HIV prevalence in children and youth living on the street and subject to commercial sexual exploitation : a systematic review

The aim of this review was to describe HIV prevalence in children and youth living on the street and subject to commercial sexual exploitation, and the studies’ characteristics in terms of place, time, population, and sample design. This was a systematic review, not a meta-analysis, based on an article search in 10 electronic databases: Science Direct, MEDLINE, OVID, LILACS, Wiley InterScience, MD Consult, Springer Link, Embase, Web of Science, and Ebsco. A complementary search was also performed in the libraries of schools of public health and webpages of U.N. agencies, besides the reference lists from the selected articles. We selected observational studies focused on children and youth living on the street and subject to commercial sexual exploitation, ranging in age from 10 to 20 years, with the results for HIV prevalence rates. A total of 9,829 references were retrieved, of which 15 met the inclusion criteria and comprise this descriptive summary. Of these 15 articles, 12 were conducted in children and youth living on the street and three in children subject to commercial sexual exploitation. All 15 were cross-sectional studies. HIV prevalence in children and youth living on the street ranged from 0% in Dallas, USA and Cochabamba, Bolivia to 37.4% in St. Petersburg, Russia. In children and youth living subject to commercial sexual exploitation, prevalence ranged from 2% in Toronto, Canada to 20% in Kolkata, India. In conclusion, HIV infection is present in children and youth living on the street and subject to commercial sexual exploitation. Measures are needed for prevention, diagnosis, and treatment as a public health priority and an ethical responsibility on the part of governments and society. Homeless Youth; Sexual Child Abuse; HIV Correspondence


Introduction
The HIV epidemic continues to be a global public health problem, and the prevention of new infections is a challenge for the world's health systems 1 .According to the Joint United Nations Programme on HIV/AIDS (UNAIDS), 36.9 million persons in the world were living with HIV in the 2014, including 2.3 million cases of new infection, while 1.2 million persons died in the world from AIDS-related causes 2 .Statistics for 2012 showed a prevalence of 0.4% in Latin America, with Guyana and Belize showing the highest prevalence rates (1.3% and 1.4%, respectively).Prevalence in North America was 0.5%, while the world's highest prevalence was in Sub-Saharan Africa (4.7%), reaching 26.5% in Swaziland 3 .
Since the beginning of the HIV epidemic, children and youth have been considered socially and physically vulnerable to the infection, given the risk of transmission during childbirth when the mother is infected with the virus 4 , situations of sexual abuse 5 , and limited access to and use of sexual health services, including condom use, due to economic, social, and cultural barriers 6 .According to global reports for the year 2012, 16% of children born to mothers with HIV acquired the infection during childbirth, a third of new HIV cases occurred in the population 15 to 24 years of age, and only 20% of female adolescents and 29% of male adolescents 15 to 19 years of age showed correct knowledge on HIV transmission and prevention 7 .
In addition to the factors mentioned above, other conditions of social vulnerability expose children and youth to increased risk of HIV infection.Such conditions include homelessness or living on the street and commercial sexual exploitation, such as child prostitution and child pornography 5,8 , which additionally restrict the guarantee of their human rights 9,10 .
In 2002, the International Labor Organization (ILO) estimated that 1.8 million children and youth in the world were sexually exploited for commercial purposes.Of these, 750,000 were in Latin America and the Caribbean, the region with the largest number of cases, followed by the Asian Pacific with 590,000 11 .The figure increased in 2006, with the United Nations Children's Fund (UNICEF) reporting 3 million children and youth victims of commercial sexual exploitation 12 .Estimates by UNICEF in 1996 pointed to 100,000 children and youth living on the street in Latin America 13 .In 2002 there were 12,000 children and youth living on the streets in Colombia, one of the Latin American countries with the largest presence of this phenomenon 14 .
While living on the street or subject to commercial sexual exploitation, children and youth are exposed to physical, mental, and social vulnerability 15,16,17,18 due to their backgrounds, circumstances, and contexts in which their daily life unfolds.Under such conditions they often abuse and develop addiction to drugs and alcohol due to the abundant supply in the street scene 19,20 .Survival becomes a daily struggle where delinquent conduct and forced labor are common 21 .Sexuality is mediated by situations of power and control by pimps or panderers over the children's bodies 22,23 , and condom use is uncommon and subject to negotiation 24 .Many children and youth also have a history of sexual abuse 23 , which in the street context places them at increased risk of further abuse and commercial sexual exploitation.
In this scenario, one might assume that children and youth living on the street and subject to commercial sexual exploitation would show high HIV prevalence rates.Despite some specific reports on the theme 25,26,27,28 , there is no compilation and summary of the scientific evidence to back such a statement.We thus conducted a systematic review with the aim of describing HIV prevalence in children and youth living on the street and subject to prostitution, as well as the studies' characteristics: place, time, population, and sampling design.

Methodology
This study is part of the project: Systematic Review: HIV and Other Sexually Transmitted Infections in Children and Youth Living on the Street and Subject to Commercial Sexual Exploitation.It was conceived as a baseline study on HIV in both populations, with a descriptive scope.The protocol was not registered in the International Prospective Register of Systematic Reviews database (PROSPERO), and the report is presented according to the Preferred Reporting Items for Systematic Reviews (PRISMA) 29 .In addition, the Cochrane manual for systematic literature reviews was consulted throughout the entire study process.
Despite not comparing two interventions in the same population, the PICOS strategy was used to define the study question and outline the study's scope in terms of population, outcome, and design.P = children and youth living on the street or subject to commercial sexual exploitation; O = HIV prevalence; S = observational studies.

Inclusion criteria
A search was conducted for scientific articles that met the following inclusion criteria: (i) observational studies (cross-sectional, cohort, casecontrol); (ii) with results for HIV prevalence, based on tests applied by the investigators or recorded in the clinical history of the children and youth; (iii) that included a sample of children and adolescents living on the street or subject to commercial sexual exploitation.For this study's purposes, living on the street considered two circumstances: those who were on the street, or who remained on the street for a large share of the time, but who had not broken off their family ties permanently, since they spent the night in their homes or slept in a specific hotel or shelter.The other situation involved children and youth of the street, that is, homeless, who had already broken off their family ties and spent 24 hours a day on the city streets, living and sleeping there 9 .In the case of commercial sexual exploitation, two situations were considered, the open or visible form, appearing on the city streets, and the disguised form, which included children and youth located in closed establishments (bars, canteens, massage parlors, nightclubs) 30 ; (iv) with an age range from 10 to 20 years (the minimum age was set at 10 years, because in some countries a sex act with a child under 10 years is considered sexual abuse, a crime, and in no case is it classified as commercial sexual exploitation, and 20 years was set as the maximum age, since various countries adopt 20 as the age of majority); (v) publications in Spanish, English, or Portuguese; (vi) no limitation on the country or year of publication.Exclusion criteria were: lack of important information on the study (sample size, sample design, HIV prevalence or number of positive cases); self-reported HIV prevalence; different publications by the authors using the same sample; or the study population failed to meet the proposed definitions, based on a reading of the full text.
To prevent loss of relevant information, we also consulted the selected articles' reference lists.A search was conducted in the "gray literature" in the principal webpages of the United Nations agencies working with childhood and adolescence.A search was also conducted in the internal reference catalogues in the libraries of the Mexican School of Public Health and the University of Antioquia, Colombia, since they were the current home institutions of the review's researchers.In the latter cases the search terms were street children, child prostitution, and HIV/ AIDS.We also consulted expert researchers on the themes of children and adolescents.

Article selection
Two authors jointly reviewed the articles' titles and abstracts, applying the inclusion criteria.In case of disagreement between the reviewers on a given article, a third author reviewed it.The reasons for excluding articles were recorded.

Data extraction
By duplicate and independently, the three authors read the full text, extracted the data, and conducted a summary, and entered the information on a Microsoft Access capture mask (Microsoft Corp., USA).The principal target outcome was HIV prevalence, including confidence intervals.In case of multicenter studies, the findings were extracted for each center.When the confidence intervals for HIV prevalence rates were not reported, they were calculated assuming a simple random sample, using Stata version 12 (StataCorp LP, College Station, USA).Data were also collected on factors associated with HIV prevalence.
Secondary data were also considered, including: (i) article's descriptive characteristics (Vancouver reference, key words, continent, city, country, and year of study, type of study population: (a) children and youth living on the street and (b) children and youth subject to commercial sexual exploitation, sample size, sex, and age.For age, when a selected study included individuals older than 20 years, whenever possible the sample size was adjusted exclusively to the population 10 to 20 years of age and a new HIV prevalence rate and confidence intervals were calculated, otherwise the study was excluded.(ii) methods (type of study, way of accessing the population, sample design, defined as probabilistic (stratified random sample, time-location, respondent-driven sampling, venue-based); nonprobabilistic (snowball, convenience sample); HIV detection test; and ethical aspects.Given this systematic review's descriptive characteristics, the risk of bias in each article was not assessed.

Data analysis
The studies' results were analyzed independently for each population: (i) children and youth living on the street; (ii) children and youth subject to commercial sexual exploitation, presented in tables, with a descriptive figure for HIV prevalence rates in children and youth living on the street using Microsoft Excel (Microsoft Corp., USA).For multicenter studies, the disaggregated results for each center are presented.

Selected articles
A total of 171 relevant references were selected and screened.Forty-eight articles were considered adequate for inclusion.Thirty-three were excluded for the following reasons: duplicates; young adult population (older than 20 years); no data on HIV prevalence; or the children and youth were not living on the street or subject to commercial sexual exploitation.Finally, 15 studies comprised this systematic review (Figure 1).Of these 15 studies, 12 focused on children and youth living on the street and three on children and youth subject to commercial sexual exploitation.The highest percentage was conducted in the Americas, four used probabilistic samples, and the highest percentage of studies were published in the 2000s (Table 1).

Studies on HIV prevalence in children and youth living on the street
Table 2 shows the characteristics of studies on HIV prevalence in children and youth living on the street.Two were multicenter studies in the United States, in five centers for youth living on the street from four cities, and in the Ukraine in three cities.The review thus includes records of 18 HIV prevalence rates in these populations.
Of the 12 studies that were included, six were conducted in North America, with the United States (four) and New York (two) as the country and city with the most studies on the topic.Seven were conducted in the 1990s.The 12 studies included a total population of 8,531 children and youth living on the street, with a median of 383 children and youth and interquartile range (IQR) of 262-569.With regard to sampling design, all the studies with probabilistic samples found positive HIV cases.The study with the highest HIV prevalence used a venue-based sample (Table 2).
HIV prevalence in children and youth living on the street ranges from 0% in cities like Dallas (USA), Cochabamba (Bolivia), and Teheran (Iran) to 37.4% in St. Petersburg (Russia).Prevalence exceeded 10% in two cities in the Ukraine (Odessa and Kiev).Two studies were found in South America, but only in Belo Horizonte (Brazil) was reported positive HIV cases, with 1%.In Asia, specifically in Kolkata (India), prevalence was 1.1%.Figure 2 shows the HIV prevalence rates and confidence intervals in children and youth living on the street according to the continent and city where the study was held.

Studies on children and youth subject to commercial sexual exploitation
Of the three studies, two were done in the Americas, in Bogotá (Colombia) and Toronto (Canada), from 2004 to 2006.They were done exclusively in females, with 336 girls and female adolescents subject to commercial sexual exploitation (median 51, IQR 30-255).With non-probabilistic samples, they found HIV prevalence ranging from 2.0% in Toronto, to 20% in Kolkata (Table 3).

Discussion
The results of this systematic review come from 15 studies, the majority of which in children and youth living on the street, with only three in children and youth subject to commercial sexual exploitation.Few studies in either group used probabilistic samples, and the largest share of studies on the HIV epidemic in these groups was in the 2000s.It is evident that the prevalence of HIV in these populations varies by region of the world: between 0% and 5% on the continent of America, 2% and 37% in Europe and between 0% and 20% in Asia.
With regard to the characteristics of time, place, type of population, and methodological aspects, the current review identified few studies in Latin America and the Caribbean, considering that it is the region of the Americas with the most children and youth living on the street and victims of prostitution 11,14,15 .This situation reflects the need for future studies in Latin America and the Caribbean to access these populations, drawing on the existing methodologies and experience elsewhere in the world.
Only four studies used probabilistic samples.This shows the persistent challenge of drawing on or generating a sampling framework for hard-toreach subgroups, partly because their status are considered illegal, stigmatized and criminalized.Likewise, the generation of sampling frameworks is costly and may prove largely unfeasible 31,32,33 .The use of samples for hidden populations, such as venue-based in St. Petersburg 26 , and time-location in India and the Ukraine 27,34 , may partially explain the high prevalence rates found there.The three studies reached sample sizes greater than 200 children and youth and prevalence rates that reached 10%.This situation reflects the samples' capacity to reach the more hidden subgroups and perhaps those at greatest vulnerability, a necessary characteristic to avoid biased estimates 32 .
In addition to the sample design, other possible explanations for the high HIV prevalence rates found in Russia and Ukraine are discussed next.In 2002, Russia estimated 3 million homeless children as the result of economic and social crises in the 1990s 35 , and by 2006 it was the country of Eastern Europe with the largest HIV epidemic concentrated in the population between 15 and 30 years of age 36 .Meanwhile, in Ukraine there were an estimated 40,000 homeless children and youth in 2006 37 , and the country had the highest HIV prevalence in Eastern Europe by 2010 38 , with young people the most heavily affected due to injection drug use and needle sharing 39 .
In the studies on HIV prevalence and child prostitution, the highest rate was in India, a country with a combined phenomenon of person trafficking and sexual exploitation of women of all ages.In 2008 there were a reported 150,000 victims of this phenomenon worldwide 40 .The three studies on commercial sexual exploitation focused exclusively on females.This indicates that the phenomenon is a form of gender-based violence, since females are the principal victims.A study in five cities of Colombia reported that of all children and youth subject to commercial sexual exploitation and assisted by government services, 91% were females 41 .
In addition to the above, the magnitude of the HIV epidemic in these populations should be analyzed in the context of the conditions of social exclusion and violence faced by children and youth 9,10 , including history of sexual abuse 23 .Condom use is mediated by drug consumption and situations of power exerted over children and youth by pimps and panderers 42,43 ; during injection use, needles are shared 26,44,45,46 .Other factors include stigma, discrimination, and invariable denial of the provision of social and health services 15 .
To respond to this HIV/AIDS problem in both populations means developing actions to prevent infection 47 .This involves mobilizing resources for monitoring, surveillance, and treatment through social and health models with comprehensive care, extending beyond individual actions 48,49,50 , where social determinants intervene in combination with the persistent inequities in these population groups 47,51 .But it also involves prosecution of sexual abusers and exploiters of children and youth 52 , drug policies with an emphasis on public health rather than criminalization, such as the creation of needle exchange programs rather than criminalization of consumers 53 .Nevertheless, the social and government response to these issues should keep international ethical principles in mind, including the child's greater interest, prevalence of their rights, and comprehensive protection 9,15,54,55,56 .
This systematic review shows some gaps between studies on HIV in children and youth living on the street when compared to the limited number of studies on studies in children and youth subject to commercial sexual exploitation.This shows how this criminal activity is clandestine and invisible, and how it violates human rights 57 .In addition, the lack of longitudinal studies in these population groups limits any causal inferences, related to injection drug use during life on the street or sexual exploitation, and the presence of HIV.Another element to consider is the scarce use of probabilistic samples and sampling techniques for hard-to-reach populations such as those included in this review.This may lead to underestimation of the epidemic, due to inability of the no probabilistic samples to capture important foci of infection, which could lead to greater spread of HIV due to the interventions' limited scope of action 32 .
The search for gray literature in only two Public Health School libraries may be a limitation to this systematic review.Universities currently have digital repositories where they store the theses and conferences available online, which could leave out studies conducted in various countries and which might contribute greater evidence for this study's purposes.Calculation of confidence intervals for the studies that did not report them restricts the data's interpretation.According to Salganik 58 , despite progress with prevalence estimates, less is known about the sample's variability for estimations, which could lead to the calculation of confidence intervals around imprecise estimates.In addition, the lack of an overall estimate of prevalence could be considered another important limitation.Nevertheless, the methodological decisions back this limiting factor: in the first place, the review's descriptive objective, which only intended to provide an overview of the characteristics of studies on HIV in children and youth living on the street and subject to commercial sexual exploitation, and the intent of not conducting a meta-analysis.The latter decision was based on the studies' heterogeneity in terms of the place where the populations were recruited, the sample designs, and different inclusion criteria.These differences could have led to a biased estimate of over- all prevalence.However, this aspect can be the object of a future systematic review and metaanalysis on HIV and other sexually transmitted infections in these population groups.The above limitations notwithstanding, this was the first systematic review of HIV prevalence in these population groups.There are few systematic reviews on groups of children and adolescents in conditions of social vulnerability.Some of the existing reviews have focused on living conditions and the effects of life on the street 21 .There is a meta-analysis on HIV risk in adolescents and adults with a history of orphanhood 59 and another on HIV prevalence in homeless adults 47 .
This study thus responded to an internationally relevant question for public health and can become a technical and policy tool for preven-tion, control, and care in the HIV epidemic, helping guarantee the right to health for children and youth living on the street and subject to sexual exploitation.Such research allows identifying methodological approaches for these populations, which are highly mobile and referred to as hidden and hard to reach 60,61 .
Based on the review's findings, we conclude: HIV is a problem in the lives of children and youth living on the street and subject to commercial sexual exploitation, requiring intervention as a global health priority with comprehensive measures to avoid spread of the epidemic and to guarantee the right to health and dignity for these population groups, with outreach strategies that include appropriate techniques for approaching hidden and hard-to-reach populations.

Contributors
C. Noreña-Herrera participated in developing the study concept and in all stages in the production of this article.He was responsible for the final version submitted for publication.C. A. Rojas participated in developing the study concept, writing of the article, critical revision of the intellectual content, and approval of the final version.L. Cruz-Jiménez participated in the data analysis and interpretation, writing of the article, and approval of the final version for publication.

Figure 2 HIV
Figure2HIV prevalence and 95% confidence interval (95%CI) in children and youth living on the street.

Table 1
Descriptive characteristics of studies on HIV prevalence in children and adolescents living on the street and subject to commercial sexual exploitation, according to type of population.

Table 2
Characteristics of HIV prevalence studies in children and adolescents living on the street.
* The study included persons > 20 years but the sample number and reported HIV prevalence refer to the population 10-20 years; ** Confidence intervals not presented in the study and calculated assuming simple random sample.