Abstract
Background Estimating healthcare needs of the homeless is associated with challenges in identifying the eligible population.
Aim To explore the demographic characteristics, disease prevalence, multimorbidity, and emergency department visits of the homeless population.
Design and setting EMIS electronic database of patient medical records and Quality and Outcomes Framework (QOF) data of all 928 patients registered with a major specialist homeless primary healthcare centre based in the West Midlands in England, from the period of October 2016 to 11 October 2017.
Method Prevalence data on 21 health conditions, multimorbidity, and visits to emergency departments were explored and compared with the general population datasets.
Results Most homeless people identified were male (89.5%), with a mean age of 38.3 (SD = 11.5) years, and of white British origin (22.1%). Prevalence of substance (13.5%) and alcohol dependence (21.3%), hepatitis C (6.3%), and multimorbidity (21.3%) were markedly higher than in the general population. A third (32.5%) had visited the emergency department in the preceding 12 months. Emergency department visits were associated with a patient history of substance (odds ratio [OR] = 2.69) and alcohol dependence (OR = 3.14).
Conclusion A high prevalence of substance and alcohol dependence, and hepatitis C, exists among the homeless population. Their emergency department visit rate is 60 times that of the general population and the extent of multimorbidity, despite their lower mean age, is comparable with that of 60–69-year-olds in the general population. Because of multimorbidity, homeless people are at risk of fragmentation of care. Diversification of services under one roof, preventive services, and multidisciplinary care are imperative.
INTRODUCTION
Homelessness is a widespread issue in the UK,1 with an estimated 250 000 people known to be homeless in England alone.2 More than 4000 people sleep rough on any given night in England, with numbers of rough sleepers rising, particularly in urban areas; in London, for example, the number of rough sleepers has doubled in the last 6 years (up to and including 2017).3
There is a dearth of literature investigating healthcare issues among homeless people in the UK. Findings from international literature suggest that those experiencing homelessness are significantly disadvantaged in achieving and maintaining a healthy lifestyle.4 They face up to 12 times higher mortality rates than the general population, mostly due to opioid overdose, accidents, heart failure, and infectious diseases.4 The negative health consequences of social exclusion are noted to be greater in females than males.4 A UK study in 2012 identified that rough sleepers and those occupying homeless shelters die at an average age of 47 years.5 Health status worsens with increasing length of time as homeless.6 Historical estimates have suggested that homelessness is independently linked with high emergency department use.7 However, there is limited literature exploring the rate of emergency department visits and the characteristics within homeless populations associated with this increased use of emergency care.
Primary healthcare service provision for homeless people
There has been an emergence of some specialist primary care support for homeless people across the UK. There is at least one such practice in most major cities in the UK that offers primary healthcare centres for homeless people and some general practices have particular expertise in homelessness.8
The lack of studies in the UK that have investigated the prevalence of key health conditions necessitates the strengthening of the evidence around the primary healthcare needs of homeless populations. Identifying the burden of disease is often challenging in socially excluded populations as social disadvantage is often not recorded in medical records and the UK general register of births and deaths. Homeless populations also have very limited coverage in routine health surveys due to their often secluded and unstable locations. There is also a need to address the current gap in the range of methodology that has been used to explore the healthcare issues of homeless people. Gathering and analysing healthcare utilisation datasets from a large specialist primary healthcare centre for the homeless can provide useful data for use by primary healthcare service providers, researchers, and decision makers to identify unmet need. It can also aid in the redesigning of services and widening preventive measures for public sector action.
How this fits in
Homeless people face extreme social exclusion. There is a dearth of literature in the UK about the healthcare needs of homeless people, and most of the literature around healthcare issues and homelessness are of international origin. This research sought to identify the demographic characteristics, disease prevalence, multimorbidity, and emergency department visits of the registrants of a specialist primary healthcare centre for the homeless, using a large sample size. The findings of this study show that homeless populations are at risk of facing fragmentation of care as a result of high levels of multimorbidity. It demonstrates the need for the provision of preventive health care and multi-sector approaches in addressing homeless people’s complex healthcare needs and minimising their use of emergency care.
The aim of this study was to explore the demographic characteristics, disease prevalence, multimorbidity, and visits to emergency departments of homeless people.
METHOD
This study was conducted in a specialist primary healthcare centre for the homeless in the West Midlands in England. The healthcare centre provides general practice services to the homeless population. Registrants have access to a GP, nurse practitioners, psychotherapy counsellor, podiatrist, alcohol dependence intervention nurse, and street outreach services. The centre does not provide treatment for substance dependence so patients are referred to a dedicated service based in the city.
The Quality and Outcomes Framework (QOF) and EMIS electronic data of patient medical records were used. QOF is an annual reward programme for general practice achievements, an aspect of which involves the building of disease registers.9,10 EMIS is an online database, which is used by most general practices across the UK to store patients’ clinical data (https://www.emishealth.com/products/emis-web/).
A search function allows the prevalence of health conditions to be gathered among the practice registrants. For disease prevalence data, all patient records were searched with relevant Read codes.
The data search was undertaken in November 2017 by staff at the general practice with routine access to the datasets using queries specific for a health condition. All data were cleaned and anonymised before being passed to the research team. The prevalence of 21 key health conditions was explored. These conditions included cardiovascular disease, mental health, infection, respiratory, neurological disorders, cancer, and endocrine disorders. For emergency department attendance, a search was run to identify patients’ EMIS datasets for the previous 12 months (October 2016 to 11 October 2017). Demographic data including age, sex, ethnicity, and smoking status were extracted. The World Health Organization definition of multimorbidity, ‘the coexistence of two or more chronic conditions in the same individual’, was used.11
All data were stored on secure password-protected computers. Data were analysed using descriptive and inferential statistics. The comparison of prevalence data across age and sex was undertaken based on the evidence from international literature that health inequality is found to affect socially excluded females and older populations more than the male population.4 Comparative data relating to the English or UK general population were taken from a variety of sources including the QOF, national statistics, and published literature. In addition, comparison was made to prevalence data as available in the international literature that related to homeless populations. Binary logistic regression analysis was conducted to identify factors that were associated with patient emergency department attendance. Emergency department attendance in the previous 12 months was used as an outcome variable. Explanatory variables related to disease areas and any demographic characteristics that showed an association (P ≤0.25)12 with the outcome ‘A&E attendance in the last 12 months’ in the univariate analysis.
DISCUSSION
Summary
This study aimed to explore the demographic characteristics, disease prevalence, multimorbidity, and visits to the emergency department by the registrants of a specialist primary healthcare centre for the homeless in the West Midlands. Datasets of all registered 928 patients were retrieved and analysed. Demographic characteristics, a range of health conditions, including alcohol and substance dependence, and emergency department attendance data were explored. This study adds to the limited evidence that exists around the prevalence of health conditions and multimorbidity in homeless people by using a large sample size. This study has demonstrated a high prevalence of multimorbidity, mental health conditions, particularly substance and drug misuse, and infectious diseases, notably hepatitis C, among the homeless population in the area studied compared with the general population.
A high rate of emergency department attendance was observed among the study population. Considering all emergency department visitors among study participants made a minimum of one visit to the emergency department, this translates to approximately 60 times the rate of emergency department attendance made by the general population (as measured in 2011).48
Strengths and limitations
The datasets presented here represents a large sample size of a homeless population and hence adds to the literature. Rigorous methods of analyses were used to explore the link between demography, diagnosed health conditions, and emergency visits among the homeless population and provides extensive comparison with existing datasets from international literature.
Similar to other studies using routinely collected datasets in investigating disease prevalence and multimorbidity, this study relied on the diagnosis of the health conditions being accurately recorded in patient medical records. Therefore, the prevalence of the health conditions and multimorbidity, as identified in this study, are likely to be an underestimation. Particularly, it was noted that health conditions such as coronary heart disease, stroke, diabetes, cancer, asthma, learning disabilities, and rheumatoid arthritis were found to be under-prevalent in the study participants compared with the findings in the literature.13
This study analysed datasets of those who presented at the specialist homeless healthcare centre. This study did not explore how much patients engaged with the practice, therefore, the actual prevalence of the included health conditions may have been under-estimated as patients may be missing scheduled appointments, which makes it likely for key health conditions to go undiagnosed and because of the inclusion of information of those who regularly attend the practice.
Comparison with existing literature
Substance and/or alcohol dependence have been cited as a cause and consequence of homelessness.49 Previous studies have looked at the extent of self-harm,50 and mortality linked to mental health conditions, including suicide, among homeless people.51 This study demonstrates that substance and alcohol dependence are important risk factors that make homeless populations seek emergency care.
This study has also demonstrated a high rate of multimorbidity among the homeless registrants. Given that the mean age of the registrants of the homeless healthcare centre was 38.3 (SD = 11.5) years, the proportion of patients with at least two long-term health conditions compares with those aged 60–69 years in the general population.52 The proportion of patients with multimorbidity was identified to be less than that reported in a Scottish study.53 The reasons for these differences should be explored; however, it is likely that, despite a small sample size in the Scottish study,53 researchers had access to individual patient medical notes. Similarly, in the current study, the prevalence of mental health conditions, particularly depression and alcohol and substance dependence, despite being higher than in the general population, was lower compared with other studies on the homeless population in the UK.16,18,20,21
The prevalence of some cardiovascular health conditions such as hypertension, as well as respiratory health conditions, diabetes, and cancer, was also noted to be lower than other studies on the homeless population in the UK. However, the literature suggests that homeless and socioeconomically disadvantaged people have higher mortality rates contributed by these health conditions than the general population and those with less deprived backgrounds.54,55 It is highly likely that some of these conditions were not appropriately coded in patient medical records or potentially underdiagnosed. Health conditions such as hypertension are asymptomatic and it may not be routine practice to record blood pressure in every consultation given the constrained resources that are available in these settings. Information on the length of time the registrants were registered at the practice was not available for this study. Registrants of similar services in other studies have demonstrated that participants also reported using mainstream general practices.18
The number of health conditions investigated for the multimorbidity analysis in this study compares favourably with other studies. There are no international standards on how many long-term conditions should feature in the measurement of multimorbidity; however, an average of 18.5 chronic health conditions was featured in a systematic review of the international literature that included 39 studies.56 The prevalence of all cardiovascular health conditions, COPD, hepatitis C, diabetes, cancer, and leg ulcers was linked to older age, and this supports the epidemiological trend in the general population.23,57–65 Repeat emergency department attendance by the study population was not investigated. A previous study has identified that homeless people, including rough sleepers, constitute approximately 8% of all repeat users of the service.7 There is a lack of research investigating in depth the reasons for such repeat attendance. Repeat attendance could be linked to poor general health and lifestyle, as well as non-access to or non-use of available primary healthcare services.66 Greater use of the emergency department may impact on patient care, as patients seeing a known and trusted clinician in primary care is imperative for ensuring the continuity of care.67
Implications for research and practice
This study provides compelling evidence that there exists a high prevalence of key chronic health conditions and multimorbidity among the homeless population. Although data of only those registered with the specialist general practice were analysed, the data can be carefully extrapolated to those not registered with such services or hidden homeless who often do not declare their fixed-abode status to their health services providers. Healthcare professionals seeing patients who are homeless are more likely to encounter multimorbidity than in mainstream healthcare centres. The extent of multimorbidity seen in this population is often only encountered in the older population and hence specialist clinical knowledge, alongside multidisciplinary management, is required for many of these patients. Diverse skill sets are imperative at these specialist healthcare centres. Patients with multimorbidity are often disadvantaged because of the fragmentation of care.53
The high level of multimorbidity in this population could be linked to socioeconomic deprivation as well as to the uptake of behaviours such as smoking, alcohol, and substance dependence, or both.53 Public health, NHS, and local government interventions, particularly preventive services in the community and primary care, can help prevent multimorbidity where such outcomes are linked to the implications of the uptake of risky behaviours. The groundwork for further collaboration between such public bodies is already being laid down in the UK, for example, through the Homelessness Reduction Act 2017.68 The act places emphasis on multi-agency approaches to preventing homelessness and provides an opportunity for public bodies to work more closely with partners and co-produce an approach to homelessness prevention through collaboration and cooperation.
Future longitudinal studies are needed to identify the contribution of key factors linked to multimorbidity. There is a continued need to improve access to mental health including for those with substance and alcohol dependence.
Community screening of bloodborne viruses, particularly opportunistic screening when presenting for other services, as has been piloted in some areas of England,69 is recommended.
The barriers associated with access and positive experiences around homeless people’s use of primary care and wider community services also needs to be addressed, given the health inequalities as demonstrated by this study. Findings of the authors’ recent study66 shows that there are organisational barriers (such as difficulty in registering with a general practice, lack of integration of services including suboptimal communications and transition of care across services) and patient-related barriers (including lack of knowledge and awareness of primary healthcare services, inadequate skills and capacity to navigate services, and level of health literacy) to access and encounter positive experiences of primary healthcare services among the homeless population. There appears to be confusion around eligibility of people who are homeless registering with a general practice, and patients have often been denied access, contrary to the guidelines that are available, which state that people do not need a fixed address or identification to register or access treatment at GP practices.70 Awareness of such policy among frontline staff, homeless people, and any partner agencies should be strengthened. Patients are often less aware of specialist services for the homeless people existing in their areas. Provision of such specialist services are often temporary solutions and are mostly located in areas with high homelessness. Long-term planning could incorporate improving capacity in mainstream general practices. Such improvement will require skills in managing multimorbidity and the communication skills required to develop rapport with homeless people, along with minimising perceived stigma and discrimination for this group in the society and healthcare settings.
Emergency department attendance data as reported in this study should be treated with caution because of the possibility of unknown confounders and the chance that visits were not linked to the conditions. It is recommended that data should be supplemented from emergency departments to identify key reasons for repeat attendance.
Future studies should consider using multiple data sources in estimating disease burden. These include consideration of aggregated datasets as used in this study, access to individual medical notes, health-related data available from other partners including housing and the voluntary sector, datasets from outreach services, surveys of homeless populations to gather self-reported data, prescribing and medicines dispensing data, and inclusion of datasets from homeless populations using mainstream services.
