Trauma-informed art and play therapy: Pilot study outcomes for children and mothers in domestic violence shelters in the United States and South Africa

Abstract

Background

While intimate partner violence (IPV) has well documented impact on women and children, few interventions have been tested for mothers and children in the domestic violence shelter system.

Objective

We used mixed methods to explore effects of a pilot intervention combining trauma focused cognitive behavior therapy (TF-CBT) (verbal therapy method) with art and play therapy (non-verbal therapy methods) in New York City and Johannesburg, to compare efficacy between high and low-middle income contexts.

Participants and setting

School-aged children and their mothers from one domestic violence shelter in each city were invited to participate.

Methods

Children were screened for depression and post-traumatic stress disorder (PTSD). Children participated in a weekly group session lasting 1–2 hours over 12 weeks and mothers received 3 group sessions. Quantitative data comprised pre-and post-intervention child self-reports (n=21) and mother’s report (n=16) of child depressive and PTSD symptoms. Qualitative in-depth interviews were conducted with children (n=11) and mothers (n=8) who completed the intervention.

Results

At baseline, children showed high rates of symptoms of probable depression and probable PTSD (33% and 66% respectively). By endline, depressive symptoms significantly reduced (mean of 13.7 to 8.3, p=0.01) and there was a non-significant trend towards PTSD improvement (40.0 to 34.4, p=0.21). Children revealed the art helped them express difficult emotions and experiences with their mothers. Multiple children felt it assisted in managing challenging behaviours.

Conclusion

This pilot trauma-informed art and play therapy group intervention holds promise in mitigating the effects of IPV for children and mothers in domestic violence shelters.

Introduction

Globally, 30% of women experience intimate partner violence (IPV) (Devries et al., 2013), defined as physical or sexual abuse by a romantic partner (Garcia-Moreno et al., 2006). A considerable body of literature implicates exposure to IPV with harmful and enduring social, psychological, and physical health consequences for both women and children. Meta-analyses confirm an association between exposure to IPV and heightened internalizing and externalizing challenges among young people (Austin et al., 2017) and mental health problems in mothers (Greene et al., 2018). In addition, residing in a household with IPV increases the odds that children will experience other forms of child maltreatment (Hamby et al., 2010). Maltreatment is a particularly troubling problem for children in low and middle income countries (LMICs) due to social determinants such as poverty, inadequate access to services, violence, food insecurity etc. (Patel et al., 2007).

Parenting in the midst of enduring IPV is notably taxing and stressful. Research utilizing maternal self-reports has highlighted that IPV adversely affects how women evaluate their own parenting (Levendosky & Graham-Bermann, 2001). In addition, children exposed to IPV report that their mothers have less patience and exhibit more frustration and anger which leads to increased tension between them (Lapierre et al., 2018). For children who witness IPV towards their mothers, their trajectory of post-traumatic stress seems to get worse over time, however, their mental health has been found to improve when their mother’s mental health improves in tandem (McFarlane et al., 2014). Resilience research indicates that crucial protective factors for children include their own mother’s mental health, her parenting style and the quality of mother-child relationships (Fogarty et al., 2019). Thus any mental health intervention for IPV exposed families should attempt to intervene systemically with both mothers and their children. However, few interventions of this kind have been examined to date in peer-reviewed literature.

The manner in which women parent is greatly influenced by their own childhood experiences as parents tend to apply parenting behaviour that is based on child-rearing practices of their own parents (Assink et al., 2018). For example, the risk for maltreating children has been demonstrated to be six times greater when parents reported a history of child maltreatment themselves (Finzi-Dottan & Harel, 2014). Mothers in the domestic violence shelter system report higher rates of adverse childhood experiences than national averages and consequently also higher rates of mental health problems (Fredland et al., 2018). Moreover, mothers with adverse childhood experiences are at greater risk of insecure parent-child attachment relationships (Cooke et al., 2019), and the IPV environment tends to put mothers and their children at risk for perpetuating this quality of attachment (Lieberman et al., 2011). In fact, violence and adversity in childhood are key risk factors for violence victimisation and perpetration against children and adults in adulthood leading to persistent intergenerational transmission of violence and compromised mental health (Kimber et al., 2018; Woollett & Thomson, 2016).

As such, women and children within the domestic violence shelter system would be a vital and vulnerable target population for mental health intervention. Although there is recognition that upon entering the system, women and children require mental health treatment, there are few services available, and limited documented interventions proven effective for them (Rizo et al., 2011). The transition to shelter living comes with its own challenges, including housing instability, poverty and psychiatric distress (Anderson et al., 2018). Children are often displaced and stressed from leaving their homes, schools and friends - all of which can bring additional strain to the mother-child relationship. The “conspiracy of silence” typical in homes characterized by IPV (Lapierre et al., 2018, pg.1026) tends to be transported to shelter environments where women and children may feel afraid to speak openly about the violence due to learned behaviour, isolation, stigma, fear that the children might be removed from the mother’s care or fear regarding repercussions from disclosure (Callaghan et al., 2017; Pifalo, 2007).

The trauma of IPV and its exposure is complex in nature. This kind of violence tends to be ongoing, cumulative and occurs within a caregiving system. Complex trauma is a persisting disorder and results in difficulties with self-regulation, attachment, interpersonal relationships, dissociation, memory and attention, and cognition (Cloitre et al., 2009). Women and children within the domestic violence shelter system require interventions that speak to both populations’ need simultaneously (Fry & Elliott, 2017) and address trauma across the family system (Hasselle et al., 2019).

Trauma Focused Cognitive Behaviour Therapy (TF-CBT) is considered an effective treatment for PTSD and complex trauma in children and youth (Foa et al., 2008). Studies in the US and in African countries have found TF-CBT to be highly effective in treating the symptoms of child and adolescent trauma (Murray et al., 2015; O’Donnell et al., 2014) with constant gains at 6 to 24 months after treatment (Mannarino et al., 2012). However, more than 30% of patients do not benefit from this treatment (Bradley et al., 2005), largely on account of prolonged, complex and multiple traumatization (Spinazzola et al., 2005), a contextual reality particularly prevalent in South Africa (Kaminer et al., 2018; Ward et al., 2018). It is important to understand qualitatively how children and mothers respond to TF-CBT approaches, though this type of research has been limited to date.

For working with children in particular new therapeutic interventions need to be investigated. Non-verbal therapeutic approaches are particularly effective in the treatment of traumatized children and adolescents (Goodman et al., 2009). Through supportive relationship these methods facilitate the safety required to express distressing events and offer discovery through access to non-verbal approaches such as art making and play. These means of communication open possibilities for imagination, creative problem solving, participatory action, and self-expression (Malchiodi, 2005).

Art therapy is defined as “a treatment in which patients, facilitated by the art therapist, use art materials, the creative process, and the resulting artwork to explore their emotions, foster self-awareness, reduce anxiety, and increase self-esteem” (American Art Therapy Association, 2019). Central to art therapy is the curative power of the creative process and the inimitable communication that occurs between the client, the artwork, and the therapist. The artwork products record the therapeutic process, facilitate their creators to ensue a dialogue with them, and are lasting objects that can be interacted with over time (Avrahami, 2006). A growing body of evidence indicates that art therapy is an effective means of working with traumatized patients in multiple contexts (Schouten et al., 2019; Tripp et al., 2019; Ugurlu et al., 2016; van Westrhenen et al., 2019). Now, more work is needed to explore how art therapy can be implemented and tested in LMICs.

Play therapy is a “structured, theoretically based approach to therapy that builds on the normal communicative and learning processes of children” (Schaefer & Drewes, 2014, pg.3). It can be defined as “the systematic use of a theoretical model to establish an interpersonal process wherein trained play therapists use the therapeutic powers of play (e.g., relationship enhancement, role-playing, communication, mastery, etc.) to help clients prevent or resolve psychosocial difficulties and achieve optimal growth and development” (Association for Play Therapy, 2019). Play has the distinctive ability to honour the child’s emotional world and personal logic whilst concurrently offering the conditions to author an improved, more adaptive narrative (Stewart et al., 2016). Play therapy is an effective and growing approach to intervention for childhood disorders and trauma (Drisko et al., 2019; Goodyear-Brown, 2019).

Timely intervention for both mothers and children who enter domestic violence shelters seems prudent, not only in terms of managing mental health symptoms, but also in curbing intergenerational transmission of violence. In this paper, we examine an intervention combining TF-CBT with art and play therapy methods. We choose to compare intervention implementation in a high-income country (HIC), the US, where there is notable evidence of effectiveness of these methods, and a LMIC, South Africa, where there is less evidence and a higher burden of disease, to assess efficacy. Nearly 20% of children are exposed to IPV in the United States (US) (Hamby et al., 2010) and between 24–58% of children in South Africa (Ward et al., 2018).

Methods

We used a mixed method design to evaluate child mental health, particularly PTSD and depression symptomology, and complemented these data with qualitative responses from children and their mothers. Research was conducted among participants living in domestic violence shelters in New York City, US (HIC) and Johannesburg, South Africa (LMIC).

Intervention components and delivery

A group model was used as participants were in each other’s company much of the time. Traumatic experiences frequently lead to isolation, loneliness and self-blame thus safely joining with others can assist children in recognizing their connectedness. The social component of group treatment (i.e. cohesion, interpersonal learning etc.) is purported to be the central mechanism through which change occurs (Yalom & Leszcz, 2005). Knowledge, awareness and attitudes towards IPV were managed within the group that ran once per week over 12 weeks and lasted 1–2 hours.

The structure of TF-CBT was used in combination with art and play therapy allowing for an intervention alternating between a pre-determined structure and client-led expression. The TF-CBT acronym of PRACTICE guides the work: “Psychoeducation and Parenting skills; Relaxation skills; Affective regulation skills; Cognitive coping skills; Trauma narrative and cognitive processing of the traumatic event(s); In vivo mastery of trauma reminders; Conjoint child-parent sessions; and Enhancing safety and future developmental trajectory” (Cohen & Mannarino, 2008).

Components of the intervention included creating a safe space (the groups were facilitated in the shelter in spaces used for other activities and upon entering them, we performed a ritual to change them into what we needed them for). A range of visual art materials was offered, including markers, crayons, pastels, clay, fingerpaint and acrylic paint. The mediums were managed according to the needs of individual participants, i.e. materials can elicit strong emotion requiring containment thus material selection was thoughtfully considered ahead of each session. Play therapy toys included: a dollhouse, police and other professionals dolls, doctor kits, handcuffs and weapons, puppets and phones. It was essential to have symbolically meaningful materials dependably available in the space, whilst being conscious of those that had the potential to trigger memories of personal trauma in harmful ways.

Psycho-education was conducted through books such as ‘A Terrible Thing Happened’ (Holmes & Pillo, 2000) and ‘Brave Bart’ (Sheppard, 1998). Stories were read and participants created response drawings to these highlighting their most frightening parts. Individual trauma narratives were also drawn, after they were enacted initially through play. Drawings were discussed and processed within the group.

Time was spent learning feelings and identifying their own and others’ feelings (children had limited emotional vocabularies and found it challenging recognizing emotions in fellow group members), as well as recognizing feeling states in their bodies. Many children exposed to IPV carry contradictory and conflicting feelings, i.e. they love and fear primary attachment figures, and have confusing messages about who is safe and unsafe. This reality needed processing. At the end of every session, the group was led in a mindfulness-based practice, including progressive muscle relaxation, guided imagery and yoga (all deliberately focusing on bodily sensations, breathwork, and sensory perceptions whilst preserving an accepting, non-judgmental attitude).

Diagnostic drawings such as ‘Person in the Rain,’ ‘House-Tree-Person’ and self-portraits were used to identify and discuss coping resources and gave insight into how children perceived themselves and others. Drawing a detailed ‘Safe Place’ fostered actively discussing and creating safety plans. Termination included participants writing what they learned and liked best about each other, a gift for each child to take upon leaving the group. They also received their own transitional object (the same bear that was used in the groupwork).

Three groups were facilitated for the mothers during the course of the children’s intervention. These groups focused on helping mothers understand how trauma shows up and impacts child development (reframing ‘bad behaviour’ in helpful ways, leading to improved empathy and less shame about their children), how their own trauma reactions can be triggering to their children, communicating the intervention whilst giving mothers the same tools utilized in the children’s group to assist in parenting, recognizing strengths in their children (emphasizing limit setting and giving positive praise), and helping mothers appreciate the value and need of playing with their children. Mothers identified with characters from books used for psycho-education, just as their children did. Mothers’ own childhood trauma was frequently revealed and required containment and processing in these groups.

The culmination of the intervention was an art exhibition hosted by the children, using drawings of their own choice for their mothers and shelter staff to witness and experience. No participants were forced to display their artwork; they chose their images and controlled their own levels of vulnerability.

Participants and procedures

School aged children and their mothers at two domestic violence shelters were invited to participate. Women’s Survival Space in Brooklyn is a Tier 1 shelter (accommodating the most acute cases in the city). Women were placed through a domestic violence hotline (1–800-HOPE) and length of stay was 3–6 months. The shelter could accommodate 30 mothers and their children. Bethany House in Johannesburg is a non-profit, semi-private shelter partially funded through the state and religious organizations. Women were placed through referral and via Department of Social Development officers. Length of stay was 3–12 months and the shelter could accommodate 10 mothers and their children.

Participants were recruited through the shelter directors and staff. Following written informed consent from mothers and children, they were enrolled in both the research and the group. Interviews with participants were conducted in person before the group intervention commenced and after completion. Mother’s and children each completed measures on child PTSD and child depression. Mother/child pairs that completed the intervention had the opportunity to complete an interview using a semi-structured interview guide, administered in person. Interviews took place at the shelters and lasted approximately an hour. Mothers and children were seen independent of each other but consecutively (children were aware their mothers gave permission for their participation but they had the right to decline). Visual aids were used with the children to assist in appropriate completion of scales. Children aged 6 years and above were eligible for inclusion considering cut-off scores for measures used.

Ethical approval was granted by New York University (NYU) Langone Hospital in New York City and the Health Professions Council of South Africa (HPCSA). Consent forms were translated into Spanish for the US group and into Zulu for the South African group. Participants were reimbursed $15 for their time per interview for the US group and R50 (approximately $5) for the South African group. In addition, all children were given a ‘stress ball’ during the interview they were able to keep afterwards.

Measures

Socio-demographics were recorded by trained researchers using paper questionnaires. Demographics included age of the mother, age and gender of the children, length of stay in the shelter, and what circumstances brought them there.

Post Traumatic Stress Disorder Reaction Index (PTSD-RI) helps screen for exposure to traumatic events as well as PTSD symptomology in school-aged children (Pynoos et al., 1998). Parent and child versions of the instrument were utilized. The instrument assesses the child’s past two-year exposure to traumatic events using a 27-item questionnaire scored with ‘yes’ and ‘no’ responses and comprises a list of 13 child exposures (including IPV). This is followed by a 20-item questionnaire scored on a 5-point likert scale (with responses of ‘none’ to ‘most’ of the time) assessing PTSD symptoms of avoidance, hypervigilence, arousal etc. The consequent score is a count of events that we treated as continuous. We also dichotomized probable PTSD using a cut-off of 38. The PTSD-RI has been utilized in multiple settings (Murray et al., 2015) including in South Africa (Williamson et al., 2017). Inter-item reliability for the current sample was α=.85.

The Children’s Depression Inventory (CDI) is designed to measure self-reported depressive symptoms, such as a child’s loss of interest in activities (Kovacs, 1992). The inventory (short form) contains 10 items each consisting of three statements. The participant is required to select the statement that best describes feelings over the previous two weeks. The short form has been shown to be valid and reliable in previous research (Kovacs, 1992). We treat the inventory as a continuous scale and dichotomize probable depression with a cut-off of 20. The CDI has been used widely, including in South Africa with this cut-off (Cluver et al., 2007). Inter-item reliability for the CDI in the current sample was α=.85.

Comparisons were made between the group in the US and in South Africa, taking into account that most standardized measures have not been normed in Sub-Saharan Africa (Flisher et al, 2007). Although cut-offs have not been validated in South Africa, they are useful in giving some indication of distress levels but should be observed with caution.

Qualitative data was obtained via an interview using a semi-structured interview guide administered in person to both mothers and children at different times. Topics included what they found most useful about the intervention, what was most challenging, what they thought should be omitted, any advice they had for other children in similar circumstances and anything they would want their mothers to know about.

Data analysis

For quantitative data, descriptive statistics for child demographics and mental health assessment data were calculated using STATA 12. Scores at the beginning (pre-test) and at the end (post-test) of the intervention were compared using the Wilcoxon signed ranks test (α = 0.05). This is a statistical test used when outcomes are non-normally distributed, as was true in the case of both CDI and PTSD-RI in our sample.

For qualitative data, thematic codes were developed deductively based on the research questions and inductively adjusted based on stories of the participants themselves. Codes were applied to chunks of text by hand by the lead author (NW), and exemplary quotes were abstracted into an analytical report that was discussed by all authors. Quotes highlighted in this manuscript are representative in nature and use pseudonyms to protect the identity of participants.

We combined the mixed method data at the point of interpretation (Creswell & Clark, 2007). We used an explanatory design to present quantitative and qualitative data in the manuscript. First, we detail basic findings as through the quantitative results. These are then interpreted and deepened using the illustrative voices of participants in the intervention.

Findings

Quantitative findings

Pre-test quantitative data was collected from 37 participants (21 children and 16 mothers) who completed measures before the group commenced. The sample included 15 (71%) children from Brooklyn and 6 children from Johannesburg (Table 1). The median age of participating children was 9.5 years (the range being 6–14 years) and half (52%) of the sample were comprised of females. At baseline, 33% of the participants screened positive for probable depression and 66% screened positive for probable PTSD.

Table 1:

Sample characteristics at baseline (n=21)

	All participants (N=21) Median (IQR) or Number (%)
Sociodemographic Characteristics
Age	9.5 (5 to 14)
Gender
Female	11 (52%)
Male	10 (48%)
Location
Brooklyn	15 (71%)
Johannesburg	6 (29%)
Mental health
Depression (CDI≥20)	7 (33%)
Post-traumatic stress (PTSD-RI≥38)	14 (66%)

IQR: Inter quartile range; CDI: Childhood Depression Inventory; PTSD-RI: Post Traumatic Stress Disorder Reaction Index

As reflected in Table 2, mothers’ assessment of PTSD symptoms in their children was significantly lower than symptoms reported by children themselves (p=0.001). Moreover, only 8 out of the 21 mothers highlighted the traumatic event their children experienced that lead to their symptoms on the scale and four mothers did not identify any trauma event at all. Children reported more physical abuse than mothers (4 vs. 2) and more children witnessed other violence outside the home than their mothers reported (6 vs. 3).

Table 2:

Differences between depression and trauma as reported by children and mothers (n=21)

	Child report	Mother report	p value
Scale	Number (%) or Mean (Std Dev)	Number (%) or Mean (Std Dev)	Fisher’s exact or Wilcoxon
Type of trauma
None	1 (5%)	4 (19%)	0.81
Physical abuse	4 (19%)	2 (10%)	0.35
Death or abandonment by caregiver	2 (10%)	3 (14%)	0.01*
Witnessed domestic violence	9 (43%)	10 (48%)	0.58
Witnessed other violence	6 (29%)	3 (14%)	0.3
Mental health
CDI (n=21)	12.8 (7.6)	CDI doesn’t call for mother’s report
PTSD-RI (n=16)	39.9 (11.8)	21.1 (10.6)	0.001*

CDI: Chilhood Depression Inventory; PTSD-RI: Post Traumatic Stress Disorder Reaction Index; Std Dev: Standard Deviation

^*:statistically significant

Eleven participants completed both pre and post assessments. Scores for probable depression showed a significant change over time from a mean of 13.7 to 8.3 over a 12-week period (p=0.01). There were also changes in mean probable PTSD scores over this time period (from 40 to 34.4) although these did not reach statistical significance (p=0.21).

Figure 1 and 2 demonstrate mental health changes by participants who completed the intervention (n=11) by country. The pooled PTSD score pre-intervention for the South African group (n=5) declined from pre-intervention (M=37.6, SD=9.81) to intervention completion (M=34.8, SD=8.84) although this was not found to be significant t(4) = 0.71, p = 0.51. The pooled PTSD score pre-intervention for the US group (n= 6) declined from pre-intervention (M=42.25, SD=12.63) to intervention completion (M=35.5, SD=24.37) although this was not found to be significant t(3) = 0.86, p = 0.45. Similarly, the depression score for the South African group declined from pre-intervention (M=13.8, SD=5.85) to intervention completion (M=7.8, SD=4.49) and this was found to be significant t(4) = 2.89, p = 0.04. The US depression score declined from pre-intervention (M=12.83, SD=9.64) to intervention completion (M=8.67, SD=8.33) although this was not found to be significant t(5) = 1.77, p = 0.14.

Figure 1.

Children’s PTSD score by country

Figure 2.

Children’s depressive symptoms score by country

Qualitative findings

Expressing and managing overwhelming feelings

From the semi-structured questionnaires completed (n=11), children reported that the intervention was helpful to them with regards to certain feelings, such as fear, anger, and sadness.

It helped me with being scared (Ayanda, 8yr female, Johannesburg)

…and for a participant with particularly high scores on the depression scale:

When we are angry or sad what we can do (Mbali, 13yr female, Johannesburg)

Drawing as a bridge to communicating difficult emotions

Several children explained that the artistic aspects of the intervention were helpful in expressing their emotions and sharing these with fellow group members with similar experiences:

How to get out anger and our sadness, especially in the drawings (Laquasia, 12yr female, Brooklyn)

When we draw and put title to the drawing so we think what it is about and then we tell the group about it, sharing our feelings with Sophie [transitional object/soft toy] (Melissa, 9yr female, Brooklyn)

The intrinsic therapeutic value of relaxation and fun

Others described that the intervention was fun and helped with physical sensations, such as feeling calmer (for one nervous and hypervigilant participant):

How to be nice and how to feel calm in our body (Melissa, 9yr female, Brooklyn)

…and feeling more alert (for participants coping with depression and exhaustion):

Yoga helped me when I was feeling sleepy all the time. You were helping me by doing stuff. Sharing our feelings with Sophie [transitional object/soft toy], doing the drawings and having fun together (Anthony, 7yr male, Brooklyn)

When you mad it helps you to be happy; it helps you not to be sleepy and it is good to have fun (Trinity, 7yrs female, Brooklyn)

Recognising changes in behaviour in self and others

Participants from both settings also recognized a change in behaviour, which affected the ways in which they perceived themselves and their potential. For one particularly impulsive participant:

The group helped me with controlling myself and thinking first (Sipho, 12yr male, Johannesburg)

It change our behaviour to be better behaviour, e.g. Troy [brother] is bad all the time but when you here he won’t be bad (Laquasia, 12yr female, Brooklyn)

Desire for emotional communication with mothers and validation of self

Children were asked if there was anything they wanted their mother to know. Several participants wanted mothers to better understand what they were going through emotionally; wanting their mothers to attune to the hardships they were experiencing to improve their appreciation of each other and positively benefit their relationship. For example, one boy who was having trouble communicating with his mother reported:

All the things we are doing so she understand (Kabelo, 8yr male, Johannesburg)

Another participant wanted to alleviate her mother’s anxiety about her attendance in the group and that a ‘therapeutic group’ was beneficial even though the content was difficult.

That you do activities with us that’s hard and that it is okay (Melissa, 9yr female, Brooklyn)

Others felt like their output in the group would demonstrate their intrinsic goodness:

I want to show my mum that I was writing well (Mpumi, 6yr female, Johannesburg)

I want my mum to know that I’m a good boy (Robert, 7yr male, Brooklyn)

To know the good work I am doing and how much I care about her (Sipho, 12yr male, Johannesburg)

Being seen through art products

During the culminating art event, all participating children chose to have their artwork displayed (note: these images were loaded, often explicit and highly personal). As artists they explained their drawings to their audience and were validated in the process. Mothers appeared present to ‘listen’ to their children, some being visibly moved and seemed to gain improved insight into the emotional lives of their children and the impact violence had on them. For many, the artwork represented a display of emotion that was otherwise challenging to share with their mothers:

I wanted my mum to look at my pictures and everything that was in them (Ayanda, 8yr female, Johannesburg)

Shelter staff also reported noticing a change in the interpersonal relationship of participating mothers and children, highlighting improved tolerance of children’s negative behaviours with more peaceful parenting responses and generally happier interactions between dyads.

Discussion

Through a pilot mixed method study, we learned that a trauma informed, art and play therapy intervention had a positive bearing on children’s mental health in domestic violence shelter settings in the US and South Africa. Though the contexts were significantly different, the outcomes for participants were remarkably comparable, highlighting that when physical and emotional safety is established, mental health needs can be successfully addressed, regardless of context. Qualitative themes of ‘expressing and managing overwhelming feelings, drawing as a bridge to communicating difficult emotions, the intrinsic therapeutic value of relaxation and fun, recognising changes in behaviour in self and others, desire for emotional communication with mothers and validation of self, and being seen through art products’ captured the value and potential of the intervention for participants. The format of the intervention (TF-CBT in conjunction with art and play therapy) offered containment through a phase-based approach while promoting flexibility and fluid expression of participant’s emotional needs. It gave direction to the therapist and participant but simultaneously offered opportunity for self-expression and creative mastery of complex issues.

Similar promising outcomes have been highlighted in previous studies utilizing the creative art therapies in traumatized populations (Stuckey & Nobel, 2010; van Westrhenen et al., 2019). We add to this knowledge with qualitative findings that participants seemed to experience themselves and each other in novel ways, consequently creating different beliefs about themselves, their potential and abilities, and challenging some negative beliefs they held upon entering the group. Many children living in contexts of IPV blame themselves for violence and have poor self-identity (Hall, 2019). The playful facet introduced through these methods may hold the child “emotionally receptive” so that powerful positive messages can sneak through psychological defenses (Pifalo, 2007).

The pilot intervention reveals that mental health symptoms of PTSD and depression decreased, but there were other benefits that deserve mention. Participants reported reductions in behaviour problems. They learned to understand, express and regulate emotions using art and play based activities. Through this they practiced alternative ways to manage difficult situations and resolve internal conflict as well as conflict between themselves. Mothers reported an improved appreciation of the effects of violence on their children (through the groupwork and also the art exhibition) and gained tools to enhance their parenting, seemingly leading to improved relationships with their children. These benefits have been highlighted in qualitative studies to be of most value to IPV exposed children and mothers engaged in interventions (Howarth et al., 2015). Arguably, parental stress decreased as children’s mental health improved, having knock on effects on mother’s mental health simultaneously. There is literature to support a bidirectional relationship between mother’s mental health and children’s; as one improves, so does the other (Greene et al., 2018).

Participants reported having fun in the group. Play therapy at its core is about having fun; learning and problem-solving creatively through this approach. Gaskill and Perry (2014) remind us that if something is not fun, it is not play, and that “it is impossible for a child to have pleasure in a relational interaction if the child’s brain is in an alarm state” (pg.186) (a state typical in traumatized children). The key, therefore, to being true to the ‘play’ in play therapy is helping the child with their own emotional regulation, and, consequently, their own safety. This reality compliments more structured verbal therapies such as TF-CBT. In addition, when children play and have fun, they usually laugh, which is therapeutically beneficial leading to improved mental health, social connection, stress reduction and acceptance (Nasr, 2013), aspects highlighted in our outcomes.

Mindfulness based practice and relaxation techniques were well received and can easily be facilitated with children. The children learned how to soothe their bodies when stressed, gaining control over their symptoms and could practice these techniques outside of the group. This practice is increasingly recognized to be valuable in the treatment of trauma, particularly with children (Pillay & Eagle, 2019) and is easily implemented in LMICs with few resources.

An interesting finding was that all children wanted to exhibit their artwork and were motivated to share their difficult experiences with their mothers. Mother’s appeared at times overwhelmed but otherwise interested in the imagery and seemed to gain insight into the emotional experience of their children. During the exhibition, dyads spoke openly together about the artwork; their gaze shared and directed to the third element of engagement, the artwork itself; helping to gain distance from the experience represented but simultaneously joining in reflection on it. This outcome is congruent with findings that maternal attunement to children’s emotion (in this case the emotion latent in the image and being expressed) is linked to better outcomes for children exposed to IPV (Fogarty et al., 2019). When a mother is aware of her child’s difficult feelings, it helps moderate the relationship between her own mental health symptoms and children’s internalizing and externalizing difficulties (Cohodes et al., 2017). Interventions that highlight a mother’s awareness of their children’s emotional experience and help contain that emotion may afford a cushion against IPV.

Qualitative research exploring children’s needs after exposure to IPV, reveals children’s desire for safety, true listening by a caregiver, and a need for deeper connection to their non-abusing parent (Buckley et al., 2007). It could be argued that in sharing their artwork and engaging their mothers, children increased their sense of being heard and the relationship with their mothers was strengthened. Through this honesty ‘shame’ might have been transformed into other skills, such as competence, communication, and empathy (Buschel & Madsen, 2006).

There was a positive effect of working in a group; activities facilitated self-expression and collectively working through similar traumatic experiences helped participants recognize they were not alone (isolation being a particularly problematic issue for those living in situations of IPV), alleviating some of the stigma of the violence experienced, and potentially reducing avoidance symptoms. The children used the group to manage some of their shared feelings about living in a shelter and what circumstances brought them there. They learned to work cooperatively but were also witness to each other’s stories, supporting one another and learning coping skills from each other while practicing these findings together. Group work is found to be particularly powerful with traumatized children (Ugurlu et al., 2016). The value of social support, even perceived social support cannot be underestimated in the work with traumatized individuals, particularly in contexts such as South Africa with high rates of disease burden and few clinical resources (Cluver et al., 2009). A powerful protective influence for children and adolescents exposed to IPV are supportive peer relationships (Howell, 2011). Even though there were challenges in facilitating a group with a variety of developmental ages represented, this method was effective and necessary and was akin to a family therapy group.

There were no perceptible differences between the two settings in terms of overall mental health improvements. Notably there were fewer services and housing options available to the South African families compared to the US ones. Mothers in South Africa generally experienced greater deprivation and structural inequalities on account of being from a LMIC (such as greater food insecurity, greater poverty, poorer service delivery, poorer criminal justice system responses to violence etc.). While mothers were in the shelters (in both countries), they and their children were safe and basic needs were being met. Arguably, this created an environment where a mental health intervention speaking to both the children and the mother’s needs was effectively enabled.

Implications for practice

When working with mothers and children with histories of high levels of exposure to violence, particularly within a shelter setting, using non-verbal methods offers another level of confidentiality or respect of their stories and the secrecy laden with the problems they have experienced. Drawing or playing out ‘secrets’ may be less terrifying and more safe than ‘telling’ them in words which is typical in most therapies.

Helping mothers through offering a child development lens to understanding children’s difficult emotions and problematic behaviours is vital. In addition, providing practical resources for parenting children in flux is helpful. There is a dearth of evidence on how parenting can reduce child maltreatment in the context of IPV in LMIC settings (Coore Desai et al., 2017). This is the first study, to our knowledge, to explore a TF-CBT approach with mother-child dyads in South Africa.

Parental play is a forecaster of the quality and creativity of children’s play (Valentino et al., 2006) and play can become inhibited and less creative in families that experience IPV (Fellin et al., 2019). Enhancing parental play as a way to improve the quality and creativity of children’s play and aiding mothers in gaining insight into children’s post-traumatic emotional state (especially their post-traumatic play (Gil, 2016)) is an untapped way for mothers and children to strengthen each other in recovery.

In quantitative findings, mother’s predominantly underreported levels of severity of both children’s violence exposure and their mental health. Similar findings are reported in other studies where mothers may be unaware of the affects of violence on their children and unable to talk about it (Buschel & Madsen, 2006). Resultantly, it is important for children to express their most significant fears to their mothers and for mothers to engage with this.

Limitations

The findings of this study are limited in several ways. The study was not experimental, making it challenging to assess whether the intervention had an impact on mental health symptoms or if other factors related to opting into the research might have influenced outcomes. There was a small sample size thus the quantitative findings should be interpreted cautiously. As this was a pilot study, future research should replicate this in larger sample sizes. The use of children’s voices in qualitative research is a novel contribution of this study, but it had important limitations in terms of gaining longer, more detailed quotes from participants. Future research can complement verbal interviews with non-verbal methods of interviewing young children to enrich their somewhat limited vocabulary. Again, the qualitative sample size was small though we did begin to reach saturation on themes highlighted in this paper. With such high rates of turnover in the shelter system, it was difficult to retain participants for the full 12-week intervention, calling for more research on shorter interventions and their impact on mental health. The study used self-report instruments vulnerable to reporting bias and there might have been underreporting on account of social desirability bias or stigma.

Conclusion

The lack of access to mental health services of sound quality is weighty in populations with restricted resources, for whom copious social risks worsen vulnerability to poor health. Children and mothers in domestic violence shelter systems require mental health interventions that improve both the individual and dyads functioning and can help mitigate the negative trajectory of potential intergenerational transmission that is likely in the face of no treatment. Non-verbal therapies push boundaries by transcending language and its defenses, intervening in accessible and safe ways. Children easily engage in non-verbal treatment, thereby improving their agency and participation in the process, and consequently increase motivation towards their symptom reduction. Engaging mothers in the art and play of their children leads to increased understanding and empathy with resultant positive gains in the relationship. Our pilot findings from combining art therapy, play therapy and CBT hold promise as a method for trauma-focused treatment for children and mothers exposed to IPV.

Table 3:

Change in self-reported depression and trauma among children completing a trauma-focused cognitive behavioural programme

	Pre-TFCBT	Post-TFCBT	Wilcoxon
Scale	Mean (Std Dev)	Mean (Std Dev)	p value
CDI (n=11)	13.7 (8.1)	8.3 (6.6)	0.01
PTSD-RI (n=9)	40.0 (10.6)	34.4 (15.6)	0.21

CDI: Childhood Depression Inventory; PTSD-RI: Post Traumatic Stress Disorder Reaction Index; TFCBT: Trauma-Focused Cognitive Behavioral Therapy; Std Dev: Standard Deviation

Highlights.

• A pilot group based intervention combining trauma focused cognitive behaviour therapy (TF-CBT) with art and play therapy had positive results on children’s mental health (PTSD and depression) in domestic violence shelters in the US and South Africa.

• Integrating non-verbal therapies into interventions with children is crucial as they push boundaries by transcending language and its defenses, intervening in accessible and safe ways.

• Engaging with children as well as their mothers improved parent-child relationships and parenting in this intervention.