When a parent dies – a systematic review of the effects of support programs for parentally bereaved children and their caregivers

Our analysis of the included studies were grouped in a table based on the following categories: type of intervention, reference (comparison), grade of evidence, study population, evaluation design, measure, outcome variable and findings as effect size within (at baseline and follow-up) and between study comparison groups.

For any ordinal or continuous variables, to be able to calculate effect size even when a means and standard deviation were not reported in studies, the standardized mean difference effect size for within-subjects design was used, which is referred to as Cohen’s d_z. The effect size estimate Cohen’s d_z. can be calculated directly from the t-value using the formula \( {d}_z=t/\sqrt{n} \). A commonly used interpretation of Cohen’s d is that value of 0.2 can be considered a small effect, 0.5 a medium effect and 0.8 a large effect [21].

The Common Language effect size (CL) [22] is also reported. The CL is also known as the probability of superiority [21], represents the probability in percent that a randomly selected person will score a different observed measurement post- than pre intervention, after controlling for individual differences. In addition when possible, the effect size of difference between groups was calculated (dm) using a method proposed by Morris in which effect size is calculated on the mean pre-post change in the treatment group minus the mean pre-post change in the control group, divided by the pooled pre-test standard deviation [23]. For categorical data, Chi-squared tests were made. Phi is reported as the effect size proposed by Fritz and colleagues using the formula \( \varphi =\sqrt{\frac{\chi 2}{N^{\prime }}} \) [24]. A value of 0.1 is considered a small effect, 0.3 a medium effect and 0.5 a large effect.

The included 17 studies were published between 1985 and 2015, the majority, 13 were published after 1999. Most studies were conducted in the United States [26, 27, 29‐39, 41]; two in England [25, 40], and another was an international collaborative study involving Iran, UK and Norway [28].

The studies differed; they were based on different study designs, contained a variety of outcome measures and varied in quality. According to our quality grading criteria (Table 1) [18‐20] 13 studies provided strong evidence. These studies were randomized controlled trials involving validated measures. Three studies provided fairly strong evidence and one study provided weaker evidence [18‐20]. Two of the included bereavement interventions were evaluated with a population of more than 100 children. Namely, “The Parent Guidance Program” [26] and “The Family Bereavement Program” [27, 29, 30, 33‐35, 37, 39, 41]. One of the interventions, Family Therapy sessions, was tested in two papers [25, 40] and one, The Family Bereavement Program, in as many as ten papers [27, 29, 30, 33‐35, 37‐39, 41].

One study employed a quasi-experimental design [31] and one study had a pre-test/post-test design [36], the others were randomized controlled trials. What the intervention was compared with varied: no intervention [25, 28, 40]; delayed treatment [31, 32]; a telephone support intervention [26]; and a self-study program [27, 29, 30, 33‐35, 37‐39, 41].

The core concepts addressed in the outcome measures were:

Fifty different outcome measures were employed. We present the most commonly reported outcomes in the included studies which focus on children’s health, behaviour, grief, self-esteem, parenting factors and caregivers’ mental health [42‐54] (see Table 2 below).

A key research question for this review is: What types of support interventions were evaluated in the studies? We found studies varied in their theoretical under-pinning and aim. They also took various forms: group interventions for the children [28, 36], family interventions [25, 27, 29, 30, 32‐35, 37‐41], parental guidance [26], and camp activities for children [31].

Some interventions were designed based on resilience, risk and protective factors for parentally bereaved children [27, 29, 30, 32‐35, 37‐39, 41]. Others were based on theory of trauma and/or the grieving process [28, 31]; psycho-education [26]; psychodynamic theory [36]; and attachment theory [25, 40]. To a large extent, the interventions were directed towards children at an early stage in their grief process. “The Family Bereavement Program” and “The Parent Guidance Program” were explicitly intended to be preventive interventions [26, 33]. However, the intervention “Writing for recovery” was directed at refugee children with high symptoms of traumatic grief [28]. For some of the refugee children, many years had passed since their parents died.

In three of the studies, the interventions were primarily directed at the bereaved child in the form of support groups and/or camp activities [28, 31, 36]. The intentions in these studies were: to provide emotional support; to normalize the children’s experiences after the loss; to provide a safe environment where the child can express emotions and thoughts; to facilitate the child’s grieving process and to aim to improve the child’s physical and mental health. For further description of the interventions, see Table 3.

In the majority of the included studies, the interventions were directed at both the child and their remaining caregiver [25‐27, 29, 30, 32‐35, 37‐41]. The intentions in the included studies were: to provide support for the children and their caregivers; to improve family communication and the caregiver–child relationship; to facilitate participants’ grieving process; to improve their health; strengthen parenting; increase stability and predictability for the children; and to reduce the occurrence of negative events among the children (see Table 3).

In general, the interventions were brief. The shortest program was “Writing for recovery”, involving two 15-min sessions in school during three consecutive days, each day consisted of two sessions (i.e. six 15-min sessions and a total of 90 min) [28]. The camp-based program “CampMAGIC” was delivered over a weekend [31, 55]. The longest, “The Parent Guidance Program” lasted a year, it began when the parent was ill, and continued during the terminal illness and at least 6 months after the parent’s death [26]. It involved at least six sessions during the terminal illness and six after the parent had died. The other interventions were based on a total of 6–14 sessions (see Table 3 for more details).

All interventions were professionally led, in most cases by social workers or counsellors with extensive experience of working with child guidance, grief or psychiatry. The highest educational attainment of professionals were those who led “The Family Bereavement Program”, who had at least a master’s degree [34]. In several studies the intervention leaders received supervision in the implementation of the support program [26, 32, 33, 36].

The included interventions in this review were directed at children from school age up to 18 years of age. This is with the exception of two studies where younger children (0–16) were involved in family therapy sessions [25, 40]. Most of the studies concerned children who had experienced a parental death from a range of causes, namely illness, accident, suicide or homicide [25, 27, 29, 30, 32‐41]. Commonly parents died because of an illness (65–82%), thereafter due to an accident (15–20%) or suicide/homicide (10–14%). In most studies there was a lack of information about what kind of illness the parent suffered from, where there was information, diseases included those of the heart and cancer [25, 32, 40]. One study compared intervention effects for children who had lost a parent to expected versus unexpected deaths [31]. One study focused on children during their parent’s terminal cancer illness as well as after the parent’s death [26]. Finally one study focused on support directed at refugee adolescents who had lost their parents in war [28]. Except for this evaluation directed at refugee children from Afghanistan, the majority of included studies had samples that were diverse in ethnicity, including for example Caucasian, Hispanic, African American, Native American, Asian/Pacific and other ethnicities [33].

In the studies, the most common deceased parent was the child’s father with the remaining caregiver being the mother. In two of the studies, women as remaining caregivers were over-represented as participants in the study populations [32, 36]. In one study 86% of the deceased parents were fathers and 14% mothers [32]. In another study, fathers as remaining caregivers only represented 5 % of the sample [36].

Another key research question for this review was: What is known about the effects of support interventions that are targeted at/or include support for parentally bereaved children? The included studies were analysed and summarized in a matrix. The results are presented in table form (see Table 4 below). There were 12 studies that analysed effects within and between trial arms, while five studies analysed moderating and mediating factors. The latter are excluded from the analysis of effects in Table 4, but are nevertheless informative and are therefore included in the article. Our focus is on comparing differences between groups, but we have also chosen to present results within groups in Table 4, as this may be relevant from a benchmarking perspective, both for researchers and clinicians [56]. The results from the analyses of included studies revealed positive effects of the support interventions both for the children and their remaining caregivers in several areas.

The included studies in this review were published within several disciplines, namely psychology, social work, medicine, psychiatry, lending weight to the argument that the subject of support for parentally bereaved children is relevant for a range of different professional groups.

One conclusion from this review of interventions is that there were studies that have shown effects for children and their caregivers. The results indicate that supportive interventions can be directed exclusively to the children or to both the bereaved child and the child’s remaining parent or caregiver. Support for the children’s caregivers can strengthen their own health and their capacity to support their children. A supportive parenting is a protective resource for parentally bereaved children [60]. Previous research indicates that when the bereaved children’s caregivers are supported, they demonstrate an enhanced capacity to support their children [60‐62].

At the same time, support also needs to be directed at the children. In the evaluation of a parental guidance program, the remaining parents expressed that they perceived a need for more support directed to their children [26]. In one of the included studies, both children and parents indicated that they wished to discuss grief-related experiences with other people who had similar experiences [32]. Being and connecting with other bereaved children can be helpful for children who attend a support group, as it can help them to feel less isolated and alone [55, 63, 64]. Simultaneous family sessions involving both children and the remaining parent may be an important component in a support program as such sessions are sometimes the first occasion that the parent and children have had the opportunity to sit down together and talk about the loss and their feelings about it [25]. Some children avoid talking about their problems or showing their feelings as they try to protect their remaining parent or other people around them. This can sometimes be misinterpreted as a sign that the child is not affected by the loss [65]. The included effect-evaluated interventions were not sufficient for all children. The majority of intervention programs were brief. Studies indicated that some children may need more intensive support or additional support [31, 35, 36, 38, 41]. Therefore, it is important to reassess children’s further needs for support at the end of an intervention [36].

Given that there are currently relatively few scientifically rigorous studies in this area, there is a clear need for further research about the effects of support interventions directed at parentally bereaved children. Indeed, there were only 17 studies that met the criteria for this review. All studies, with the exception of one [28], were comprised of studies about English language interventions that were evaluated in the USA or UK (see Table 4). It is evident that there is a need for more effect studies with longer follow-up, with the Family Bereavement Program being a notable exception, as children’s problems can appear later and it may also take time before changes in the participant families stabilize post intervention and have an effect for the children [33]. Furthermore, there is a need for studies with populations sufficiently large enough to make comparisons of the effects for various categories, so that the interventions can be modified to various children’s needs. Some studies for example, showed differences in the efficacy of interventions for children at different ages [35, 41], for girls and boys [26, 33, 35, 37], for mothers and fathers [26] and for children with different levels of problems at baseline [35, 41]. In the majority of included studies the sample were diverse in ethnicity, but did not analyse effects for different ethnic minority groups. The sample sizes of minority groups were too small to allow the testing of program effects for various groups [34]. In the studies, the most common deceased parent was the child’s father with the remaining caregiver the mother. This is consistent with mortality statistic rates as children under the age of 18 are more likely to experience the death of a father than the death of a mother [1].

This systematic review highlights that interventions evaluated with a focus on effects for children have almost exclusively been directed at school age children, while the bereavement research shows increased risks for the youngest children when one or both parents dies [4]. The younger children are especially vulnerable as they are totally dependent on their caregivers. In addition, they often find it more difficult to comprehend what has happened to their deceased parent and what this means [66]. Consequently, development of supportive interventions and evaluation of bereavement interventions for younger children is an important issue for further research. Involving younger children in evaluations of interventions may require innovative methods, where the children are given the opportunity to express themselves in a way that is adapted to their capacity and cognitive development. Such evaluations may also include qualitative interviews where the children can express themselves in their own words or through creative methods such as art or play [63, 67]. Further, children need to be enabled to participate in the research to develop knowledge about their experiences, to explore with children what they themselves perceive as helpful in the grieving process and what kinds of outcome measures are most important from their perspective. For example, few of the outcome measures in the included studies concerned children’s physical health and somatic symptoms, their situation in school and their peer relationships. It is also important that children have the opportunity to be involved in evaluations of support programs as parental reports have a tendency to underestimate children’s problems and report less symptomatology in their children than do the children themselves [68]. Qualitative data from evaluations could also be helpful to identify opportunities to improve current bereavement interventions.

Finally, studies of bereavement interventions for children are more generally focused on children that are living in a nuclear family, where one parent dies and the other parent is the child’s remaining caregiver. However, there are also children who have lived with a single parent who dies, and there are children who lose both their parents through death. These children have to change caregivers and residence. The death of a parent engenders secondary losses that occur as a result of the primary loss. When the child’s only parent or both parents die, the secondary losses are increased, in number and complexity [69]. Therefore, special attention is merited towards these groups of children. One explanation why these children are underrepresented in evaluation studies is that the largest proportion of children in the western world live together with both their parents. It is difficult to conduct evaluation studies with this vulnerable group of children.