Developing an Understanding of Complex Trauma Among Child Sexual Assault Survivors in South Africa: Implications for Practice

ABSTRACT

South Africa has extraordinarily high levels of CSA with profound immediate and long‐term mental health outcomes. This study aimed to investigate the mental health adjustment of children who experienced CSA and accessed support services using mixed methods. Structured interviews using validated mental health screening tools were conducted with children, between the ages of 7–18 years at three intervals over a 10‐month period. Semi‐structured qualitative interviews were conducted with primary caregivers to explore their views on the child's adjustment, and a validated parenting measure was used to assess parenting practices over time. Nearly 20% of all children met the clinical criteria for depression, but younger children showed increased recovery from depression over time. More than a third (37%) of younger children had full‐symptom PTSD at baseline, with about a quarter (24%) of adolescents meeting the full‐symptom PTSD criteria. We found that although full‐symptom PTSD decreased to 26% for younger children, there was a slight increase (32%) in full‐symptom PTSD among adolescents. Children's experiences of poly‐victimisation were highlighted through caregiver interviews combined with caregivers' own experiences of unresolved trauma. Therapeutic models need to address the complex nature of trauma that children experience and unresolved trauma experienced by many families to break this intergenerational cycle.

Summary.

• Child sexual assault victims require context‐specific therapeutic models of care that consider high levels of poly‐victimisation in high violence contexts.

• Therapeutic practice needs to address the complex nature of trauma that children experience as well as the intergenerational trauma many families encounter when children disclose CSA.

• Therapeutic practice needs to consider when children remain in unsafe contexts, as this requires strengthening the caregiver–child relationship to improve the caregivers' participation in ensuring the child's safety.

1. Introduction

Child sexual abuse (CSA) is a global phenomenon that affects children and young people of all ages and in different settings, across countries, geographies and development levels (Veenema, Thornton, and Corley ²⁰¹⁵). It is estimated that 18 million girls aged 15–19 years had ever experienced sexual abuse, with sexual violence of girls most prevalent in East and Southern Africa (Akobirshoev and Nandakumar 2017). Although both boys and girls experience sexual violence, sexual abuse and exploitation of boys is a neglected area (UNICEF ²⁰²⁰). Although children of all ages are at risk, adolescence is a period of increased vulnerability, especially for girls (UNICEF ²⁰¹⁷).

South Africa has similar excessive lifetime experience of CSA. A national prevalence study on sexual abuse of children in South Africa (household survey) estimated that 10% of boys and 15% of girls reported lifetime sexual victimisation (Ward et al. 2018). However, the school‐based survey of this study reported higher estimates that 35.4% of children are sexually victimised before the age of 18 years (Artz et al. 2016). This estimate is consistent with an earlier community‐based survey that reported 38% of women and 17% of men had been sexually victimised before the age of 18 years (Jewkes et al. 2010).

A review of the long‐term outcomes of CSA shows that it is associated with an increase in behavioural problems, alcohol and substance abuse; deleterious educational outcomes; sexual risk‐taking behaviour; and an increase in risk of revictimisation (MacIntosh and Ménard ²⁰²¹). Other risk factors such as gender, age when abused, type and severity of abuse, cognitive abilities and relationship to the perpetrator also influence the child's immediate and long‐term mental health response to victimisation (Maniglio 2009). A study on the consequences of violence against children (VAC) in South Africa found that depression and anxiety could be reduced by 6%, alcohol abuse by 10% and interpersonal violence by 7% for women, whereas drug abuse can be reduced by 14% and self‐harm by 12% for the entire population, if CSA was prevented (Hsiao et al. 2018). Using a socio‐ecological approach, we understand that determining the risks and outcomes of CSA is due to a dynamic interplay between a child, the developmental stage and the social context that all influence long term outcomes (Maternowska and Fry 2018). These factors also interact with larger structural drivers such as economic and social policy environments, as well as institutional factors such as existence of services addressing violence that may create an enabling or protective environment (Maternowska and Fry 2018). Thus, higher prevalence of CSA can be attributed to a web of interrelated factors that lead to increased risk for children living in the Global South, in particular Africa.

In South Africa, violence is pervasive, and CSA often co‐occurs with other forms of violence and individual‐level risk factors coalescing with family and community‐level risks, influencing long‐term outcomes for children. The Birth to Twenty Plus Cohort revealed the cumulative nature of VAC and showed that CSA was incremental across childhood, from 10% among primary school‐aged children to just over 30% among adolescents and young adults (Richter et al. 2018). In addition, approximately 40% of children in this birth cohort had been poly‐victimised either through exposure to or experiences of multiple forms of violence over their life course (Richter et al. 2018). Poly‐victimisation leads to a higher likelihood of experiencing poor mental health with feelings of powerlessness and depression, which affect school and social adjustment compared to children with no victimisation experience (Cyr, Clément, and Chamberland ²⁰¹⁴). These repeated experiences of violence are better described by the term ‘complex trauma’, defined as the experience of CSA or other forms of violence that is multiple or chronic in nature (van der Kolk ²⁰⁰⁵). The concept of ‘complex trauma’ is useful when applied to CSA in the South African context, where CSA often occurs within a relationship of trust, as the perpetrator is known to the child (Ward et al. 2018). Psychologically, the child experiences a crisis of loyalty and is unable to interpret what is happening, impacting on emotional self‐regulation and impulse control, resulting in limited awareness of their behaviour and feelings (van der Kolk ²⁰⁰⁵). The effects of complex trauma are particularly damaging when the perpetrator is a close caregiver or parent, disrupting healthy attachment and sensitive, responsive caregiving and likely leading to serious developmental consequences (McCrory, De Brito, and Viding ²⁰¹⁰). Accessing support services early for CSA survivors is important to reduce the risk of long‐term mental health and psycho‐social difficulties.

The South African government has a constitutional obligation to prevent VAC, to protect children from further harm and to support and treat children who have experienced violence to restore their physical and psychological well‐being. The Children's Act of 2005 makes provision for services including therapeutic services for children and their families to reduce the long‐term impact of abuse. However, evidence suggests that few children who have experienced CSA have access to adequate therapeutic support through public services (Mathews, Abrahams, and Jewkes ²⁰¹³) and child survivors often fall through the cracks and remain at risk of continued abuse (Jamieson, Sambu, and Mathews ²⁰¹⁷). In addition, services are predominantly located in urban centres, creating an additional barrier for children in rural areas. This paper explores ‘complex’ trauma among CSA survivors and its impact on children's recovery post‐sexual abuse, considering the mental health impact and the implications for therapeutic service approaches in South Africa.

2. Methods

2.1. Study Design

This study employed a mixed methods longitudinal follow‐up study design to explore mental health adjustment in a sample of children who had experienced CSA. The study sample completed interviews thrice between June 2015 and May 2016. Ethical approval was obtained from the University of Cape Town's Health Sciences Faculty human research ethics committee (HREC REF 133/2015).

A set of validated instruments were used to measure the psychological and behavioural adjustment of the children over time, and the parenting behaviours of the primary caregivers. ¹ Qualitative interviews were also conducted with the caregivers of the children. Data collection instruments were translated into the local languages and pilot‐tested for applicability to the local setting.

2.2. Study Procedures

The study occurred in two provinces in South Africa, in two districts of KwaZulu‐Natal and three districts of the Eastern Cape. These regions have similar characteristics including being predominantly rural with high rates of poverty and unemployment and are plagued by several social problems, inadequate infrastructure and associated poor access to basic amenities and limited access to social services.

The strength of the study is the combined data from children and their caregivers. A convenience sample of children between the ages of 7 and 18 years was included in the study together with their primary caregivers, recruited from the provincial state departments and a non‐governmental organisation, both offering child protection and support services to children who had experienced sexual assault. Children with severe intellectual or behavioural difficulties, severely traumatised children, those who had been removed from their home through statutory processes before the first interview and those younger than 7 and 18 years or older were excluded from the study.

The Child and Youth Care Workers from the non‐governmental organisation introduced the study to participants; similarly state social workers assisted researchers with the initial contact. If caregivers and children showed an interest in the research, trained research assistants proceeded to obtain informed assent and consent from the child and caregiver. Child and caregiver dyads received compensation for travel costs and were provided with a gift voucher. After obtaining the necessary consents, research assistants conducted a combination of structured and semi‐structured interviews with the children and their caregivers in their homes or safe community venues or at the department of social development offices.

At the baseline assessment, 80 children and their caregivers participated. Follow‐up assessments took place at 4 months (midpoint) and again at 10 months (endpoint) after baseline. There was slight attrition at the second and third assessment points, with 74 and 77 child–caregiver dyads interviewed, respectively. Loss to follow‐up resulted because of families' migration to other provinces, harsh weather conditions resulting in inaccessible routes in rural terrain during fieldwork and challenges with tracing families. We discuss only the baseline and endpoint findings in this paper.

2.3. Therapeutic and Support Services

All children were offered therapeutic care and support; however, the NGO and state‐delivered programmes used different approaches. The NGO's therapeutic model was delivered through a residential programme delivered through both individual and group therapy over a 1‐week period using qualified therapists. It used an eclectic model of therapy, drawing on play therapy modalities, debriefing, Rogers' person‐centred approach and psycho‐educational therapy. Where feasible, primary caregivers were present at the programme and participated in limited parenting education sessions to learn strategies for enabling their children's recovery process.

The state‐delivered programme includes child‐protection services and support from the local social services office. They offer general psychosocial services that may include home visits, counselling and psycho‐educational support (Mathews, Abrahams, and Jewkes ²⁰¹³). It is assumed that therapeutic intervention is ongoing (as opposed to the intensive residential programme), and therefore, psychological and behavioural change should be evident over time. The child survivor is the focus of these services, with no or limited focus on the caregiver's role in facilitating their child's recovery.

2.4. Measuring Children's Mental Health Adjustment and Parenting Practices

PTSD symptomology was measured using the child PTSD Checklist, a 28‐item scale using a 4‐point Likert scale to measure severity of symptoms. The PTSD checklist rates the degree to which each of the 17 symptoms of PTSD has been present over the past month. The scale has previously been used in South Africa (Cluver and Gardner 2006; Seedat et al. ²⁰⁰⁴) and found to be diagnostically useful. The scale shows test–retest reliability (r = 0.91) and internal consistency; Cronbach's α ranges from 0.82 to 0.95 (Cluver, Fincham, and Seedat ²⁰⁰⁹).

Depression was assessed using the Child Depression Inventory (CDI short form), which consists of 10 items (Kovacs 1985) and had previously been used in South Africa with good psychometric properties (Cluver and Gardner 2006). Each item has three responses with scores of 0 to 2 for a total composite score of 0 (not depressed) to 20 (very high risk of depression). A cut‐off of 8 was used for this study as this is equivalent to a standardised t‐score of 62–66, which corresponds to identifying the upper 10% of the distribution in a non‐clinical sample (Cluver and Gardner 2006).

Parenting practices was measured using the Alabama Parenting Questionnaire (short form), a nine‐item shortened scale of the full Alabama Parenting Questionnaire (42 items), which measures three elements of parenting behaviour: positive parenting, inconsistent discipline and poor supervision (Elgar et al. 2007).

Table 1 further describes the measures used to assess the study outcomes. Demographic information was also collected on the children, caregivers and their households.

TABLE 1.

Child and caregiver: research instruments and administration intervals.

Outcomes: Psychological, social and behavioural	Children: age 7–12 years	Children: age 13–17 years	Caregiver	Administration at intervals
Post‐traumatic stress	The Child PTSD Checklist	The Child PTSD Checklist	N/A	Baseline, midpoint and endpoint
Depression	The Child Depression Inventory	The Child Depression Inventory		Baseline, midpoint and endpoint
Parenting practices/parent–child relationship	N/A	N/A	The Alabama Parenting Questionnaire ‐ 9	Baseline, midpoint and endpoint

2.5. Data Analysis

STATA Version 13 was used as the primary quantitative analysis tool for this study. Demographic characteristics were assessed using frequency and descriptive statistics. Paired sample t‐tests were used to explore the changes in mental health symptomatology, behaviour and parenting across the assessment intervals and to assess whether changes were significant, with a significance level set at p < 0.05. We present descriptive statistics combined with qualitative data from caregiver interviews to support the data interpretation and consider how the combined data addresses the study questions. All transcripts were inductively analysed through a process of coding and sub‐coding. After broad codes had been established (loosely based on interview questions), thematic sub‐categories were developed. These were analysed as stand‐alone categories, in addition to exploring and analysing the relationships between them.

3. Results

This study included both girls and boys (N = 80), but only three boys were enrolled. This mirrors global patterns of CSA as boys are less likely to report experiences of sexual violence possibly due to stigma (Know Violence ²⁰¹⁷). Table 2 shows that there were slightly more younger children than adolescents (53.1% vs 46.9%) in the sample. The mean age of the sample was 12.4 years (CI 11.7–13.1 years). Most of the children (56.3%) lived with either a parent or both parents, followed by grandparents (25.5%) then relatives (12.5%) which includes aunts, uncles and extended families. More than half (55.2%) of the children lived in a household where no one was employed, indicating the severe levels of poverty and deprivation these children experienced. Nearly two‐thirds (63%) of children received cash transfers linked to the state's social assistance programme—with most in receipt of the primary poverty alleviation measure, the child support grant (75%) and 7 children in formal foster care received the higher value foster care grant. Only one child received a care dependency grant, available to children with a disability in need of permanent care.

TABLE 2.

Demographic characteristics of the sample.

	N = 80 (%)
Age (years)
7–12	43 (53.1)
13–17	37 (46.9)
Primary caregiver/s
Parent/s	45 (56.3)
Relative/s	10 (12.5)
Grandparent/s	22 (25.5)
Unknown	3 (3.8)
Is someone employed in the household?
Yes	35 (44.1)
No	44 (55.2)
Unknown	1 (0.7)
Receipt of a social grant
Yes	63 (81.8)
No	13 (16.9)
Unknown	1 (1.3)
When did the abuse occur?
< 6 months	16 (20.0)
6–1 year	5 (6.3)
> 1 year	59 (73.8)

Nearly two‐thirds of the sample (73.8%) experienced the sexual abuse incident more than a year prior to inclusion in the study. Disclosure of CSA is a complex process; a South African study found that nearly half of the children presenting at a clinical setting failed to disclose the sexual abuse immediately, particularly fearing caregivers' reactions (Mathews, Hendricks, and Abrahams ²⁰¹⁶). In addition, it is hypothesised that children seek a confidante to ‘tell’ as a strategy to transfer responsibility to a trusted person whom they perceive can ‘end’ the sexual abuse.

3.1. Mental Health Recovery of Children Post‐CSA Disclosure

Table 3 shows the outcomes of the mental health self‐reports by children at baseline and endline. Using the clinical cut‐off for depression, 19% of both the younger children and adolescents scored in the clinically depressed range at baseline. Nearly a year post CSA disclosure, we found a significant decrease in depression scores (10%) for younger children, but for older children, the decrease (14%) was not significant.

TABLE 3.

Mental health scores by age over time.

	Baseline		Endpoint
	N = 80		N = 77		p
	7–12 years	13–17 years	7–12 years	13–17 years	7–12 years	13–17 years
	N = 43	N = 37	N = 41	N = 36
Depression inventory					0.0158	0.115
< 8	35 (81%)	30 (81%)	37 (90%)	31 (86%)
≥ 8	8 (19%)	7 (19%)	4 (10%)	5 (14%)
PTSD mean
Symptom scores:
Re‐experiencing (5)	2.6	2.2	1.3	1.8
Avoidance (7)	2.6	2.11.6	1.5	2.6
Hyper‐arousal (5)	2.0	1.6	1.1	1.3
PTSD diagnosis
Full symptom	16 (37%)	9 (24%)	8 (26%)	10 (32%)	0.292	0.0725
Partial symptom	21 (49%)	19 (51%)	23 (56%)	21 (58%)	< 0.000	< 0.000

A conservative symptom threshold of ‘most of the time’ was used to measure symptom presence of PTSD when the symptoms persist for at least a month post CSA disclosure. In the majority (80%) of the sample the incident occurred more than 6 months before reporting, the PTSD measure was therefore used across the study period. More than a third (37%) of younger children had full‐symptom PTSD at baseline, and about a quarter (24%) of older children met full‐symptom PTSD criteria. We found that full‐symptom PTSD decreased to 26% for younger children at endline, although not statistically significant. However, as a slight increase (32%) in full‐symptom PTSD was found for the older children at endline, it approaches statistical significance. This needs to be treated with caution as the sample is small.

Importantly, approximately half of younger (49%) and older (51%) children fulfilled the criteria for partial‐symptom PTSD, meaning that they were symptomatic in one of the three symptom clusters. This shows that overall, approximately 80% of children met the clinical criteria for either full‐symptom or partial‐symptom PTSD at baseline. At endpoint, 56% of younger and 58% of older children met the partial‐symptom criteria with significant increases (p < 0.000) in symptoms for both younger and older children.

3.2. Unravelling Children's Mental Health Recovery

To understand mental health recovery for these CSA survivors, qualitative interviews with caregivers provide important insights into children's experiences. The narratives of caregivers highlight the turmoil in the lives of these children post‐CSA disclosure. A caregiver describes:

Her (the child) father's family disowned her because they said I was feeding her with lies and she fell for it. (The father) is walking free. We don't know how he won the case. We were only told the case was closed. (Caregiver of 12‐year‐old)

An important factor in the continuation of trauma symptoms is the continued presence of perpetrators. South Africa is known for a weak criminal justice system (Jamieson, Sambu, and Mathews ²⁰¹⁷) resulting in few perpetrators charged or convicted for CSA, leaving children vulnerable to continued abuse and trauma. This is also exacerbated by some families' protection of perpetrators, particularly when the perpetrator is a close relative, and the criminal case gets withdrawn to maintain family honour. However, the continued presence of perpetrators impacts on children's ability to recover. Narratives by caregivers describe that children continue to be fearful for their safety as the risk of continued abuse is real. This is noted by a parent of a 16‐ year old who said: “It has been hard … because the perpetrator is free”.

Similarly, another caregiver talked about the presence of the perpetrator in the community:

The perpetrator is walking free, he got out on bail, but after three years there hasn't been any development in the case. I have been attending court, but nothing ever happens. My child is traumatised because she saw this man and he blew his hooter at her, she was terrified …. (Caregiver of 12‐year‐old)

Caregivers described the need to develop strategies to protect their children, as one mother explained:

I never leave xxx with him (referring to the perpetrator). I take her with me everywhere I go. If I go to a funeral I leave her with the neighbour. I don't want to test him by leaving my child with him and have the same thing happen again. (Caregiver of 8‐year‐old)

The continued presence of perpetrators has a negative effect on children's recovery and can result in continuous and complex forms of trauma. Importantly, for any psychosocial support or therapeutic programme to be effective, the safety of the child is paramount (Kaminer, Seedat, and Stein ²⁰⁰⁵). Therefore, ensuring a child's safety post‐disclosure is the first step towards recovery and needs to be prioritised by any child protection programme.

Fearful behaviour and continuing to be ‘scared’ of perpetrators were common among the children and can lead to continued symptomatology. For younger children caregivers describe changes in behaviour such as not wanting to play outdoors and other forms of separation anxiety. But other children reverted to behaviours such as wanting to sleep with the caregiver at night and bedwetting. An inability to process the trauma can lead to an increased risk for internalising symptoms, behavioural problems and suicidal ideation for children (Watters and Wojciak 2020). A caregiver discussed her daughter's suicidal ideation and her feelings of helplessness:

Things have become easier, but I am troubled by xxxx, she does not talk when she is sad, I can see tears in her eyes. I don't know if it is because of what happened or is she naturally like that. … One day a young child came and called me, saying that xxxx is at home saying she wants to kill herself. (Caregiver of 17‐year‐old)

This was not the only child in the study who attempted suicide, but another adolescent girl was admitted to a psychiatric unit for an attempted suicide. Threats of suicide and withdrawal should be noted as risk factors not to be minimised, but caregivers struggle to make sense of the changes in children's behaviour.

3.2.1. Caregiver Experiences and the Interconnections With Parenting Practices

Emotional support by caregivers is important for mental health recovery post‐CSA (Hyman, Gold, and Cott ²⁰⁰³). However, caregivers struggle to understand children's change in behaviour and experience difficulties in managing their behaviours. Caregivers described the following:

I found out, when she went out 'til late and came back with a woman from the neighbourhood … I remember shouting at her, and I hit her. I don't want to lie, she's very respectful and she listens; one time I sat her down and I explained it to her that the reason for disciplining her in that harsh manner was because I don't want them going through the hardship I had experienced growing up. (Caregiver of 14‐year‐old)

Another said:

Besides not listening … she doesn't listen at all. If you say to her, ‘xxxx stop what you are doing, she will say, okay.’ … but she won't stop what she's doing. I am also forced to take a stick and beat her. In terms of respect, she respects (me), but sometimes she becomes naughty. (Caregiver of eight‐year‐old)

Caregivers' negative perceptions and inability to make sense of the child's behaviour and the feelings of not being able to manage the situation resulted in several caregivers resorting to punitive forms of discipline. Caregivers' experiences of unresolved trauma emerged through their interviews, adding to the complexity of parenting children who have experienced trauma. Experiences of adversity including personal experiences of sexual abuse or intimate partner violence shaped the ability of caregivers to be responsive to the needs of their children. A mother shared:

Eish it is hard … (pause) sometimes I think of killing my children because of what happened to me, now it is happening to xxxx (crying). I did not plan to have her. I lost both my parents and I was then abused. Now they (her children) are abused as well. What did I do to the Lord? (silence, crying) I even think about killing myself …. (Caregiver of a 13 year old)

Caregiver interviews also highlighted conflict and the intergenerational effects of trauma, with mothers blaming children for being victimised, and how their own experience of abuse affects their emotional availability to the child.

We fight a lot and I end up saying bad things, like telling her that what happened to her (referring to the sexual abuse) was no accident, she wanted it as the perpetrator was her boyfriend. I regret saying those things to her, but it is because of the hurt and pain I am feeling especially when she shuts me off, I can tell what I say hurts her too. (Caregiver of 13‐year‐old)

Similarly, another caregiver reported:

I am better now (parent speaking about her mental health) but when she is not around, I get very sick. I will go and look for her; I only get rest when she is at school … I get headaches thinking about what had happened to me (referring to her own sexual abuse). This will not go away. (Caregiver of 13‐year‐old)

This is a complex area as parenting practices intersect with parents' own experience of adversity during childhood that can impact on their sense of parenting competence, resulting in increases in parenting stress and reliance on punitive discipline strategies (Schuetze and Eiden 2005), which is considered normative practice (Lachman et al. 2016).

Table 4 shows caregivers' parenting practices at baseline and at endpoint. By endline parenting practices had deteriorated in almost all domains. Unresponsive parenting behaviour worsened over time showing a significant difference in both younger and older children, with more than a third of caregivers not responding appropriately to the needs of these young people. A caregiver described her threats, an indication of inappropriate discipline, and possibly over‐protective behaviour:

TABLE 4.

Parenting practices of caregivers of children participating in the study.

	Baseline		Endpoint
	N = 80 (%)		N = 77 (%)		p
	7–12 years	13–17 years	7–12 years	13–17 years	7–12 years	13–17 years
	N = 43	N = 37	N = 41	N = 36
Unresponsive parenting	13 (30.2)	7 (18.9)	16 (39.0)	12 (34.3)	< 0.00	< 0.00
Poor supervision	39 (90.7)	36 (97.3)	41 (100.0)	34 (97.1)	0.044	0.824
Inconsistent discipline	33 (76.7)	36 (97.3)	37 (90.2)	31 (88.6)	0.0437	0.0536

… And this playing with boys now has appeared because as the schools are closed those boys sit here at my yard and call xxx outside the yard. I said, ‘I will kill you with my hands if you go and play with those boys’, and they would run away and never come back [the boys]. (Caregiver of 8‐year‐old)

Poor supervision was very high at baseline for both the younger and older group, but it significantly worsened among younger children by endline. Similarly inconsistent discipline worsened among younger children, but it marginally improved for older children. However, we must be cautious comparing the data over time, as nearly half (46%) of the caregivers interviewed at endpoint were not the same caregiver interviewed at baseline, but these data still provide us with insights into the children's caregiving experiences, which influences their mental health recovery.

4. Discussion

This study found that although this cohort of children accessed support services post sexual abuse disclosure, nearly a year later, their mental health recovery was limited. We observed that full‐symptom PTSD levels remained high and increased to about a third of adolescents but younger children showed some recovery from PTSD symptomatology. The persistence of high levels of PTSD (26% younger children; 32% adolescents) pose significant risk for long‐term psycho‐social outcomes (Maniglio 2009). Levels of depression also persisted (14%) among adolescents, placing them at increased risk for a number of poor psychosocial outcomes, including lower educational attainment, unemployment and lower perceived social support (Naicker et al. 2013).

Drawing on the qualitative findings, we show that the poor mental health recovery in this cohort of children intersects with multiple factors such as children's continued exposure to perpetrators, a weak criminal justice system, unresolved caregiver trauma, poor parenting practices that include poor emotional and social support. The therapeutic approaches used in the South African context is a further pertinent factor as discussed below.

Child and adolescent mental health are greatly impacted by the family context, particularly the parent–child relationship. Especially, when children and adolescents experience trauma, they need parental warmth, acceptance and nurturance to restore a sense of safety and security to assist in recovery post CSA. Most of the children in this study did not receive emotionally supportive, positive parenting. We found that caregivers displayed unresponsive caregiving, poor supervision and inconsistent and harsh discipline. The lack of supportive parenting can heighten children's levels of emotional distress and influence behavioural and psychological adjustment (Meyerson et al. 2002), contributing to the slow mental health recovery post‐CSA. In addition, we found that care arrangements in this sample were influx post CSA disclosure. Such changes in care arrangements can create emotional insecurities in an already vulnerable group of young people, particularly younger children, and can compromise their mental health recovery.

Caregivers' own mental health, including past trauma, was highlighted by their narratives, which has bearing on their parenting abilities. Parental trauma is likely to impact on parenting attitudes and behaviours; this includes a diminished ability to show empathy and to provide emotional support and security to the victimised child (Chiesa et al. 2018; Reese et al. ²⁰²²). Parenting behaviour is also influenced by the parent's own emotional struggles for example depression can lead to an inability to emotionally connect with the child, making it difficult to manage child behaviour (Craig and Sprang 2007). Parents affected by trauma can demonstrate an emotional instability, which their children are likely to emulate and embed in their own behaviours, increasing the risk of transmitting trauma to their children (Reese et al. 2022).

The limited mental health recovery in this group of children suggest that current psycho‐social support services, including an intensive residential therapeutic programme, are not having the desired outcome on trauma particularly among the older children. Ensuring that a traumatised child is safe is crucial in any therapeutic approach, yet in South Africa only 12% of CSA cases result in a conviction of the perpetrator (Machisa et al. 2017). Therapeutic approaches need to consider when children remain in unsafe contexts, strengthening the caregiver–child relationship and improving the caregivers' participation in enabling the child's physical, emotional and psychological safety, is critical. Building the capacity of caregivers to respond appropriately to the needs of their children is vital. Caregiver support and parenting interventions that promote responsive, sensitive parenting are recognised as a valuable contribution to supporting and strengthening families and to preventing violence (Knerr, Gardner, and Cluver ²⁰¹³). It is also critical that our treatment approaches consider the caregivers' own trauma exposure as this is important in a society like South Africa, where there are high levels of trauma experienced by the population (Gregorowski and Seedat 2013).

5. Conclusion

The ubiquitous nature of all forms of VAC in South African requires the development of trauma‐informed, context‐specific therapeutic models of care that take account of the high levels of poly‐victimisation and inability to ensure CSA survivors' safety. The complex trauma working group of the National Child Traumatic Stress Network (NCTSN) in the United States has developed treatment guidelines focused on reducing a child's emotional distress while simultaneously working with the child and caregiver using a strengths‐based approach to build resilience (Gregorowski and Seedat 2013). However, public mental health services in South Africa are limited; we therefore urgently need to develop cost‐effective treatment approaches to manage trauma early on to prevent long‐term adverse psychological and biological outcomes (Flisher et al. 2012). The feasibility of using a task‐shifting approach by training lay counsellors to deliver Trauma‐Focused‐Cognitive Behavioural Therapy (TF‐CBT) has been shown to be effective in Zambia and Tanzania, with children and families with high levels of trauma (Murray, Nguyen, and Cohen ²⁰¹⁴; O'Donnell et al. ²⁰¹⁴). Testing its efficacy for sexually abused children in South Africa, where there are high levels of violence and trauma is important. Furthermore, trauma‐informed practices have utility as it uses multiple strategies such as, empowerment, enhancing self‐regulation, building healthy relationships and safe environments to restore a sense of safety (Steele and Malchiodi 2012).

This study is not without its limitations, but it has provided us with important insights into the mental health recovery of children accessing support services in two provinces in South Africa. The high rates of trauma among CSA survivors in a context of poly‐victimisation and continued exposure to perpetrators suggest that we urgently need to reconsider our therapeutic models to address poly‐victimisation, the complex nature of trauma that children experience and the intergenerational trauma that many families endure. Current therapeutic approaches are mainly adapted from high‐income countries aimed to reduce psychological symptomatology such as PTSD and internalising and externalising behaviour problems, but these approaches have limited efficacy when trauma is continuous, and children are poly‐victimised. The current training of frontline staff like social workers and teachers do not adequately equip them to deal with the trauma that children have experienced. We need to integrate trauma‐informed approaches into the delivery of support services and across settings like schools; health; social services and the criminal justice system (police and courts). It is only through adopting a holistic approach to addressing trauma that we will be able to break the intergenerational cycle of violence.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of the study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions.