisis over the last few years.11 The number of acute-care hospital beds decreased by 14% between 1992 and 2002, despite increasing admission rates.12 These shortages have given way to access block and overwork in Emergency Departments (EDs), factors that are associated with ambulance diversion,13 increased length of stay14 and increased mortality.15 The NSW Department of Health has implemented an 'Avoidable Admissions' strategy to ameliorate this situation. This acknowledges that it may be more cost effective and preferable to treat certain low-complexity medical conditions, including cellulitis, community-acquired pneumonia and deep vein thrombosis in an ambulatory setting. Eight of these identified conditions were listed amongst the top 20 diagnoses for patients admitted to NSW public hospitals. This suggests that potentially avoidable admissions are frequent. It has been estimated that there have been 25 000 potentially avoidable admissions during the 2003-04 period alone.16 Several of these conditions, including cellulitis and community-acquired pneumonia, are common amongst homeless patients presenting to the ED.6 The propensity for this vulnerable population to present frequently to EDs and their utilisation of 'a disproportionate amount of health resources' compared with the general population has been well described.4,17-21 This occurs in the context of a high level of mental health, substance use and physical health comorbidities. In addition, it is recognised that a small number of frequent ED users account for a high percentage of all ED presentations.18 This pattern is reflected in the experience of the ED at St Vincent's Hospital in Darlinghurst, Sydney, where the most frequent presentations are by homeless patients.22 The majority of conditions presented by this population are uncomplicated and potentially manageable in an ambulatory care setting.23

Housing insecurity has been shown to be a factor associated with poor access to outpatient care and subsequently high rates of acute care in other countries.24 A recent randomised control trial conducted in Chicago aimed to assess the impact of a combined case management and housing program in decreasing emergency medical service use amongst homeless people. The study concluded that such a program was useful in decreasing hospitalisation and ED visits, with a 29% decrease in hospital days and a 24% decrease in ED visits.25 This suggests that integrated and novel approaches need to be explored in the provision of healthcare to the homeless. The recently released Government Homeless White Paper expressed a need for further research into this area to determine how best to approach this challenging issue.26 An in-hospital count of homeless patients at St Vincent's Hospital was conducted on two occasions to gain demographic and health data and to ascertain the level of health resource utilisation to inform future health services.

Method

Data collection

The City of Sydney homeless street counts took place in August 2008 and February 2009. An inpatient hospital count was conducted of homeless patients aged 16 years or over within St Vincent's Hospital to coincide with each street count. The areas covered included all inpatient wards at St Vincent's Hospital, the Sacred Heart Hospice rehabilitation and palliative care wards, psychiatric ward, the substance-use detoxification unit and the emergency department. All wards were informed of the purpose of the counts before the day.

The counts were performed by visiting each ward and speaking with the Nurse Unit Manager or nursing staff to identify any patients who were homeless or at risk of being homeless. The medical record numbers of identified patients were noted. Patients were also assigned a unique identifier to preserve confidentiality. The number of identified patients in each ward was noted, as was the total bed occupancy. The files were reviewed retrospectively for demographic and health profile data. Patients meeting the criteria for homelessness1 were further stratified into primary, secondary or tertiary levels of homelessness. Principal diagnoses were grouped into broad diagnostic categories of medical, surgical, mental health and substance-use issues based on the Australian Refined Diagnosis Related Groups (ARDRGs) 27 where available, or principal diagnosis identified on the discharge summary pertaining to the episode of care concerned. Other data collected included triage categories as per the Australasian Triage Score,28 mode of arrival to hospital, level of bed occupancy within the hospital and linkages to services following discharge from hospital.

Identification of homeless persons

Accommodation status was obtained from the patient medical record, using the address information supplied on the specified episode of care. Patients who were identified as having 'no fixed address' or having an address that corresponded to crisis accommodation such as a hostel, intoxicated persons unit or boarding house without a secure lease, were classified as being homeless for the purposes of the study. If there were uncertainties regarding accommodation status, the medical record was examined in greater detail with a focus on entries made by hospital social workers regarding classification during that episode of care.

Analysis

Data were analysed usingSPSSV17.0. Descriptive statistics were obtained on demographic and other variables of interest. Before combining data from two periods, comparisons between periods were carried out. For continuous variables comparisons were conducted using Student's t-test, and Chi-square tests were used for categorical variables. When the expected value in any cell was less than 5, the Fisher's Exact Test was used. A 5% significance level was used for all tests of hypothesis.

Results

No significant differences were found between the patients in the two counts (n1 = 29, n2 = 32). Data from the two counts were combined, given the small sample size. One file from 2008 was unavailable for audit and excluded from the study.

Demographic data

In the 2008 inpatient count, 47 patients were identified by nursing staff as being homeless or at risk ofhomelessness.Of these, 29were found to be homeless (62%). In the 2009 count, 48 patients were identifiedbyward staff as being homeless or disadvantaged, and, of these, 32 satisfied the criteria for being homeless (66%). The age range was from 16 to 70 years, and the mean age was 42.4 years. Males were over-represented, totalling 75% of homeless patients counted. Of 61 patients, 27 met the criteria for primary homelessness (44%), 22 were classed as being secondary homeless (39%) and 10 were found to be tertiary homeless (16%) (Table 1).

Comparison with the City of Sydney street count

The proportion of homeless people who were in hospital was 3.7% of the total number of homeless persons counted from the street count of rough sleepers and Supportive Accommodation Assistance Program (SAAP) occupied beds in the Sydney LGA.

Health profile

Presentations with a diagnosis of mental health or substance-use issues were by far the most common, accounting for 84%. Most homeless hospital attendees either self-presented or were transported by ambulance services. The most common triage category for patients presenting toEDswas 3(41%of attendees). However, 39% of inpatients attended hospital via a non-emergency route and most of these had self-presented to the detoxification unit for substance-use issues. The most common mode of arrival was by ambulance (34%) and self-presentation (33%) (Fig. 1).

Health resource utilisation

The proportion of hospital beds occupied by homeless persons compared with the bed occupancy was calculated as an indicator of health resource utilisation. Services with the highest percentage of resource utilisation by homeless persons included the mental health unit (56%) and the drug and alcohol unit (52%). Inpatient beds occupied by homeless people at St Vincent's Hospital during both counts totalled 8% (Table 2).

Linkages after discharge

Most patients (95%) were linked to at least one service upon discharge. These included mental health services, drug and alcohol services, ambulatory care (defined as a healthcare visit in a non-inpatient setting) or a general practitioner. The most common linkages found were to mental health (57%), nongovernment organisations (56%), drug and alcohol (30%) and other (34%), which included pastoral care and community palliative care. Several patients (18%) were transferred to ongoing care at another facility. Of note, only 23% of homeless patients were linked to a GP or primary-care service upon discharge, and 12% of patients were linked to an ambulatory service for ongoing care.

Discussion

A disproportionate number of homeless people using hospital resources was found in an inner city public hospital when compared with the average number of housed people of Sydney. There was an 8% cumulative level of bed occupancy over the two inpatient counts. There are ~2.9 public beds per 1000 population in NSW.29 From the data obtained through the counts at St Vincent's Hospital, the equivalent of 13 acute ward beds per 1000 population was being utilised by the homeless. This is over four times the expected number of beds per 1000 population for NSW. Although this figure may be extrapolated only to other health services also servicing high numbers of homeless patients, it is still useful for demonstrating the marked disparity between the housed and homeless people in terms of health resource utilisation. This is not surprising given the poor health profile of homeless people and their high levels of mental health, substance use and physical health comorbidities.

The average triage category was 3 (urgent), which indicates that most presentations to the ED were appropriate, as opposed to Categories 4 (semi-urgent) or 5 (non-urgent), which might be more appropriate for management by a general practice. Most Category 3 presentations were for mental health (64%). However, 39% of patients presented via a non-ED pathway, and most of these were for drug and alcohol issues. As anticipated, there was high utilisation of mental health and drug and alcohol services by homeless persons, indicating that there is a need for ongoing resource allocation in these areas.

There were high levels of linkages to post-discharge services, particularly to mental health and drug and alcohol services, which was encouraging. However, relatively poor rates of linkage to general practitioner and ambulatory care services were noted. This finding suggests that if homeless patients had improved linkages to general practice and outpatient services they might seek these out in preference to the ED when requiring healthcare. However, there may be other factors that influence whether patients present to the ED, such as economic reasons and needing to access healthcare outside of normal business hours. These questions are outside the scope of this study.

Study limitations

This study involved several limitations. St Vincent's Hospital is subject to a relatively high level of homeless presentations due to its mission of caring for the disadvantaged, its location and its accessibility for the homeless people of Sydney. We also excluded the homeless inpatients at Sydney Hospital, which is also subject to high utilisation by the homeless people of inner Sydney.

Although the SAAP crisis bed data from the City of Sydney street count is reliable, it is likely that the numbers obtained from the street counts of rough sleepers in inner Sydney were an underestimation. Only people bedded down in a set number of local government areas within a short period were counted, resulting in a minimum estimation of rough sleeping, rather than the full extent. This is the same methodology used for conducting street counts internationally in cities including London and Toronto, which has been criticised for lacking sufficient rigour. 30,31 In addition, this study does not include the numbers of 'hidden homeless': people staying with friends and relatives, or in temporary hotel or motel accommodation. Despite these limitations, these data are still the most accurate that we could obtain, due to the logistical difficulties of collecting reliable data on the homeless who often rely on invisibility and transience in order to survive.

Conclusions

This descriptive study reflects the disproportionate number of homeless people using acute public hospital services compared with the average housed population of Sydney. The main reasons for presentation to the ED were mental health and drug and alcohol related. This is consistent with current literature and indicates that further resources are needed in this area.

The current economic climate may continue to contribute to homelessness in the foreseeable future. Given the poor health status of this population, targeted interventions and ongoing research are required to address the marked health inequalities between the homeless and the housed. There is growing interest in integrated health management approaches, such as offering health services in combination with supported housing.

The relatively poor levels of linkages to ambulatory care and general practice services upon discharge were noted. This is an area with potential for improvement. One question that arose from this was whether a doctor and nurse team acting as a general practice service for the homeless would assist in decreasing ED presentations. We anticipate that this will form the basis for the next part of our study.

Acknowledgements

This research was supported by community donations to the Urban Health Trust fund at St Vincent's Hospital. Caroline Chin's work was supported by the University of Notre Dame, Sydney. We sincerely thank Elizabeth Giles and her team from the Homelessness Unit, City of Sydney Council, and Adrienne Lucey, Area Coordinator for Homelessness Health, South Eastern Sydney and Illawarra Health for their assistance. We are also indebted to Associate Professor Lawrence Lam (Biostatistics and Public Health) at the University of Notre Dame for his continued advice and support.

[Reference]

References

1 Census of Population and Housing. ABS; 2009. Available at http://www. abs.gov.au/websitedbs/D3310114.nsf/Home/census?opendocument? utm_id=GT [verified June 2009].

2 Homelessness Services. City of Sydney; 2009. Available at http:// www.cityofsydney.nsw.gov.au/Community/HomelessnessServices/ HomelessPersonsInformationCentre.asp [verified April 2009].

3 Chamberlain C, MacKenzie D. Understanding contemporary homelessness: issues of definition and meaning. Aust J Soc Issues 1992; 4(27): 274-97.

4 Forchuk C, Brown SA, Schofield R, Jensen E. Perceptions of health and health service utilization among homeless and housed psychiatric consumer/ survivors. J Psychiatr Ment Health Nurs 2008; 15(5): 399-407. doi:10.1111/j.1365-2850.2007.01246.x

5 Ensign J, Santelli J. Shelter-based homeless youth - health and access to care. Arch Pediatr Adolesc Med 1997; 151(8): 817-23.

6 Hwang SW. Homelessness and health. CMAJ 2001; 164(2): 229-33.

7 Hwang SW, Tolomiczenko G, Kouyoumdjian FG, Garner RE. Interventions to improve the health of the homeless: a systematic review. Am J Prev Med 2005; 29(4): 311-9. doi:10.1016/j.amepre.2005.06.017

8 Kershaw A, Singleton N, Meltzer H. Survey of the health and well-being of homeless people in Glasgow. Int Rev Psychiatry 2003; 15(1-2): 141-3. doi:10.1080/0954026021000046065

9 Power R, French R, Connelly J, George S, Hawes D, Hinton T, Klee H, RobinsonD, Senior J, Timms P, Warner D. Health, health promotion, and homelessness. BMJ 1999; 318: 590-2.

10 Hwang SW, Bugera AL. Barriers to appropriate diabetes management among homeless people in Toronto. CMAJ 2000; 163(2): 161-5.

11 Crisis in Crowded Emergency Departments. Australian Medical Association; 2007. Available at http://www.ama.com.au/node/2913 [verified April 2009].

12 Braitberg G. Emergency department overcrowding: dying to get in? Med J Aust 2007; 187: 11-2.

13 Fatovich DM, Nagree Y, Sprivulis P. Access block causes emergency department overcrowding and ambulance diversion in Perth, Western Australia. Emerg Med J 2005; 22: 351-4. doi:10.1136/emj.2004.018002

14 Richardson DB. The access-block effect: relationship between delay to reaching an inpatient bed and inpatient length of stay. Med J Aust 2002; 177: 492-5.

15 Richardson DB. Increase in patient mortality at 10 days associated with emergency department overcrowding. Med J Aust 2006; 184: 213-6.

16 Iemma M. A new direction for NSW. State Plan. Premiers Department; 2006. Available at www.stateplan.nsw.gov.au [verified August 2009].

17 Kessell ER, Bhatia R, Bamberger JD, Kushel MB. Public health care utilization in a cohort of homeless adult applicants to a supportive housing program. J Urban Health 2006; 83(5): 860-73. doi:10.1007/s11524-006- 9083-0

18 Kushel MB, Perry S, Bangsberg D, Clark R, Moss AR. Emergency department use among the homeless and marginally housed: results from a community-based study. Am J Public Health 2002; 92(5): 778-84. doi:10.2105/AJPH.92.5.778

19 Kushel MB, Vittinghoff E, Haas JS. Factors associated with the health care utilization of homeless persons. JAMA 2001; 285(2): 200-6. doi:10.1001/jama.285.2.200

20 Kushel MB, Yen IH, Gee L, Courtney ME. Homelessness and health care access after emancipation: results from the Midwest evaluation of adult functioning of former foster youth. Arch Pediatr Adolesc Med 2007; 161(10): 986-93. doi:10.1001/archpedi.161.10.986

21 Van Ness PH, Davis WR, Johnson BD. Socioeconomic marginality and health services utilization among central Harlem substance users. Subst Use Misuse 2004; 39(1): 61-85. doi:10.1081/JA-120027766

22 Fulde GWO. The homeless and the emergency department: a special relationship. Med J Aust 2003; 179: 651-2.

23 Esplin P, Mahony A, Wilson SF. The Community Outreach Medical Emergency Team (COMET): Setting up a new homeless service, an evaluation of the first six months. Nursing Monograph. Sydney: St Vincent's and Mater Health; 2007: 42-47.

24 Kushel MBMD, Gupta RMD, Gee LMPHJD, Haas JSMDM. Housing instability and food insecurity as barriers to health care among lowincome Americans. J Gen Intern Med 2006; 21(1): 71-7. doi:10.1111/ j.1525-1497.2005.00278.x

25 Sadowski LS, Kee RA, VanderWeele TJ, Buchanan D. Effect of a housing and case management program on emergency department visits and hospitalisations among chronically ill homeless adults. JAMA 2009; 301: 1771-8. doi:10.1001/jama.2009.561

26 The Road Home - Homelessness White Paper; Chapter Six: Research. Department of Housing; 2008. Available at http://www.fahcsia.gov.au/ sa/housing/progserv/homelessness/whitepaper/Documents/ch6.htm [verified May 2009].

27 History of the AR-DRG Classification Development. Department of Health and Ageing; 2009. Available at http://www.health.gov.au/ internet/main/publishing.nsf/Content/AR-DRG-History [verified July 2009].

28 Policy on the Australasian Triage Scale. Australasian College for Emergency Medicine; 2006. Available at http://www.acem.org.au/infocentre. aspx?docId=59 [verified April 2009].

29 Australian Hospital Statistics. Australian Institute of Health and Welfare, 2009. Available at http://www.aihw.gov.au/publications/index.cfm/title/ 10776 [verified July 2009].

30 Hulchanski JD. A New Canadian Pastime? Counting homeless people: addressing and preventing homelessness is a political problem, not a statistical or definitional problem. Toronto, ON: University of Toronto, Centre for Urban and Community Studies; 2000. Available at http://www.urbancenter.utoronto.ca/pdfs/researchassociates/2000_ Hulchanski_Counting-Homeless-People.pdf [verified 24 January 2011].

31 Shapcott M. Backgrounder: Counting Toronto's homeless people. Toronto, ON: The Wellesley Institute; 2006.

Manuscript received 12 August 2009, accepted 18 May 2010

[Author Affiliation]

Caroline N. Chin1 MBBS, Registrar

Kate Sullivan2

Stephen F. Wilson3,4 MBBS, PhD, FRACGP, FAFRM, Director Population Health

1Notre Dame University, Sydney School of Medicine, 160 Oxford Street, Darlinghurst, NSW 2010, Australia. Email: aoi.sakana@gmail.com

2School of Medicine, University of New South Wales, Sydney, NSW 2052, Australia.

3St Vincent's Hospital, 390 Victoria Street, Darlinghurst, NSW 2010, Australia. Email: knhsullivan@gmail.com

4Corresponding author. Email: stwilson@stvincents.com.au