JECH Online First, published on March 9, 2016 as 10.

1136/jech-2015-205860
Research report

Physical, emotional and sexual adolescent abuse

victimisation in South Africa: prevalence, incidence,
perpetrators and locations
Franziska Meinck,1 Lucie D Cluver,1,2,3 Mark E Boyes,1,4 Heidi Loening-Voysey5

▸ Additional material is ABSTRACT victimisation.10 Furthermore, adolescence is a

published online only. To view Background Physical, emotional and sexual abuse of time where the intergenerational transmission of
please visit the journal online
(http://dx.doi.org/10.1136/jech- children is a major problem in South Africa, with severe poverty, violence victimisation and perpetration,
2015-205860). negative outcomes for survivors. To date, no known gender inequalities and educational disadvantage
studies have used data directly obtained from manifest themselves.9
For numbered affiliations see
end of article. community-based samples of children to investigate Evidence of the epidemiology of adolescent
prevalence, incidence, locations and perpetrators of child abuse in South Africa is limited and inconsistent.
Correspondence to abuse victimisation. This study aims to investigate Studies differ in sampling, and definitions or mea-
Dr Franziska Meinck, Centre prevalence and incidence, perpetrators, and locations of sures for severity of abuse.11 Consequently,
for Evidence-Based
Intervention, Department of child abuse victimisation in South Africa using a reported prevalence rates vary greatly, 6.7–32% for
Social Policy & Intervention, multicommunity sample. physical abuse,12 13 11.9–35.5% for emotional
Oxford University, Oxford OX1 Methods 3515 children aged 10–17 years (56.6% abuse12 14 and 1.6–60% for sexual abuse.15 16 No
2ER, UK; female) were interviewed from all households in published data are available on multiple abuse vic-
Franziska.Meinck@spi.ox.ac.uk
randomly selected census enumeration areas in two timisation. Furthermore, existing evidence focuses
Received 2 April 2015 South African provinces. Child self-report questionnaires on particular subpopulations of children.
Revised 4 November 2015 were completed at baseline and at 1-year follow-up Specifically, past studies sampled selectively by
Accepted 22 February 2016 (96.7% retention). including only one gender or by sampling exclu-
Results Prevalence was 56.3% for lifetime physical sively university or high school students.17 These
abuse (18.2% past-year incidence), 35.5% for lifetime studies may be unrepresentative, given the low par-
emotional abuse (12.1% incidence) and 9% for lifetime ticipation in higher grades of secondary school and
sexual abuse (5.3% incidence). 68.9% of children tertiary education.18 Moreover, children who have
reported any type of lifetime victimisation and 27.1% dropped out of school may be at particularly high
reported lifetime multiple abuse victimisation. Main risk of abuse.12 Other studies assessed only high-
perpetrators of abuse were reported: for physical abuse, risk children, such as orphans and vulnerable chil-
primary caregivers and teachers; for emotional abuse, dren (OVC)12 13 and small-scale clinical case
primary caregivers and relatives; and for sexual abuse, studies that examine clinical reviews of social ser-
girlfriend/boyfriends or other peers. vices cases or children referred for psychological
Conclusions This is the first study assessing current treatment.19 While these studies provide valuable
self-reported child abuse through a large, community- evidence on subpopulations, wider community
based sample in South Africa. Findings of high rates of studies can help to understand broader prevalence
physical, emotional and sexual abuse demonstrate the rates and patterns. There is valuable research on
need for targeted and effective interventions to prevent outcomes of abuse and risk factors, using retro-
incidence and re-victimisation. spective studies from adult samples,15 but these
may need to be interpreted with caution in terms
of determining prevalence rates due to risks of
INTRODUCTION recall bias, particularly with regard to childhood
Evidence shows that victims of child abuse in memories of abuse.20 Finally, all known studies are
sub-Saharan Africa have consistently poorer phys- either cross-sectional or retrospective. Therefore,
ical and mental health outcomes than other chil- they cannot measure incidence of abuse.
dren.1 They are at increased risk for HIV To the best of the authors’ knowledge, this study
infection,2 exposure to transactional sex,3 risk for provides the first descriptive examination of preva-
re-victimisation,4 bullying5 and higher levels of lence, frequency, incidence, location, perpetrators
depression and suicidal ideation.6 Little is known, and severity of child abuse in a longitudinal,
however, about prevalence rates, perpetrators and community-based sample in South Africa. The
locations of child abuse victimisation in South study will focus on physical, emotional and sexual
Africa, particularly among adolescents. abuse victimisation. Community violence and other
Adolescence is an important decade in a child’s traumatic experiences outside of the home were
development, marking the period of transition excluded due to the study’s use of the WHO defin-
To cite: Meinck F, from childhood to adulthood.7 Adolescents are a ition of child abuse ‘in the context of a relationship
Cluver LD, Boyes ME, et al.
J Epidemiol Community
particularly vulnerable group, experiencing a of responsibility, trust or power’.21 Childhood
Health Published Online third of all new HIV infections worldwide,8 high neglect was not included due to high poverty
First: [ please include Day levels of violence, lower school attendance and within the sample, which made distinction between
Month Year] doi:10.1136/ enrolment than primary schoolchildren, early neglectful behaviour and poverty-related inability
jech-2015-205860 marriage and higher levels9 of sexual abuse to provide difficult.22
Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860 1
Copyright Article author (or their employer) 2016. Produced by BMJ Publishing Group Ltd under licence.
 Research report

Objectives additional items were designed and tested for follow-up data
This study aims to assess physical, emotional, sexual and mul- collection with the help of local social workers, NGO staff
tiple child abuse victimisation within a large, random, working with OVC, and adults and children from the local com-
community-based sample of adolescents in terms of: (1) preva- munity (all items are listed in online supplementary material 1;
lence and frequency; (2) incidence; (3) differences by gender, for frequencies of responses to individual items, please see
age and rural/urban location; and (4) perpetrators and locations Meinck et al24). All follow-up items also measured the relation-
of victimisation. ship of the perpetrator to the child and the location of the
abuse. The overall reliability for this 14-item scale was α=0.74.
METHODS Child sexual abuse victimisation at baseline was measured
Participants using two items designed by social workers in South Africa and
Recruitment took place in two urban and rural health districts one item from the National Survey of HIV and Risk Behaviour
in two provinces: Mpumalanga and the Western Cape. In each Amongst Young South Africans.25 All baseline sexual abuse
district, census enumeration areas were randomly selected. items measured lifetime exposure with a no/yes response code.
Within these areas, every resident child aged 10–17 years Sexual abuse victimisation at follow-up was measured in more
(n=3514) was included through door-to-door sampling of all detail using five items from the Juvenile Victimization
households. In households with several children, one participant Questionnaire.26 Items were modified to fit the cultural context
was selected randomly. Baseline data were collected between with the help of experienced social workers and were then pre-
January 2010 and June 2011. A year later, 3401 (96.7%) parti- piloted with children in South Africa (all items are listed in
cipants were traced and reinterviewed. online supplementary material 2). Contact sexual abuse was
defined as any unwanted touching or kissing, touching of
private parts and/or forced sex. Exposure to sexual harassment
Procedures and forced watching of pornographic material were also
Child participants completed confidential questionnaires which measured.
were translated into Xhosa, Swati, Tsonga, Sotho and Zulu and For all abuse items administered at follow-up and for physical
checked with back translation. Interviewers assisted participants and emotional abuse at baseline, participants were asked to state
in filling in the questionnaires, which took 60 min to complete. frequency of abuse in the past year (never, happened but not
Children were interviewed in the language and locations of past year, at least once, monthly and weekly). All follow-up
their choice, such as spare classrooms in schools or under a items also measured the relationship of the perpetrator to the
secluded tree, to ensure confidentiality. Interviewers received child and the location of the abuse.
intensive training in working with vulnerable children and in To estimate the incidence of the three types of abuse, individ-
administering standardised questionnaires. Vulnerable youth pre- ual scales for physical, emotional and sexual abuse were coded
piloted the survey to assess age-appropriateness and into dichotomous variables. If the child did not report victimisa-
cultural-appropriateness. Data were checked for quality and tion at baseline or at follow-up, or if the child mentioned vic-
missing data were <0.5%. timisation at both baseline and follow-up, the response was
Ethical approval was granted by the Universities of Oxford, coded as 0. If victimisation was mentioned at follow-up but not
Cape Town and KwaZulu-Natal; the National Department of at baseline, this was coded as 1. To allow comparability across
Social Development; and the Western Cape and Mpumalanga the two measurement points, exactly the same items were used
provincial Departments of Health and Education. Informed at baseline and follow-up to determine incidence for physical
consent was sought from both children and their caregivers. and emotional abuse. Similarly, dichotomous variables were
Within sub-Saharan Africa, many children are not looked after created for prevalence at baseline and follow-up in each abuse
by their biological parents. The term ‘caregivers’ is therefore category (0: not abused; 1: abused). Prevalence was further
used and refers to the person who provides primary parenting divided to reflect lifetime abuse, past year and monthly or more
responsibilities. Caregivers can be biological parents, aunts and frequent victimisation.
uncles, siblings, other relatives or foster carers. Information and Perpetrators and locations of abuse victimisation were identi-
consent sheets were read out loud and questions answered in fied by participants for each abusive act they had experienced.
the light of low literacy in the sampled population group. The options were: caregiver, teacher, relative, neighbour or
Participation was voluntary, and all participants received a cer- ‘other’. For ‘other’, further clarification on the perpetrator’s
tificate of completion and light refreshments. relationship with the child was requested. Participants also had
Children at risk of significant harm, as well as those with past the option of indicating the location in which they had been vic-
experiences of abuse or who requested help, were referred to timised. The options were: home, community, school, home of
local child protection services, counselling centres and neighbour and ‘other’. For ‘other’, further clarification of the
HIV-testing services with follow-up support from the inter- location was requested.
viewers. These options were always discussed with the child, Multiple abuse victimisation was defined as two or more cat-
otherwise strict confidentiality was maintained. In total, 664 egories of abuse, and measured as two or more concurrent
referrals were made. abuse types (of sexual, physical and emotional abuse).
Sociodemographic information on gender, age, province and
Measures location (urban or rural) was measured using items modelled on
Child physical and emotional abuse victimisation were measured the South African census.
(at both baseline and follow-up assessments) using five items
from the UNICEF Measures for National-level Monitoring of
OVC.23 Participants were asked to state frequency of abuse in Analyses
the past year (never, happened but not past year, at least once, Descriptive analyses were conducted using SPSS V.22. Estimates
monthly and weekly). The scale had been used in South Africa of incidence and prevalence of physical, emotional, sexual abuse
previously and showed good reliability of α=70.12 Seven and multiple abuse victimisation were assessed. Perpetrators and
2 Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860
 Research report

locations of perpetration were also examined. Prevalence rates lifetime rape, 0.8% reported past-year rape, and 0.3% reported
were tested for demographic differences using Pearson’s χ2-tests. frequent monthly rape victimisation. Past-year incidence of
contact sexual abuse was 5.3%; past-year rape incidence was
RESULTS 2.1%. Since sexual harassment and exposure to pornography
The sample included 3515 children at baseline (56.7% female, were not measured at baseline, incidence of either was not
mean age 13.45 years) and 3401 (96.7% retention rate) at calculated.
follow-up (54.5% female, 14.67 years). Refusal rates were 2.8% Multiple victimisation: Up to 27.1% reported being victims of
at baseline and <0.5% at follow-up. Approximately half of the two or more types of abuse victimisation in their lifetime, with
participants lived in urban areas (50.6% at baseline, 48.6% at physical and emotional abuse most commonly co-occurring. Up
follow-up) and in Mpumalanga Province (47.3%). to 19.6% reported frequent multiple victimisation.

Prevalence and incidence Children at higher risk for abuse

In total 68.9% of children reported at least one type of victimisa- Demographic differences in abuse victimisation were also
tion in their lifetime and 54.9% reported some type of victimisa- observed (table 2). Younger children were more likely to experi-
tion in the past year. In total 32.3% of children reported at least ence physical abuse ( p<0.001), while older children were more
one type of frequent monthly abuse victimisation (table 1). likely to report emotional ( p<0.05) and sexual abuse
Physical abuse: At follow-up, 56.3% of children reported life- (p<0.001). Girls were more likely than boys to report emo-
time physical abuse, 37.9% reported past-year physical abuse tional abuse ( p<0.005), sexual harassment ( p<0.001), contact
and 16.6% reported frequent monthly physical abuse victimisa- sexual abuse ( p<0.001) and rape ( p<0.001). In addition, chil-
tion. Past-year incidence of physical abuse was 18.2%. dren in rural areas were more likely to report physical
Emotional abuse: In total 35.5% of children reported lifetime (p<0.005), emotional (p<0.05) and contact sexual ( p<0.05)
emotional abuse, 31.6% reported past-year emotional abuse and abuse victimisation than those in urban areas. However, the dif-
20.7% reported frequent monthly emotional abuse victimisa- ferences were not as distinct as for other demographic factors.
tion. Past-year incidence of emotional abuse was 12.1%. There were no differences between genders for physical abuse
Sexual abuse: At follow-up, 14.8% of children reported life- and exposure to pornography. Likewise, no differences between
time sexual harassment and 12.8% reported sexual harassment urban and rural sites could be observed for sexual harassment,
in the past year. Up to 2.4% reported lifetime forced exposure exposure to pornography and rape victimisation. Perpetrators
to pornography, and 2% reported forced exposure to pornog- and locations of abuse.
raphy in the past year. Nine per cent of children reported life- Perpetrators of physical abuse victimisation were most com-
time contact sexual abuse, 5.9% reported past-year contact monly primary caregivers, followed by teachers and relatives.
sexual abuse exposure, and 2.8% reported frequent monthly Perpetrators of emotional abuse were most commonly primary
sexual abuse victimisation. Up to 3.3% of children reported caregivers, followed by relatives and teachers. Perpetrators of

Table 1 Prevalence rates of physical, emotional and sexual child abuse victimisation
Never % (n) Lifetime % (n) In the past year %(n) Monthly % (n) Incidence % (n)

Physical abuse
Baseline 59.9 (2036) – 40.1 (1365) 18.2 (619) –
Follow-up 43.7 (1485) 56.3 (1916) 37.9 (1289) 16.6 (564) 18.2 (641)
Emotional abuse
Baseline 65.8 (2237) – 34.2 (1160) 19.2 (654) –
Follow-up 64.5 (2195) 35.5 (1206) 31.6 (1076) 20.7 (704) 12.1 (410)
Sexual harassment
Baseline – – – – –
Follow-up 85.7 (2899) 14.8 (502) 12.8 (437) 8.1 (276) –
Forced exposure to pornography
Baseline – – – – –
Follow-up 97.6 (3319) 2.4 (82) 2 (69) 0.8 (26) –
Contact sexual abuse
Baseline 96.3 (3276) 3.7 (125) – – –
Follow-up 89.8 (3054) 9 (306) 5.9 (201) 2.8 (94) 5.3 (181)
Rape
Baseline 98.8 (3360) 1.2 (41) – – –
Follow-up 96.7 (3290) 3.3 (111) 0.8 (28) 0.3 (10) 2.1 (70)
Any type of victimisation experienced
Baseline 46.1 (1567) 53.9 (1834) – – –
Follow-up 31.1 (1057) 68.9 (2344) 54.9 (1868) 32.3 (1100) –
Multiple abuse victimisation
Physical and emotional 72.9 (2479) 27.1 (922) 19.6 (667) 9.2 (312) –
Physical and sexual 93.9 (3192) 6.1 (209) 3.6 (122) 0.8 (28) –
Emotional and sexual 94.4 (3211) 5.6 (190) 3.6 (123) 1.6 (56) –
Physical, emotional and sexual 95.4 (3246) 4.6 (155) 3.2 (109) 0.6 (22) –

Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860 3

 Research report

Table 2 Differences in abuse victimisation by gender, location and age

Male (n=1475) Female (n=1926) Urban (n=1720) Rural (n=1681) Age <15 (n=2234) Age 15+ (n=1167)

Physical abuse
Past year 17% (577) 20.9% (712) 17.6% (600) 20.3% (689)** 20.3% (691) 17.6% (589)***
Lifetime 25.2% (856) 31.2% (1060) 27.8% (946) 28.5% (970) 29.1% (988) 27.3% (928)***
Emotional abuse
Past year 12.6% (428) 19.1% (648)** 15% (509) 16.7% (567)* 14% (475) 17.7% (601)*
Lifetime 14.3% (458) 21.2% (721)** 16.9% (576) 18.5% (630)* 15.8% (538) 19.6% (668)*
Sexual harassment
Past year 3.1% (106) 9.7% (331)*** 6.1% (207) 6.8% (230) 3.6% (122) 9.3% (315)***
Lifetime 3.7% (127) 11% (375)*** 7% (239) 7.7% (263) 4.3% (145) 10.5% (357)***
Pornography
Past year 0.7% (25) 1.3% (44) 0.8 (27)% 1.2% (42) 0.7% (24) 1.3% (45)
Lifetime 0.9% (29) 1.6% (53) 1.1% (36) 1.4% (46) 0.9% (29) 1.6% (52)*
Contact sexual abuse
Past year 1.8% (61) 4.1% (140)*** 2.5% (84) 3.4% (117)* 1.8% (60) 4.1% (141)***
Lifetime 3% (101) 6% (205)*** 4.1% (140) 4.9% (166) 2.6% (87) 6.4% (210)***
Rape
Past year 0.1% (3) 0.7% (25)*** 0.4% (14) 0.4% (14) 0.3% (9) 0.6% (19)
Lifetime 0.9% (29) 2.4% (82)*** 1.8% (60) 1.5% (51) 0.8% (27) 2.5% (84)***
Pearson’s χ2 test (two-tailed): p<0.001 ***, p<0.01**, p<0.05*.

sexual abuse (harassment, forcing participants to watch pornog- What sociodemographic factors put children at higher risk
raphy, unwanted sexual touching or kissing, unwanted genital of abuse?
touching) were most commonly peers and intimate partners. While younger children experience more physical abuse,
Forced sex was mainly perpetrated by strangers, relatives and older children were more likely to report emotional and
intimate partners. The most common locations for physical and
emotional abuse were the home, followed by schools and com-
munities. The locations for sexual harassment, being forced to Table 3 Perpetrators and locations of child abuse victimisation
watch pornography, unwanted sexual touching and unwanted Perpetrators Percentage Locations Percentage
genital touching were primarily the community and school.
Forced sex was mostly reported to have happened in the home Physical abuse
and at school (table 3). Caregiver 47.4 Home 58.6
Teacher 32.4 School 36.1
Relative 11.4 Community 4.7
DISCUSSION Neighbour 1.7 Home of neighbour 0.2
This is the first large-scale community-based study examining Other 7.1 Other 0.1
the incidence and prevalence of current self-reported child Emotional abuse
abuse victimisation in South Africa. It adds valuable information Caregiver 56.9 Home 76.2
to the existing literature regarding perpetrators, locations, inci- Teacher 5.9 School 12.2
dence and prevalence rates of physical, emotional and sexual Relative 19.3 Community 9.8
child abuse. In addition, no other published study in South Neighbour 6.4 Home of neighbour 1.3
Africa has investigated multiple abuse victimisation. This is Other 11.5 Other 0.5
important, since research in other countries has shown that Sexual abuse
large numbers of child abuse victims are subject to multiple Caregiver 2.2 Home 11
types of abuse27 and have especially poor physical and mental Teacher 0.9 School 32.1
health outcomes.28 Relative 7.5 Community 52.6
Neighbour 13.2 Home of neighbour 1.2
What are the prevalence and incidence rates of physical, Boyfriend/girlfriend 24.1 Other 3.1
emotional and sexual child abuse in South Africa? Peers 30.9
Overall, participants reported a high prevalence of lifetime, Stranger 14.9
past-year and frequent physical, emotional and contact sexual Other 6.3
abuse victimisation. A high incidence of physical, emotional and Rape
contact sexual abuse was also reported. Up to one-third of parti- Caregiver 1.9 Home 58.6
cipants reported lifetime multiple abuse victimisation. Girls and Teacher 0 School 36.1
older children were found to be at particular risk for sexual and Relative 23.1 Community 4.7
emotional abuse. Younger children were at higher risk for phys- Neighbour 9.6 Home of neighbour 0.4
ical abuse victimisation. Boyfriend/girlfriend 23.1 Other 0.3
Comparisons of child abuse victimisation are notoriously dif- Peers 7.9
ficult due to measurement issues.29 However, the present study Stranger 28.8
and official reports suggest very high rates of child abuse victim- Other 5.8
isation compared to those in high-income countries.30
4 Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860
 Research report

sexual abuse. This finding seems broadly in line with the lit- abuse victimisation rather than polyvictimisation. Results should
erature from other studies in Africa,11 such as the Violence be interpreted with this in mind.
Against Children studies,31 32 as well as studies from high-
income countries.33–35
Gender also matters in abuse victimisation. Girls reported Implications
more emotional and lifetime sexual abuse than boys. These find- Current child abuse prevention efforts often focus on
ings correspond to global36 and African evidence.31 32 For emo- younger children, but our results suggest that adolescents are
tional abuse, only one study in South Africa has previously also vulnerable to abuse victimisation (with 32.3% of the
examined gender differences13 (most surveys have been single adolescents in this sample experiencing at least one type of
gender), and evidence from other countries is inconclusive.33 frequent abuse victimisation). This suggests that broadening
Interestingly, substantial differences in abuse victimisation child protection efforts is necessary to ensure that this age
between children in rural and urban areas could not be found, group is reached. The findings also demonstrate that child
and previous findings were also inconclusive.11 abuse appears to be prevalent across genders, locations and
settings despite some differences between boys and girls and
between younger and older adolescents. While the current
Who are the perpetrators, and what locations do they use? child protection system in South Africa mainly focuses on
Findings show differences between the perpetrators of physical, responding to child abuse, policymakers and practitioners
emotional and sexual abuse. Caregivers and teachers committed should consider investing in child abuse prevention pro-
the majority of physical abuse. Since caregivers are usually the grammes in the light of these findings. Home visiting,
child’s main disciplinarian, spend a large amount of time with evidence-based parenting programmes and multicomponent
the child and may consider physical discipline a behavioural interventions have been shown to be effective in other parts
measure,37 it is not surprising that they are the main perpetra- of the world.45 Given the wide number of contexts in which
tors of physical and emotional abuse.38 Consistent with previous abuse occurs, these services should be integrated with educa-
evidence, physical punishment by teachers remains common in tion, family health services such as maternal health, early
schools across the country despite the abolition of corporal pun- childhood development, immunisations and adolescent health
ishment in schools.39 services as suggested by a recent Child Maltreatment
The most common perpetrators of rape were strangers, peers Readiness Assessment in South Africa.46
and relatives, a finding which corresponds to findings from
other studies in South Africa.15 17 Contact sexual abuse and
sexual harassment were carried out mostly by peers/friends or Directions for future research
relatives, findings which are consistent with other South African Research investigating risk and protective factors for child abuse
studies.39 Contrary to other studies in South Africa, few chil- victimisation is needed to inform intervention design and pro-
dren reported teachers as sexual abusers.40 gramming. Further research is required to examine social and
familial predictors of abuse in order to identify modifiable path-
ways. Since existing family interventions focus primarily on
Limitations parents with infants or young children, future research should
This study was subject to a number of limitations. First, it took consider the design and evaluation of child abuse prevention
place in low-income black African communities, and thus the interventions for families with older children and adolescents.
observed prevalence rates cannot be generalised across high- Furthermore, the role of gender (for both the child and the per-
income areas and other ethnic groups within South Africa. petrator) in sexual abuse victimisation needs to be investigated.
However, 54% of black South Africans live under the poverty Future research should also investigate risk factors for mul-
line,41 and the study benefited from in-sample variation such as tiple abuse victimisation in South Africa. Child abuse prevention
the inclusion of five African language groups and the administer- efforts should consider the importance of co-occurrence of
ing of questionnaires in rural and urban areas in two provinces. several types of abuse, since studies have shown that multiple
Second, data were collected using child self-report only. childhood abuse victims are at a higher risk of developing
Evidence suggests that children routinely under-report their trauma symptomology and are more likely to experience severe
experiences of abuse.42 Studies that include parents have shown, abuse.47
however, that parents are even more likely to under-report
abusive behaviours towards their children,43 and use of social
services cases or substantiated court reports are unreliable in CONCLUSION
contexts where services are unable to reach the majority of The incidence and prevalence rates of physical, emotional and
abuse victims. Furthermore, current child self-report is prefer- sexual abuse in South Africa are high in comparison to Western
able to the routinely used retrospective measures.44 In addition, samples and similar to rates shown in other sub-Saharan African
measures were carefully piloted, included in a removable confi- countries. Many children also report multiple abuse victimisa-
dential section of the questionnaire, and interviews were con- tion. Perpetrators of physical abuse are mainly caregivers and
ducted with particularly sympathetic interviewers. Third, no teachers; perpetrators of emotional abuse are mainly caregivers
information on the perpetrators’ gender was collected. Future and relatives; and perpetrators of sexual abuse tend to be intim-
studies should investigate perpetrator gender in order to inform ate partners and peers.
targeted interventions aimed specifically at sexual abuse perpe- The findings of this study have implications for policy and
trators. Finally, the study focused on the prevalence and inci- intervention design in South Africa. They suggest the need for
dence of physical, emotional and sexual abuse victimisation only targeted child abuse prevention interventions and service provi-
and did not measure neglect or any forms of community vio- sion. Such programmes should take into account the heterogen-
lence. Experience of more than one type of child abuse victim- eity in victims, perpetrators and locations across the different
isation as defined in this study is therefore classified as multiple types of abuse.
Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860 5
 Research report
Provenance and peer review Not commissioned; externally peer reviewed.
What is already known on this subject Data sharing statement Data can be shared if needed and in collaboration.
Please contact Dr Lucie Cluver lucie.cluver@spi.ox.ac.uk. More information on the
study is available on http://www.youngcarers.org.za.
Physical, emotional and sexual child abuse victimisation has Open Access This is an Open Access article distributed in accordance with the
been shown to influence health in later life. Recent studies have terms of the Creative Commons Attribution (CC BY 4.0) license, which permits
found that the prevalence of child abuse victimisation is high others to distribute, remix, adapt and build upon this work, for commercial use,
across sub-Saharan Africa, but existing studies focus exclusively provided the original work is properly cited. See: http://creativecommons.org/licenses/
on school or university students, or on only one gender. They by/4.0/
are cross-sectional and therefore cannot determine incidence.
REFERENCES
1 Reza A, Breiding MJ, Gulaid J, et al. Sexual violence and its health consequences
for female children in Swaziland: a cluster survey study. Lancet 2009;373:1966–72.
What this study adds 2 Jewkes R, Dunkle K, Nduna M, et al. Associations between childhood adversity and
depression, substance abuse and HIV and HSV2 incident infections in rural South
African youth. Child Abus Negl 2010;34:833–41.
This longitudinal study examined physical, emotional, sexual 3 Cluver L, Orkin M, Boyes M, et al. Transactional sex amongst AIDS-orphaned and
and multiple abuse victimisation in a community-based sample AIDS-affected adolescents predicted by abuse and extreme poverty. J Acquir
of South African adolescents. It identified incidence and Immune Defic Syndr 2011;58:336–43.
4 Jewkes R, Levin J, Penn-Kekana L. Risk factors for domestic violence: findings from
prevalence of abuse victimisation as well as perpetrators and a South African cross-sectional study. Soc Sci Med 2002;55:1603–17.
locations. The results emphasise the great importance of 5 Cluver L, Bowes L, Gardner F. Risk and protective factors for bullying victimization
targeted interventions for abuse response services and among AIDS-affected and vulnerable children in South Africa. Child Abus Negl
prevention. 2010;34:793–803. http://www.sciencedirect.com/science/article/B6V7N-51430PX-1/
2/df0da929cb2b11f2ccf09430874bc140
6 Brown D, Riley L, Butchart A, et al. Exposure to physical and sexual violence and
adverse health behaviours in African children: results from the Global School-based
Author affiliations Student Health Study. Bull World Health Organ 2009;87:447–55.
1
Department of Social Policy & Intervention, Centre for Evidence-Based Intervention, 7 Steinberg L, Morris AS. Adolescent development. Annu Rev Psychol
University of Oxford, Oxford, UK 2001;52:83–110.
2
Department of Psychiatry and Mental Health, University of Cape Town, Cape Town, 8 UNICEF. Towards an AIDS-Free Generation—children and AIDS: Sixth Stocktaking
South Africa Report, 2013. New York: Unicef, 2013.
3
Health Economics and HIV/AIDS Research Division, University of KwaZulu-Natal, 9 UNICEF. The State of the World’s Children: adolescence, an age of opportunity.
Durban, South Africa New York: Unicef, 2011.
4
Health Psychology and Behavioural Medicine Research Group, School of Psychology 10 Meinck F, Cluver LD, Boyes ME. Longitudinal Predictors of Child Sexual Abuse in a
and Speech Pathology, Curtin University, Perth, Western Australia, Australia Large Community-Based Sample of South African Youth. J Interpers Violence 2015.
5
UNICEF Office of Research Innocenti, Florence, Italy Published Online First:
11 Meinck F, Cluver LD, Boyes ME, et al. Risk and protective factors for physical and
Twitter Follow Mark Boyes at @me_boyo, the research project at @YoungcarersS sexual abuse of children and adolescents in Africa: a review and implications for
and Franziska Meinck at @AresQui practice. Trauma Violence Abus 2015;16:81–107.
Acknowledgements The authors wish to thank the children who participated 12 Meinck F, Cluver LD, Boyes ME, et al. Risk and protective factors for physical and
in the study, their families, all the fieldworkers working tirelessly to interview as emotional abuse victimisation amongst vulnerable children in South Africa. Child
many children as possible, and the Rural AIDS Development Action Research Abus Rev 2015;24:182–97.
programme (RADAR) at the University of the Witwatersrand and Cape Town 13 Thurman T, Kidman R. Child maltreatment at home: prevalence among orphans and
Child Welfare. The authors would also like to thank Jennifer Rabedeau, Dr Thees vulnerable children in KwaZulu-Natal, South Africa. New Orleans: Tulane University
Spreckelsen, Dr Lucy Bowes and Professors Lorraine Radford and Frances School of Public Health, 2011.
Gardner for comments and edits. The authors have no conflicts of interest to 14 Madu S, Idemunda S, Jegede A. Some perceived parental undesirable behaviours
disclose. predicting child abuse: a study among a sample of university students in South
Africa. J Soc Sci 2003;7:111–19.
Contributors LDC and MEB had responsibility for the overall study design and 15 Jewkes R, Levin J, Mbananga N, et al. Rape of girls in South Africa. Lancet
management. FM had responsibility for conceptualising and writing the paper. FM 2002;359:319–20.
conducted part of the fieldwork and data collection, and also led the analyses. LDC 16 Madu S, Peltzer K. Risk factors and child sexual abuse among secondary school
and MEB contributed to the analyses. FM, LDC, MEB and HL-V worked together on students in the Northern Province (South Africa). Child Abuse Negl
the interpretation of findings and writing of the paper. All authors reviewed and 2000;24:259–68.
approved the final version. 17 Anderson N, Ho-Foster A. 13,915 reasons for equity in sexual offences legislation: a
Funding This work was supported by the Economic and Social Research Council national school-based survey in South Africa. Int J Equity Health 2008;7:20.
(UK; http://www.esrc.ac.uk) and the National Research Foundation (South Africa; 18 Council on Higher Education. Vital Stats Pulic Higher Education 2010. Pretoria:
http://www.nrf.ac.za) [grant number RES-062–23-2068], the National Department of Council on Higher Education, 2012.
Social Development (South Africa; http://www.dsd.gov.za), the Claude Leon 19 Collings S. Physically and sexually abused children: a comparative analysis of 200
Foundation (South Africa; http://www.leonfoundation.co.za), the Nuffield Foundation reported cases. Soc Work 1993;29:301–6.
(UK, http://www.nuffieldfoundation.org) [grant number OPD/31598], the Health 20 Chu JA, Frey LM, Ganzel BL, et al. Memories of childhood abuse: dissociation,
Economics and HIV/AIDS Research Division at the University of KwaZulu-Natal amnesia, and corroboration. Am J Psychiatry 1999;156:749–55. http://ajp.
(South Africa; http://www.heard.org.za) [grant number R14304/AA002], the John psychiatryonline.org/cgi/content/abstract/156/5/749 http://dx.doi.org/10.1176/ajp.
Fell Fund (UK, http://www.admin.ox.ac.uk/pras/jff ) [grant number 103/757], the 156.5.749
Leverhulme Trust (UK, http://www.leverhulme.ac.uk) [grant number PLP-2014–095], 21 Organisation WH. Report of the Consultation on Child Abuse Prevention. Geneva:
the University of Oxford’s ESRC Impact Acceleration Account, and the European WHO, 1999.
Research Council under the European Union’s Seventh Framework Programme [FP7/ 22 Slack KS, Holl JL, McDaniel M, et al. Understanding the risks of child neglect: an
2007–2013, grant agreement 313421]. FM was funded by an ESRC doctoral exploration of poverty and parenting characteristics. Child Maltreat
studentship [grant number OSSID 454387]. The funders had no role in the study 2004;9:395–408.
design, data collection and analysis, decision to publish or preparation of the 23 Snider L, Dawes A. Psychosocial vulnerability and resilience measures for
manuscript. national-level monitoring of orphans and other vulnerable children:
recommendations for revision of the UNICEF psychological Indicator. Cape Town:
Competing interests None declared. UNICEF, 2006.
Ethics approval Ethical approval was granted by the Universities of Oxford, Cape 24 Meinck F, Cluver LD, Boyes ME. Household illness, poverty and physical and
Town and KwaZulu-Natal; the National Department of Social Development; and the emotional child abuse victimisation: findings from South Africa’s first prospective
Western Cape and Mpumalanga provincial Departments of Health and Education. cohort study. BMC Public Health 2015;15:444.

6 Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860

 Research report
25 Pettifor AE, Rees HV, Kleinschmidt I, et al. Young people’s sexual health in South 36 Pereda N, Guilera G, Forns M, et al. The prevalence of child sexual abuse in
Africa: HIV prevalence and sexual behaviors from a nationally representative community and student samples: a meta-analysis. Clin Psychol Rev
household survey. AIDS 2005;19:1525–34. http://pubget.com/paper/16135907 2009;29:328–38.
http://gateway.ovid.com/ovidweb.cgi?T=JS&PAGE=fulltext&D=ovft&MODE= 37 Dawes A, De Sas Kropiwnicki Z, Kafaar Z, et al. Survey examines South Africa’s
ovid&NEWS=N&SEARCH=0269-9370.is+and+19.vo+and+14.ip+and+1525.pg attitude towards corporal punishment. Article 19 2005;1:1–3.
26 Finkelhor D, Hamby SL, Ormrod R, et al. The juvenile victimization questionnaire: 38 Cicchetti D, Carlson V. Child maltreatment: theories and research on the causes and
reliability, validity, and national norms. Child Abus Negl 2005;29:383–412. consequences of child abuse and neglect. 1989.
27 Finkelhor D, Ormrod R, Turner H. Poly-victimization: a neglected component in child 39 Morrell R. Corporal punishment and masculinity in South African schools. Men
victimization. Child Abus Negl 2007;31:7–26. Masc 2001;4:140–57.
28 Felitti V, Anda M, Nordenberg D, et al. Relationship of childhood abuse and 40 Prinsloo S. Sexual harassment and violence in South African schools. South Afr J
household dysfunction to many of the leading causes of death in adults: the Educ 2006;26:305–18.
Adverse Childhood Experiences (ACE) study. Am J Prev Med 1998;14:245–58. 41 Statistics South Africa. Poverty Trends in South Africa. An examination of absolute
29 Besharov DJ. Toward better research on child abuse and neglect: making poverty between 2006 and 2011. Report. 2014. http://beta2.statssa.gov.za/
definitional issues an explicit methodological concern. Child Abuse Negl publications/Report-03-10-06/Report-03-10-06March2014.pdf (accessed 11 Mar
1981;5:383–90. 2015).
30 UNICEF. Hidden in plain sight: a statistical analysis of violence against children. 42 Gilbert R, Kemp A, Thoburn J, et al. Recognising and responding to child
New York: UNICEF, 2014. maltreatment. Lancet 2009;373:167–80.
31 UNICEF. Violence against children in Tanzania: findings from a National Survey, 43 Johnsona R, Kotch J, Catellier D, et al. Adverse behavioural and emotional
2009. Dar es Salaam: UNICEF Tanzania,Centers for Disease Control and Prevention outcomes from child abuse and witnessed violence. Child Maltreat 2002;7:179–86.
& Muhimbili University of Health and Allied Sciences, 2011. 44 Schwarz N. Retrospective and concurrent-self-reports: the rationale for real-time
32 UNICEF. Violence against children in Kenya: findings from a 2010 National Survey. data capture. In: Stone A, Shiffman S, Atienza A, et al, eds. The science of
Nairobi: UNICEF, Centers for Disease Control and Prevention & Kenya National real-time data capture: self-reports in health research. New York: Oxford University
Bureau of Statistics, 2012. Press 2007:11–26.
33 Black DA, Smith Slep AM, Heyman RE. Risk factors for child psychological abuse. 45 Mikton C, Butchart A. Child maltreatment prevention: a systematic review of
Aggress Violent Behav 2001;6:189–201. reviews. . Bull World Heal Organ 2009;87:305–24.
34 Black DA, Heyman RE, Smith Slep AM. Risk factors for child physical abuse. Aggress 46 Makoae M, Roberts H, Ward CL. Child maltreatment prevention readiness
Violent Behav 2001;6:121–88. assessment: South Africa. Cape Town: Human Sciences Research Council (HSRC) &
35 Black D, Heyman R, Smith Slep A. Risk factors for child sexual abuse. Aggress World Health Organization (WHO), 2012.
Violent Behav 2001;6:203–29. http://www.sciencedirect.com/science/article/ 47 Vranceanu A-M, Hobfoll S, Johnson R. Child multi-type maltreatment and
B6VH7-42JYV6K-3/2/7800a5cbfd9c28e6e6309445732230fa http://dx.doi.org/10. associated depression and PTSD symptoms: the role of social support and stress.
1016/S1359-1789(00)00023-9 Child Abus Negl 2007;31:71–84.

Meinck F, et al. J Epidemiol Community Health 2016;0:1–7. doi:10.1136/jech-2015-205860 7