Experiences of homeless individuals using ambulance services: a narrative review

Information sources

The Cumulative Index to Nursing and Allied Health Literature (CINAHL) Plus, Allied and Complementary Medicine Database (AMED), MEDLINE and Emcare databases were chosen for their significant inclusion of allied health profession literature (Holland et al, 2021). Additional literature was identified from AMBER (Ambulance Research Repository), OpenGrey via Data Archiving and Networked Services (DANS), Google/Google Scholar and by reviewing reference lists. All sources were searched in May 2022.

Search strategy

An approach looking at two distinct populations was found to achieve the fullest understanding of the subject area rather than using other strategies such as a population, intervention, comparison, and outcome (PICO) methodology.

Experimental searching was undertaken before the search strategy was finalised. Search terms included Medical Subject Headings (MeSH) terms and alternative free text words/phrases (Table 1).

Literature published between 2002 and 2022 was set as the paramedic profession has changed significantly over the past 20 years, especially with the development of specialist and advanced practice (Newton, 2011; Petter and Armitage, 2012; Brown 2017).

Eligibility criteria

Studies relating to homeless individuals, the paramedic profession and ambulance/prehospital settings were included for synthesis.

Globally, the definition of a paramedic is not universal. Therefore, paramedics were included as a professional group as well as related roles such as emergency medical technician.

Non-paramedic groups were excluded, along with mixed population studies where paramedic and homeless groups could not be identified as individual populations.

Studies outside ambulance and prehospital settings were not included, except for ED studies, which were excluded only after full-text review if they showed that paramedics had not interacted with the patient. Table 2 details the full inclusion and exclusion criteria.

Study selection

As this review was undertaken as part of MSc study, a single reviewer screened studies against eligibility criteria. Studies were excluded at abstract if not relevant before full texts were reviewed. Methodological quality and bias were assessed using Critical Appraisal Skills Programme (CASP) (2024) tools with relevant biases discussed narratively.

Data synthesis

Qualitative data and descriptive statistics were extracted manually. Themes were developed through thematic synthesis (Thomas and Harden, 2008) with coding undertaken manually.

Overview

Five hundred and eighty-three records were identified, with 23 taken forward for review (Figure 1). Table 3 summarises study characteristics and relevant findings.

While the number of studies identified was larger than expected, the amount of relevant information within each study is generally small. Qualitative data and descriptive statistics were extracted.

Six main themes were identified through inductive thematic synthesis.

Frequency of ambulance use

Eleven studies explored ambulance use by homeless individuals.

Abramson et al (2021) identified that homeless individuals called the emergency services at 14 times the rate of non-homeless individuals. Coe et al (2015) found homeless individuals were three times as likely as non-homeless individuals to arrive at ED by ambulance. Ku et al (2010) identified homeless individuals arrived by ambulance in 35.7% of cases compared to 16.0% for non-homeless individuals.

Oates et al (2009) identified homeless individuals were significantly more likely to arrive at EDs by ambulance than non-homeless (34.12% compared to 15.42%). Pearson et al (2007) found homeless individuals arrived by ambulance in 51% of cases compared to 29% for non-homeless.

Tadros et al (2016) reported that, in 2010, in the later of two time periods, homeless individuals arrived by ambulance in 48.0% of cases compared to 16.0% for non-homeless. Ambulance use by homeless individuals increased from 34.1% during the earlier period in 2005, which has not been explored further.

For psychiatric presentations, Lamparter et al (2020) found that homeless individuals arrived by ambulance in 46% of cases compared to 16% for non-homeless. For lower-acuity conditions, Durant and Fahimi (2012) identify homeless individuals were more likely to arrive at hospital by ambulance than non-homeless (30.05% compared to 6.4%).

For frequent callers, Tangherlini et al (2010) found that, in older patients (aged >65 years), homelessness was the biggest predictor for having ‘high’ or ‘very high’ ambulance usage, accounting for 23% and 33% of transports respectively, compared to 2% for the ‘low’ usage control group. Tangherlini et al (2016) found that being homeless was the biggest individual characteristic in the study population of frequent callers (38.9%).

Finally, Brown et al (2010) identified that 63.1% of homeless individuals arrived at ED by ambulance during cold weather conditions. Taken in isolation, this finding does not provide meaningful comprehension. However, it is useful for contextual consideration.

Clinical characteristics

Six studies examined common presentations to ambulance services.

Abramson et al (2021) listed the most common presentations as behavioural/psychiatric for homeless individuals (12.8%) and traumatic injury for non-homeless (13.9%). The lists are comparable apart from alcohol intoxication, overdose and pain/swelling to extremities, which do not feature on the non-homeless list. No other studies compare homeless and non-homeless individuals.

Yeniocak et al (2017) identify ‘clouded consciousness’ as the most common condition (29.8%). This has not been defined and the term's unfamiliarity may be down to the study having been conducted in a non-western setting. It may be comparable to the findings of Abramson et al (2021) as, when specialist referral was required, the most common reason was for psychiatric support (23.35%). Medical presentations are more common than trauma. Tadros et al (2012) noted that ‘other medical’ conditions is the most common presentation for homeless individuals (23.4%). Mackelprang et al (2014) identified medical illness as the most common presentation (48%).

Brown and Steinman (2013) suggested that older homeless individuals (aged ≥50 years) were more likely to attend ED for injuries compared to alcohol/substance misuse in younger patients who are homeless. Whether this correlates with ambulance use is unclear as the study assessed all ED attendances and ambulances were used for only 48% for older and 36% for younger individuals. Leggio et al (2020) investigated the experiences of individuals in homeless shelters; the main condition requiring an ambulance was seizure followed by other medical and traumatic reasons. As this is a qualitative study, quantitative statistics are not presented and it is difficult to compare this to other studies.

Specifically regarding alcohol intoxication, Flower et al (2011) identified that 57% of patients who attended EDs by ambulance for this reason were homeless. These findings do not offer a strong insight but may suggest that alcohol misuse is a common presentation.

Regarding clinical acuity, Abramson et al (2021) concluded that, based on the requirement for an ambulance equipped to provide advanced life support, calls from homeless individuals were more likely to be of lower acuity than non-homeless (31.4% compared to 42.5%). Pearson et al (2007) reported that homeless individuals were 7% less likely than non-homeless to be transported urgently and 11% less likely to be hospitalised.

Through qualitative interviews, Alunni-Menichini et al (2020) identified that homeless individuals called ambulances for long-term conditions because they had difficulty in accessing and navigating alternative services. This could be interpreted as lower acuity. Tadros et al (2016), however, found no significant difference in clinical acuity between homeless and non-homeless individuals.

Moore et al (2019) undertook interviews investigating the experiences of homeless individuals who frequently used the ED. The relevance of this study is limited; however, one patient discussed that they called for an ambulance only when ‘on death's doorstep’, which could suggest higher clinical acuity.

Why homeless individuals use ambulance services

Two studies explored homeless individuals' motivations to use ambulances to some extent.

Elwell-Sutton et al (2017) found homeless individuals who were refused registration with a GP or dentist were less likely to access healthcare through ambulance services. Conversely, homeless individuals who accessed outreach teams and day centres were more likely to use ambulances. This suggests not being able to access healthcare services does not necessarily lead to ambulances being used to compensate.

Alunni-Menichini et al (2020) disagreed and established several themes illustrating that emergency services are commonly used to compensate for other services not meeting the needs of homeless individuals, including services that are complex to navigate, fragmented and not responsive to service users' needs. Paramedics were considered to be gatekeepers to other services. One consideration is that emergency services within this study include police, paramedics and hospital workers. However, some responses focused on paramedics as an exclusive group so data can be extracted for the purposes of the present study.

Why homeless individuals do not use ambulance services

Seven studies discuss why homeless individuals may not require an ambulance response.

Abramson et al (2021) found that 27.3% of calls for homeless individuals were made by third-party callers compared to 14.4% for non-homeless, and homeless individuals were more likely to refuse treatment (4.9% compared to 3.9%). The reasons for these findings are not explored. Leggio et al (2020) found that homeless individuals frequently do not have a choice as to whether an ambulance is called and often do not call themselves.

Regarding substance misuse, Koester et al (2017) identified barriers including fear of police involvement and impacting hostel acceptance. Also, people may have already recovered so no longer need an ambulance. Alunni-Menichini et al (2020) agreed that mistrust may result in homeless individuals not calling ambulance services, although little supporting data were presented. As already discussed, Elwell-Sutton et al (2017) found homeless individuals use ambulances less often if they are not registered with a GP or dentist, so being disengaged with services may be relevant.

Mackelprang et al (2014) found that providing housing reduced ambulance contacts. Providing settled accommodation may therefore be a reason for not using ambulance services. Whether this reduction continues if an individual becomes homeless again has not been assessed. Parsell et al (2017) disagree, finding that use of government services, including ambulances, did not change when homeless individuals were allocated housing.

Negative experiences with paramedics have also been identified by Alunni-Menichini et al (2020), Koester et al (2017) and Leggio et al (2020) as reasons for not calling ambulances; this is discussed within the next theme.

Experiences of homeless individuals

Four studies reported the experiences of homeless individuals when interacting with ambulance services.

Zakrison et al (2004) conducted qualitative interviews and converted responses to a quantitative survey, although the authors did not adequately explain the method for this. In relation to paramedics, 89.5% of participants reported previous positive interactions while 23.7% reported negative experiences. Most (90.5%) would call an ambulance in an emergency compared to 6.0% who would not. An explanation for the findings was not presented. One respondent reported being assaulted by a paramedic.

Alunni-Menichini et al (2020) found homeless individuals consider paramedics to be approachable and have positive experiences. Participants in Koester et al's (2017) study agreed that paramedics always give care and never refuse treatment. However, overall experiences reported were negative, which may be owing to substance misuse rather than homelessness.

Leggio et al (2020) reported mixed experiences. Positive experiences included professionalism, kindness and being non-judgemental. Several direct quotes from homeless individuals highlighted the high regard in which paramedics are held:

Negative experiences were also reported including biased and judgemental attitudes and feeling they were not taken seriously. Inconsistent treatment between different ambulance crews was reported, including assumptions they had called because of alcohol or substance misuse, having to walk to the ambulance, not being able to lie down on a bed and paramedics trying to convince them not to travel to hospital. This led to individuals feeling ashamed of their circumstances.

Koester et al (2017) detailed similar experiences, especially judgemental and disparaging attitudes. Participants reported ‘brutal stigmatising language’ related to homelessness and substance misuse.

Experiences of paramedics

Limited data were identified regarding the experiences of paramedics. Alunni-Menichini et al (2020) detail barriers including services being unavailable when needed, leading to an ED being the only option. The criteria and procedures for referrals were also often considered to be complicated and time-consuming, with changes made without any notification.

Frequency of ambulance use

The most common theme was frequency of ambulance use by homeless individuals. All studies looking at frequency concluded homeless individuals are more likely to use ambulances than non-homeless. The extent of this differs between studies. However, where comparable data can be extracted, usage is in a range of 30.05–63.1% for homeless individuals and 6.4–29.0% for people who are not homeless (Pearson et al, 2007; Oates et al, 2009; Brown et al, 2010; Ku et al, 2010; Durant and Fahimi, 2012; Coe et al, 2015; Tadros et al, 2016; Lamparter et al, 2020).

What is unclear is whether all homeless individuals use ambulance services more frequently or whether a specific cohort have disproportionate use. Evidence from other studies indicates that homelessness alone is unlikely to be the main reason for ambulance contacts (Søvsø et al, 2019). Evidence relating to whether homeless individuals are frequent callers is limited and a conclusion cannot be reached (Tangherlini et al, 2010; 2016).

Previous research has found that complex and chronic physical and mental health conditions are the most common characteristic of frequent callers, with homelessness not identified as an independent variable (Edwards et al, 2015). Therefore, it is reasonable to suggest that homeless individuals are more likely to call ambulances because they carry a disproportionate burden of chronic illness (Lewer et al, 2019) and health inequalities (Marmot, 2010) rather than because of homelessness itself. This means that not all homeless individuals call more frequently than non-homeless individuals.

Clinical characteristics

Evidence suggests that physical and mental health issues rather than trauma are common reasons for homeless individuals calling ambulances. Whether this differs from non-homeless individuals is uncertain as only one study (Abramson et al, 2021) offers a comparison. The authors concluded that clinical characteristics are different and that non-homeless individuals are more likely to present with traumatic injury. However, other studies suggest medical presentations are more common than trauma (Pittet et al, 2014; Edwards et al, 2015; Møller et al, 2015). Considering that homeless individuals are more likely to experience chronic illness (Lewer et al, 2019), it is rational to suggest that physical and mental health conditions rather than trauma are the most likely presentations. However, without further research, it cannot be confirmed that presentations differ between homeless and non-homeless individuals.

Most studies concluded that homeless individuals present with lower acuity conditions than non-homeless. These studies determined acuity based on clinician assessment, which is heavily reliant on judgement and potentially subject to inaccuracies, bias and variability (Mistry et al, 2018). Bias regarding homeless individuals is well documented and may result in under triage (Gilmer and Buccieri, 2020; Schaffer et al, 2020). Triage is a tool used to assess immediate priority but someone categorised as a lower level may still require intensive and prolonged treatment, whereas someone who is higher priority may need lot of immediate treatment but then be discharged quickly. For example, while Abramson et al (2021) identified lower acuity in homeless individuals, they found they are also more likely to be transported to ED, and Pearson et al (2007) suggested that they spend longer as hospital inpatients. While initial triage is low, there is an increased overall clinical need.

Reports of homeless individuals presenting with no medical need are low across all studies. Therefore, it cannot be definitively concluded that homeless individuals present with lower acuity conditions. Chronic illness is likely to be one of the main reasons for ambulance use. Studies assessing overall outcomes in terms of interventions and discharge status would be a better indication of clinical need for comparison between groups (Drynda et al, 2020).

Why homeless individuals use ambulance services

Stakeholder engagement is vital to understand the experiences of homeless individuals. McCormack et al (2022) and Alunni-Menichini et al (2020) give the biggest insight into why they use ambulances rather than other services. Explanations include that homelessness services are not responsive to their needs, may be difficult to navigate, have restrictive referral or admission criteria and are not available in the areas or time when needed. This correlates with other studies identifying problems registering with other health services may result in increased ED use (Gunner et al, 2019).

However, Alunni-Menichini et al (2020) highlighted that homeless services are often widely available but not well known. Newly homeless individuals in particular have difficulty finding out what is available. However, homeless individuals and health professionals indicate that, once services have been accessed for the first time, needs are generally met and referrals were made to other services. Alunni-Menichini et al (2020) suggested that paramedics are gateways to other services. It could be proposed that paramedics have access to more information or are more familiar with services than homeless individuals and are able to refer or signpost appropriately.

Elwell-Sutton et al (2017) found that not being registered with a GP or dentist led to reduced ambulance use. It could be surmised that not being registered with primary care services indicates a general lack of engagement in all healthcare services, including ambulance services. This was identified by Purkey and MacKenzie (2019), who found that negative experiences with one service may lead to avoidance of other services by homeless individuals.

Overall, evidence that a lack of availability or difficulty accessing services leads to increased ambulance use has not been demonstrated.

Why homeless individuals do not use ambulance services

Abramson et al (2021) and Leggio et al (2020) highlighted that homeless individuals may not want or need an ambulance when others call on their behalf. Although not statistically significant, Elwell-Sutton et al (2017) reported that rough sleepers are more likely to use ambulance services than other homeless individuals.

It could be suggested that the visibility of some homeless individuals may lead to increased calls. This is likely to be for genuine concern and data for one city show the public regularly make calls for the welfare of rough sleepers (South East Coast Ambulance Service, 2018). Ambulances may not be wanted or needed because the patient has recovered from their initial issue. This was suggested by Koester et al (2017) for substance misuse and may also apply to other presentations.

Mistrust may be a reason for not using ambulance services although evidence here is weak. For homeless individuals who misuse substances, ambulances may not be called because they fear police involvement (Koester et al, 2017). This mistrust is likely to be owing to substance misuse rather than homelessness itself as other studies with non-homeless substance users have identified similar fears (Bergstein et al, 2021; Smiley-McDonald et al, 2022). Alunni-Menichini et al (2020) also suggested there may be mistrust in paramedics but did not explain why this would be. Other studies exploring this have not been identified and positive experiences reported by Leggio et al (2020) and Zakrison et al (2004) suggest that paramedics are trusted overall.

Experiences of homeless individuals

Studies detailing experiences of homeless individuals largely report positive interactions with paramedics including that they are professional, caring and non-judgemental. This correlates with positive experiences of the general public for all health professionals (Birkhäuer et al, 2017) and paramedics specifically (Brydges et al, 2016). Other studies relating to homeless individuals show that experiences with health professionals are often negative (Purkey and MacKenzie, 2019; Ramsay et al, 2019). It may be that paramedics are held in higher regard and trusted more than other health professionals or services but a definitive conclusion cannot be drawn.

Not all experiences were positive. Leggio et al (2020) reported some experiences led to feelings of being judged, stigmatised and not taken seriously. Koester et al (2017) identified similar experiences, although the evidence here may relate to substance misuse rather homelessness.

Although this was identified in only one study, it is alarming that one homeless individual reported being assaulted by a paramedic (Zakrison et al, 2004).

The overall conclusion is that interactions are inconsistent and may suggest individual prejudices and biases rather than systematic or cultural issues. However, it would be naive to reach this conclusion as other studies identify discrimination relating to homeless individuals throughout society, including in healthcare (Johnstone et al, 2015). Negative experiences lead to poorer health outcomes (Ramsay et al, 2019). Therefore, reports of negative experience need to be taken seriously.

Experiences of paramedics

Alunni-Menichini et al (2020) identified practical difficulties when referring patients to homeless and healthcare services. This may lead to the ED being the only option which could explain the findings of Abramson et al (2021) that homeless individuals are more frequently transported to ED than other people. This may not be unique to homeless services as difficulties in making referrals to other services have been reported in other studies (Blodgett et al, 2021).

No study explored the attitudes or behaviours of paramedics towards homeless individuals and whether specific enablers or other barriers to care exist.

The evidence for this theme is weak and definitive conclusions cannot be drawn.