This study compares chronic diseases and multi-morbidity between patients of different housing circumstances, who sought medical care at clinics of a humanitarian network in the wake of the COVID-19 pandemic. Our findings showed that homelessness and housing exclusion were not associated with a higher likelihood of chronic diseases and multi-morbidity. However, we report poorer self-rated health and a higher prevalence of psychiatric illnesses amongst PEH, as well as noteworthy socio-demographic differences between the three housing groups.
The majority of patients who sought care at the humanitarian clinics had either no or limited health coverage (92.3%) and were below the poverty line (96.7%). While selection bias contributes to this finding, it highlights the role of health insurance and social exclusion as important barriers to accessing regular healthcare services in Germany. Interestingly, a higher proportion of PEH had health coverage (33.03%) compared to persons experiencing housing exclusion (12.1%). A possible explanation is that a larger proportion of PEH in our study originated from EU/EEA member states and may have been eligible for insurance entitlements from their respective countries. Despite their entitlements, they preferred the services of a humanitarian organisation, which may reflect the multitude of individual barriers faced by this population, but also the benefit of the trauma-informed care and social counselling services provided by the clinics.
About three times as many men (74.2%) in our study experienced homelessness in comparison to women. The predominance of men among PEH has been widely reported [14, 19, 33, 34]. However, more than half (54.6%) of persons experiencing housing exclusion were women. This is consistent with previous reports that women are more likely to rely on relatives, friends, and other informal systems when they fall into homelessness, only approaching homeless and welfare services when these supports are exhausted . Women are shown to be present in larger proportions when definitions of, and data collection frameworks on homelessness extend beyond persons living in emergency shelters and sleeping rough . Homelessness among women is also commonly triggered by experiences of intimate partner violence (IPV) , which underwent a dramatic increase as a result of the lockdowns and restrictions imposed during the COVID-19 pandemic [36, 37]. A higher proportion of PEH also originated from EU/EEA countries (56.8%), while non-EU/EEA nationals (60.5%) and undocumented migrants (39.6%) were more highly represented in persons facing housing exclusion. According to previous reports by the BAG W, EU/EEA citizens make up most rough sleepers in Germany, consistent with our findings . The overall high migrant population in our study population also highlights the difficulties faced by individuals with a migrant background in accessing regular health services in Germany. Moreover, it confirms a previous statement by the European Observatory on Homelessness describing migration as a new structural risk factor for the development of homelessness .
Chronic diseases made up 45.7% of the total number of diagnoses in patients visiting the humanitarian clinics in 2020. The predominance of hypertension, diabetes and chronic ischemic heart disease in our study population mirrors the most common chronic diseases in the general German population . However, the prevalence of hypertension (10%, 14.2% and 8.2% in PEH, persons experiencing housing exclusion and living in secure housing, respectively) was considerably lower than the 12-month prevalence of 31.8% in the general population in Germany, reported by the German Health Update (GEDA), a population-representative health survey conducted by the Robert Koch Institute (RKI) . Similarly, the prevalence of chronic ischemic heart disease is higher in the general population; 5.8% compared to 1.82%, 3.16% and 4.1% in PEH, persons experiencing housing exclusion and with secure housing, respectively . The 12-month prevalence of diabetes (8.9%) was comparable to our population experiencing housing exclusion (8.4%), but higher than the two other housing groups . These findings could be attributed to the majority of the diagnoses being new rather than previously diagnosed and self-reported by the patients. Furthermore, since chronic diseases develop over time and the manifestations are often intermittent, a “point” prevalence based on a single examination and clinic visit is likely to underestimate disease frequency.
Our analysis found that age was the only factor associated with a higher risk of chronic disease, reflecting the epidemiological nature of most chronic diseases. In contrast to our findings, multiple cohort studies have demonstrated higher rates of chronic diseases and multi-morbidity in persons experiencing homelessness [14, 41,42,43]. A study analysing electronic health records (EHRs) in the UK between 1998 and 2019 found that persons experiencing homelessness were 1.8 times more likely to have baseline prevalence of cardiovascular diseases (CVD), and suffered a higher burden of comorbidities in comparison to housed controls . A systematic review of 17 observational studies on CVD in homeless versus housed individuals also found that hypertension was more likely to occur in persons experiencing homelessness . However much of the research draws comparisons between PEH and the general or generally deprived population, without accounting for the additional barriers faced by these populations. A study examining the unmet health needs in homeless versus vulnerably housed adults in three Canadian cities, reported no significant differences between the two, suggesting they are intersecting populations with similar health status and experienced barriers .
However, we describe a positive association between psychiatric illness and homelessness, consistent with numerous studies describing higher rates major depression, anxiety disorders, bipolar disorder and alcohol and drug dependency among PEH [1, 4, 9, 18, 44]. While the association between alcohol-related disorders and housing situation was not significant, only PEH in our study were diagnosed with the former. Two systematic reviews spanning studies published in 1979–2005 and 2007–2021, and altogether comprising 13,733 individuals, described substance-use disorders, in particular alcohol-related disorders as the most common psychiatric illness among PEH, with random effects pooled prevalence of 37.9% and 36.7%, respectively [44, 45]. Interestingly, the latter study reported studies conducted in Germany were associated with higher prevalence rates in multivariable analysis . Similarly, a systematic review of studies conducted in Germany reported a pooled prevalence of alcohol dependency of 55.4%; higher than any other psychiatric illness, and 22 times higher than the prevalence in the general German population . The low prevalence of alcohol-dependency in our study (4%) may result from multiple factors. Our study population included first presenters to the clinics, and potentially excluded a larger population who visited the psychiatric outpatient services of the clinic, or were linked into the regular mental health services after their first visit. Furthermore, clinicians at the humanitarian clinics may not have been specifically trained at diagnosing mental illness, and relied on clinical examination only to reach a diagnosis, which could have resulted in potential underdiagnoses.
Psychiatric illnesses are an important contributor to the increased mortality rates among PEH, from suicide and substance abuse, but also through higher rates of criminalisation and violent victimisation . Psychiatric illnesses also further complicate the treatment of chronic diseases by acting as an important barrier to seeking and maintaining contact with health services .
Strengths and limitations
There are several strengths and limitations to this study. Through our cooperation with the clinics of “Ärzte der Welt”, we were able to study a population which is generally hard to reach and often excluded from operational and medical research. Our study also includes persons experiencing housing exclusion, a population often excluded in studies examining homelessness. While our findings cannot be generalised to the entire population experiencing social exclusion and homelessness in Germany, we provide insight into a population deprived of access to regular health services at one point in time. We acknowledge that the characteristics of our study population are likely influenced by public health measures implemented within the period of our study, and due to the lockdowns and restrictions imposed throughout 2020, it is likely that it excludes an even larger population experiencing marginalisation who might have even higher mental and physical health needs. Furthermore, some humanitarian clinics underwent restructuring, temporary closure, or experienced understaffing, affecting data collection and resulting in underreporting. Systematic data collection regarding referrals and testing were also not fully implemented in the clinics until late 2020, with some clinics not undertaking SARS-COV-2 testing at all, contributing to the scarcity of data on SARS-COV-2. However, this underlines an important effect of the COVID-19 pandemic on service provision for persons experiencing social exclusion.
Another limitation is the much smaller population of patients with secure housing in comparison to the other groups, affecting the power of our statistical analysis. This is an expected finding, as persons visiting the humanitarian clinics are more likely to originate from low socio-economic backgrounds, and therefore experience precarious housing. Due to a hesitancy to disclose sensitive information, some patients also did not provide answers to all questions, resulting in missing data. Additionally, since we relied on the physician’s coding of a disease as “acute” or “chronic” during medical consultations, some conditions may not have been coded or coded inappropriately, resulting in underdiagnoses. Physicians work at the humanitarian clinics on a voluntary and rotational basis and may not have the necessary training to ask for specific mental health experiences, such as exposure to domestic violence, adverse childhood experiences and addiction disorders, which fit within ICD-10 diagnoses but were not identified in our population.