1082767 HPQ0010.1177/13591053221082767Journal of Health PsychologyAmissah et al. research-article2022 Article Journal of Health Psychology Assessing psychosocial distress 1 – 12© The Author(s) 2022 associated with homelessness Article reuse guidelines: sagepub.com/journals-permissions DhttOpsI::/ /1d0o.i1.o1r7g/71/01.13157971/10355392102513028221708627767 in Ghana: A springboard for journals.sagepub.com/home/hpq interventional policy design Nelly BF Amissah1 , Christopher M Amissah2 and Benjamin Amponsah1 Abstract This study investigated the causes of homelessness in Ghana and associated psychosocial distress. A sample of 86 homeless participants listed perceived causes of their homelessness and completed measures of psychosocial distress, and 97 non-homeless participants completed the same measures psychosocial distress. Causes of homelessness among the participants included poverty (30.1%), migration (10.4%), unemployment (2.2%), parental demise (2.2%), parental neglect (0.5%), and parental divorce (0.5%). Multivariate analysis revealed higher psychosocial distress among the homeless than the non-homeless. Homeless females reported higher levels of stress and suicidality than their male counterparts. The study demonstrates the need for timely and effective implementation of interventions such as provision of affordable housing, financial assistance, job-creation, and skill training for the homeless directly related to known causes of homelessness and accounting for gender differences. Keywords Ghana, homelessness, causes, interventions, psychosocial distress Introduction severe challenges to policymakers and research- ers alike (Byrne et al., 2013). It is a serious global Background problem with the situation affecting about 2.5– Homelessness is an undesirable life circum- 3.5 million people annually (Balasuriya et al., stance that results in psychosocial distress. It is 2020) and about 650,000 people sleeping on the characterized by economic deprivations, social streets on daily basis in high income countries marginalization, poor mental health, and nega- such as the United States (Corrigan et al., 2015; tive self-concept (Embleton et al., 2016; Osei Fox et al., 2016). Asibey et al., 2020; Schütz, 2016; Smartt et al., 2019). Homelessness usually limits individuals’ access to resources, employment opportunity, 1University of Ghana, Ghana and social network (Calvo et al., 2018; Nishio 2Morgan State University, USA et al., 2017; Vázquez et al., 2018). It can affect Corresponding author: every part of a person’s daily life. Nelly BF Amissah, Department of Psychology, University In the past three decades, homelessness has of Ghana, P.O. Box LG84, Legon, Ghana. become a major public health concern that poses Email: nellybfamissah@gmail.com 2 Journal of Health Psychology 00(0) Nature of homelessness in Ghana (Cleverley and Kidd, 2011; Osei Asibey et al., 2020). Several studies have shown that home- Homelessness is considered a major social and less individuals are at extreme risk of abuse and public health issue in developing countries victimization, poor physical and mental health, (Abekah-Carter and Oti, 2022; Osei Asibey et al., and high premature mortality (e.g., Balasuriya 2020), and more so in developed countries (Toro, et al., 2020; Fox et al., 2016; Osei Asibey et al., 2007). Research on homelessness and associated 2020; Urbanoski et al., 2018). Compared to the health challenges have received increased atten- non-homeless, homeless people suffer worse tion in developed countries. However, the con- mental health (Hwang, 2001; Lippert and Lee, cept of homelessness appears to be relatively new 2015; Public Health Agency of Canada, 2006). in developing countries. In Ghana, for example, Among the mental health conditions frequent in the term “homelessness” has no direct local ren- the literature are depression, traumatic symp- dition in any Ghanaian language. This suggests toms, anxiety, stress, and somatization (Asante that homelessness is an emerging issue in devel- et al., 2015; Votta and Manion, 2003). oping countries. However, De-Graft Aikins and Chronically homeless individuals in particular Ofori-Atta (2007) have reported that homeless are reported to have higher rates of mental individuals in Ghana are exposed to elevated lev- health problems than episodically homeless els of mental health challenges that threaten their people or new-entry homeless people (Lippert life and well-being. and Lee, 2015). By large, the nature of homelessness in Research evidence further demonstrates that developing countries like Ghana is signifi- homeless individuals are at an elevated risk of cantly different from homelessness in devel- substance abuse (Baggett et al., 2015), health- oped countries (Komla, 2013). The Ghana risk behaviors and delinquency especially Statistical Service (2002) defined homeless- among homeless youth (Chondraki et al., 2014). ness not only in housing terms but also lack of Studies have cited substance use as normative belongingness to a household. However, it of homelessness, with evidence indicating as must be noted that while the concept of home- high as 69% to 71% of homeless youth report- lessness is well understood in Ghana, there is ing abuse of alcohol or drugs (Fox et al., 2016; no formal definition rendered by any state Urbanoski et al., 2018). Abuse of drugs and department. Previous studies on homelessness alcohol places homeless youths at an increased in Ghana conceptualized homelessness in risk of negative health consequences such as terms of the quality of residential facilities – risk of disease infection, depression, suicide, primarily identifying people living in slums as physical victimization, and illegal activities homeless (Asante et al., 2015; De-Graft Aikins (Chondraki et al., 2014; Fekadu et al., 2014). and Ofori-Atta, 2007). It must be pointed out that such conceptualization of homelessness in Ghana is highly superficial and limited in Rationale for the present study focus. A valid definition of homelessness Much of the existing research has focused on should take into consideration lack of settle- examining the risk factors for homelessness ment structures and enormity of financial (Chamberlain and Johnson, 2013; Fox et al., stress. These two defining elements are highly 2016). Some researchers, however, are shifting conspicuous in the streets of Accra. attention to examining physical and mental health consequences that homeless individuals Impact of homelessness confront on daily basis (Chondraki et al., 2014). Due to the substantial body of literature on risk Homelessness poses a consistent challenge to factors of homelessness, there is a need to shift health systems, especially to mental health attention to mental health outcomes of home- practitioners, policymakers, and researchers less individuals, especially those in developing Amissah et al. 3 countries (Smartt et al., 2019). Given the translation to other languages. A pilot study was increasing number of homeless people in carried out by the researchers to test the reliability Ghana, the current study aimed at examining of the measuring instruments (questionnaire). In the psychosocial consequences of homeless- line with ethical requirements, the researchers ness in Ghana. The psychosocial consequences explained the nature and purpose of the study to include depression, anxiety, stress, somatiza- prospective participants and sought their consent tion, loneliness, traumatic symptoms, and suici- for participation. Participants were asked to com- dality. Other psychosocial symptoms such as plete informed consent forms prior to answering hyperactivity, emotional difficulty, relationship the questionnaire. Participants took an average of problems, substance abuse, violent behaviors, 18 minutes to complete the questionnaire. and conduct disorders have previously been investigated in Ghana (Asante et al., 2015; Participants De-Graft Aikins and Ofori-Atta, 2007). In addition to assessing the psychosocial con- The study participants (n = 183) were purpo- sequences of homelessness, the study sought to sively drawn from the Greater Accra Region of examine gender differences in the psychosocial Ghana. They comprised of 86 homeless people impact of homelessness. Previous research had (47.0%) and 97 non-homeless people (53.0%). reported high rates of violence against homeless The non-homeless sample served as a compari- women (Vaughn, 2017), yet minimal research son group. The ages of the participants ranged had focused on how these experiences impact on from 14 to 49 years (M = 24.95, SD = 8.75). In homeless women’s mental health. In line with terms of gender composition, there were more these objectives, two hypotheses were formu- females (52.5%) than males (47.5%). Expectedly, lated. First, it was hypothesized that there are there was a significant educational gap between higher levels of psychosocial distress among homeless participants and non-homeless partici- homeless people compared to non-homeless pants. Majority of homeless participants were people. Second, it was hypothesized that female either uneducated (41.9%) or had received only homeless people experience higher levels of psy- basic education (50.0%). On the contrary, non- chosocial distress than male homeless people. In homeless participants mostly had high school the context of this study, psychosocial conse- education (61.9%) or tertiary education (35.1%). quences of homelessness are conceptualized as Table 1 provides further details on the demo- psychosocial distress and operationalized as graphic characteristics of the participants. depression, anxiety, stress, somatization, loneli- ness, traumatic symptoms, and suicidality. Measures Structured questionnaires were used for data Methods collection. They were used to collect partici- Design and procedure pants’ demographic information, causes of homelessness, and the state of their mental The research involved a quantitative cross-sectional health. The mental health mental variables were survey. Structured questionnaires were used to cap- loneliness, depression, anxiety, stress, post- ture self-report information of the participants’ traumatic symptoms, suicidality, and somatic mental health experiences. The questionnaires were symptoms. These variables were assessed using translated into a major Ghanaian language (Twi) standardized instruments whose psychometric and adapted to fit the Ghanaian culture. In situa- descriptions are provided below. tions where a participant could not comprehend neither the English language nor the Twi language, Revised UCLA Loneliness Scale. Russell et al. the assistance of a native speaker of a particular (1978) developed the UCLA Loneliness Scale Ghanaian language was sought to facilitate local consisting of 20 negatively worded items. 4 Journal of Health Psychology 00(0) Table 1. Demographic characteristics of the respondents. Variables Homeless, n = 86 Non-homeless, Total, N = 183 (47.0%) n = 97 (53.0%) (100%) Gender Males 41 (47.7%) 46 (47.4%) 87 (47.5%) Females 45 (52.3%) 51 (52.6%) 96 (52.5%) Age range 14–49 years M = 27.37, SD = 9.53 M = 22.80, SD = 7.41 M = 24.95, SD = 8.75 Educational background No formal education 36 (41.9%) 1 (1.0%) 37 (20.2%) Basic education 43 (50.0%) 2 (2.1%) 45 (24.6%) Secondary education 7 (8.1%) 60 (61.9%) 67 (36.6%) Tertiary education 0 (0.0%) 34 (35.1%) 34 (18.6%) Marital status Never married 63 (73.3%) 76 (78.4%) 139 (76.0%) Married 14 (16.3%) 21 (21.6%) 35 (19.1%) Divorced 6 (7.0%) 0 (0.0%) 6 (3.3%) Widowed 3 (3.5%) 0 (0.0%) 3 (1.6%) Religious affiliation Christianity 40 (46.5%) 89 (91.8%) 129 (70.5%) Islam 46 (53.5%) 8 (8.2%) 54 (29.5%) Russell et al. (1980) revised the instrument by depression subscale is labeled DASS-D and reversing half of the test items to positive. measures devaluation of life, dysphoria, lack Items 1, 4, 5, 6, 9, 10, 15, 16, 19, and 20 are the of interest, hopelessness, self-deprecation, positively worded items and the remaining 10 anhedonia, and inertia. The anxiety subscale items are negatively worded. Example items is labeled DASS-A and evaluates situational include “There is no one I can turn to,” “I am anxiety, skeletal muscle effects, autonomic no longer close to anyone,” and “I feel left arousal, and subjective experience of anxious out.” Each item on the scale is rated as Never affect. The stress subscale is labeled DASS-S (1), Rarely (2), Sometimes (3), and Often (4). and measures nervousness, difficulties in Total scores can range from 20 to 80. Higher relaxation, and irritability. Example items scores represent greater levels of loneliness. include “I couldn’t seem to experience any Regarding the psychometric properties of the positive feeling at all,” “I tended to over-react UCLA loneliness scale, Russell (1996) to situations,” and “I found it difficult to reported Cronbach’s alpha internal reliability relax.” Each item on the scale is rated Never coefficient range of 0.89–0.94, and 1-year test- (0), Sometimes (1), Often (2), and Almost retest reliability coefficient of 0.73. The Cron- always (3). The cumulative score ranges from bach’s alpha for the current Ghanaian sample 0 to 21 for each sub-scale; and 0 to 63 for the was 0.84. whole scale. High Cronbach’s alpha internal reliability has been reported for DASS-D Depression, Anxiety and Stress Scale-21 (DASS- (0.90), DASS-A (0.83), and DASS-S (0.86). 21). The DASS-21 (Lovibond and Lovibond, Among the current sample, the Cronbach’s 1995) assesses the psychosocial symptoms of alpha for the whole DASS was 0.88, DASS-D depression, anxiety, and stress. It consists of was 0.86, DASS-A was 0.75, and DASS-S three subscales with seven items each. The was 0.77. Amissah et al. 5 Modified PTSD Symptom Scale-Self Report (MPSS- Table 2. Causes of homelessness among the SR). The MPSS-SR (Falsetti et al., 1993) is a participants. 17-item measure that employs a 5-point scale to Causes of homelessness Frequency Percentage measure the frequency and severity of post- (n = 86) (%) traumatic stress disorder symptoms. Response options range from “not at all” (rated 0), “a little Unemployment 4 2.2 bit” (rated 1), “moderately” (rated 2), “quite a Migration 19 10.4 bit” (rated 3), “to extremely” (rated 4). Total Death of parent(s) 4 2.2 rating scores range from 0 to 68. Example items Poverty 55 30.1 are, “Have you had repeated or intrusive upset- Domestic violence 2 1.1 Parental neglect 1 0.5 ting thoughts or recollections of the event(s)?”, Divorce of parents 1 0.5 “Have you been having repeated bad dreams or nightmares about the event(s)?”, and “Have you had the experience of suddenly reliving the event(s), flashbacks of it or acting or feeling as The total rating scores range from 0 to 32. The if the event were happening again?”. The SSS-8 is reported to have a Cronbach’s alpha MPSS-SR has Cronbach’s alpha internal relia- internal reliability of 0.81. Among the local bility of 0.96. The Cronbach’s alpha for the sample, Cronbach’s alpha was 0.86. local Ghanaian sample was 0.87. Data analysis Suicidal Behaviors Questionnaire-Revised (SBQ-R). The SBQ-R (Osman et al., 2001) is a 4-item rat- Data analysis was conducted in SPSS version ing scale that measures the history of suicide 27. Descriptive analyses were performed on behaviors, suicidal ideation, frequency of sui- demographic variables and causes of homeless- cidal ideation, previous suicide attempts, and ness. The results of the descriptive analyses are the probability of future suicidal attempts. Sam- reported in frequencies and percentages (see ple items from the scale are “Have you ever Tables 1 and 2). Multivariate analysis of vari- thought about or attempted to kill yourself?” ance (MANOVA) test was used to analyze the and “How likely is it that you will attempt sui- two research hypotheses. Each hypothesis com- cide someday?” Responses to the SBQ-R differ pared two independent groups of participants across its items but are summed up to generate on seven dependent variables whose data were a single measure of suicidality for each respond- continuous and normally distributed. Normal ent ranging from 4 to 23. Higher scores suggest distribution analysis was conducted on each of greater risk of suicidality. The SBQ-R is known the seven dependent variables using moment- to have a Cronbach’s alpha internal reliability based measures (i.e. skewness and kurtosis). It of 0.88. The Cronbach’s alpha among the cur- has been argued that data with skewness rent sample was 0.85. between −2 to +2 and Kurtosis between −7 to +7 are acceptable for normal distribution Somatic Symptom Scale-8 (SSS-8). The SSS-8 (Byrne, 2010; Hair et al., 2010). However, for (Gierk et al., 2014) is a short version of the analysis in SPSS, skewness and kurtosis values PHQ-15 questionnaire (Kroenke et al., 2002). It ranging from −2 to + 2 are considered as satis- was designed to measure fatigue, pain, cardio- factory (George and Mallery, 2010). In the cur- pulmonary and gastrointestinal elements of the rent study, the SPSS analysis produced values somatic symptom burden. Sample items include ranging from −0.95 to 1.68 for both skewness “back pain,” “headaches,” and “trouble sleep- and kurtosis, thus satisfying normality assump- ing.” Response options comprise not at all tion for multivariate analysis in the hypothesis (rated 0), a little bit (rated 1), somewhat (rated testing. The results of the hypothesis testing are 2), quite a bit (rated 3), and very much (rated 4). presented in Tables 3 and 4 respectively. 6 Journal of Health Psychology 00(0) Table 3. Psychosocial distress among the participants. Psychosocial distress Homeless Non-homeless F Sig. η2 (n = 86) (n = 96) Mean (SD) Mean (SD) Loneliness 53.43 (4.89) 51.51 (6.28) 5.21 0.024 0.03 Depression 10.65 (3.97) 4.08 (3.35) 146.46 0.000 0.45 Anxiety 5.53 (3.78) 4.39 (3.25) 4.86 0.029 0.03 Stress 7.09 (3.93) 5.20 (3.33) 12.40 0.001 0.06 Traumatic symptoms 22.19 (11.98) 16.47 (10.97) 11.29 0.001 0.06 Suicidal behavior 7.69 (3.85) 4.11 (2.94) 50.11 0.000 0.22 Somatic symptoms 13.52 (8.20) 9.32 (7.45) 13.28 0.000 0.07 Pillai’s trace: V = 0.52; F(7, 174) = 27.20; p = 0.000; Partial eta squared = 0.52; df = 1/181. Table 4. Gender differences in psychosocial distress among the homeless participants. Psychosocial distress Female (n = 45) Male (n = 41) F Sig. η2 Mean (SD) Mean (SD) Loneliness 52.51 (5.92) 54.44 (3.20) 3.44 0.067 – Depression 10.98 (3.65) 10.29 (4.31) 0.64 0.428 – Anxiety 5.16 (4.32) 5.95 (3.08) 0.95 0.333 – Stress 8.29 (3.89) 5.78 (3.57) 9.64 0.003 0.10 Traumatic symptoms 22.69 (12.92) 21.63 (11.00) 0.17 0.686 – Suicidal behaviors 8.62 (4.01) 6.66 (3.41) 5.92 0.017 0.07 Somatic symptoms 14.64 (7.20) 12.29 (9.10) 1.78 0.186 – Pillai’s trace: V = 0.18; F(7, 78) = 2.50; p < 0.05; Partial eta squared = 0.18; df = 1/84. Data sharing statement of the homeless participants. Unemployment and death of parent(s) each affected 2.2% of the The current article includes the complete raw homeless participants. Finally, parental neglect dataset collected in the study including the par- and divorce of parents affected just a few (0.5% ticipants’ data set, syntax file and log files for each). analysis. Pending acceptance for publication, Table 3 shows the MANOVA results on the all the data files will be automatically uploaded to the Figshare repository. effects of homelessness on psychosocial distress. Using Pillai’s trace, there was a significant impact of homelessness on psychosocial distress Results (V = 0.52, F(7, 174) = 27.20, p < 0.001, Partial η2 = Table 2 displays causes of homelessness as 0.52). Separate univariate analyses (ANOVAs) reported by the 86 homeless participants. In all, on the outcome variables revealed significant seven causes of homelessness were reported by effects of homelessness on loneliness (F(1, 181) = the homeless participants based on their indi- 5.21, p < 0.05, Partial η2 = 0.03), depression vidual circumstances. Among them, poverty (F(1, 181) = 146.46, p <0.001, Partial η 2 = 0.45), emerged as the leading cause of homelessness anxiety (F(1, 181) = 4.86, p < 0.05, Partial η 2 = for 30.1% of the homeless participants. It was 0.03), stress (F(1, 181) = 12.40, p = 0.001, Partial followed by migration which implicated 10.4% η2 = 0.06), traumatic symptoms (F(1, 181) = 11.29, Amissah et al. 7 p = 0.001, Partial η2 = 0.06), suicidal behavior levels of psychosocial distress than male home- (F = 50.11, p < 0.001, Partial η2(1, 181) = 0.03), and less participants. somatic symptoms (F(1, 181) = 13.28, p = 0.000, Partial η2 = 0.06). The mean scores show that Discussion homeless participants felt lonelier (M = 53.43, SD = 4.89) than their non-homeless counterparts The purpose of the study was to investigate psy- (M = 51.51, SD = 6.28). They also felt more chosocial distress among the homeless in depressed (M = 10.65, SD = 3.97) than the non- Ghana. As indicated, psychosocial distress was homeless (M = 4.08, SD = 3.35), reported higher defined as depression, anxiety, stress, somatiza- anxiety level (M = 5.53, SD = 3.78) than non- tion, loneliness, somatic symptoms, and suici- homeless participants (M = 4.39, SD = 3.25), and dality. It emerged that financial distress, had higher level of stress (M = 7.09, SD = 3.93) dysfunctional family structure, and personal than non-homeless participants (M = 5.20, SD = motivation to migrate were the major forces 3.33). Again, homeless participants reported that resulted in the unfortunate incident of higher traumatic symptoms (M = 22.19, SD = homelessness. This observation is consistent 11.98) than their non-homeless counterparts with previous reports (Calvo et al., 2018; (M = 16.47, SD = 10.97), engaged in more sui- Embleton et al., 2016; Vázquez et al., 2018). cidal behaviors (M = 7.69, SD = 3.85) than the The nature of the causes of homelessness, as non-homeless (M = 4.11, SD = 2.94), and had reported by the participants, suggests that higher level of somatic symptoms (M = 13.52, homelessness should not only be understood as SD = 8.20) than their non-homeless counterparts the result of dysfunctions in individual personal (M = 9.32, SD = 7.45). These results confirm that character, but more so, the resultant effect of there was higher psychosocial distress among weak national economy and ineffective institu- the homeless than the non-homeless. tional structures in the larger socio-economic Table 4 displays the MANOVA results for context (McLeroy et al., 1988; Nooe and gender differences in psychosocial distress Patterson, 2010). This is consistent with the call among the homeless participants. Using Pillai’s to shift perspective from a single-model trace, there was a significant gender differences approach to a multilateral-model approach in in psychosocial distress (V = 0.18, F(7, 78) = 2.50, efforts to resolve homelessness (Anderson and p < 0.05, Partial η2 = 0.18). Analyses of variance Christian, 2003). on individual distress variables reveal significant The findings supported the first research gender differences in stress (F(1, 84) = 9.64, hypothesis. Consistent with prediction, there p < 0.01, Partial η2 = 0.10), and suicidal behavior were significantly higher levels of psychosocial (F(1, 84) = 5.92, p < 0.05, Partial η 2 = 0.07). Female distress among the homeless participants than homeless participants reported higher stress lev- the non-homeless participants. Compared to els (M = 8.29, SD = 3.89) than did their male non-homeless, homeless participants felt lone- counterparts (M = 5.78, SD = 3.57). Similarly, lier, were more depressed, more anxious, more there was greater suicidality among female home- stressed, more traumatized, more suicidal, and less participants (M = 8.62, SD = 4.01) than the had higher level of somatic symptoms. The male homeless participants (M = 6.66, SD = 3.41). high incidence of psychosocial distress among However, there were no significant gender differ- the homeless participants supports the assertion ences in loneliness (F(1, 84) = 3.44, p > 0.05), of the Gelberg-Andersen behavioral model for depression (F(1, 84) = 0.64, p > 0.05), anxiety (F(1, vulnerable populations (Padgett et al., 1990, 84) = 0.95, p > 0.05), traumatization, (F(1, 84) = .17, 1995; Swanson et al., 2003) which outlines the p > 0.05), and somatic symptoms (F(1, 84) = 1.78, vulnerabilities common to homeless individu- p > 0.05). These results provide partial confirma- als such as environmental hazards, drug and tion for the prediction that female homeless par- alcohol use, and lack of access to health care ticipants will experience significantly higher (Kushel et al., 2006). The present study 8 Journal of Health Psychology 00(0) confirmed that these common vulnerabilities The second research objective was to deter- among the homeless undermines psychosocial mine gender differences in psychosocial dis- wellbeing, thus the need for health care service tress among the homeless. Pursuant to this providers to focus attention on reducing or objective, the researchers predicted that female eliminating these vulnerabilities in order to homeless participants are likely to experience enhance the mental health of the homeless. significantly higher levels of psychosocial dis- The above finding is also consistent with pre- tress than male homeless participants. This viously reported findings in the literature hypothesis was partially supported. The study (Hwang, 2001; Lippert and Lee, 2015; Public revealed no significant gender differences in Health Agency of Canada, 2006). The Canadian loneliness, depression, anxiety, traumatization, Institute for Health Information (2007), for and somatic symptoms. However, female example, reported that homeless adults in Canada homeless participants reported higher stress and experienced about 24% higher stress level than greater suicidal behavior than the male home- their non-homeless counterparts. Similarly, Votta less participants. The observed gender differ- and Manion (2003) found that Canadian home- ences in some aspects of psychosocial distress less male youth had twice as high stress level may have resulted from socio-cultural norms in than Canadian non-homeless male youth. Ghana that sometimes tend to put females at a Research on homelessness in the U.S. confirmed disadvantaged position and cause them to suffer higher depressive symptoms, higher stress, unreasonably greater amounts of hardships. poorer health, and greater use of drugs and alco- Homeless females may also have greater vul- hol among youth who were homeless than those nerability to various forms of harassment who were not (Unger et al., 1998). including sexual harassment and physical In the African context, it has been reported assaults. These vulnerabilities together with that as many as 90% of homeless adults in adverse socio-cultural norms in the Ghanaian Ethiopia had experienced one form of mental society may have contributed to the higher disorder or another (Fekadu et al., 2014). stress level and greater suicidality among the Within the Ghanaian context, the study vali- homeless female participants. dates the earlier findings of De-Graft Aikins Johnson et al. (2017) found the intergenera- and Ofori-Atta (2007) that showed higher tional impact of homelessness to be highly gen- prevalence of physical and psychological dered. They found a stronger relationship stressors among homeless participants in between homelessness and physical health for Accra. It should be emphasized that the defini- females than for males. According to them, the tion of homelessness of De-Graft Aikins and indispensable role of women as heads of home- Ofori-Atta (2007) focused on individuals liv- less families made the impact of homelessness ing in squatter settlements. However, the pre- on children proportionally remarkable. In their sent study adopted a deeper definition of study, children under the care of homeless homelessness which incorporates a lack of women experienced poorer health outcomes. access to settlement structures. The present Those in schools exhibited lower achievement finding also agrees with the conclusion of on cognitive tests. Although the current study Asante et al. (2015) which indicated that home- did not assess intergenerational impact of less individuals in Ghana experience severe homelessness it is reasonable to assume that in psychosocial problems including hyperactivity, situations where homeless women in Ghana emotional difficulty, relationship problems, experience higher stress and suicidality, the and conduct disorders. Beyond these, the pre- wellbeing of children under their care will most sent study reveals higher incidence of trau- likely be compromised. The impact of women matic symptoms, depression, anxiety, stress, homelessness implies that any planned inter- somatization, loneliness, and suicidality among vention should seek to prioritize homeless homeless individuals in Ghana. women over their homeless male counterparts, Amissah et al. 9 especially in situations where homeless women Given that financial difficulty is one of the carry the burden of childcare in addition to their key causes of homelessness, it is recommended daily struggles. that job creation and skill training be integrated The findings of this study should be under- into interventional policies in Ghana. With relia- stood and interpreted within the socio-cultural ble employment, individuals can have the finan- context of the Ghanaian society. Generalized cial ability to afford housing and meet their daily implications of the findings beyond the confines needs. The task of creating jobs for the homeless of this study must be drawn with caution since the should be a shared responsibility between gov- investigated sample was, at best, only representa- ernment, the private sector, and other stakehold- tive of the Ghanaian population. In addition, it ers like non-government organizations (NGOs), should be noted that inherent differences between religious institutions, and philanthropic individu- the homeless group and the non-homeless com- als. This mission has the potential to see a com- parison group such as significant gap in educa- prehensive transformation of lives among the tion, family structure, and economic conditions homeless population in Ghana. may have contributed to the higher psychosocial Unarguably, homelessness and poverty are distress of the homeless. However, the current threaded together. Although financial aid may study did not examine the influence of these vari- not be a sustainable intervention, it can serve as ables. Therefore, future studies should examine a head start for talented homeless individuals correlates of homelessness that result into higher with unflinching determination to succeed. psychosocial distress. Sample size should also be Financial aid can also be tied to specific efforts increased in future studies to enhance the reliabil- such as housing, skill training, and entrepre- ity and generalizability of findings. neurial adventure. When tied to efforts, finan- Provision of housing should be fundamental cial aid can yield the most useful and lasting in the institution of any intervention that aims to results to homeless individuals. provide lasting and effective solution to home- lessness in Ghana. The request for housing inter- Conclusion vention is consistent with the housing-first policy (Atherton and Nicholls, 2008; U.S. The study has revealed the gravity of psychoso- Department of Housing and Urban Development cial distress that homeless people in Ghana expe- Office of Policy Development and Research, rience. The homeless, compared to non-homeless, 2007). Housing-first policy intervention does reported significantly higher levels of depression, not only entail provision of housing facilities, stress, anxiety, traumatic symptoms, loneliness, but more importantly the provision of essential somatic symptoms, and suicidality. Although no services that meet the social and healthcare significant gender differences were observed on needs of the individual. It has been proven that loneliness, depression, anxiety, traumatization, the application of the housing-first model has and somatic symptoms, female homeless partici- facilitated the coping abilities of many homeless pants were found to have higher stress and greater individuals in managing the myriad of problems suicidality than their male counterparts. The associated with homelessness (Atherton and observed gender differences in psychosocial dis- Nicholls, 2008). Although the housing-first pol- tress may have emanated from the biased socio- icy intervention may be expensive to implement cultural norms and discrimination against females in developing countries, the government of in the Ghanaian society. Given the gravity and Ghana through partnership with the private sec- pervasiveness of the psychosocial consequences tor can have a laid down plan to make housing of homeliness in Ghana, it is recommended that facilities both accessible and affordable to dis- the government of Ghana and other stakeholders advantageous groups in the population. This implement immediate interventions that will help may serve as a protective intervention for the at- to improve the wellbeing of homeless individuals risk population. in the Ghanaian society. 10 Journal of Health Psychology 00(0) Authors’ note Atherton IM and Nicholls CCM (2008) ‘Housing Nelly BF Amissah was previously known as Nelly First’ as a means of addressing multiple needs Betty Fosu. and homelessness. European Journal of Homelessness 2: 289–303. Baggett TP, Chang Y, Singer DE, et al. (2015) Consent to participate Tobacco-, alcohol-, and drug-attributable deaths Informed consent was obtained from all individual and their contribution to mortality disparities in participants included in the study. a cohort of homeless adults in Boston. American Journal of Public Health 105(6): 1189–1197. Declaration of conflicting interests Balasuriya L, Buelt E and Tsai J (2020) The never- ending loop: Homelessness, psychiatric disor- The author(s) declared no potential conflicts of inter- der, and mortality. Psychiatric Times, 37(5), est with respect to the research, authorship, and/or 12–14. publication of this article. Byrne BM (2010) Structural Equation Modeling With AMOS: Basic Concepts, and Programming, 2nd Ethics approval edn. New York: Routledge. This study was performed in line with the principles Byrne T, Munley EA, Fargo JD, et al. (2013) New of the Declaration of Helsinki. Approval was granted perspectives on community-level determinants by the University of Ghana Ethics Committee for of homelessness. Journal of Urban Affairs Humanities (Ref No. ECH/030/17-18). 35(5): 607–625. Calvo F, Carbonell X and Badia M (2018) Funding Homelessness and unemployment during the economic recession: The case of the City of The author(s) received no financial support for the Girona. European Scientific Journal 14(13): 59. research, authorship, and/or publication of this article. Canadian Institute for Health Information (2007). Improving the health of Canadians: Mental ORCID iDs health and homelessness. Ottawa: CIHI. Chamberlain C and Johnson G (2013) Pathways into Nelly BF Amissah https://orcid.org/0000-0001- adult homelessness. Journal of Sociology 49(1): 8948-8783 60–77. Benjamin Amponsah https://orcid.org/0000-0002- Chondraki P, Madianos MG, Dragioti E, et al. (2014) 2597-4229 Homeless mentally ill in Athens area: A cross- sectional study on unmet needs and help-seek- Supplemental material ing. International Journal of Social Psychiatry Supplemental material for this article is available 60(6): 544–553. online. Cleverley K and Kidd SA (2011) Resilience and suicidality among homeless youth. Journal of References Adolescence 34(5): 1049–1054. Corrigan P, Pickett S, Kraus D, et al. (2015) Abekah-Carter K and Oti GO (2022) Perspectives of Community-based participatory research exam- community members on homeless people with ining the health care needs of African Americans mental illness in Nsawam, Ghana. International who are homeless with mental illness. Journal Journal of Social Psychiatry 68: 196–202. of Health Care for the Poor and Underserved Anderson I and Christian J (2003) Causes of home- 26(1): 119–133. lessness in the UK: A dynamic analysis. Journal De-Graft Aikins A and Ofori-Atta AL (2007) of Community & Applied Social Psychology Homelessness and mental health in Ghana: 13(2): 105–118. Everyday experiences of Accra’s migrant Asante KO, Meyer-Weitz A and Petersen I (2015) squatters. Journal of Health Psychology 12(5): Correlates of psychological functioning of 761–778. homeless youth in Accra, Ghana: A cross-sec- Embleton L, Lee H, Gunn J, et al. (2016) Causes tional study. International Journal of Mental of child and youth homelessness in developed Health Systems 9(1): 1–9. and developing countries: A systematic review Amissah et al. 11 and meta-analysis. JAMA Pediatrics 170(5): Lovibond SH and Lovibond PF (1995) Manual for 435–444. the Depression Anxiety & Stress Scales, 2nd Falsetti SA, Resnick HS, Resnick PA, et al. (1993) edn. Sydney: Psychology Foundation. The Modified PTSD Symptom Scale: A brief McLeroy KR, Bibeau D, Steckler A, et al. (1988) self-report measure of posttraumatic stress dis- An ecological perspective on health promotion order. Behavior Therapy 16: 161–162. programs. Health Education Quarterly 15(4): Fekadu A, Hanlon C, Gebre-Eyesus E, et al. (2014) 351–377. Burden of mental disorders and unmet needs Nishio A, Horita R, Sado T, et al. (2017) Causes of among street homeless people in Addis Ababa, homelessness prevalence: Relationship between Ethiopia. BMC Medicine 12(1): 138–150. homelessness and disability. Psychiatry and Fox AM, Mulvey P, Katz CM, et al. (2016) Untangling Clinical Neurosciences 71(3): 180–188. the relationship between mental health and home- Nooe RM and Patterson DA (2010) The ecology of lessness among a sample of arrestees. Crime and homelessness. Journal of Human Behavior in Delinquency 62(5): 592–613. the Social Environment 20(2): 105–152. George D and Mallery M (2010) SPSS for Windows Osei Asibey B, Conroy E and Marjadi B (2020) Step by Step: A Simple Guide and Reference, Health problems and healthcare service utilisa- 10th edn. Boston, MA: Pearson. tion amongst homeless adults in Africa- a scop- Ghana Statistical Service (2002) 2000 population ing review. BMC Public Health 20(1): 594. and housing census. Accra: Ghana Statistical Osman A, Bagge CL, Gutierrez PM, et al. (2001) Service. The Suicidal Behaviors questionnaire-Revised Gierk B, Kohlmann S, Kroenke K, et al. (2014) (SBQ-R): Validation with clinical and nonclini- The somatic symptom scale-8 (SSS-8): A brief cal samples. Assessment 8(4): 443–454. measure of somatic symptom burden. JAMA Padgett D, Struening EL and Andrews H (1990) Internal Medicine 174(3): 399–407. Factors affecting the use of medical, mental Hair JF, Black WC, Babin BJ, et al. (2010) health, alcohol, and drug treatment services by Multivariate Data Analysis, 7th edn. Upper homeless adults. Medical Care 28: 805–821. Saddle River: Pearson. Padgett DK, Struening EL, Andrews H, et al. (1995) Hwang SW (2001) Homelessness and health. Predictors of emergency room use by home- Canadian Medical Association Journal, 164(2), less adults in New York City: The influence of 229–233. predisposing, enabling and need factors. Social Johnson G, Ribar DC and Zhu A (2017) Women’s Science & Medicine 41(4): 547–556. homelessness: International evidence on causes, Public Health Agency of Canada (2006) The human consequences, coping and policies. In: Averett face of mental health and mental illness in SL, Argys LM and Hoffman SD (eds) Oxford Canada. Ottawa, ON: Government of Canada. Handbook on the Economics of Women. New Russell D, Peplau LA and Cutrona CE (1980) The York: Oxford University Press, pp.1–9. revised UCLA Loneliness Scale: Concurrent Komla (2013) Homelessness in Ghana, is it my prob- and discriminant validity evidence. Journal of lem? Available at: https://www.ghanaweb.com/ Personality and Social Psychology 39: 472– GhanaHomePage/features/Homelessness-in- 480. Ghana-Is-it-my-Problem-269760 (accessed 10 Russell D, Peplau LA and Ferguson ML (1978) May 2021). Developing a measure of loneliness. Journal of Kroenke K, Spitzer RL and Williams JBW (2002) Personality Assessment 42(3): 290–294. The PHQ-15: Validity of a new measure for Russell DW (1996) UCLA Loneliness Scale (Version evaluating the severity of somatic symptoms. 3): Reliability, validity, and factor structure. Psychosomatic Medicine 64(2): 258–266. Journal of Personality Assessment 66: 20–40. Kushel MB, Gupta R, Gee L, et al. (2006) Housing Schütz CG (2016) Homelessness and addiction: instability and food insecurity as barriers to health Causes, consequences, and interventions. Current care among low-income Americans. Journal of Treatment Options in Psychiatry 3: 306–313. General Internal Medicine 21(1): 71–77. Smartt C, Prince M, Frissa S, et al. (2019) Lippert AM and Lee BA (2015) Stress, coping, and Homelessness and severe mental illness in low- mental health differences among homeless peo- and middle-income countries: Scoping review. ple. Sociological Inquiry 85(3): 343–374. BJPsych Open 5(4): e57. 12 Journal of Health Psychology 00(0) Swanson KA, Andersen R and Gelberg L (2003) to homeless persons with serious mental illness. Patient satisfaction for homeless women. Final report. Washington, DC: U.S. Department Journal of Women's Health 12: 675–686. of Housing and Urban Development Office of Toro PA (2007) Toward an international understand- Policy Development and Research. ing of homelessness. Journal of Social Issues Vaughn RE (2017) Interpersonal Violence and 63(3): 461–481. Homelessness. Doctoral dissertation, Adler Unger JB, Kipke MD, Simon TR, et al. (1998) Stress, University, Chicago, IL. coping, and social support among homeless youth. Vázquez JJ, Panadero S and Zúñiga C (2018) Journal of Adolescent Research 13(2): 134–157. Attributions about homelessness in homeless Urbanoski K, Veldhuizen S, Krausz M, et al. (2018) and domiciled people in Madrid, Spain: “Why Effects of comorbid substance use disorders on are they homeless people?”. American Journal outcomes in a Housing First intervention for of Orthopsychiatry 88(2): 236–247. homeless people with mental illness. Addiction Votta E and Manion IG (2003) Factors in the psy- 113(1): 137–145. chological adjustment of homeless adolescent U.S. Department of Housing and Urban Development males: The role of coping style. Journal of the Office of Policy Development and Research American Academy of Child and Adolescent (2007) The applicability of housing first models Psychiatry 42(7): 778–785.