Morbidity and housing status 10 years after shelter use—follow-up of homeless men in Helsinki, Finland

Morbidity and housing status 10 years after shelter use—follow-up of homeless men in Helsinki,... Abstract Background Homelessness is associated with increased mortality, morbidity and social difficulties and periods of homelessness are sometimes prolonged or repeated. However, there are no long-term follow-up studies focusing upon housing status among homeless people. The aim of this study was to examine morbidity and housing outcomes and to identify factors predicting being independently housed 10 years after shelter use. Methods By combining data from several registers we followed all 552 homeless men who stayed in shelter in Helsinki during 2004 and determined their housing situation and morbidity 10 years later. Their situation was compared with an age-matched control group from the general population (N = 946). Using logistic regression analysis, we assessed the predictive effects of socioeconomic factors and health service use at baseline on becoming independently housed. Results By the end of the follow-up 52.0% of the formerly homeless study group had died, compared with 14.6% of the controls. At 10 years, 6.0% were independently housed, 37.5% lived in supported housing and 4.5% were still or again homeless. Psychiatric disorders, including substance use disorder, were present in 77.5% of the homeless, compared with 16.1% among the controls. Being married (OR 8.3, 95% CI 3.0 to 23.2) and having less than four shelter nights in year 2004 (OR 9.1, 95% CI 2.7 to 30.8) strongly predicted being independently housed 10 years later. Conclusions Homeless staying in shelters have high mortality and morbidity and most of those surviving, are in need of support in their everyday lives even years after the shelter period. Introduction Homelessness is associated with chronic disease and social problems that hamper entering the housing market and staying there.1,2 There is evidence that while most homeless persons at any given time are temporarily homeless, a minority experience repeated or prolonged, long-term homelessness.3,4 A study on shelter users in Helsinki metropolitan area showed that 55% of them had been homeless for over a year and 23% for over 5 years.5 Studies suggest that chronic and repeated homelessness is associated with poor health outcomes.6 Tracing homeless people for interview-based follow-up studies is difficult: most studies have rather short follow-up periods.6–12 Hence, the long-term outcome of homelessness in terms of housing situation is not known. Clinical experience and smaller studies have shown that prolonged and repeated homelessness is common and that many previously homeless people require different levels of support also after being housed.2,13,14 Different efforts such as Housing First, an evidence-based intervention offering permanent supported housing for highly-vulnerable homeless individuals without requiring clients to meet preconditions related to treatment completion or treatment success,15 have been introduced in North America and several European countries to address the housing needs of homeless.16–18 In 2008–15, the Finnish government implemented large investment and cooperation projects (PAAVO I and II) to eradicate long-term homelessness. Shelter type accommodation was replaced with housing units based on Housing First principles. The Finnish model is a slightly adjusted version of the American Housing first model in order to better suit the Finnish setting.19 Finland is among the few countries where long-term homelessness decreased between 2008–15.19 However, despite efforts to provide permanent housing for long-term homeless, homelessness also in Finland is often renewed and suitable housing not always found. A Finnish study that followed a cohort of shelter users between 2009 and 2012 found that during the study period very few got rental contracts in the private housing market and the majority moved to some form of supported housing.2 One fourth of the housed persons became homeless again during the study period and 12% remained homeless during the whole period.2 Both somatic and psychiatric morbidity is common among homeless.1,8,20,21 In our previous study examining the primary health care records of homeless shelter users in Helsinki metropolitan area we found that 89% of them had some mental health disorder, including Substance Use Disorders (SUD).5 The mortality of homeless in Helsinki is also high compared to mortality rates in North America.22 Previous studies have found that younger age,10,11 female sex,7,23 income,7,10 current or recent employment,10 family support and size of social network,7,10,23 absence of SUD,7,10 absence of arrests10 and access to subsidized housing7 predict being stably housed or exiting homelessness in a shorter perspective of 1–3 years. All of the above studies have used exiting homelessness as an outcome, not looking at the type of housing and its level of support. By combining information from several registers, using Personal Identification Codes (PICS) for exact linkages, we were able to avoid lost cases due to persons declining to participate or lost to follow-up that interview-based studies face. We traced and assessed the housing situation and level of support in a representative sample of 552 homeless men at 10 years after shelter use. The aims of the study were: First, to assess how many of the homeless entered the housing market in terms of being independently housed after 10 years, and how many were still homeless or staying in supported housing. Second, to assess the morbidity of the homeless depending on the housing situation at the end of follow-up and compared to the general population. Third, to estimate the effects of educational attainment, marital status, time in shelter, employment and health care use on being independently housed. Methods Study population The study population consists of a total sample of all 624 homeless men that stayed in Herttoniemi shelter in Helsinki during 2004. Their PICs were retrieved from the shelter register, which contains information on the total amount of nights each person spent in the shelter during 2004, but not their exact dates. Thus, the start of follow-up was set at 1 July 2004 for everyone (i.e. halfway through the year). The Population Register Center provided a control group in a 2 : 1 ratio (N = 1248), matching criteria being place of domicile in 2004 and month of birth. Three persons in the control group were also cases in the homeless sample and were therefore excluded. We also excluded those who emigrated or deceased before the start of follow-up (N = 7 among homeless and N = 5 among controls group). Follow-up continued until 31 December 2014. We determined the housing status at the end of follow-up and because we only had access to the registers of City of Helsinki the 65 (10.4%) homeless and 294 (23.6%) control persons who had migrated out of Helsinki before the end of the follow-up were excluded, giving a final study sample of 552 in the study group and 946 control persons (flow-chart presented in Supplementary Appendix S1). Registers used and housing categories Registers at Statistics Finland were used for information on socio-economic status, education and time of death. Marital status at baseline and place of domicile at the end of follow-up were obtained from the Population Register Centre. The Care Register for Health Care, kept by National Institute for Health and Welfare and including data on all hospital inpatient and outpatient episodes, was used to get data on health service use prior to baseline and diagnoses during follow-up to determine morbidity. The housing situation at the end of follow-up was determined using the Social Service Client Register and Primary Health Care Register from the City of Helsinki. PICs were used for register linkages. We defined the housing status at the end of 2014 for all cases and controls and grouped them into those (1) independently housed in the free rental market or having a council apartment or owned home, (2) living in supported housing, (3) still homeless (including people staying temporarily with friends and family, in shelters and in institutions without a permanent address) and (4) deceased. The supported housing services (group 2) were further grouped into (2a) nursing homes, (2b) housing-first unit homes with staff present 24/7 and (2c) low intensity support housing. The housing categories and method used for defining the housing status at the end of follow-up are presented in Supplementary Appendix S1. Psychiatric and somatic morbidity We used the diagnoses in the Care Register for Health Care between 2004 and 2014 to gather information about morbidity. The register contains physician-given diagnoses as ICD-10 codes for each contact with specialized health care in any Finnish hospital. The prevalence of psychiatric morbidity was defined as presence of an ICD-10 code between F10-F69 as primary or secondary diagnosis during the study period. Similarly, the prevalence of the following psychiatric conditions was calculated: psychotic disorders (F20–F29), SUD (F10–F19) and dual diagnosis including both psychotic disorder and SUD. Somatic morbidity was assessed using the Charlson’s Comorbidity Index (CCI).24 The Index includes seventeen chronic conditions, each assigned a weight of 1, 2, 3 or 6, where HIV and metastatic carcinoma give the highest weight of 6 and less severe conditions such as mild liver disease and diabetes without complications give a score of 1. All diagnoses present in the Care Register for Health Care during the study period were used to calculate the CCI. Predictors of interest To identify predictors of independent housing, we examined the following socio-demographic factors at baseline: age (continuous variable), education (higher than the 9 years basic level vs. not), employment (employed or student vs. unemployed) and marital status (married/in a registered partnership vs. not) and the total number of nights in shelters during year 2004 (as 1–3 nights vs. at least four nights). We also controlled for health care contacts with specialized care 6 months prior to baseline (1 January 2004–30 June 2004) to examine the predictive effect of health care use on being independently housed after 10 years. The health care contacts separately examined were emergency-room visits, outpatient visits at specialist care level, hospitalizations and inpatient detoxification treatment. All predictors apart from age were used as dichotomous factors in the regression analysis (i.e. contacts vs. no contact for the health care variables). Statistical analysis The characteristics of the study population are presented as means with standard deviations (SD) and counts with percentages. Statistical comparisons between the groups were performed by t-test or bootstrap type t-test, Chi-square test or Fisher-Freeman-Halton test when appropriate. To determine the predictive effects of our chosen variables on being independently housed 10 years later, we performed univariate logistic regression analysis on the homeless study group for all variables separately and multivariate forward stepwise regression analysis including all variables. In the analysis, those independently housed at the end of follow-up were compared to all other end-point groups combined. All analyses were performed using STATA 15.0 (StataCorp LP, College Station, TX). Results The mean age of the homeless at baseline was 49.4 years (range 21–78 years; table 1). The proportion of single and unemployed and persons on pension was high among the homeless and they had significantly more health care service use, compared with the control group (P < 0.05). Of the homeless, 51.5% had stayed 1–3 nights in shelter in 2004, while the rest of them had four or more nights (up to 366 nights). Table 1 Baseline characteristics of study group (homeless men) and controls in Helsinki, Finland, year 2004 Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Table 1 Baseline characteristics of study group (homeless men) and controls in Helsinki, Finland, year 2004 Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Housing status after 10 years By the end of the follow-up over half (52%) of the study group had died, compared to 15% in the control group (P < 0.001) (figure 1). In the study group 6% were independently housed, 38% lived in supported housing and 5% were homeless at 10 years. The corresponding numbers in the control group were 82%, 3% and 0.1%, respectively. Of the 25 persons in the study group, who were homeless after 10 years, 9 stayed in shelters, 7 in correctional institutions or psychiatric hospitals and 3 with friends or family and in six cases their place of stay could not be verified. One person in the control group was homeless and staying in shelters at the end of the follow-up. Figure 1 View largeDownload slide Status at 10 years after homeless shelter use in the study group (N = 552) and their general population controls (N = 946), Helsinki, Finland, percentages with 95% confidence intervals Figure 1 View largeDownload slide Status at 10 years after homeless shelter use in the study group (N = 552) and their general population controls (N = 946), Helsinki, Finland, percentages with 95% confidence intervals Of the 207 previously homeless men that lived in supported housing, 84 persons (15% of the total sample) stayed in nursing homes, 77 (14%) in housing first units and 46 (8%) in low intensity supported housing (Supplementary Appendix S1). Correspondingly, in the control group, 18 persons (2% of the control group) stayed in nursing homes, 5 (0.5%) in housing first units and 9 (1%) in low intensity supported housing. Psychiatric and somatic morbidity The 10-year prevalence of psychiatric morbidity in the study group was 78% (95% CI 74–81%) compared with 16% (95% CI 14–18%) in the control group (Supplementary Appendix S2). SUD was the most prevalent psychiatric diagnosis in the study group; 75% of the previously homeless and 8% of the controls had a SUD. Further, psychotic disorders and dual diagnosis presented more often in the study group compared to controls (13% and 12% in the study group and 2% and 0.5% among the controls, respectively). The somatic morbidity was also significantly higher in the study group with a mean CCI of 1.3 (SD 2.0) in the study group, compared with 0.7 (SD 1.6) among the controls, (P < 0.001). For those in supported housing at the end of follow-up, the CCI index and psychiatric morbidity increased with the level of support among the previously homeless (Supplementary Appendix S2). The prevalence of psychiatric morbidity among those independently housed and those in nursing homes was higher in the study group compared with the control group (P < 0, 001, respectively). We also compared the prevalence of SUD, psychotic disorders and dual diagnosis in the homeless group according to the housing status at end of follow-up (figure 2). The prevalence of SUD varied significantly depending on the housing status at the end of follow-up, being lowest among those independently housed (55%) and highest among those homeless (89%) at 10 years after shelter use (P = 0.012). The differences in the prevalence of psychotic disorders and dual diagnosis between the different housing situations at the end of follow-up was even bigger (P < 0.001 in both groups), and also here, the prevalence figures are highest among those homeless with 40% having psychosis and dual diagnosis compared to 3% among those living independently at the end of follow-up. Figure 2 View largeDownload slide The 10-year prevalence of SUD, psychotic disorders and dual diagnosis among previously homeless men (N = 552) by housing status after 10-year follow-up, Helsinki, Finland, percentage Figure 2 View largeDownload slide The 10-year prevalence of SUD, psychotic disorders and dual diagnosis among previously homeless men (N = 552) by housing status after 10-year follow-up, Helsinki, Finland, percentage The predictive effects of socioeconomic factors and health service use among shelter users on being independently housed 10 years later In the multivariate analysis assessing factors predicting independent housing 10 years after shelter use within the study group (including those deceased during follow-up) we found that being married (OR 8.34, 95% CI 3.01–23.15) and staying only briefly in shelter (OR 9.08, 95% CI 2.68–30.76) were strongly associated with independent housing (table 2). Having no emergency room visits (OR 3.55, 95% CI 1.17–10.77) and having more than basic level education (OR 2.33, 95% CI: 1.09–4.99) also increased the odds of being independently housed at the end of follow up. The results remained when those deceased were removed from the analysis (data not shown). Table 2 Factors predicting independent housing at 10 years after stay in shelter among homeless men (N = 552) in Helsinki, Finland Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) a Forward stepwise selection including all variables. Only those variables are shown which entered the model. Table 2 Factors predicting independent housing at 10 years after stay in shelter among homeless men (N = 552) in Helsinki, Finland Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) a Forward stepwise selection including all variables. Only those variables are shown which entered the model. Discussion We studied the housing situation and morbidity at 10 years after shelter use and found that besides having high mortality as shown in our previous study,22 only 6% of the previously homeless were independently housed at the end of follow-up. Further, regardless of the housing situation the psychiatric morbidity in the homeless cohort was high. The somatic morbidity was also higher in the study group compared with the controls. Being married and staying only briefly in shelters were strongly associated with independent housing after 10 years. Previous studies with follow-ups of 1–3 years have found that, depending on the setting and methods, 19–61% are still homeless at the end of follow-up.6,–12 The fact that in our study only 5% (9% of those alive) were still homeless after 10 years, can be seen as a result and success of the efforts to eliminate long-term homelessness that have been implemented in the last decade in Finland.19 Considering for instance findings from Stockholm, a society with similar social support system, where 75% of those alive were still homeless 5 years later,14 it is not unlikely that many more also in our sample would still be homeless had it not been for the intensive housing efforts in the City of Helsinki. The fact that only 6% (12% of those alive) in our study were independently housed at the end of follow-up, however, supports findings from previous studies2 showing that very few homeless shelter users manage to enter the free housing markets. We can conclude that a variety of supported housing services are needed to end homelessness in this highly vulnerable population with complex needs. The prevalence of psychiatric diagnoses and especially SUD in our cohort was high compared with previous studies. Swedish and Danish studies using hospital discharge registers to define prevalence of mental disorders among homeless reported 42% and 49% rates of SUD, respectively,20,21 compared to 75% in our cohort. This finding is in line with the high prevalence of SUD in Finnish shelter users5 and offers a further explanation to the high mortality rate in the cohort. Our finding that the psychiatric morbidity was highest among those still or again homeless indicates that there still is a group of homeless with multiple needs for which the housing services on offer are not providing adequate or enough support. Several factors, such as being married and staying only briefly in shelters were associated with independent housing. Interestingly, these same factors were not associated with mortality in a previous study of the same cohort.22 We can speculate that death, especially when related to SUD is more random and difficult to predict while factors that are associated with housing skills and not being institutionalized, such as being married and staying only briefly in shelters are more easily identified among homeless shelter users. Previous studies have found that the size of social network predicts exiting homelessness7,10,23 and this is supported by our findings: those that find other accommodation within a few days have usually solved their housing situation by themselves with the help of their social network. There are some limitations to this study. First, we only had access to the social service and primary health care records from the City of Helsinki and consequently we had to exclude the 10% of the cohort that had migrated. It can be speculated that some of them had gained independent housing somewhere else. However, considering that the morbidity of the excluded group did not differ from those in supported housing (data not shown) it is likely that the main part of the men excluded from the analysis also lived in some form of supported housing in an other municipality. Second, we only had access to the information that could be extracted from the registers and there are probably many other factors associated with independent housing including length of homelessness prior to shelter use, housing skills, family and social networks that we could not examine. Third, being limited to the diagnosis registered in specialized health care, there is most likely a problem of under-diagnosis, especially in the deceased group, where many died of external causes already in the first years of follow-up.22 A longer follow-up prior to baseline would have improved our morbidity data. Using a registered-based approach to study the homeless situation also offers several strengths, and thus we were able to, for the first time, describe the housing situation of a cohort of homeless men after 10 years, without the problems of cases lost to follow-up and persons declining to participate that interview-based studies struggle with. Having a matched control group further validates the results. The long perspective of the study offers new valuable information to policy makers and care workers that deal with homeless. In conclusion, we showed that very few of the homeless that survive manage to enter the free housing market and that this population requires supported housing services with different levels of support. Further research is needed to study the permanence of the housing solutions and their effect on quality of life, mortality and service use. Ethical consideration Ethics approval for this study was granted by the ethics committee of the Hospital district of Helsinki and Uusimaa (HUS), and research permits were obtained from all the register keepers in the study. According to the Finnish legislation no informed consent from participants was needed because only existing register data were used, the registered persons were not contacted and the data were analyzed anonymously. Supplementary data Supplementary data are available at EURPUB online. Acknowledgements The authors thank the staff at the City of Helsinki for helping to gather the data, the National Institute for Health and Welfare for technical support and all the funders of this study. Funding The study has been funded by Samfundet Folkhälsan, Finnish Medical Association (Finska läkaresällskapet), the City of Helsinki and the Wilhelm and Else Stockmann’s foundation. The funders have not played any role in the design of the study and collection, analysis and interpretation of data or in writing the manuscript. Conflicts of interest: None declared. Key points While it is known that most homeless staying in shelters are there only briefly and then move on, the long-term prognosis of homelessness in terms of housing is not previously known. We found that after 10 years over half of the shelter users had died, only 6% had independent housing, 38% lived in supported housing and 5% were still or again homeless, showing that most of those surviving are in need of support in their everyday lives even years after the shelter period. Among the formerly homeless, psychiatric disorders, including substance use disorder, were present in 78% of the study group, compared with 16% in the general population. Among those still alive at the end of follow-up the psychiatric morbidity was highest among those still or again homeless followed by those in intensively supported housing. Being married and staying only briefly in shelters is strongly associated with being independently housed 10 years later. References 1 Fazel S , Geddes JR , Kushel M . The health of homeless people in high-income countries: descriptive epidemiology, health consequences, and clinical and policy recommendations . Lancet 2014 ; 384 : 1529 – 40 . Google Scholar CrossRef Search ADS PubMed 2 Sunikka S . Pathways to homelessness of people who stayed overnight in hietsu–Mobility between homelessness and the dwelling population. (Professional Licentiate thesis in Finnish). University of Helsinki, 2016 . Available at: http://www.sosnet.fi/loader.aspx? id=971cb92e-9f9b-4f46-9cc7-436e4ef7c966 (16 November 2017, date last accessed). 3 Kuhn R , Culhane DP . Applying cluster analysis to test a typology of homelessness by pattern of shelter utilization: results from the analysis of administrative data . Am J Community Psychol 1998 ; 26 : 207 – 32 . Google Scholar CrossRef Search ADS PubMed 4 Busch-Geertsema V , Benjaminsen L , Filipovic Hrast M , Pleace N . Extent and Profile of Homelessness in European Member States: A Statistical Update, 2014 . Available at: http://housingfirstguide.eu/website/wp-content/uploads/2016/04/feantsa-studies_04-web2.pdf (16 November 2017, date last accessed). 5 Stenius-Ayoade A , Haaramo P , Erkkila E , et al. Mental disorders and the use of primary health care services among homeless shelter users in the Helsinki Metropolitan Area, Finland . BMC Health Serv Res 2017 ; 17 : 428 – 017 , 2372–3. Google Scholar CrossRef Search ADS PubMed 6 Kertesz SG , Larson MJ , Horton NJ , et al. Homeless chronicity and health-related quality of life trajectories among adults with addictions . Med Care 2005 ; 43 : 574 – 85 . Google Scholar CrossRef Search ADS PubMed 7 Aubry T , Duhoux A , Klodawsky F , et al. A longitudinal study of predictors of housing stability, housing quality, and mental health functioning among single homeless individuals staying in emergency shelters . Am J Community Psychol 2016 ; 58 : 123 – 35 . Google Scholar CrossRef Search ADS PubMed 8 Schanzer B , Dominguez B , Shrout PE , Caton CL . Homelessness, health status, and health care use . Am J Public Health 2007 ; 97 : 464 – 9 . Google Scholar CrossRef Search ADS PubMed 9 Spicer B , Smith DI , Conroy E , et al. Mental illness and housing outcomes among a sample of homeless men in an Australian urban centre . Aust N Z J Psychiatry 2015 ; 49 : 471 – 80 . Google Scholar CrossRef Search ADS PubMed 10 Caton CL , Dominguez B , Schanzer B , et al. Risk factors for long-term homelessness: findings from a longitudinal study of first-time homeless single adults . Am J Public Health 2005 ; 95 : 1753 – 9 . Google Scholar CrossRef Search ADS PubMed 11 Johnson G , Scutella R , Tseng Y , et al. Entries and Exits from Homelessness: A Dynamic Analysis of the Relationship between Structural Conditions and Individual Characteristics, 2015 : 1 – 67 . Available at: https://www.ahuri.edu.au/__data/assets/pdf_file/0013/2155/AHURI_Final_Report_No248_Entries-and-exits-from-homelessness.pdf (16 November 2017, date last accessed). 12 Fichter MM , Quadflieg N . Three year course and outcome of mental illness in homeless men: a prospective longitudinal study based on a representative sample . Eur Arch Psychiatry Clin Neurosci 2005 ; 255 : 111 – 20 . Google Scholar CrossRef Search ADS PubMed 13 McQuistion HL , Gorroochurn P , Hsu E , Caton CL . Risk factors associated with recurrent homelessness after a first homeless episode . Community Ment Health J 2014 ; 50 : 505 – 13 . Google Scholar CrossRef Search ADS PubMed 14 Beijer U , Andreasson A , Agren G , Fugelstad A . Mortality, mental disorders and addiction: a 5-year follow-up of 82 homeless men in stockholm . Nord J Psychiatry 2007 ; 61 : 363 – 8 . Google Scholar CrossRef Search ADS PubMed 15 Tsemberis S . Housing First: The Pathways Model to End Homelessness for People with Mental Illness and Addiction . Center City (MN ): Hazelden Press , 2010 . 16 Aubry T , Tsemberis S , Adair CE , et al. One-year outcomes of a randomized controlled trial of housing first with ACT in five Canadian cities . Psychiatr Serv 2015 ; 66 : 463 – 9 . Google Scholar CrossRef Search ADS PubMed 17 Kertesz SG , Austin EL , Holmes SK , et al. Housing first on a large scale: fidelity strengths and challenges in the VA’s HUD-VASH program . Psychol Serv 2017 ; 14 : 118 – 28 . Google Scholar CrossRef Search ADS PubMed 18 Busch-Geertsema V . Housing first Europe–results of a European social experimentation project . Eur J Homelessness 2014 ; 8 : 13 – 28 . 19 Pleace N , Culhane D , Granfelt R , Knutagård M . The Finnish Homelessness Strategy–an International Review. 2015-02-10. Available at: https://helda.helsinki.fi/handle/10138/153258 (16 November 2017, date last accessed). 20 Nielsen SF , Hjorthoj CR , Erlangsen A , Nordentoft M . Psychiatric disorders and mortality among people in homeless shelters in Denmark: a nationwide register-based cohort study . Lancet 2011 ; 377 : 2205 – 14 . Google Scholar CrossRef Search ADS PubMed 21 Beijer U , Andreasson S . Gender, hospitalization and mental disorders among homeless people compared with the general population in Stockholm . Eur J Public Health 2010 ; 20 : 511 – 6 . Google Scholar CrossRef Search ADS PubMed 22 Stenius-Ayoade A , Haaramo P , Kautiainen H , et al. Mortality and causes of death among homeless in Finland: a 10-year follow-up study . J Epidemiol Community Health 2017 ; 71 : 841 – 8 . Google Scholar CrossRef Search ADS 23 Zlotnick C , Tam T , Robertson MJ . Disaffiliation, substance use, and exiting homelessness . Subst Use Misuse 2003 ; 38 : 577 – 99 . Google Scholar CrossRef Search ADS PubMed 24 Charlson ME , Pompei P , Ales KL , MacKenzie CR . A new method of classifying prognostic comorbidity in longitudinal studies: development and validation . J Chronic Dis 1987 ; 40 : 373 – 83 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The European Journal of Public Health Oxford University Press

Morbidity and housing status 10 years after shelter use—follow-up of homeless men in Helsinki, Finland

Loading next page...
 
/lp/ou_press/morbidity-and-housing-status-10-years-after-shelter-use-follow-up-of-gRFC1ElwO7
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
ISSN
1101-1262
eISSN
1464-360X
D.O.I.
10.1093/eurpub/cky038
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Homelessness is associated with increased mortality, morbidity and social difficulties and periods of homelessness are sometimes prolonged or repeated. However, there are no long-term follow-up studies focusing upon housing status among homeless people. The aim of this study was to examine morbidity and housing outcomes and to identify factors predicting being independently housed 10 years after shelter use. Methods By combining data from several registers we followed all 552 homeless men who stayed in shelter in Helsinki during 2004 and determined their housing situation and morbidity 10 years later. Their situation was compared with an age-matched control group from the general population (N = 946). Using logistic regression analysis, we assessed the predictive effects of socioeconomic factors and health service use at baseline on becoming independently housed. Results By the end of the follow-up 52.0% of the formerly homeless study group had died, compared with 14.6% of the controls. At 10 years, 6.0% were independently housed, 37.5% lived in supported housing and 4.5% were still or again homeless. Psychiatric disorders, including substance use disorder, were present in 77.5% of the homeless, compared with 16.1% among the controls. Being married (OR 8.3, 95% CI 3.0 to 23.2) and having less than four shelter nights in year 2004 (OR 9.1, 95% CI 2.7 to 30.8) strongly predicted being independently housed 10 years later. Conclusions Homeless staying in shelters have high mortality and morbidity and most of those surviving, are in need of support in their everyday lives even years after the shelter period. Introduction Homelessness is associated with chronic disease and social problems that hamper entering the housing market and staying there.1,2 There is evidence that while most homeless persons at any given time are temporarily homeless, a minority experience repeated or prolonged, long-term homelessness.3,4 A study on shelter users in Helsinki metropolitan area showed that 55% of them had been homeless for over a year and 23% for over 5 years.5 Studies suggest that chronic and repeated homelessness is associated with poor health outcomes.6 Tracing homeless people for interview-based follow-up studies is difficult: most studies have rather short follow-up periods.6–12 Hence, the long-term outcome of homelessness in terms of housing situation is not known. Clinical experience and smaller studies have shown that prolonged and repeated homelessness is common and that many previously homeless people require different levels of support also after being housed.2,13,14 Different efforts such as Housing First, an evidence-based intervention offering permanent supported housing for highly-vulnerable homeless individuals without requiring clients to meet preconditions related to treatment completion or treatment success,15 have been introduced in North America and several European countries to address the housing needs of homeless.16–18 In 2008–15, the Finnish government implemented large investment and cooperation projects (PAAVO I and II) to eradicate long-term homelessness. Shelter type accommodation was replaced with housing units based on Housing First principles. The Finnish model is a slightly adjusted version of the American Housing first model in order to better suit the Finnish setting.19 Finland is among the few countries where long-term homelessness decreased between 2008–15.19 However, despite efforts to provide permanent housing for long-term homeless, homelessness also in Finland is often renewed and suitable housing not always found. A Finnish study that followed a cohort of shelter users between 2009 and 2012 found that during the study period very few got rental contracts in the private housing market and the majority moved to some form of supported housing.2 One fourth of the housed persons became homeless again during the study period and 12% remained homeless during the whole period.2 Both somatic and psychiatric morbidity is common among homeless.1,8,20,21 In our previous study examining the primary health care records of homeless shelter users in Helsinki metropolitan area we found that 89% of them had some mental health disorder, including Substance Use Disorders (SUD).5 The mortality of homeless in Helsinki is also high compared to mortality rates in North America.22 Previous studies have found that younger age,10,11 female sex,7,23 income,7,10 current or recent employment,10 family support and size of social network,7,10,23 absence of SUD,7,10 absence of arrests10 and access to subsidized housing7 predict being stably housed or exiting homelessness in a shorter perspective of 1–3 years. All of the above studies have used exiting homelessness as an outcome, not looking at the type of housing and its level of support. By combining information from several registers, using Personal Identification Codes (PICS) for exact linkages, we were able to avoid lost cases due to persons declining to participate or lost to follow-up that interview-based studies face. We traced and assessed the housing situation and level of support in a representative sample of 552 homeless men at 10 years after shelter use. The aims of the study were: First, to assess how many of the homeless entered the housing market in terms of being independently housed after 10 years, and how many were still homeless or staying in supported housing. Second, to assess the morbidity of the homeless depending on the housing situation at the end of follow-up and compared to the general population. Third, to estimate the effects of educational attainment, marital status, time in shelter, employment and health care use on being independently housed. Methods Study population The study population consists of a total sample of all 624 homeless men that stayed in Herttoniemi shelter in Helsinki during 2004. Their PICs were retrieved from the shelter register, which contains information on the total amount of nights each person spent in the shelter during 2004, but not their exact dates. Thus, the start of follow-up was set at 1 July 2004 for everyone (i.e. halfway through the year). The Population Register Center provided a control group in a 2 : 1 ratio (N = 1248), matching criteria being place of domicile in 2004 and month of birth. Three persons in the control group were also cases in the homeless sample and were therefore excluded. We also excluded those who emigrated or deceased before the start of follow-up (N = 7 among homeless and N = 5 among controls group). Follow-up continued until 31 December 2014. We determined the housing status at the end of follow-up and because we only had access to the registers of City of Helsinki the 65 (10.4%) homeless and 294 (23.6%) control persons who had migrated out of Helsinki before the end of the follow-up were excluded, giving a final study sample of 552 in the study group and 946 control persons (flow-chart presented in Supplementary Appendix S1). Registers used and housing categories Registers at Statistics Finland were used for information on socio-economic status, education and time of death. Marital status at baseline and place of domicile at the end of follow-up were obtained from the Population Register Centre. The Care Register for Health Care, kept by National Institute for Health and Welfare and including data on all hospital inpatient and outpatient episodes, was used to get data on health service use prior to baseline and diagnoses during follow-up to determine morbidity. The housing situation at the end of follow-up was determined using the Social Service Client Register and Primary Health Care Register from the City of Helsinki. PICs were used for register linkages. We defined the housing status at the end of 2014 for all cases and controls and grouped them into those (1) independently housed in the free rental market or having a council apartment or owned home, (2) living in supported housing, (3) still homeless (including people staying temporarily with friends and family, in shelters and in institutions without a permanent address) and (4) deceased. The supported housing services (group 2) were further grouped into (2a) nursing homes, (2b) housing-first unit homes with staff present 24/7 and (2c) low intensity support housing. The housing categories and method used for defining the housing status at the end of follow-up are presented in Supplementary Appendix S1. Psychiatric and somatic morbidity We used the diagnoses in the Care Register for Health Care between 2004 and 2014 to gather information about morbidity. The register contains physician-given diagnoses as ICD-10 codes for each contact with specialized health care in any Finnish hospital. The prevalence of psychiatric morbidity was defined as presence of an ICD-10 code between F10-F69 as primary or secondary diagnosis during the study period. Similarly, the prevalence of the following psychiatric conditions was calculated: psychotic disorders (F20–F29), SUD (F10–F19) and dual diagnosis including both psychotic disorder and SUD. Somatic morbidity was assessed using the Charlson’s Comorbidity Index (CCI).24 The Index includes seventeen chronic conditions, each assigned a weight of 1, 2, 3 or 6, where HIV and metastatic carcinoma give the highest weight of 6 and less severe conditions such as mild liver disease and diabetes without complications give a score of 1. All diagnoses present in the Care Register for Health Care during the study period were used to calculate the CCI. Predictors of interest To identify predictors of independent housing, we examined the following socio-demographic factors at baseline: age (continuous variable), education (higher than the 9 years basic level vs. not), employment (employed or student vs. unemployed) and marital status (married/in a registered partnership vs. not) and the total number of nights in shelters during year 2004 (as 1–3 nights vs. at least four nights). We also controlled for health care contacts with specialized care 6 months prior to baseline (1 January 2004–30 June 2004) to examine the predictive effect of health care use on being independently housed after 10 years. The health care contacts separately examined were emergency-room visits, outpatient visits at specialist care level, hospitalizations and inpatient detoxification treatment. All predictors apart from age were used as dichotomous factors in the regression analysis (i.e. contacts vs. no contact for the health care variables). Statistical analysis The characteristics of the study population are presented as means with standard deviations (SD) and counts with percentages. Statistical comparisons between the groups were performed by t-test or bootstrap type t-test, Chi-square test or Fisher-Freeman-Halton test when appropriate. To determine the predictive effects of our chosen variables on being independently housed 10 years later, we performed univariate logistic regression analysis on the homeless study group for all variables separately and multivariate forward stepwise regression analysis including all variables. In the analysis, those independently housed at the end of follow-up were compared to all other end-point groups combined. All analyses were performed using STATA 15.0 (StataCorp LP, College Station, TX). Results The mean age of the homeless at baseline was 49.4 years (range 21–78 years; table 1). The proportion of single and unemployed and persons on pension was high among the homeless and they had significantly more health care service use, compared with the control group (P < 0.05). Of the homeless, 51.5% had stayed 1–3 nights in shelter in 2004, while the rest of them had four or more nights (up to 366 nights). Table 1 Baseline characteristics of study group (homeless men) and controls in Helsinki, Finland, year 2004 Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Table 1 Baseline characteristics of study group (homeless men) and controls in Helsinki, Finland, year 2004 Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Study group Controls N = 552 N = 946 Age at baseline, mean (years) (SD) 49.4 (10.8) 50.3 (10.7) Age group, n (%)     18–34 years 54 (9.8) 81 (8.6)     35–49 years 210 (38.0) 336 (35.5)     50–64 years 240 (43.5) 433 (45.8)     65–78 years 48 (8.7) 96 (10.2) Marital status, n (%)     Married or in registered partnership 26 (4.7) 496 (52.4)     Years of education, mean (SD) 10.2 (1.8) 12.7 (3.3)     Only basic level, n (%) 354 (64.1) 302 (31.9) Employment status     Employed, n (%) 28 (5.1) 592 (62.6)     Nights in shelter 2004, mean (SD) 30 (64)     1–3 nights, n (%) 284 (51.4)     4–10 nights, n (%) 88 (15.9)     11–30 nights, n (%) 57 (10.3)     31–90 nights, n (%) 64 (11.6)     91–366 nights, n (%) 59 (10.7) Persons with health care contacts 6 months prior to baseline     Hospitalizations, n (%) 171 (31.0) 49 (5.2)     Outpatient visits to hospitals, n (%) 106 (19.2) 133 (14.1)     Emergency room visits, n (%) 194 (35.1) 53 (5.6)     Detoxification treatments, n (%) 157 (28.4) 7 (0.7) Housing status after 10 years By the end of the follow-up over half (52%) of the study group had died, compared to 15% in the control group (P < 0.001) (figure 1). In the study group 6% were independently housed, 38% lived in supported housing and 5% were homeless at 10 years. The corresponding numbers in the control group were 82%, 3% and 0.1%, respectively. Of the 25 persons in the study group, who were homeless after 10 years, 9 stayed in shelters, 7 in correctional institutions or psychiatric hospitals and 3 with friends or family and in six cases their place of stay could not be verified. One person in the control group was homeless and staying in shelters at the end of the follow-up. Figure 1 View largeDownload slide Status at 10 years after homeless shelter use in the study group (N = 552) and their general population controls (N = 946), Helsinki, Finland, percentages with 95% confidence intervals Figure 1 View largeDownload slide Status at 10 years after homeless shelter use in the study group (N = 552) and their general population controls (N = 946), Helsinki, Finland, percentages with 95% confidence intervals Of the 207 previously homeless men that lived in supported housing, 84 persons (15% of the total sample) stayed in nursing homes, 77 (14%) in housing first units and 46 (8%) in low intensity supported housing (Supplementary Appendix S1). Correspondingly, in the control group, 18 persons (2% of the control group) stayed in nursing homes, 5 (0.5%) in housing first units and 9 (1%) in low intensity supported housing. Psychiatric and somatic morbidity The 10-year prevalence of psychiatric morbidity in the study group was 78% (95% CI 74–81%) compared with 16% (95% CI 14–18%) in the control group (Supplementary Appendix S2). SUD was the most prevalent psychiatric diagnosis in the study group; 75% of the previously homeless and 8% of the controls had a SUD. Further, psychotic disorders and dual diagnosis presented more often in the study group compared to controls (13% and 12% in the study group and 2% and 0.5% among the controls, respectively). The somatic morbidity was also significantly higher in the study group with a mean CCI of 1.3 (SD 2.0) in the study group, compared with 0.7 (SD 1.6) among the controls, (P < 0.001). For those in supported housing at the end of follow-up, the CCI index and psychiatric morbidity increased with the level of support among the previously homeless (Supplementary Appendix S2). The prevalence of psychiatric morbidity among those independently housed and those in nursing homes was higher in the study group compared with the control group (P < 0, 001, respectively). We also compared the prevalence of SUD, psychotic disorders and dual diagnosis in the homeless group according to the housing status at end of follow-up (figure 2). The prevalence of SUD varied significantly depending on the housing status at the end of follow-up, being lowest among those independently housed (55%) and highest among those homeless (89%) at 10 years after shelter use (P = 0.012). The differences in the prevalence of psychotic disorders and dual diagnosis between the different housing situations at the end of follow-up was even bigger (P < 0.001 in both groups), and also here, the prevalence figures are highest among those homeless with 40% having psychosis and dual diagnosis compared to 3% among those living independently at the end of follow-up. Figure 2 View largeDownload slide The 10-year prevalence of SUD, psychotic disorders and dual diagnosis among previously homeless men (N = 552) by housing status after 10-year follow-up, Helsinki, Finland, percentage Figure 2 View largeDownload slide The 10-year prevalence of SUD, psychotic disorders and dual diagnosis among previously homeless men (N = 552) by housing status after 10-year follow-up, Helsinki, Finland, percentage The predictive effects of socioeconomic factors and health service use among shelter users on being independently housed 10 years later In the multivariate analysis assessing factors predicting independent housing 10 years after shelter use within the study group (including those deceased during follow-up) we found that being married (OR 8.34, 95% CI 3.01–23.15) and staying only briefly in shelter (OR 9.08, 95% CI 2.68–30.76) were strongly associated with independent housing (table 2). Having no emergency room visits (OR 3.55, 95% CI 1.17–10.77) and having more than basic level education (OR 2.33, 95% CI: 1.09–4.99) also increased the odds of being independently housed at the end of follow up. The results remained when those deceased were removed from the analysis (data not shown). Table 2 Factors predicting independent housing at 10 years after stay in shelter among homeless men (N = 552) in Helsinki, Finland Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) a Forward stepwise selection including all variables. Only those variables are shown which entered the model. Table 2 Factors predicting independent housing at 10 years after stay in shelter among homeless men (N = 552) in Helsinki, Finland Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) Univariate Multivariatea OR (95% CI) OR (95% CI) Age 0.98 (0.95–1.01) Married or in registered partnership 8.90 (3.53–22.45) 8.34 (3.01–23.15) More than basic level education 2.26 (1.11–4.59) 2.33 (1.09–4.99) Employed (incl. students) 3.85 (1.36–10.88) Less than four nights in shelter during year 2004 10.43 (3.14–34.61) 9.08 (2.68–30.76) No hospitalizations 3.43 (1.19–9.94) No outpatient visits to hospitals 1.77 (0.61–5.16) No emergency room visits 4.19 (1.45–12.09) 3.55 (1.17–10.77) No detoxification treatments 0.78 (0.37–1.66) a Forward stepwise selection including all variables. Only those variables are shown which entered the model. Discussion We studied the housing situation and morbidity at 10 years after shelter use and found that besides having high mortality as shown in our previous study,22 only 6% of the previously homeless were independently housed at the end of follow-up. Further, regardless of the housing situation the psychiatric morbidity in the homeless cohort was high. The somatic morbidity was also higher in the study group compared with the controls. Being married and staying only briefly in shelters were strongly associated with independent housing after 10 years. Previous studies with follow-ups of 1–3 years have found that, depending on the setting and methods, 19–61% are still homeless at the end of follow-up.6,–12 The fact that in our study only 5% (9% of those alive) were still homeless after 10 years, can be seen as a result and success of the efforts to eliminate long-term homelessness that have been implemented in the last decade in Finland.19 Considering for instance findings from Stockholm, a society with similar social support system, where 75% of those alive were still homeless 5 years later,14 it is not unlikely that many more also in our sample would still be homeless had it not been for the intensive housing efforts in the City of Helsinki. The fact that only 6% (12% of those alive) in our study were independently housed at the end of follow-up, however, supports findings from previous studies2 showing that very few homeless shelter users manage to enter the free housing markets. We can conclude that a variety of supported housing services are needed to end homelessness in this highly vulnerable population with complex needs. The prevalence of psychiatric diagnoses and especially SUD in our cohort was high compared with previous studies. Swedish and Danish studies using hospital discharge registers to define prevalence of mental disorders among homeless reported 42% and 49% rates of SUD, respectively,20,21 compared to 75% in our cohort. This finding is in line with the high prevalence of SUD in Finnish shelter users5 and offers a further explanation to the high mortality rate in the cohort. Our finding that the psychiatric morbidity was highest among those still or again homeless indicates that there still is a group of homeless with multiple needs for which the housing services on offer are not providing adequate or enough support. Several factors, such as being married and staying only briefly in shelters were associated with independent housing. Interestingly, these same factors were not associated with mortality in a previous study of the same cohort.22 We can speculate that death, especially when related to SUD is more random and difficult to predict while factors that are associated with housing skills and not being institutionalized, such as being married and staying only briefly in shelters are more easily identified among homeless shelter users. Previous studies have found that the size of social network predicts exiting homelessness7,10,23 and this is supported by our findings: those that find other accommodation within a few days have usually solved their housing situation by themselves with the help of their social network. There are some limitations to this study. First, we only had access to the social service and primary health care records from the City of Helsinki and consequently we had to exclude the 10% of the cohort that had migrated. It can be speculated that some of them had gained independent housing somewhere else. However, considering that the morbidity of the excluded group did not differ from those in supported housing (data not shown) it is likely that the main part of the men excluded from the analysis also lived in some form of supported housing in an other municipality. Second, we only had access to the information that could be extracted from the registers and there are probably many other factors associated with independent housing including length of homelessness prior to shelter use, housing skills, family and social networks that we could not examine. Third, being limited to the diagnosis registered in specialized health care, there is most likely a problem of under-diagnosis, especially in the deceased group, where many died of external causes already in the first years of follow-up.22 A longer follow-up prior to baseline would have improved our morbidity data. Using a registered-based approach to study the homeless situation also offers several strengths, and thus we were able to, for the first time, describe the housing situation of a cohort of homeless men after 10 years, without the problems of cases lost to follow-up and persons declining to participate that interview-based studies struggle with. Having a matched control group further validates the results. The long perspective of the study offers new valuable information to policy makers and care workers that deal with homeless. In conclusion, we showed that very few of the homeless that survive manage to enter the free housing market and that this population requires supported housing services with different levels of support. Further research is needed to study the permanence of the housing solutions and their effect on quality of life, mortality and service use. Ethical consideration Ethics approval for this study was granted by the ethics committee of the Hospital district of Helsinki and Uusimaa (HUS), and research permits were obtained from all the register keepers in the study. According to the Finnish legislation no informed consent from participants was needed because only existing register data were used, the registered persons were not contacted and the data were analyzed anonymously. Supplementary data Supplementary data are available at EURPUB online. Acknowledgements The authors thank the staff at the City of Helsinki for helping to gather the data, the National Institute for Health and Welfare for technical support and all the funders of this study. Funding The study has been funded by Samfundet Folkhälsan, Finnish Medical Association (Finska läkaresällskapet), the City of Helsinki and the Wilhelm and Else Stockmann’s foundation. The funders have not played any role in the design of the study and collection, analysis and interpretation of data or in writing the manuscript. Conflicts of interest: None declared. Key points While it is known that most homeless staying in shelters are there only briefly and then move on, the long-term prognosis of homelessness in terms of housing is not previously known. We found that after 10 years over half of the shelter users had died, only 6% had independent housing, 38% lived in supported housing and 5% were still or again homeless, showing that most of those surviving are in need of support in their everyday lives even years after the shelter period. Among the formerly homeless, psychiatric disorders, including substance use disorder, were present in 78% of the study group, compared with 16% in the general population. Among those still alive at the end of follow-up the psychiatric morbidity was highest among those still or again homeless followed by those in intensively supported housing. Being married and staying only briefly in shelters is strongly associated with being independently housed 10 years later. References 1 Fazel S , Geddes JR , Kushel M . The health of homeless people in high-income countries: descriptive epidemiology, health consequences, and clinical and policy recommendations . Lancet 2014 ; 384 : 1529 – 40 . Google Scholar CrossRef Search ADS PubMed 2 Sunikka S . Pathways to homelessness of people who stayed overnight in hietsu–Mobility between homelessness and the dwelling population. (Professional Licentiate thesis in Finnish). University of Helsinki, 2016 . Available at: http://www.sosnet.fi/loader.aspx? id=971cb92e-9f9b-4f46-9cc7-436e4ef7c966 (16 November 2017, date last accessed). 3 Kuhn R , Culhane DP . Applying cluster analysis to test a typology of homelessness by pattern of shelter utilization: results from the analysis of administrative data . Am J Community Psychol 1998 ; 26 : 207 – 32 . Google Scholar CrossRef Search ADS PubMed 4 Busch-Geertsema V , Benjaminsen L , Filipovic Hrast M , Pleace N . Extent and Profile of Homelessness in European Member States: A Statistical Update, 2014 . Available at: http://housingfirstguide.eu/website/wp-content/uploads/2016/04/feantsa-studies_04-web2.pdf (16 November 2017, date last accessed). 5 Stenius-Ayoade A , Haaramo P , Erkkila E , et al. Mental disorders and the use of primary health care services among homeless shelter users in the Helsinki Metropolitan Area, Finland . BMC Health Serv Res 2017 ; 17 : 428 – 017 , 2372–3. Google Scholar CrossRef Search ADS PubMed 6 Kertesz SG , Larson MJ , Horton NJ , et al. Homeless chronicity and health-related quality of life trajectories among adults with addictions . Med Care 2005 ; 43 : 574 – 85 . Google Scholar CrossRef Search ADS PubMed 7 Aubry T , Duhoux A , Klodawsky F , et al. A longitudinal study of predictors of housing stability, housing quality, and mental health functioning among single homeless individuals staying in emergency shelters . Am J Community Psychol 2016 ; 58 : 123 – 35 . Google Scholar CrossRef Search ADS PubMed 8 Schanzer B , Dominguez B , Shrout PE , Caton CL . Homelessness, health status, and health care use . Am J Public Health 2007 ; 97 : 464 – 9 . Google Scholar CrossRef Search ADS PubMed 9 Spicer B , Smith DI , Conroy E , et al. Mental illness and housing outcomes among a sample of homeless men in an Australian urban centre . Aust N Z J Psychiatry 2015 ; 49 : 471 – 80 . Google Scholar CrossRef Search ADS PubMed 10 Caton CL , Dominguez B , Schanzer B , et al. Risk factors for long-term homelessness: findings from a longitudinal study of first-time homeless single adults . Am J Public Health 2005 ; 95 : 1753 – 9 . Google Scholar CrossRef Search ADS PubMed 11 Johnson G , Scutella R , Tseng Y , et al. Entries and Exits from Homelessness: A Dynamic Analysis of the Relationship between Structural Conditions and Individual Characteristics, 2015 : 1 – 67 . Available at: https://www.ahuri.edu.au/__data/assets/pdf_file/0013/2155/AHURI_Final_Report_No248_Entries-and-exits-from-homelessness.pdf (16 November 2017, date last accessed). 12 Fichter MM , Quadflieg N . Three year course and outcome of mental illness in homeless men: a prospective longitudinal study based on a representative sample . Eur Arch Psychiatry Clin Neurosci 2005 ; 255 : 111 – 20 . Google Scholar CrossRef Search ADS PubMed 13 McQuistion HL , Gorroochurn P , Hsu E , Caton CL . Risk factors associated with recurrent homelessness after a first homeless episode . Community Ment Health J 2014 ; 50 : 505 – 13 . Google Scholar CrossRef Search ADS PubMed 14 Beijer U , Andreasson A , Agren G , Fugelstad A . Mortality, mental disorders and addiction: a 5-year follow-up of 82 homeless men in stockholm . Nord J Psychiatry 2007 ; 61 : 363 – 8 . Google Scholar CrossRef Search ADS PubMed 15 Tsemberis S . Housing First: The Pathways Model to End Homelessness for People with Mental Illness and Addiction . Center City (MN ): Hazelden Press , 2010 . 16 Aubry T , Tsemberis S , Adair CE , et al. One-year outcomes of a randomized controlled trial of housing first with ACT in five Canadian cities . Psychiatr Serv 2015 ; 66 : 463 – 9 . Google Scholar CrossRef Search ADS PubMed 17 Kertesz SG , Austin EL , Holmes SK , et al. Housing first on a large scale: fidelity strengths and challenges in the VA’s HUD-VASH program . Psychol Serv 2017 ; 14 : 118 – 28 . Google Scholar CrossRef Search ADS PubMed 18 Busch-Geertsema V . Housing first Europe–results of a European social experimentation project . Eur J Homelessness 2014 ; 8 : 13 – 28 . 19 Pleace N , Culhane D , Granfelt R , Knutagård M . The Finnish Homelessness Strategy–an International Review. 2015-02-10. Available at: https://helda.helsinki.fi/handle/10138/153258 (16 November 2017, date last accessed). 20 Nielsen SF , Hjorthoj CR , Erlangsen A , Nordentoft M . Psychiatric disorders and mortality among people in homeless shelters in Denmark: a nationwide register-based cohort study . Lancet 2011 ; 377 : 2205 – 14 . Google Scholar CrossRef Search ADS PubMed 21 Beijer U , Andreasson S . Gender, hospitalization and mental disorders among homeless people compared with the general population in Stockholm . Eur J Public Health 2010 ; 20 : 511 – 6 . Google Scholar CrossRef Search ADS PubMed 22 Stenius-Ayoade A , Haaramo P , Kautiainen H , et al. Mortality and causes of death among homeless in Finland: a 10-year follow-up study . J Epidemiol Community Health 2017 ; 71 : 841 – 8 . Google Scholar CrossRef Search ADS 23 Zlotnick C , Tam T , Robertson MJ . Disaffiliation, substance use, and exiting homelessness . Subst Use Misuse 2003 ; 38 : 577 – 99 . Google Scholar CrossRef Search ADS PubMed 24 Charlson ME , Pompei P , Ales KL , MacKenzie CR . A new method of classifying prognostic comorbidity in longitudinal studies: development and validation . J Chronic Dis 1987 ; 40 : 373 – 83 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The European Journal of Public HealthOxford University Press

Published: Mar 23, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off