Determinants of health-related quality of life (HRQoL) among homeless individuals during the COVID-19 pandemic

Homelessness is an important challenge in Germany and beyond. Estimates from 2022 showed that about 263,000 people were homeless in Germany [1]. The number has increased in the last few years [2]. In the light of the COVID-19 pandemic, it is important to note that the population of homeless individuals is shown to have a higher risk for outbreaks of SARS-CoV-2 due to their living conditions such as crowded homeless shelters and limited access to hygienic supply [3]. In addition, measures to counteract the transmission of SARS-CoV-2 taken by the general population, as for example social distancing, contact tracing and personal hygienic precautions, are more difficult to apply for homeless individuals. These conditions make the population of homeless individuals even more vulnerable, which underlines the relevance to examine their situation during the COVID-19 -pandemic specifically.

Being homeless is associated with multimorbidity [4]. For example, disease prevalence in the categories infections, cardiovascular and respiratory conditions are higher among this population compared to the general adult population [4]. Moreover, homelessness is associated with a higher likelihood of mental health issues [4].

During the pandemic, studies indicate that homeless people are a particularly disadvantaged group when it comes to physical and mental health. For instance, a study reported high levels of depression amongst residents of homeless shelters compared to the general population during the COVID-19 pandemic in France [5]. Another study showed that homeless individuals have a higher mortality risk due to social inequality during the pandemic [6]. A study by Fields et al. found higher hospitalization rates due to COVID-19 itself for symptomatic homeless individuals compared to the general population [7].

Various studies have investigated the determinants of health-related quality of life (HRQoL) in general adult samples, whereas very little is known about the HRQoL in the population of homeless individuals, especially during the COVID-19-pandemic. A Swedish study evaluated the health status among homeless individuals in 2006 and 2018 and reported poor health in all EQ-5D-3L dimensions and the EQ VAS score in 2018 was even significantly lower than in 2006 [8]. In contrast to this, a study that investigated the HRQoL during the first wave of the pandemic (May/June 2020) in Hamburg (second largest city in Germany with now the highest number of homeless people nationally) among homeless individuals found remarkably high EQ-VAS scores particularly compared with the general population [9]. Also, higher age and lower education was positively associated with a lower EQ-5D-5L index.

For a better understanding it may be useful to describe the pandemic situation in Germany during the time of data collection. An incidence rate of about 14 cases per 100,000 population per week was measured in July 2021 (start of data collection). The incidence rate significantly increased in the following weeks, rising up to 74.7 cases per 100,000 population per week by the end of data collection in September 2021. During this period, Delta lineage of the virus became the predominant variant. At the time in Germany the 3G rule was applied (access to certain facilities and companies only for vaccinated, recovered or tested persons) in an effort to prevent further spreading of the coronavirus.

Due to the limited knowledge, the aim of this study was to describe HRQoL and to examine the determinants of HRQoL among the population of homeless individuals during later stages of the COVID-19 pandemic multicentrically. This is of particular relevance to address homeless individuals at risk of poor HRQoL during the COVID-19 pandemic. Moreover, this topic is important because HRQoL affects morbidity and mortality [10, 11].

First, sample characteristics were depicted (stratified by problems in the five dimensions of the EQ-5D). Second, multiple logistic regressions and multiple linear regressions, as appropriate, were conducted to examine the determinants of the EQ-5D outcomes. Please see the next two sections.

The aim of this multicenter study was to describe HRQoL and determine factors associated with HRQoL of homeless individuals in a later stage of the COVID-19 pandemic. In sum, our study revealed a roughly comparable HRQoL among homeless individuals during the COVID-19 pandemic in Germany (average EQ-5D-5L index: 0.85, SD: 0.24)) compared to the general adult population nationally (average EQ-5D-5L index: 0.88, SD 0.18) in 2019 [13]. Also the average EQ-VAS score measured in our study (68.97, SD: 23.83), was similar to the average EQ-VAS score in the general adult population in Germany, which was 71.59 (SD: 21.36) in 2019 [13]. Considering the small amount of data of homeless individuals in general, our study, which is the largest homeless study in Germany, adds important knowledge to the factors being associated with HRQoL in this special cohort. Most of the preceding studies found the HRQoL among homeless people to be relatively low in comparison with the HRQoL of the general population [8, 14]. However, a cross-sectional study conducted in Hamburg by van Rüth et al. which investigated the HRQoL of homeless in May/June 2020 [9] also found a remarkably high HRQoL (average EQ-5D-5L index: 0.84, SD: 0.23)—which is well in line with our here presented findings.

The conditions of homeless individuals are certainly different from the general adult population. Thus, it is quite surprising that the average HRQoL is quite high. One possible explanation for the quite high HRQoL identified in our study could be a habituation of the homeless to their (health) conditions—which may be why their EQ-5D-5L index is similar to that of the general adult population although they are surrounded by special living conditions. Another possible reason could be that there might be a lower expectation within the group studied. Studies suggest that adverse childhood experiences is a risk factor for homelessness [15] and it may be the case that individuals who experienced adverse childhood experiences may have low expectations. The results of previous studies investigating the EQ-VAS among homeless somewhat vary, showing both low [8] and high [9] EQ-VAS scores. A Swedish study compared the EQ-5D-3L health status among homeless people in Stockholm in 2006 and 2018 [8], observing a mean EQ-VAS score of 53.4 among the homeless in 2018, which was not significantly different from the measured value a decade before. However, the EQ-VAS score determined by Van Rüth et al. [9] was 75.34 (SD: 22.23, ranging from 1 to 100). One reason for the inconclusive evidence could be that the study design (e.g., single center vs. multicenter studies; differences in sample size and period of data collection) varies between the studies.

Generally, as homeless individuals are often neglected in research, only scarce data exists concerning the HRQoL of individuals experiencing homelessness. Thus, our findings are rather difficult to compare with prior studies conducted among homeless individuals. However, the age association – which was found in our study—is in line with previous research [9]. It is also consistent with expectation as with age functional abilities decrease. This can result in an impairment of mobility and more pain/discomfort, as reported here.

The significantly higher likelihood of having problems in anxiety/depression in women compared to men in this study is consistent with the higher prevalence among women in the general adult population in Germany [16]. A study investigating depression during the COVID-19 pandemic amongst residents of homeless shelters in France also found that being female is associated with a higher likelihood of depression [5].

Being married and living together was associated with higher EQ-VAS as well as a higher EQ-5D-5L Index in our study. This is also well in line with prior research based on general population samples before the pandemic [17].

The negative correlation between existing health insurance and problems with mobility reported in our study could be the result of better access to medical treatment. Thus, individuals with health insurance may have more regular physician visits to treat their ailments. In contrast, having a health insurance was also significantly associated with a higher likelihood of having problems with anxiety/depression. This could be justified by the awareness of one`s physical state when in medical treatment, as in this cohort the prevalence of multimorbidity and chronic diseases is extremely high compared to the general population—as an Australian study investigating multimorbidity among people experiencing homelessness also found (75% and 68%) [18].

Some strengths and limitations are worth mentioning. There are only very few studies examining the health-related quality of life among homeless individuals, particularly during the COVID-19 pandemic. More precisely, this is the second study reporting on the HRQoL of homeless individuals during the COVID-19 pandemic using the EQ-5D questionnaire. Thus, this study markedly adds to the current knowledge concerning HRQoL among homeless individuals. One strength of the present study is its large sample. Moreover, this study is multi-centered, which adds to the generalizability. Another strength is the EQ-5D questionnaire, which is a common and valid tool to assess the HRQoL. Furthermore, our survey collected data regarding a special, vulnerable group, which is usually difficult to reach, because a considerable percentage of the suitable individuals might be aversed to participate due to illnesses, cognitive impairments, and general distrust of official institutions.

In contrast, some weaknesses should be noted. Although the overall sample was large, the proportion of women was relatively small, which should be acknowledged when interpreting the results for women separately. The proportion of homeless women in our sample may reflect the actual proportion of homeless women in Germany [1]. In addition, it may be difficult to generalize our findings. As our study only included homeless individuals who use certain social services, some selection bias might be present, as it may be difficult to transfer our findings to homeless individuals who do not make use of those institutional accommodations. Furthermore, participation bias and exclusion criteria might have led to the underrepresentation of individuals with bad physical and mental conditions. Also, the cross-sectional character makes it difficult to clarify the directionality between the factors examined in this study. As we contacted all homeless aid institutions in the respective regions and those who agreed to participate took part in the study, it was neither a random selection nor based on fixed criteria. Moreover, some aid facilities could not be included and a response rate could not be calculated. Thus, the generalizability of our results should be treated with caution. Furthermore, it should be noted that the questionnaires were translated by native speakers and therefore the gold standard of forward and backward translation was not carried out. Additionally, it may be of interest to examine the association between income (or amount of donations) and HRQoL among homeless individuals in future studies.

Compared to the general adult population [19] during the COVID-19 pandemic, our study findings showed a quite high HRQoL among homeless individuals during the COVID-19 pandemic. Some determinants of HRQoL were identified (e.g., age or marital status).

This knowledge is important to characterize homeless individuals at risk for low HRQoL. In turn, this is important because HRQoL and subjective well-being affect morbidity and mortality [10, 11]. However, longitudinal studies are required to confirm our findings.