Disclosure of physical, emotional and sexual child abuse, help-seeking and access to abuse response services in two South African Provinces

Abstract

Physical, emotional and sexual child abuse are major problems in South Africa. This study investigates whether children know about post-abuse services, if they disclose and seek services, and what the outcomes of help-seeking behaviour are. It examines factors associated with request and receipt of services. Confidential self-report questionnaires were completed by adolescents aged 10–17 (n = 3515) in South Africa. Prevalence of frequent (>weekly) physical abuse was 7.4%, frequent emotional abuse 12.4%, and lifetime contact sexual abuse 9.0%. 98.6% could name one suitable confidante or formal service for abuse disclosure, but only 20.0% of abuse victims disclosed. Of those, 72% received help. Most common confidantes were caregivers and teachers. Of all abuse victims, 85.6% did not receive help due to non-disclosure or inactivity of services, and 14.4% received help: 4.9% from formal health or social services and 7.1% through community vigilante action. Emotional abuse, sexual abuse and female gender were associated with higher odds of help-seeking. While children in South Africa showed high knowledge of available services, access to and receipt of formal services among abused children was low. Notably fewer children received help from formal services than through community vigilante action. Urgent action is needed to improve service access for child abuse victims.

Keywords:

• Child abuse
• sexual abuse
• physical abuse
• emotional abuse
• service access
• victim services
• children

“I went to the police to tell them that that boy who had raped me was harassing me. They laughed and said I should not bother them. So now, every time he rapes me I won’t go to the police because they won’t believe I am telling the truth.” (Girl, 17, Mpumalanga)

Background

South Africa has a high prevalence of physical, emotional and sexual child abuse (Artz et al., 2016), and this abuse is associated with negative long-term and short-term health outcomes (Norman et al., 2012; Paolucci, Genuis, & Violato, 2001). While national policy frameworks and legislation for child protection is rigorous, problems have been identified with the implementation of services. The national government, international agencies such as UNICEF, and non-governmental services have identified the urgent need for integrated health, social, police and criminal justice services to support victims of child abuse (UNICEF South Africa, 2013). However, limited studies about access to services for abused children exist globally and within South Africa, and those that do focus mostly on victims of sexual abuse.

It is essential that victims disclose their abuse in order to access support services. Research on sexual abuse disclosure suggests that it is a multi-stage process that may include stages of reluctance, denial and re-affirmation (Bradley & Wood, 1996). Globally, research indicates that disclosure of sexual abuse is often delayed: the younger the victim, the longer the delay between onset and disclosure (London, Bruck, Ceci, & Shuman, 2005). In the majority of cases, initial disclosures are made to informal confidantes such as parents, teachers or friends, rather than professionals (Allnock & Miller, 2013; McElvaney, 2015). Where children do disclose to professionals, swift action is usually experienced as helpful (Allnock & Miller, 2013), while matter-of-fact responses, not being told about the plan of action and not being listened to further exacerbated the trauma (Mudaly & Goddard, 2006).

Reasons for non-disclosure have been identified as lacking vocabulary to describe the abuse, feeling threatened by the perpetrator, and experiencing shame, embarrassment and fear of not being believed (Allnock & Miller, 2013; McElvaney, Greene, & Hogan, 2014). Factors promoting disclosure are being asked about it, growing more mature and being able to vocalise the abuse, feeling unable to cope anymore, protecting others from similar abuse and enduring an escalation of abusive acts (Allnock & Miller, 2013). Evidence is thus far unclear about whether severity of the assault increases disclosure rates (London et al., 2005). Throughout the disclosure process, many children describe not being believed, lack of support and inadequate protective action (Allnock & Miller, 2013; Hershkowitz, Lanes, & Lamb, 2007; Mudaly & Goddard, 2006).

A recent qualitative study from South Africa suggests similar findings. The multi-stage process of disclosure is often delayed due to the child fearing that they would be blamed or punished by the caregiver or not believed. Most initial disclosures are made to close confidantes such as parents. However, some of these disclosures are elicited through threats or beatings where caregivers realise something is wrong with the child. Responses following disclosure range from being supportive to blaming and punishing the child (Mathews, Hendricks, & Abrahams, 2016).

A small number of quantitative studies examine abuse response services in South Africa with a focus on initial disclosure of rape, service delivery following disclosure and progression of court cases. Almost all those investigating service delivery have used official records. They are therefore limited to sexual abuse cases reported and responded to by available services. As children were selected from services records, these data give little insight into the views and experiences of those who do not have access to services.

In a study in KwaZulu-Natal, one in four children who disclosed rape (25.9%) reported non-supportive reactions by confidantes such as disbelieving or blaming the child when they inititally disclosed sexual abuse victimisation. Non-supportive disclosure was higher where children chose community members as confidantes compared to guardians, other relatives or professionals (Collings, 2007).

There is, however, an important distinction between supportive disclosure and receipt of actual help. In another study amongst children referred to socio-medical services following sexual assault, only 49% received counselling and social work services. These services were delayed anywhere between two days and six months (61% seen within the first week), and in all but one case, the support was limited to a single appointment. Children living in informal settings or those who sought help after hours were much less likely to receive any counselling or social services (Collings, 2009). In Gauteng, a suspect was arrested in only 57.2% of child rape cases. Of those, 18.1% went to trial, 7.4% were convicted and 4% sentenced to imprisonment. Documented injuries were predictive of cases going to trial but not of a conviction. The presence of obtainable DNA, which was only available in 1.4% of all reported cases was not a predictor either (Jewkes et al., 2009). These rates are much lower than in the United States, where a meta-analysis found that between 28% and 95% of alleged child abusers were charged. A mean of 79% of these cases were carried forward, 82% pleaded guilty and 94% were convicted (Cross, Walsh, Simone, & Jones, 2003).

In a recent, nationally representative sample of South African adolescents, the majority of children disclosed sexual abuse to their parents, although boys reported much lower levels of initial disclosure (15.2% for rape) than girls (36.2%). Of those who disclosed, 72.5% experienced a supportive disclosure when the perpetrator was a known adult. When the perpetrator was an unknown adult, 59.1% experienced a supportive disclosure. Much lower rates of supportive disclosure were recorded for those who experienced rape. Social services investigations were carried out in only 28.6% of rape cases (Artz et al., 2016).

Trends are similar across the sub-Saharan region as investigated by the nationally representative Violence Against Children Studies. In Malawi and Kenya, less than 10% of child sexual abuse victims sought formal services, while in Tanzania, less than 20% surveyed disclosed their abuse experience to formal service providers or family members. The percentage of children receiving services following sexual violence exposure varied from 2.7% in Zimbabwe to 25% in Swaziland (Sumner et al., 2015).

These existing studies provide valuable evidence of rates of disclosure, disclosure responses, access to services and receipt of services for sexual abuse victims who have reported their experiences to official services. However, evidence is urgently needed from community-based studies of all child abuse victims, including those who experience physical and emotional abuse and sexual abuse victims who have not sought services. Such studies would investigate the different stages in the process, from knowledge of services via disclosure to help-seeking behaviours to receipt of services. This information is essential to identifying whether there are gaps in knowledge of services, barriers to access and help-seeking, or difficulties with service delivery. Increased information can support plans for an adequate national implementation of child protection policies (DSD, DWCPD, & UNICEF, 2012).

Objective

This paper aims to establish (1) whether South African children know where to seek help in case of abuse (physical, emotional or sexual); (2) what proportions of abused children disclose and seek help; (3) the outcomes reported by children who accessed post-abuse services; and (4) the sociodemographic factors associated with requesting and receiving help.

Methods

This paper reports on a longitudinal, community-based household survey carried out in two different provinces of South Africa. In each province, two health districts with high deprivation were selected and census enumeration areas were chosen randomly. Within these census enumeration areas, every household was visited to recruit children aged 10–17 (n = 3514) for participation. The majority of children (n = 3401, 96.8%) were followed up a year later and this analysis reports on the follow-up sample.

Procedure

Children completed confidential 60–80-min study-specific self-report questionnaires with the help of trained local interviewers. Questionnaires were translated into five local languages and checked with back translation. Children participated in the language of their choice. All survey items were pre-piloted with vulnerable youth to investigate age-appropriateness and cultural sensitivity. Data were collected using paper questionnaires with the help of experienced research assistants. Stringent quality checks were in place so that missing data were <.05%.

Ethics statement

Ethical approval was granted by the University of Oxford (SSD/CUREC2/09–52).

Informed consent for child participation was sought from children and their caregivers. Due to low literacy in the sampled population groups, information and consent sheets were read aloud to children and caregivers, and clarification questions were answered until participants were satisfied and gave written consent. Participation was voluntary and children were able to stop the interview at any time. All participants received a certificate and a participant pack irrespective of completion of the questionnaire.

Confidentiality was maintained throughout the study unless participants were considered at risk of significant harm or requested help from the research team. In this case, the project manager and interviewer discussed options for referrals with the child, and immediate referrals were made to local child protection or health services as necessary. A total of 664 referrals were made throughout the duration of the research.

Measures

Physical and emotional abuse was measured with the UNICEF Measures for National-level monitoring of orphans and other vulnerable children (Snider & Dawes, 2006), and seven additional items devised with the help of local social workers and children. All items were modified to fit the cultural context in consultation with service providers and children. The measure was previously used in the Western Cape and showed good reliability of α = .70 (Meinck, Cluver, Boyes, & Ndhlovu, 2015). Items measured hitting or slapping with a hard item or so that the child had marks, standing/kneeling in an uncomfortable position for a long period of time, being singled out to do household chores, being insulted, threatened to be hurt or abandoned, told one is a burden, felt unwelcome in the home and having a meal withheld. For the full measure see (Meinck, Cluver, & Boyes, 2015a). Frequent physical and emotional abuse was defined as occurring weekly or more frequently. Contact child sexual abuse and rape were measured using the Juvenile Victimisation Questionnaire (Finkelhor, Hamby, Ormrod, & Turner, 2005). Items included unwanted touching or kissing, touching of private parts and forced sex. Contact sexual abuse was defined as lifetime exposure to contact sexual abuse or rape. For individual items please see (Meinck, Cluver, & Boyes, 2015b).

Access to services was measured using four items developed with local social workers examining knowledge of services, help-seeking behaviours and the outcomes if reporting occurred. One dichotomous item measured whether the child sought help. Three items were free-text responses without prescribed answer categories: ‘Where would you tell victims of violence to get help? Who did you ask for help? What actions were taken?’ The answers were then re-coded into categorical variables for this analysis.

Potential demographic covariates such as gender, age, and urban or rural location were measured. Poverty was measured using an index of the eight highest socially-perceived necessities for children in South Africa (Barnes & Wright, 2012) and was defined as lacking three or more necessities.

Analyses

Descriptive analyses were conducted using SPSS 22. Knowledge of post-abuse services, prevalence rates of help-seeking behaviour and receipt of services were examined. Multivariate logistic regression analyses were employed to investigate factors associated with help-seeking behaviour and receipt of services, as sample size allowed for subgroup analyses. Interactions between gender and the different abuse types were examined for each model.

Results

Prevalence of abuse

Socio-demographic characteristics of the sample are displayed in Table 1. Prevalence of frequent (weekly or more) physical abuse was 7.4%, frequent emotional abuse was 12.4% and lifetime contact sexual abuse was 9%. Overall, 22.0% of children were exposed to at least one form of violence (Table 1).

Table 1. Socio-demographic characteristics of the samples.

	(N = 3401)
Age	14.67 SD 2.22 SE .04
Female	57.0% (1937)
Province	Mpumalanga 48.5% (1648)
	Western Cape 51.5% (1753)
Poverty (missing 3 or more necessities)	45.4% (1543)
Rural location	49.8% (1692)
Weekly physical	7.4% (250)
Weekly emotional	12.4% (427)
Contact sexual abuse and rape	9.0% (306)
Multi-victims	6% (203)
Any type of victimisation	22.0% (749)
Knowledge of available services	98.6% (3353)
Of those abused: disclosed and asked for help	20.0% (150)
Of those abused: child received help	14.4% (108)

Knowledge of where to report abuse

98.6% of children could name at least one suitable formal service or confidante for abuse disclosure. 48.6% recommended that abuse victims seek help from the police while trusted adults, social workers, and the clinic or hospital were mentioned less often. Only 1.6% of children were not able to name a service or person that could provide help (Figure 1).

Figure 1. Children’s knowledge of services where child abuse victims could seek help.

Disclosure and help-seeking

20.0% of abused children had disclosed and asked for help while 80.0% did not (Table 1). Confidantes were the caregiver (27%), teachers (22%), other family members (12.2%), friends (9.6%), siblings (6.8%), social workers (6.8%), police (6.8%), community organisations (4%), health care professionals (2.7%) and neighbours (2.7%).

Outcomes of help-seeking and disclosure

Overall, 85.6% of abused children received no help due to a combination of non-disclosure and inactivity of services. Figure 2 demonstrates differences in knowledge about service, requests for help by abused children and help received through formal services. 28.0% (n = 42) of children who had been abused and requested help from a confidante or service were denied it. In 89.1% (n = 37) of these cases, no further action was taken while 11.9% (n = 5) were re-victimised by being subjected to harassment by the perpetrators or punished by their parents.

Figure 2. Child abuse victim’s access to services: attrition in the process from knowledge of services to outcomes of help-seeking.

Of those who were abused, 2.8% (n = 21) reported being assisted by the police, 2.1% (n = 16) by medical or social services, and 2.2% (n = 18) in other, not specified ways. However, a large proportion of those who were abused and received services, 7.1% (n = 53), received assistance in the form of community vigilante retribution and not from government or NGO services. Community vigilante retribution is defined as actions that are carried out by members within the community who chase the perpetrator away, beat the perpetrator or arrange some form of financial recompense to the family of the victim. This was the outcome of a free-text answer option and not anticipated by the researchers.

Is type of abuse associated with help-seeking and receipt of help?

Girls and victims of emotional and contact sexual abuse were more likely to seek help. Type of abuse was not significantly associated with receipt of help (Table 2). No significant interactions between type of abuse and gender with respect to help-seeking were found.

Table 2. Multivariate regression analyses of factors associated with help-seeking behaviour among abused children.

	Help seeking (n = 749)	Receiving help (n = 749)
Physical abuse	1.22 (.69–2.14)	1.58 (.49–5.11)
Emotional abuse	1.71^* (1.15–2.89)	.37 (.12–1.11)
Sexual abuse	10.08^*** (5.53–18.37)	.41 (.11–1.51)
Female gender	2.42^*** (1.45–4.04)	.73 (.22–2.33)
Age	.95 (.85–1.03)	.82 (.65–1.04)
Poverty	1.01 (.64–1.58)	.62 (.23–1.68)
Rural location	.67 (.42–.1.03)	1.64 (.64–4.22)

^*** p < .001

^* p < .05.

Discussion

This study identified high rates of frequent child abuse victimisation in two provinces of South Africa. Severe disparities between children’s knowledge of services and children’s reluctance to disclose and seek help from formalised services were found.

Knowledge of services and disclosure

The vast majority of children knew about services that can help victims of abuse. It may be that education in schools and communities about formal services is successful in generating knowledge about their availability. However, in this large sample, only 20% of children who experienced contact sexual abuse or frequent physical or emotional abuse disclosed their abuse or requested help.

Similar low rates of disclosure of sexual abuse in child samples have been reported in studies from high-income countries (London et al., 2005; Mcgee, Garavan, & Byrne, 2002) and South Africa (Artz et al., 2016) and may be due to fear of the perpetrator, not being believed, or being shamed or blamed (Allnock & Miller, 2013; Mathews et al., 2016). International research is limited on disclosure rates of physical and emotional abuse. Where it is available, disclosure rates are much higher (40–80%), but studies use retrospective self-report in adult samples and thus recall memory of when disclosure happened may be affected (Bottoms et al., 2016; Foynes, Freyd, & DePrince, 2009).

Choice of confidante

In this sample, the majority of children disclosed to a caregiver, teacher or other family member rather than professionals. Research from high-income countries corroborates this finding and adds friends to the list of close confidantes (Bottoms et al., 2016). In Sweden, the majority of adolescents disclosed to their peers (Priebe & Svedin, 2008), while in Ireland, confidantes were parents and friends (Mcgee et al., 2002). Confidante’s response to the disclosure of abuse was not directly measured in this study. Although the confidante was a trusted adult in most cases, 11.9% of children who disclosed abuse reported re-victimisation as an outcome while 28% reported that no further action was taken. This tallies with evidence from Israel where 20% of children reported unsupported disclosure (Hershkowitz et al., 2007). It is thus important to distinguish non-supportive disclosure, supportive disclosure and receipt of service provision as these can be mutually exclusive (Mudaly & Goddard, 2006).

Service provision

Notably, the number of cases in which community vigilante action took place was greater than the combined number of cases where responses were provided by social, health and police services. The perception that formal services are inaccessible or ineffective may be the reason that crimes are reported to community leaders instead of professional services. This finding is in line with a decrease of reported cases of maltreatment to the police since 2003 (South African Police Service, 2010), even though surveys suggest continuing high past-year prevalence rates of abuse (Artz et al., 2016; Meinck, Cluver, Boyes, & Loening-Voysey, 2016). Similar findings are also reported in Kenya, where village elders are perceived to be most effective in responding to child abuse reports (UNICEF, 2011); Sierra Leone, where structural barriers and poor training of child protection committees are reported (Wessells et al., 2012); and Tanzania, where distances, slow responses, and corruption of police and other service providers prevent victims from seeking help (Abeid et al., 2014). Within South Africa, qualitative research suggests that professional services are not designed to facilitate abuse disclosure and service access for children (Bray, Gooskens, Kahn, Moses, & Seekings, 2010). Some of the perceived barriers are logistical and practical obstacles, such as transport or money (Lankowski, Siedner, Bangsberg, & Tsai, 2014); inadequate and badly designed services and poor service delivery (Roehrs, 2011); lack of faith in a timely and positive outcome (Smith, Bryant-Davis, Tillman, & Marks, 2010); fear of repercussions such as re-victimisation (Akal, 2005); and stigma and coercion within the family (Akal, 2005; Bray et al., 2010). The low numbers of children disclosing directly to professionals and receiving services in our present study support these qualitative findings.

Even in high-income countries, the number of sexual abuse cases reported to formal services is strikingly low with 4.6%. Half of respondents stated that they were not given satisfactory information about further support services or procedures, and a significant minority reported insensitivity to their feelings (Mcgee et al., 2002).

Factors associated with help-seeking

In this study, factors associated with increased odds of requesting help were gender, sexual abuse and emotional abuse, whereas poverty, location and age were not associated with help-seeking. In contrast, a similar Canadian study with a much larger sample size (N = 23,000) found associations between contact with child protection organisations and younger age, poverty and gender, as well as exposure to physical, emotional and sexual abuse using adult self-report (Afifi et al., 2015). Differences in the findings may be due to differences in sample size and methodology as well as vastly different social welfare systems and resourcing.

Notably, girls were more likely to seek help for abuse than boys. Although girls reported higher rates of sexual abuse in this sample, interactions between abuse types and gender were not significant. This may demonstrate an increased vulnerability for male victims of sexual abuse who are not disclosing and accessing services (Artz et al., 2016). Several qualitative studies similarly find that male childhood sexual abuse victims experience similar barriers to disclosures as females and, in addition, face negative stereotypes and lack of acceptance of male sexual victimisation (Gagnier & Collin-Vézina, 2016; Sorsoli, Kia-Keating, & Grossman, 2008). Future qualitative research in South Africa should examine this further.

Cultural factors that were not measured in this study may also play a role: lower rates of disclosure in physical abuse cases may reflect high levels of societal acceptance of corporal punishment (Dawes, De Sas Kropiwnicki, Kafaar, & Richter, 2005). Lower rates in sexual abuse disclosure by males may be influenced by patriarchal societies and traditional values of strong males who are not victimised (Bridgewater, 2016).

Implications for future research

In addition to identifying a need for urgent action to improve services in South Africa, these findings also suggest a number of further research needs. (1) Examine service bottlenecks: although South Africa has a robust and well-considered legal framework for child protection, it would be valuable to explore how the application of this framework is failing abused children in these two provinces. (2) Investigate ways in which services can be made more accessible to children. (3) Explore the ways that services can be made more appealing and accessible to adult confidantes following child disclosure. (4) Examine how services can be more gender- and child-sensitive. (5) Explore knowledge, attitudes and practices of mandated service providers to identify potential for changes in service delivery. (6) Investigate outcomes of community vigilante actions and why this approach is a preferred response of victims and their families, taking power dynamics among victims, confidantes and formal services into account. In addition, research should evaluate whether and in which ways use of each respective post-abuse service type impacts on the psychosocial well-being of victims.

Limitations

This study is subject to a number of limitations. First, data are not representative of the South African population as the study took place in low-income, black African communities only. However, the study benefitted from in-sample variation such as the inclusion of five African language groups from rural and urban areas. Second, the research used child-self report, which may be subject to social-desirability bias. However, self-report is preferable to social services data, considering the scope of the study and the variation across provinces in the provision of social services. Third, the study did not measure response to first disclosure and this makes it difficult to separate disclosure from help-seeking. Fourth, the authors had not anticipated vigilante action as a major source of help. Further research is needed to elucidate why families use vigilante action as a preferred response to child abuse. Finally, low numbers of children who had disclosed to a confidante did not allow for sub-group analyses due to a lack of statistical power on those who had received particular services or were re-victimised.

Implications for policy and programming

This study identifies a number of key implications for policy and programming in response to child abuse in South Africa. First, while children know where help is theoretically available when they experience abuse, few disclose and request help. Urgent action at national, provincial and municipal levels is needed to establish why this is the case and to determine how access to services and confidence in disclosure of abuse can be improved. In particular, stigmatisation of male sexual abuse victims needs to be addressed to increase levels of disclosure in boys. Second, the low number of abused children who receive formal services is concerning and has the potential to increase risk of re-victimisation as perpetrators are not prosecuted. Furthermore, there are increased health risks, particularly for sexually-abused children, as post-exposure prophylaxis for HIV-prevention and emergency contraception are not administered. These shortcomings need to be addressed urgently as they have the potential to lead to further severe traumatisation and poor health outcomes. Policy makers and child protection professionals who deliver services could enlist community support expressed by the vigilante movement to create safe and protective environments, thereby increasing our chances of circumventing the perpetuation of violence.

Funding

This work was supported by the following funding agencies: Economic and Social Research Council (UK) and the National Research Foundation [RES-062-23-2068]; the National Department of Social Development [27/2011/11 HIV AND AIDS]; the Claude Leon Foundation [F08 559/C]; the Nuffield Foundation [PD/31598]; the Health Economics and HIV/AIDS Research Division at the University of KwaZulu-Natal [R14304/AA002]; the John Fell Fund [103/757]; the Leverhulme Trust [PLP-2014-095]; the University of Oxford’s ESRC Impact Acceleration Account [1311-KEA-004]; and the John Fell Fund [103/757]. FM was funded by an ESRC studentship [OSSID 454387]. FM and LC’s time were funded by the European Union’s Seventh Framework Programme [FP7/2007-2013 313421]. The paper was specifically provided for, and supported by, the Know Violence in Childhood initiative.

Acknowledgements

The authors wish to thank the children who participated in the study, their families and all the field staff working tirelessly to interview as many children as possible. We further thank Colleen Kelly, Sarah Hoeksma, Yulia Shenderovich, the Rural AIDS Development Action Research programme (RADAR) at the University of the Witwatersrand, Cape Town Child Welfare, and Jennifer Rabedeau for comments and edits. We also thank the Know Violence in Childhood initiative.

Disclosure statement

The authors have no conflicts of interest to disclose.