Health and well-being of the homeless in South African cities and towns

Abstract

Little is known about the health and well-being of people who live on the streets although their lifestyle involves health risks. This study used qualitative and quantitative methods to explore the health status and health service needs of homeless people in several South African cities and towns. It was found in some cases that their own or family members’ ill-health had contributed to their homelessness. Illnesses associated with poor living conditions or lifestyles were more common in the homeless than in the general population in certain age and sex categories. Access to healthcare was generally good in urban centres and most respondents were satisfied, although some reported discriminatory treatment. The risk profile of the homeless shows that systematic health promotion is required and that some health service providers need to be made more aware of the health needs of the homeless.

Keywords:

• health
• well-being
• homelessness
• HIV/AIDS
• alcohol and substance abuse
• mental illness
• violence
• disability
• access to health services
• South Africa

1. Introduction

Homeless people are frequently stigmatised because people assume they have alcohol or other substance abuse problems, and many are labelled as mentally ill. However, there is relatively little hard data on the health status of the homeless, especially in Africa. This information is important for planning health services. We also need to know more about the direction of the relationship between homelessness and health. In other words, do certain health problems contribute to people becoming homeless or does homelessness itself contribute to ill-health? For example, alcoholism may cause people to become homeless because they have lost their job or been rejected by family members, but the desperation of being homeless may in turn cause them to turn to alcohol. Similarly, mental illness may lead to a person being expelled from the housed population but being homeless may itself contribute to anxiety and depressive disorders. There is a growing body of evidence that homelessness should not be regarded as a homogeneous health risk but that it is more appropriate to consider street people as part of a continuum of risk profiles associated with poverty. Panter-Brick (2002) provides a comprehensive review of these issues and evidence that homeless people may be just as healthy (or just as sick) as the poorer housed population.

Apart from the direct health impacts of homelessness, such as those resulting from exposure to the elements (e.g. respiratory infections, hypothermia, skin diseases), both direct and indirect consequences of substance abuse are common in this population (Hwang, 2000, 2001; Frankish et al., 2005). Direct consequences include liver disease associated with excessive alcohol consumption and the toxic effects of various narcotics. Indirect effects are less well known, but include risky sexual behaviour – multiple partners, prostitution and sexual abuse (affecting women and youth disproportionately) – which is likely to result in a high prevalence of sexually transmitted infection (STI) and HIV in the homeless population (Manzon et al., 1992; Allen et al., 1994; Haley et al., 2002). Alcohol and drug abuse is also strongly correlated with violence and unintentional injury, which can be another feature of the lives of homeless people (O'Toole et al., 2004).

Much of the literature on homelessness is based on observations from developed countries, notably the US (Gelberg & Linn, 1989; Brickner et al., 1993; Kuhn & Culhane, 1998; Kushel et al., 2001) and Canada (Hwang, 2000, 2001; Frankish et al., 2005), and although there are studies from developing countries, the majority tend to be of small samples (for example, Nzimakwe & Brookes, 1994; Olufemi, 2000, 2002). The causal pathways for First-World homelessness differ from those seen in developing countries, but many of the health problems (notably alcohol and other substance abuse, and aspects of HIV/AIDS and tuberculosis [TB] risk) are likely to be similar.

Arriving at a reliable definition of homelessness is critical for making meaningful comparisons between areas or different studies (see Cross & Seager et al., this issue). More importantly, if we are to understand the relationship between homelessness and health, we need to know more about the lives of homeless people in order to understand the health risks involved. Various definitions, and a wide variety of survey methods, have been used in the US and Canada (Peressini et al., 1995). In many ways, homelessness can be regarded as a continuum, ranging from people who may be at risk of becoming homeless to those who currently have absolutely no shelter of their own and live and sleep ‘on the streets’ (Frankish et al., 2005). Chung (1991) divides the at risk homeless into five subgroups: at-risk renters, at-risk homeowners, street homeless, shelter homeless, and by-choice homeless. This last group is perhaps the most challenging in terms of intervention strategies. Another important consideration is that homelessness is not normally a permanent condition but may be subject to lifecycle effects (more prevalent for certain age groups) and seasonality.

From a policy perspective, the at-risk homeless – whether renting or owning homes, but liable to lose them – are potentially the most interesting, since successful interventions at this level should help to prevent further homelessness. However, when it comes to immediate needs for health and other services, the ‘absolute homeless’ have to be given priority. As with all public health interventions, prevention is the preferred intervention strategy but where there are large populations of sick people, resources must be brought to bear on current problems, while at the same time addressing prevention of further ‘cases’.

For the purposes of this study, we focused on the absolute homeless. This definition includes people who sleep in the open, one or more nights per week, and those making use of shelters specifically for the homeless. Sleeping ‘in the open’ includes sleeping in doorways, cars, temporary cardboard or plastic shelters, and so on, but excludes any form of permanent residential structure such as a shack. Those sleeping in derelict buildings, under bridges and in public areas of railway, bus or taxi stations are included.

Given the relative paucity of information on the health of the homeless in South Africa, the first phase of this study opted for a formative research approach using key informant interviews and focus groups, with some additional information derived from secondary data sources. The first phase was carried out in Cape Town. Subsequently, a wider scale quantitative survey in cities and secondary towns in Gauteng, Mpumulanga and Limpopo was carried out to obtain more detailed information on the health of homeless people and their use of health services.

2. Aims and objectives

The primary aim of the project was to explore the health characteristics of homeless adults and children in South African urban areas. The secondary aim was to assess health service access for the homeless. Specific objectives were to assess risk behaviours for sexually transmitted infections, including HIV, alcohol and other substance abuse, mental illness, violence and disability; and to quantify self-reported health problems for homeless adults and children.

3. Methods

Key informant interviews were held in Cape Town and Johannesburg with people having responsibility for the homeless, such as Social Development Facilitators, and representatives from non-governmental organisations providing services to the homeless. Homeless people living on the streets or in homeless shelters in Cape Town and Johannesburg were invited to participate in focus groups of 10–16 people to explore experiences of homelessness and health. For those living on the streets, homeless shelter outreach workers were asked to contact people and to request them to come to the shelter for the discussion. The categories of respondents are summarised in Table 1.

Table 1: Summary of key informants and focus group participants

Description	Key informants	Focus groups (10–16 in each group)
Municipal officials	4
Shelter staff (executive, managers, fieldworkers)	12
Homeless support groups (feeding, training)	8
Adults living in shelters (male and female separately)		3
Adults living on the street (male and female separately)		3
Children/youth living in shelters (boys and girls separately)		3
Children/youth living on the street (boys and girls separately)		3

Males and females were interviewed in separate groups because there are gender-specific aspects to life on the streets that people might not have been comfortable discussing in mixed groups. The focus group participants ranged in age from 12 years (selected as the minimum age for ethical reasons) to 73 years.

Among the topics addressed were health issues, including recent illness and chronic conditions, disability, HIV/AIDS and health-related factors contributing to leaving home, access to health services, safety and security, sexual risk behaviour, and alcohol and substance abuse.

Although the participants were volunteers, and therefore demonstrate some self-selection bias, it was notable that the majority of homeless adults appeared quite willing to talk about their experiences. They seemed to appreciate ‘being listened to’ and welcomed the opportunity to express their views. However, restrictions on topics covered with children were imposed on the research team by some facility managers, who felt that certain questions were ‘too sensitive’ for street children to answer. Most of the themes causing concern for the caregivers were related to home background and reasons for being on the street. This situation was regrettable but had to be accommodated in the formative stage of the research. Given more time, it is hoped that a more trusting relationship with shelter staff, and the children they care for, could be achieved and more in-depth interviews completed. Children who are on the streets are inevitably vulnerable but their resilience should not be underestimated by those who are looking after them. In most cases, these children are far from helpless and demonstrate remarkable resourcefulness. This being so, it may be that they would be the best placed to decide which questions they prefer not to answer.

On the basis of the results of the qualitative research, a more detailed questionnaire was developed for a survey of people living on the streets. As far as possible, questions on health issues were taken from previously validated questionnaires used in international surveys such as the South Africa Demographic and Health Survey (Department of Health et al., 1998) and instruments developed by the World Health Organization for issues such as alcohol and other substance abuse (WHO, 1992), mental health and disability (United Nations Economic and Social Council, 2007). People who appeared to be homeless were approached, on the streets, by trained fieldworkers, and asked if they were willing to take part in an interview. The quantitative survey was carried out in Gauteng, Mpumalanga and Limpopo, and the sampling strategy is described in more detail in Kok et al. (this issue). In total, 942 adults and 305 children were interviewed. While the questionnaire addressed a broad range of issues to do with homelessness, this paper addresses only the health and well-being sections. These included topics such as reasons for homelessness, violence, disability, alcohol and other substance abuse, access to and use of health services and sexual risk behaviour.

While every effort was made to include all types of homeless people in the quantitative sample, the circumstances of homeless people made it impossible to draw a fully representative sample. Although there were few refusals, participation was voluntary and required coherent respondents who were not severely impaired by alcohol, drugs or mental illness, which resulted in some selection bias. For this reason, the analysis is based on the proportional prevalence of various conditions. Comparisons with national representative data is made in general terms and statistical comparisons were considered inappropriate owing to the different nature of the samples involved.

The study was approved by the Human Sciences Research Council Research Ethics Committee and was conducted in accordance with the principles of the ethical guidelines for conducting research involving homeless people (Asfour, 2004). Interviewers were trained to explain the purpose of the survey to potential respondents, inform them about their rights and benefits, obtain informed consent or assent (according to age), and conduct interviews in a neutral way, without any coercion. Confidentiality measures were put in place to protect respondents’ privacy. Homeless adults were compensated for their time (up to 1 hour), with a donation of R20 (about US$3), and an equivalent donation was given to children's shelters for each child participant. Shelter managers dealing with street children were adamant that younger children should not receive cash incentives since this could place them at added risk (e.g. exposure to violent theft or drug abuse).

4. Results

4.1 Socio-demographic profile of homeless people in Cape Town

According to city council officials, the numbers of homeless people in Cape Town and the surrounding area have increased during recent years. They range in age from about eight years to over 90 years of age but no reliable estimates for the total number of people currently living on the street were available. Data from the Haven group's night shelters, which only cater for adults, indicate that the median age for male and female shelter residents was almost identical (54 and 55 years, respectively) but there were almost three times as many homeless males (74 per cent) as females (26 per cent) living in the shelters (Figure 1). Street people in Cape Town represent all race groups and are predominantly from the surrounding suburbs and townships, but some come from the Northern and Eastern Cape, KwaZulu-Natal and Gauteng.

Figure 1: Age distribution of the Cape Town shelter population

For homeless adults, the main reasons for homelessness were retrenchment, alcohol abuse, divorce and domestic violence, especially for women. Health factors, including mental illness, and in a few cases HIV infection or AIDS, were reported to be contributory factors for becoming homeless. A senior manager of a homeless shelter elaborated:

When people have mental problems they are thrown out of their homes. They were in the past sent to mental institutions, but since the government closed down such facilities, they have nowhere to go, except to stay on the street.

For children, domestic violence, negligence due to alcohol abuse and economic hardships, and sexual abuse of young girls were common reasons for leaving home. Some families were driven out of their homes by insolvency.

Conditions at home were generally such that very few people said they would be willing to return to their homes. As an outreach worker for homeless adults said:

homeless people do not like to talk about their backgrounds and few of them would ever consider to rejoin the family members.

4.2 Demographic profile of the sample from Gauteng, Limpopo and Mpumalanga

The sample included 942 adults (≥18 years) and 305 children (12–17 years), and 88 per cent were male. Ages ranged from 12 years (a requirement of the sampling protocol) to 75 years, with adults spread fairly evenly across the full age range but children showing a negatively skewed distribution peaking at 17 years of age. Ninety-six per cent of the sample was black African, 3 per cent coloured, 1 per cent white and there was one Indian respondent (0.1 per cent). The vast majority (95 per cent) answered ‘Yes’ to the question ‘Do you consider yourself homeless?’

4.3 Health-related activities of organisations working with the homeless

Most of the organisations providing services to the homeless do not see their primary function as health related, but by addressing the basic needs of food and shelter they clearly contribute to their clients’ health. This is in line with the WHO's definition of health as ‘a state of complete physical, mental, and social wellbeing, and not merely the absence of disease or infirmity’ (WHO, 1986). In addition, most formal shelters address direct health needs by referring clients to state health centres, such as clinics and day hospitals, treating minor injuries and skin conditions such as lice and scabies, and assisting with treatment for chronic conditions such as TB. There are social workers attached to most of the homeless shelters.

While catering for both males and females, most adult shelters provide only single-sex sleeping facilities. This was reportedly a factor that discouraged homeless couples and families from entering the shelters since they would naturally prefer to stay together. The manager of one homeless shelter commented:

There are too many families living on the street, we cannot cope. The reason many families do not like to stay in homeless shelters is because they do not like to be separated from their spouses. Many homeless shelters do not have arrangements for families.

The adverse social and health consequences of single-sex accommodation are well documented and have been strongly criticised in the context of workers’ hostels (Ramphele, 1993). Similar adverse impacts on the health of shelter residents are likely. There is awareness of this shortcoming of the current facilities and at the time of the study a new shelter was being built by the Haven Group in Cape Town, which will have some family accommodation.

Organisations serving homeless children or teenagers are divided along gender lines. This practice is said to help prevent potential sexual abuse but it can have the negative consequence of separating siblings. A 12-year-old girl at one homeless shelter asked the researcher:

Do you have the telephone numbers for Khayelitsha children's home for me? I would like to call my younger brother who stays there. I haven't spoken to him for long time now.

Nor do single-sex shelters guarantee an end to the abuse of younger boys by older ones. One of the shelter managers said that:

many boys do not like to go to stay in homeless shelters. They complain about abuse by staff and other children. One boy has been chased away from a shelter because he molested the younger ones. This young boy was apparently also molested, so he was seemingly acting this abuse out on others… It's a sad story.

4.4 Safety and security

Homeless people are very vulnerable to injuries and assaults (including sexual assault). Young boys complained about beatings and robbery carried out by older boys. Older boys mentioned gang fights, drugs and alcohol-related violence, as well as abuse by members of the public. A 15-year-old boy described the ordeals faced by younger street children:

Life on the street is not nice at all. We get beaten by older boys and drunk people. They sometimes just pee on you when you are sleeping.

Regarding gang-related violence, young men had the following to say:

It is not safe at all. There are lots of gangs that are fighting against each other. We are also targeted by the police. Whenever there is a reported theft or robbery, we are the first suspects. We live in fear.

It is not safe at all. I was once beaten with a knobkerrie while asleep.

Street children fight a lot. One of our close friends was recently stabbed to death. Another one was critically injured in a fight over parking money.

It was also reported by outreach workers that young boys were sexually assaulted by older ones, especially during the period immediately after their arrival in the city centre. However, the boys denied sexual assaults (possibly through embarrassment at making such disclosures publicly).

While girls are generally protected by their boyfriends from assault, fights between girls over boyfriends were reportedly common. Many of these fights reportedly resulted in serious injuries.

Stab wounds and gunshot injuries were reported by both outreach workers and street boys to be common. These injuries resulted from fighting among homeless children, and assaults on children by members of the public, who ascribe criminal intent to them. For some children, these injuries prove fatal. Stories such as the ones below were common:

I was shot with two bullets while gambling with friends. I won most sessions on that particular day.

I once attempted to rob somebody with a toy gun. He produced his, and shot me on my legs. I couldn't run away after that. He swore at me and left me for dead.

I got injured by a cross bullet. Some people were shooting at each other and I got hit in the process. My friends called an ambulance for me. We help one another when we are sick.

I was also shot with two bullets last year by a friend of a street boy whom I shot dead. He was apparently revenging the death of his friend.

We are easily targeted by motorists. My friend and I were once shot at by a motorist who suspected that we were trying to rob him. Unfortunately, my friend died from the wounds.

We were once playing with crickets [a toy that makes a loud clicking noise] as it was nearing year end. A passer-by thought we were shooting at him and he shot back at us, I was injured in the process.

Children reported harassment and assault at the hands of members of both the South African Police Service and the Metro Police. They claimed that whenever a criminal offence is being reported, they are the first suspects, and that the police chase them away from vacant flats and other buildings in the Johannesburg central business district.

Twenty-two per cent of the children and 22.5 per cent of adults interviewed in the survey said that they had been injured or assaulted within the past six months, and although some of these injuries may have occurred in an abusive home environment, it appears that life on the streets is frequently violent. Table 2 shows that about one-half of the injuries reported for both adults and children were the result of being ‘beaten up’ and a substantial number had been stabbed. Given that those reporting these injuries and assaults are the ones who survived, the real figure could be considerably higher. While fatal incidents will always tend to receive more attention, the frequency with which violent deaths of street people, including children, were mentioned by both shelter staff and focus group participants suggests that this is indeed a relatively frequent occurrence. Sexual assault may have been under-reported but appears to be slightly more common among youth than adults.

Table 2: Summary of types of injury and assault (in the past six months)

Causes of injury	Child (<18 years)	Adult (≥18 years)
Beaten up	34 (51.5%)	100 (47.4%)
Stabbed	16 (24.2%)	46 (21.8%)
Hit by vehicle while a pedestrian	5 (7.6%)	23 (10.9%)
Vehicle accident while passenger or driver	0	12 (5.7%)
Sexually assaulted	5 (7.6%)	11 (5.2%)
Injured at work	0	12 (5.7%)
Shot	1 (1.5%)	8 (3.8%)
Other (including undefined injuries and falls)	6 (9.1%)	13 (6.2%)
Total	66	211
Percentage of overall sample injured or assaulted	22.0%	22.5%

4.5 Health

The initial health risk that street people face is lack of shelter and the resultant exposure to cold and damp. Other risks are lack of access to hygiene facilities (Nzimakwe & Brookes, 1994), inadequate nutrition, substance abuse, sexually risky behaviour (Swart-Kruger & Donald, 1994), vulnerability to traffic accidents (often when intoxicated), and high levels of violence and abuse.

In response to questions about health problems, outreach workers mentioned STIs, TB, HIV infection and AIDS as the most common health problems experienced by street people, other than injuries. Prevalence rates are beyond the scope of this study but, going on the reports of sexually risky behaviour, are likely to be high. Skin diseases and malnutrition were also reported to be common. In the literature, skin diseases identified include sores, rashes and scabies (Nzimakwe & Brookes, 1994).

Pregnancy among homeless girls was cited as a common problem. One programme coordinator reported that, in 2005 alone, six children were born on the streets of Hillbrow, Johannesburg. While pregnancy should not normally be a health risk, it has serious implications for the health of both mother and child in the context of a street lifestyle with its own inherent risks, and where adequate nutrition and antenatal care are unlikely to be available. It should also be noted that fewer of the young people we interviewed than might be expected from this information had their children with them while on the streets. This suggests that they had placed the children in some form of care and then returned to the streets themselves.

4.6 Disability

A surprisingly high proportion of children and adults reported some degree of limitation in mobility caused by pain – 14 per cent of children and 19 per cent of adults. Experience of disability is summarised in Tables 3 and 4. Difficulties in seeing were reported by about 13 per cent of children and 19 per cent of adults. The figure for adults (19 per cent) is similar to the estimate from a population based survey carried out by Statistics South Africa (2007) – 20.4 per cent – although their sample was slightly younger, including those from 15 years of age. Difficulty hearing was reported by 9 per cent of children and 11 per cent of adults, compared with only 7.3 per cent of those over 15 years in the national survey. About 5 per cent of children and 15 per cent of adults had difficulty climbing stairs or walking, compared with 11.7 per cent reported for the national survey. Difficulties with self-care (washing all over or dressing) were reported by 5 per cent of children and 14 per cent of adults. This figure is far in excess of the 1.7 per cent figure reported in the Statistics South Africa survey, which suggests that the question might have been misinterpreted to mean ‘not having the means to wash easily’, rather than difficulties in doing the actual activity. About 6 per cent of children and 9 per cent of adults reported difficulty concentrating or remembering things, which may be indicative of a mental health disorder, a formal learning disability or substance abuse. However, this figure was lower than the 12.9 per cent found in the national survey. About 5 per cent of children and 8 per cent of adults had difficulty communicating owing to physical, mental or emotional conditions. These figures are substantially higher than the 2.6 per cent found in the national survey and could be indicative of emotional difficulties, which may also put people at risk for becoming homeless.

Table 3: Disability reported by children under 18 years old

Description of disability	No difficulty	Yes, some difficulty	Yes, a lot of difficulty	Cannot do at all
Difficulty seeing, even if wearing glasses?	267 (87.5%)	29 (9.5%)	9 (3.0%)	0
Difficulty hearing, even if using a hearing aid?	278 (91.1%)	18 (5.9%)	9 (3.0%)	0
Difficulty walking or climbing stairs?	290 (95.1%)	13 (4.3%)	2 (0.7%)	0
Difficulty remembering or concentrating?	287 (94.1%)	17 (5.6%)	0	0
Difficulty with self-care such as washing all over or dressing?	291 (95.4%)	11 (3.6%)	3 (1.0%)	0
Difficulty communicating because of a physical, mental or emotional health condition (e.g. understanding others or others understanding you)?	290 (95.1%)	15 (4.9%)	0	0

Table 4: Disability reported by adults aged 18 and over

Description of disability	No difficulty	Yes, some difficulty	Yes, a lot of difficulty	Cannot do at all
Difficulty seeing, even if wearing glasses?	759 (80.7%)	141 (15.0%)	40 (4.3%)	0
Difficulty hearing, even if using a hearing aid?	835 (88.6%)	90 (9.6%)	16 (1.7%)	1 (0.1%)
Difficulty walking or climbing stairs?	804 (85.4%)	98 (10.4%)	38 (4.0%)	1 (0.1%)
Difficulty remembering or concentrating?	847 (90.7%)	78 (8.4%)	9 (1.0%)	0
Difficulty with self-care such as washing all over or dressing?	807 (85.9%)	97 (10.3%)	33 (3.5%)	3 (0.3%)
Difficulty communicating because of a physical, mental or emotional health condition (for example, understanding others or others understanding you)?	869 (92.3%)	61 (6.5%)	11 (1.2%)	0

Using a fairly strict cut-off point that classifies only those respondents referring to ‘a lot of difficulty’ and ‘cannot do at all’ as disabled for the activities of daily living shown in Table 3 and Table 4, nearly 7 per cent of the children and 12 per cent of adults could be classified as disabled (excluding the possibly misunderstood question relating to self-care). This degree of disability is described by the Statistics South Africa (2007) report as reflecting ‘people who are in need of services such as social assistance, welfare, health, reasonable accommodation at work and inclusive education … and who are more likely to experience significant disadvantage due to their difficulty’. The prevalence of disability in our homeless sample is slightly higher than that reported in the national survey (10 per cent), although the slightly older age range could explain this. However, there is cause for concern where people with disabilities are living under such difficult circumstances.

Ironically, disabled shelter residents (15 per cent of the Haven Group's shelter population) were seen as being in a more secure position than many of the able-bodied persons because they were more likely to be receiving a government grant. Conversely, the sad reality for some disabled children is less positive: unscrupulous adults sometimes take advantage of them to facilitate begging. A shelter manager observed:

They make the younger ones do their begging for them, because people are more sympathetic towards the younger kids. And it is the ‘hoppies’ [the disabled ones] who get more money.

4.7 Mental illness

Nearly one-half of the children (45 per cent) and 58 per cent of adults reported experiencing symptoms of depression during the past 30 days (Table 5), although only 3 per cent of children and 6 per cent of adults said that they had been diagnosed with a psychiatric illness, such as schizophrenia or mood disorder. This may represent an under-count, since many might not have remembered the diagnosis or might not have been able to access services (and therefore not received a diagnosis), but these figures suggest a substantial unmet need for mental health services among this population.

Table 5: Self-reported symptoms of mental illness

Respondents reporting symptoms of depression (‘At any time in the last 30 days have you felt sad, empty or depressed for a period lasting several days?’)	Child (<18 years)	Adult (≥18 years)
Never	169 (55.4%)	393 (41.7%)
Some of the time	84 (27.5%)	279 (29.6%)
A lot of the time	37 (12.1%)	133 (14.1%)
Most of the time	10 (3.3%)	92 (9.8%)
All the time	5 (1.6%)	45 (4.8%)
Total	305	942

Shelter staff did not generally identify mental illness when describing children's health problems. However, one South African study found that approximately one-third of street children sampled showed signs of anxiety or depression, and one-third had some kind of physical deformity, perceptual problem or manifest psychological disorder, and one-third had received a blow to the head, which is a risk factor for a number of neuropsychiatric problems (Richter, 1991). This failure by shelter staff to identify mental health problems may indicate that they are not aware of the signs of these problems, or that they face difficulties in obtaining mental health services for children in need. In Cape Town, shelter staff reported that mental health services are difficult to access for children in their care. High rates of mental health problems may be associated with substance abuse.

4.8 Substance abuse

Rates of substance abuse are reported to be high among street people, and this could easily be observed in the course of the fieldwork. Many young people came to focus group discussions visibly dizzy from sniffing glue. This was particularly true of those children who were still on the streets and not living in shelters. Many readily admitted to the practice. When asked whether homeless children ever take drugs, facing her friends, an 18-year-old girl responded affirmatively:

We all sniff glue, daily, let's be honest …

And another added:

I sniff glue, smoke dagga and cigarettes daily. I sometimes take tik, cocaine and Mandrax.

Access to rehabilitation centres, however, was reported to be severely limited. Key informants stated that there were no rehabilitation centres for young people in the Johannesburg city centre (where other services for street children are concentrated). The only facilities near the city centre were private, expensive, and thus inaccessible to homeless people.

Thirty-seven per cent of children and 23 per cent of adults in the street survey reported having used recreational drugs at some time. Current drug use was similar, with a few adults saying they had previously used but were no longer using drugs. Use of the so-called harder drugs, during the past three months, was reported relatively infrequently. Mandrax (methaqualone combined with diphenhydramine), heroin, ‘tik’ (crystal methamphetamine) and ecstasy (MDMA) were all being used by less than 5 per cent of children or adults (Table 6). Much more frequent use of dagga (marijuana) was reported (23.3 per cent for children and 18.5 per cent for adults). Solvents, such as glue and thinners, were the substances most commonly abused by children (25.9 per cent). While it is difficult to obtain directly comparable age-specific data for comparison with the general population, these rates indicate high levels of substance abuse (personal communication, Charles Parry, Director, Alcohol & Drug Abuse Research Unit, Medical Research Council, Cape Town, 10 June 2008). Pettifor et al. (2004) reported that 9.1 per cent of 15–24-year-olds had used dagga at some time.

Table 6: Self-reported drug use

	Child (<18 years)	Adult (≥18 years)
Those who had ever used recreational drugs	121 (37.0%)	217 (23.1%)
Current drug use; that is, during the past three months (multiple responses possible)	121 (37.0%)	206 (21.9%)
Dagga (marijuana)	71 (23.3%)	174 (18.5%)
Glue, thinners, petrol	79 (25.9%)	33 (3.5%)
Mandrax	15 (4.9%)	14 (1.5%)
Heroin	8 (2.6%)	8 (0.8%)
Tik	8 (2.6%)	11 (1.2%)
Ecstasy	6 (2.0%)	9 (1.0%)
Other	14 (4.6%)	14 (1.5%)

In contrast, regular use of alcohol appears to be relatively uncommon among children, with only 3.7 per cent admitting to using alcohol more than two or three times per week. Alcohol consumption was more common in adults, with 21.4 per cent reporting regular use of alcohol. Of the drinkers, over two-thirds reported bouts of heavy drinking (six or more drinks on the same occasion), although this was a relatively infrequent occurrence (Table 7).

Table 7: Self-reported alcohol use

	Child (<18 years)	Adult (≥18 years)
Drinking alcohol two to three times per week or more	11 (3.7%)	201 (21.4%)
Number of drinks consumed when drinking
1 or 2	20 (36.4%)	119 (27.1%)
3 or 4	19 (34.5%)	163 (37.1%)
5 or 6	12 (21.8%)	78 (17.8%)
7, 8 or 9	0	38 (8.7%)
10 or more	1 (1.8%)	27 (6.2%)
Don't know	3 (1.0%)	14 (3.2%)
Frequency of drinking six or more drinks on one occasion for those who drink
Never	17 (30.9%)	133 (30.3%)
Less than monthly	8 (14.5%)	84 (19.1%)
Monthly	13 (23.6%)	88 (20%)
Weekly	13 (23.6%)	88 (20.0%)
Daily or almost daily	1 (1.8%)	50 (11.4%)
Don't know	3 (3.5%)	9 (2.1%)

On the basis of the WHO's Alcohol Use Disorders Identification Test (WHO, 1992), included in the questionnaire for adults, 25 per cent of males and 19.8 per cent of females could be classified as potentially hazardous drinkers (using a cut-off point of 10, the recommended cut-off point for higher specificity of the screening instrument) (see Table 8).

Table 8: Assessment of potentially hazardous drinking among adults

	Male	Female
Hazardous drinkers (AUDIT≥10)	203 (25.0%)	25 (19.8%)
Non hazardous drinkers (AUDIT<10)	610 (75.0%)	101 (80.2%)
Total	813	126
AUDIT, Alcohol Use Disorders Identification Test (WHO, 1992).

4.9 Sexual risk behaviour

The young men who participated in the focus groups all said they used condoms every time they had sex. A 21-year-old man stated, waving a condom pulled from his back pocket:

I do not have a girlfriend, but I use a condom every time I have sex. Here is my condom. I always carry it with me.

However, outreach workers, who knew that pregnancy and STIs were a common occurrence, said there was reason to doubt these claims. A programme coordinator had the following to say:

STI cases are very common, as well HIV/AIDS. They sleep around a lot without using protection. Another thing – when a girl's boyfriend is sent to jail, one of the [other] boys gets involved with her.

Girls complained that boys were dishonest about the use of condoms – they said they either refused to use them or deliberately tore them, in order to get the girls pregnant.

The questionnaire survey provided more reliable estimates for condom use, with 36.7 per cent of adults and 22.9 per cent of adolescents reporting condom use at last intercourse. These rates of condom use are broadly comparable with those reported by Shisana et al. (2005) for a nationally representative sample of adults 25–49 years of age. Just over 7 per cent of adults and 5.5 per cent of young people reported having had a sexually transmitted infection in the past 12 months.

For girls, it appeared that having multiple partners was a livelihood and protection strategy, while for boys it was proof of manhood. When asked whether they were forced to have boyfriends when living on the streets, young women responded affirmatively, saying they offered physical protection and financial security:

Yes. You get protection that way … against rape by other street boys.

Yes. Who will look after you? Where will you get money to buy soap, clothes …?

It's in your own interest to have another man, sooner.

Access to condoms was reported to be fairly easy. Key informants said most health centres in the city, as well as shelters and some drinking places, offer free condoms – and the respondents agreed, saying they knew where to find them.

Street people are at high risk of contracting HIV or other STIs, since they tend to be sexually active from a young age, have more partners, are less likely to use condoms, tend to use alcohol and drugs more frequently than other groups (which may impair their decision making about safe sex), and they are vulnerable to rape and ‘survival sex’ (sex in exchange for food and other essentials). Although studies suggest that street children's AIDS knowledge is comparable with other hard-to-reach groups, they also suggest that worrying about infection is not a high priority for children who have more immediate needs for food, clothing and shelter (Swart-Kruger & Richter, 1997).

4.10 Access to health services

Homeless people reported that they were able to access health facilities, and the vast majority were satisfied with the treatment they received. However, street children said they were not always well treated at the health centres if they were not accompanied by a shelter staff member. Cape Town shelter staff echoed this, describing how they sprayed children with deodorant before they took them to the clinic, as nurses often did not treat street children well because they took exception to the way they smell.

A young woman aged 18 years complained bitterly about the way she was treated when she went to give birth at Hillbrow hospital in Johannesburg:

They [health officials] do not treat you well at all. When you go to the clinic and you are pregnant, they make all funny comments about you, and they laugh at you. When I went into labour for my child, I nearly lost my baby and my own life, due to the bad treatment at the hands of officials. They did not attend to me when I was in pain. They left me alone to suffer and only came at their own time. By this time, I nearly gave birth. I was almost holding my baby.

Another young man, with a fresh plaster cast on his arm, described his treatment:

When I was taken to hospital after the shooting incident, nurses were shouting at me. They refused me hospital's bedding facilities saying I was drunk and dirty. I slept with no bedding the whole night. I was hurting badly.

By contrast, homeless youth reported, with enthusiasm, that they were always treated well at the facilities run by non-governmental organisations.

The survey confirmed that contact with health services was overwhelmingly at government facilities, primarily clinics and day hospitals, with adults using a range of complementary services such as pharmacies and traditional healers. The literature indicates that tooth decay is a common problem for street people and it is therefore of concern that so few respondents mentioned using dental services (Table 9).

Table 9: Reported use of and satisfaction with health services during the past three months

Health services	Children making use of service (<18 years)	Of those children using service, proportion dissatisfied (<18 years)	Adults making use of service (≥18 years)	Of those adults using service, proportion dissatisfied (≥18 years)
Government hospital/government clinic	45 (14.8%)	3 (6.8%)	238 (25.3%)	25 (11.2%)
Day hospital	10 (3.3%)	2 (20%)	87 (9.3%)	14 (16.3%)
Retail pharmacy	6 (2.0%)	1 (16.7%)	55 (5.8%)	4 (7.4%)
Faith healer	–	–	21 (2.2%)	0
Traditional healer or herbalist	–	–	27 (2.9%)	1 (3.7%)
Home-based care services/house visit/treatment at a shelter	–	–	10 (1.1%)	1 (10.0%)
Dentist/oral hygienist/oral therapist	–	–	12 (1.3%)	0
Private doctor	–	–	13 (1.4%)	0

4.11 Illness and treatment

Although some shelters assist with treatment supervision (e.g. directly observed therapy for TB), most health problems are referred to the social worker attached to the shelter or directly to a local clinic or day hospital. Some staff raised a concern that they need more information about precautionary measures to be taken when working with infectious diseases (such as TB or HIV). In terms of services to be offered, more health promotion talks may be needed, for both residents and staff. A common comment was that when they identified a health problem, such as alcohol or drug abuse, they could find few accessible rehabilitation facilities.

The reported history of diagnosed illness was broadly comparable with that of the general population (see South Africa Demographic and Health Survey 2003 in Department of Health et al., 2004), although for many of the conditions – notably heart attack/angina, high cholesterol levels and diabetes – there appears to be excess morbidity for certain age groups of homeless women (see boxed and bolded cells in Table 10). Conversely, for several conditions, age-specific prevalence rates among homeless men were slightly lower than might be expected. This could imply that homeless men are relatively healthier than the general population, which seems unlikely, or that there is undiagnosed illness in this population. Given that the prevalence of chronic illness differs markedly by race, the prevalence of common illnesses by race was also analysed (Table 11). TB is typically much more prevalent in coloured and African people than whites, and is a conspicuous problem among the homeless. The homeless were two to five times more likely to have been diagnosed with TB than people of the same age group in the general population (Figure 2) and two to three times more likely than those of the same race (Table 11).

Figure 2: Prevalence of tuberculosis in homeless males compared with the general population

Table 10: Self-reported chronic illness by age group (percentage of persons reporting they had been told by a doctor or nurse that they had the condition)

Age group (years)	High blood pressure		Heart attack/angina		Stroke		High blood cholesterol		Diabetes or blood sugar		Emphysema or bronchitis		Asthma		Tuberculosis		Cancer
	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS
Male
15–24	2.1	1.7	1.4	1.2		0.1	0.7	0.2	1.4	1.3	2.1	0.8	0.7	1.9	2.8	1.8	0.7	0.5
25–34	2.3	2.9	2.3	2.0	1.5	0.7	2.3	0.6	1.1	0.4	1.9	1.1	3.1	2.7	5.7	3.4	1.9	0.1
35–44	8.2	10.6	2.1	2.4	0.5	0.5	2.1	3.5	3.1	1.8	2.1	1.7	3.1	2.4	8.8	3.0	1.0	0.2
45–54	10.9	18.2	3.9	4.4	2.3	1.0	1.6	6.3	3.9	5.9	3.9	4.3	3.1	4.1	8.5	6.1		1.1
55–64	20.6	17.9	1.6	5.1		4.8	1.6	1.7	1.6	7.0	6.3	5.2	7.9	5.5	11.1	3.9		1.0
65+	14.3	26.2	4.8	7.0		1.5		2.9	9.5	5.2		2.3		5.4	9.5	6.7		1.8
Female
15–24	3.0	3.1		0.9	6.1	0.0		0.8		0.4		1.0	3.0	2.7		1.9		0.4
25–34	7.9	8.8	2.6	2.5		0.1	5.3	0.9	2.6	1.1	2.6	1.6	7.9	3.5	5.3	1.2		0.9
35–44	23.5	19.2	8.8	2.9		1.0	8.8	2.3	2.9	2.9	5.9	2.2	2.9	5.1	14.7	2.6	5.9	0.6
45–54	46.7	33.5	13.3	7.8	6.7	1.6	13.3	3.7	13.3	6.7		4.1		5.8		3.2		1.1
55–64	50.0	40.1	16.7	8.2	16.7	3.6	16.7	4.8		12.5	16.7	4.8	16.7	6.7		1.9		2.0
65+		41.4		7.6		2.3		3.3		8.6		5.4		5.5		3.5		0.9
Comparison of the homeless population (Human Sciences Research Council Survey) with a nationally representative sample (South Africa Demographic and Health Survey 2003; Department of Health et al., 2004). H = HSRC homelessness survey; DHS = South Africa Demographic and Health Survey 2003. Boxed and bolded cells indicate excess morbidity in the homeless population relative to the general population (South Africa Demographic and Health Survey 2003). Excess morbidity in females aged 55–64 should be interpreted with caution owing to the small sample size (n = 6).

Table 11: Self-reported chronic illness by population group (percentage of persons reporting that they had been told by a doctor or nurse that they had the condition)

Population group	High blood pressure		Heart attack/angina		Stroke		High blood cholesterol		Diabetes or blood sugar		Emphysema or bronchitis		Asthma		Tuberculosis		Cancer
	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS	H	DHS
Male
African	6.7	6.9	2.4	2.8	1.0	0.9	1.7	1.0	2.2	1.8	2.6	1.7	3.1	2.5	6.9	3.6	1.0	0.3
Coloured	5.9	15.3	0.0	2.1	0.0	1.5	5.9	3.8	11.8	3.1	5.9	3.2	0.0	3.9	11.8	4.4	0.0	0.8
White	15.4	23.3	0.0	0.9	0.0	0.3	0.0	9.8	0.0	8.6	0.0	3.7	0.0	8.0	0.0	1.2	0.0	4.6
Female
African	16.2	18.9	6.0	4.1	3.4	0.9	6.8	1.5	3.4	3.6	3.4	1.9	4.3	3.8	5.1	2.1	1.7	0.7
Coloured	33.3	24.0	0.0	2.8	0.0	1.0	0.0	2.9	0.0	5.5	0.0	5.1	11.1	8.5	11.1	5.1	0.0	1.4
Comparison of the homeless population (Human Sciences Research Council Survey) with a nationally representative sample (South Africa Demographic and Health Survey 2003; Department of Health et al., 2004). H = HSRC homelessness survey; DHS = South Africa Demographic and Health Survey 2003. Boxed and bolded cells indicate excess morbidity in the homeless population relative to the general population (South Africa Demographic and Health Survey 2003).

4.12 Relationship between illness and homelessness

While most respondents said illness was not the main reason for their leaving home, many said it had been a contributory factor – either their own or family members’ illness. A 22-year-old woman said:

My grandmother had a stroke and it was my responsibility to look after her… I left home.

Another woman said:

My mother had liver cancer and I could not handle it.

The following were some teenage girls’ comments about the association between disability or disease and leaving home:

My friend moved out of their house because of her child – they did not want her as she was disabled. Other children were laughing at her.

My friend ran away from home because she was HIV positive. Nobody liked her. She ended up living here on the street with us.

A close friend of ours died of AIDS last year. She used to stay here with us. She used to cough a lot. She went to hospital where she died.

Sophie [not her real name] was also HIV positive. She was very sick; coughing a lot. She used to stay with us at the homeless shelter.

4.13 Mortality

Both homeless young people and key informants confirmed that premature deaths of the homeless, including youth, are a common occurrence, often caused by violent injuries. By the time of the study, in February, three deaths of homeless children had already been reported to shelters for that year. Another death took place early in the week of the research exercise: a child was fatally stabbed during a street fight. An outreach worker said he had handled eight deaths in the previous year, of which at least two were thought to be HIV/AIDS related.

Death registrations are notoriously inaccurate in South Africa (Botha & Bradshaw, 1985) and usually only record place of death, without any reliable indication of residential address. Registrations showing ‘address unknown’ are frequent but are no indication of whether the deceased was homeless. Similarly, mortuary and cemetery records do not distinguish between deaths of persons with no known address and deaths of people with no permanent address (i.e. the homeless). The register of ‘pauper burials’ is also an unreliable source of data on deaths of the homeless, since poor families may fail to claim a body for burial, if they know that this may result in expense that they simply cannot afford. Thus, there is no way of knowing the proportion of pauper burials that are of homeless people. These various factors conspire to make mortality estimates specifically for homeless people virtually impossible to establish from secondary data.

5. Conclusions and recommendations

The results indicate that health factors contribute to people becoming homeless and that some illnesses associated with poor living conditions appear to be more common among the homeless in certain age and sex categories. While access to healthcare appears to be quite good, there was evidence that preventive efforts should be increased and special attention appears to be necessary for diagnosing and managing chronic illnesses.

A case can be made for more systematic health promotion among the homeless, who are habitually exposed to risky behaviour, including alcohol and drug abuse, unsafe sexual practices leading to STIs, injuries and violence. An integrated approach is necessary to address the many facets of homelessness and health, and it should recognise the potential contributions of all those providing services for the homeless, including departments of health, social welfare, housing and labour, and the many non-governmental and faith-based organisations.

There is a need for better awareness of the health needs of the homeless and training for health professionals, so as to encourage more sympathetic approaches. More low-cost rehabilitation facilities for alcohol and other substance abuse appear to be needed and measures to reduce the risk of sexually transmitted infections should be encouraged.

The reality, however, is that what homeless people want most is to have their basic needs taken care of, to give them the hope of a better life. A lasting impression when talking to homeless people, especially the older ones, is that they are vulnerable people living in fear of what the world has in store for them.

Acknowledgements

This work forms part of the Human Sciences Research Council's 2005–2008 study of homelessness. Funding from the National Department of Social Development, the Human Sciences Research Council, the Gauteng Department of Social Development, and the Swiss Agency for Development and Cooperation is gratefully acknowledged.

The authors wish, in particular, to thank the homeless shelter managers and staff for their advice and cooperation, the homeless people for sharing their experiences with us, Annemarie Booyens for formatting the questionnaire, our fieldworkers for working hard under often difficult conditions, and Cathy Ward and Margie Schneider for providing invaluable comments on an earlier draft of this paper.

Notes