Implementation of resiliency interventions for caregivers of pediatric cancer patients: A systematic review

Eligibility criteria

Studies were selected based on the criteria outlined in Table 1. Articles were included if they presented original data published between January 1, 2010, to February 8, 2020, were available in English language, tested the effects of an intervention targeting parents or caregivers of pediatric cancer patients and included a primary or secondary outcome of resiliency. Related outcomes, including anxiety, quality of life, burnout, psychological distress, and problem-solving, were also reported but did not determine eligibility for inclusion.

OPEN IN VIEWER

Table 1.

Population, intervention, comparator, exposure, and study design framework for eligible studies.

PICOS strategy	Inclusion criteria	Exclusion criteria
P–Population	Interventions targeting parent or caregiver of a pediatric cancer patient.	N/A
I–Intervention	Studies must have tested the effects of an intervention with or without a control group.	N/A
C–Comparison	Among studies with a comparison group, all comparison groups were accepted.	N/A
O–Outcome	Studies must have included a primary or secondary outcome/ related to resiliency or post-traumatic growth, including anxiety, quality of life, burnout, psychological distress, and resilience, in the parent or caregiver of pediatric cancer patients	N/A
S – Study design	Randomized and non-randomized controlled trials, quasi-experimental (pre/post studies)	Reviews, observational studies, case reports, and case series

To capture a wide range of interventions, exclusion criteria were kept to a minimum. No restrictions were established based on country of origin.

Search strategy

A literature search was conducted by one study team member (RL) on February 9th, 2021, in PubMed using the following MeSH (Medical Subject Headings) terms: resiliency, post-traumatic growth, resilience, stress, caregivers, family, spouse, pediatric, infant, child, teen, youth, young adult, neoplasia or neoplasm, tumor or tumor, cancer, and malignancy or malignant. The search string was developed based on discussions with study team members with expertise in interventions for cancer patients and families, mental health and wellbeing, clinical health psychology, psychosocial oncology, and cancer survivorship. The PubMed search strategy used for this review is provided in Table 1. The original search was not supplemented with reports after this date to avoid potential confounding due to the COVID-19 pandemic, which likely influenced participation, type, and frequency of delivery of resiliency programs for this target population due to the necessary adaptation of programs during this period.

Data management

Records were managed and de-duplicated using EndNote version X9 (The EndNote Team, 2013) and subsequently uploaded to Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia) for title/abstract screening and full text review.

Study selection

Title/abstract screening and full text review were independently completed by two members of the research team (CM and RL); disagreements were resolved by discussion with other study team members until consensus was met.

Data extraction and synthesis

A data extraction spreadsheet was developed and piloted on one included study by the study team and iteratively refined through review and discussion. The inquiry for the abstraction was guided by the Consolidated Framework for Implementation Research (CFIR; Damschroder et al., 2022). CFIR is an implementation science framework widely used to both assess and help explain factors that affect the implementation of evidence-based factors. By using CFIR to guide the abstraction, barriers and facilitators of the scale-up and spread of effective approaches to resiliency-building programs can be identified. CFIR posits that factors related to the intervention itself (e.g. complexity of intervention components and active ingredients, qualifications of staff necessary to deliver, mode of delivery, protocol requirements), the inner setting (e.g. target population needs and characteristics), and external supports such as training and technical assistance can affect implementation and dissemination. Extracted data for these elements included the year of publication, study design, target population and setting, sample size, underlying theoretical components or conceptual model, description of the intervention and control conditions, characteristics of the intervention (interventionist, mode/frequency of delivery), protocols, training, or other supports for intervention fidelity, main study findings, reasons for attrition, and other considerations reported by authors relevant to implementation. Theoretical underpinnings and frameworks were also abstracted, if noted by the authors. Included studies were dual coded by three study team members (SH, CM, BR). After coding, the extracted data was reviewed and tabulated by KM.

Due to the clinical and methodological heterogeneity in the type of interventions and outcomes assessed across the included studies, a meta-analysis was not possible. Data were tabulated and qualitatively synthesized within the text.

Study characteristics

Of the 11 included studies, publication dates ranged from 2012 to 2020, with seven (64%) of the studies published in 2017 or later (Supplemental File Table 2). The majority of studies were conducted in the U.S., while others took place in Canada, Australia, Iran, and Belgium; one study was a collaboration between the U.S. and Canada. Nearly half (45%) of the studies were pilot trials—either randomized controlled trials (RCTs) or pre-post designs—while the remainder included five RCTs and one non-randomized controlled trial. Among the eight two-arm trials, control groups typically received usual care (counseling, education and resource referrals), with one study not reporting details about the control condition. Study sample sizes ranged from 13 to 342 participants, with pilot studies tending to have fewer participants. Detailed study characteristics, including design, location, and control conditions, are summarized in the Supplemental File (Table 2).

Setting and participant characteristics

Participants were primarily recruited from hospitals, transplant centers, or cancer centers, with one study not reporting the recruitment setting. Most studies enrolled only parents or caregivers of pediatric cancer patients. However, one study included patient-parent dyads, and another involved both patients and their parents, with a modified intervention for the latter. Pediatric patients ranged in age from infancy to 25 years. Nearly all studies focused on families of children undergoing active treatment. Several studies limited participation to English-speaking families. Full details on recruitment settings, participant groups, and eligibility criteria are presented in the Supplemental File (Table 2).

Intervention characteristics and strategies

Intervention strategies across the 11 included studies primarily operated at the individual or interpersonal level; however, two interventions incorporated provider-level strategies through electronic summaries of family psychosocial risk information shared with the child’s care team. The majority of interventions targeted improving knowledge and skills among parents and caregivers. Several studies featured unique components, such as spiritual guidance, personalized coaching to promote physical activity, peer mentoring by veteran parents, and weekly visits from therapy dogs. Delivery modalities varied and included fully in-person formats, mixed in-person and phone-based sessions, and, in one case, a web-based platform that facilitated peer interaction and activity tracking.

Three interventions tested the effects of a group-based interventions, focusing on feasibility or resiliency-related outcomes. In contrast, one 12-week, web-based intervention invited parents to join a private group on a FitBit™ Website to share data with researchers and other parents (Halliday et al., 2017). Interventionists varied broadly across studies and included members of the research or treatment team (e.g. oncologists, nurses, psychologists, and social workers), religious advisors, trained peer parents, and dog handlers. Detailed information on interventionists and delivery models is provided in the Supplemental File (Table 3).

Underlying theoretical components or conceptual models

Among the 11 studies, four (36%) did not clearly specify a guiding theory or conceptual framework. The remaining studies drew on various theoretical underpinnings, including resilience and stress-and-coping theories, social support theory, and principles from positive psychology. Two studies were informed by the Pediatric Psychosocial Preventive Health Model—a biopsychosocial framework designed to align psychosocial support with family risk and need.

Study outcomes and data collection tools

Outcomes across studies primarily focused on changes in psychosocial outcomes in parents or caregivers (e.g., anxiety, quality of life, burnout, mood, well-being, psychological distress, burden, perceived stress, uncertainty related to child’s illness, resilience), assessed pre- and post-intervention with self-reported questionnaires. Only a few studies reported the internal reliability of these measures. Outcomes related to feasibility and acceptability were usually among the primary outcomes for pilot studies. Detailed information on the study outcomes is provided in the Supplemental File (Table 2).

Results of trials

Given differences in outcomes in pilot versus non-pilot studies (Tickle-Degnen, 2013), we have summarized trial results separately for these studies.

Pilot studies

Among the four pilot studies that assessed intervention feasibility, two demonstrated feasibility based on recruitment, retention, and participation metrics. Evidence for low feasibility in the remaining two pilot interventions included: (1) participants not wearing their FitBit™ device for the duration of the study and not meeting their weekly step goal of 10,000 steps per day; and (2) lower-than-expected completion rates for intervention sessions Despite these challenges, all pilot studies reported high acceptability among participants.

In addition to pilot outcomes, all five pilot interventions assessed preliminary effectiveness of the intervention on parentor caregiver psychosocial outcomes. One pilot RCT found significant reductions in psychosocial difficulties at 6 months following the provision of family psychosocial information to the child’s care team. Another study testing an adapted skill-based intervention reported improvements in resilience and distress among parents who completed the program. A third pilot, targeting physical activity, showed positive effects on mood and psychological distress despite low feasibility. Additional pilot interventions demonstrated gains in emotion regulation, well-being, and a potential buffering effect against worsening symptoms over time. Detailed findings for each pilot study are summarized in the Supplemental File (Table 2).

Non-pilot studies

Among the six non-pilot studies, five tested the effects of interventions designed to provide therapy, social/emotional support, or to enhance parental knowledge and coping skills related to their child’s cancer diagnosis. Several demonstrated positive psychosocial outcomes for parents or caregivers, though effects varied in strength and specificity.

One study found that parents participating in a group-based educational and spiritual intervention reported significantly lower burnout post-intervention and at 1-month follow-up. Another combined educational sessions with massage and relaxation therapy but found no added benefit from the addition of a parent intervention to the child-only intervention, with improvements observed across all groups. In a trial comparing one-on-one versus group delivery of a skills-based intervention, only the individual format produced significant gains in resilience and benefit-finding compared to control.

Other interventions showed mixed or limited quantitative effects. A therapy dog intervention showed no significant psychosocial improvements compared to usual care, though reductions in parenting stress were noted over time. The study design, however, allowed control group participants access to therapy dogs, which may have resulted in potential treatment contamination. Similarly, a peer-mentor program involving veteran parents yielded no significant change in resilience scores. Yet, qualitative feedback indicated parents felt supported and empowered, particularly among Hispanic participants, who emphasized the value of knowledge and empowerment.

A notable pattern observed across studies, including those with control arms, was an overall improvement in parent or caregiver psychosocial functioning over time, regardless of intervention assignment. Full results from each non-pilot study are summarized in the Supplemental File (Table 2).

Common reasons for attrition

The following reasons for attrition were reported by included studies: 1) difficulty reaching participants (Barrera et al., 2014); passive refusal (Lindwall et al., 2014); busy or scheduling conflicts (Barrera et al., 2014; McCullough et al., 2018; Mullins et al., 2012; Rosenberg et al., 2019); issues with learning how to use technology included in intervention (Halliday et al., 2017); declining after being assigned to control (McCullough et al., 2018); and reasons related to illness or treatment or child death (Grégoire et al., 2019; Mullins et al., 2012; Rosenberg et al., 2019). Parents who were unmarried and had lower income at baseline were also found to be more likely to drop out (Rosenberg et al., 2019).

Implementation considerations

Given the in-clinic setting of many interventions reviewed, we also extracted information reported by the authors that could inform future efforts to implement interventions targeting resiliency and related outcomes in parents or caregivers of pediatric cancer patients. This included information related to the roles and responsibilities of researchers and treatment team members, the training necessary to deliver the intervention, challenges, and considerations for conducting interventions in different settings with varying availability and types of resources, and intervention fidelity.

Roles and responsibilities of study team

Overall, studies provided varying levels of detail related to who was responsible for performing research activities. For example, studies often did not provide information on who was responsible for identifying eligible individuals, consenting interested individuals, and collecting outcome information. In some cases, information related to who conducted study activities was mentioned for certain activities (e.g. recruitment), but not all (Barrera et al., 2014, 2020; Mullins et al., 2012).

Training of study team

Only a small number of studies provided a description of study team member training (Baron Nelson et al., 2018; Rosenberg et al., 2019; Yi-Frazier et al., 2017). In the two studies testing the effect of the PRISMA-P intervention, authors included information/citation related to the program manual (Yi-Frazier et al., 2017) and one additionally included the program protocol as a supplement (Rosenberg et al., 2019). Other studies stated briefly that interventionists were trained but did not provide a further description (Grégoire et al., 2019; Halliday et al., 2017; Mullins et al., 2012).

Challenges reported

Examples of challenges reported by included studies included: (1) the potential for differences in resources across sites to impact the timing of assessments/data collection (McCullough et al., 2018) or level of care provided to control (usual care) participants (Barrera et al., 2014); (2) generalizability of findings and ability to implement interventions in settings that do not have ample resources (Lindwall et al., 2014); (3) re-scheduling by participants to impact length of intervention period (Mullins et al., 2012); (4) managing busy participant schedules which can influence their ability to find the time to complete intervention sessions (Yi-Frazier et al., 2017); (5) the potential for treatment to influence the ability to detect intervention effects (Barrera et al., 2020); (6) managing risk of treatment contamination given it may not be ethical to withhold certain resources to control participants (McCullough et al., 2018); (7) balancing efforts to increase reach with age-specific criteria of psychological outcome data collection instruments (McCullough et al., 2018); (8) weighing the benefits and potential drawbacks of group-based intervention activities (Rosenberg et al., 2019); and (9) balancing flexibility to increase engagement and study rigor and fidelity (Yi-Frazier et al., 2017).

Findings of review and comparison to published reviews of pediatric cancer caregivers

Being a caregiver for a child with cancer presents unique challenges including a navigating a variety of treatment, understanding complex medical information and providing emotional support to the child and siblings while handling multiple life roles. Further, caregivers often face severe financial strain and social isolation (Lewandowska, 2022). Evidence from the 11 studies reviewed suggests that a variety of intervention strategies may be feasible, acceptable, and effective when delivered to this population. The majority of interventions in the field focused on improving knowledge and skills to cultivate resiliency, manage or cope with emotions, and ultimately decrease feelings of distress and emotional burden related to the stressor (in this case, the child’s illness). Interventions in our review, similar to others, were primarily delivered to individuals/families as opposed to group settings and were heavily tailored to the emotional and informational support needs of the participant. Interventional components most often included individual counseling sessions and sometimes other components such as physical activity or pet or supportive therapies. The most common need addressed was related to psychological well-being, and overall, the findings suggest modest but positive effects in a variety of outcomes, including distress, mood (e.g. depression, anxiety), resiliency, and coping skills.

Intervention participants were caregivers, and studies overwhelmingly reported the effect on mothers; a few studies included families (i.e. siblings). Of the few that included fathers, the impact of the father was usually not reported separately, despite some evidence of the unique role of the father’s distress in the child’s experience which may be due to an enhanced sense of responsibility for family members’ well-being and finances (Pai et al., 2006). Our finding that most studies included non-Hispanic White, English-speaking mothers’ parallels that of a previously published review of the literature (Shudy et al., 2006). Similar to other studies, differences in effect by age and gender, and the need for research of other caregivers besides mothers were noted. These are especially important considerations to inform the translation of findings to the general population to understand the need for different intervention strategies in different groups facing a variety of stressors.

Mechanistic pathways of interventions

A unique aspect of our review is the examination of the theoretical basis for the mechanism of action of resiliency interventions reported in the literature. Our review found that the mechanistic pathway of four of the eleven studies included was not defined. The other seven studies described causal models of distress and resiliency and identified improvements in distress, anxiety, mood, quality of life, and psychological well-being as primary outcomes. However, these studies did not directly equate interventional components to specific constructs or domains within these models, suggesting that the mechanistic pathways for resiliency interventions are not well-defined in the literature. This observation is further supported by the high degree of variability of outcome measures deployed in these studies, which could further hinder the development of the evidence base for resiliency interventions. Model-mechanism-outcome specificity and correspondence could potentially increase the ability of researchers, and eventually clinicians, to choose specific interventions for specific problems and populations. For example, the report by Sahler et al. (2013) links specific intervention components such as problem-solving and information to well-defined outcomes, and proposes a model to explain the effects, serving as a potential guide for future research.

Notably, nearly half of the studies summarized in this review focused on resiliency outcomes were pilot studies, indicating that intervention research focused on understanding factors influencing the design and implementation of interventions to build resiliency in caregivers of pediatric patients may still be in nascent stages. Similarly, few studies used randomized or well-designed comparison groups in their evaluations. These conclusions about the need for more robust studies with defined mechanistic pathways for interventions and inclusion of diverse populations to improve the evidence base were supported by other published meta-analyses and reviews (Bonanno et al., 2015; Luo et al., 2021b; Park et al., 2022; Rosenberg et al., 2013a; Shudy et al., 2006).

Knowledge translation and implementation considerations for resiliency interventions in the general population

This review offers some important lessons from pediatric cancer caregivers that can inform the design, study, and delivery of future resiliency interventions to the general population. First, these results suggest that coping, along with several other domains of psychosocial well-being, can be positively impacted through intervention in highly distressed populations. This observation was consistent despite the lack of clarity in interventional components and mechanisms of action, suggesting that there are likely intervention activities that positively impact multiple overlapping and inter-related aspects of well-being. Thus, what is currently known about resiliency suggests that the continued delivery of these interventions, along with the development of research and program evaluations to enhance our understanding of how they work and why, is warranted.

Second, although few of the interventions included in this review involved group-level delivery, there were either no findings of significant benefit with groups or a clear preference for individual versus group interventions among participants. This suggests that future interventions should include a tailored, individualized component as a de minimis component and thoroughly evaluate the need for and degree of tailoring going forward. This examination should include these needs change over time and with varying stressors. Given some studies reported improvements over time regardless of treatment group, considerations regarding intervention timing and the inclusion of a usual care comparator group are warranted to assess effects beyond those that may occur temporally as treatment demands lessen and some individuals adjust to the crisis and learn how to cope.

Third, implementers should recognize that the literature describes a high degree of resources including time and staff training to deliver these tailored interventions. Additionally, many studies reported extensive control of delivery through either program-specific training or the use of a select pool of clinical staff. In the one study that deployed non-clinical staff, positive results in resiliency and distress were observed, suggesting that clinical staff may not necessarily be required. Notably this study did report the use of a program manual and supervision of staff by a licensed psychologist to support implementation.

Fourth, regarding the mode of delivery, a variety of routes, including telephone and in-person, were used, with no indication of a difference in effect or participation. Hocking and colleagues (Hocking et al., 2014) conducted a review of family-based interventions and found a variety of intervention delivery modes were acceptable to participants including internet chat rooms, interactive websites, and webcams, though viewed as supplemental to in-person components by participants.

Fifth, intervention strategies that are low burden to caregivers who are already overwhelmed demonstrate effectiveness and may address commonly reported barriers to participation. A significant emphasis was noted in multiple studies regarding participant burden, and the need for flexibility in scheduling and delivery was common; for example, one study noted it took 50% more time to deliver the scheduled sessions than originally planned due to rescheduling based on participant needs and experiences. Several noted that not all sessions were eventually completed. Related to this idea is the observation that participants’ needs change over time and range from high informational needs initially to emotional support over the long term. Thus, existing evidence suggests that planning for this flexibility, and adjusting content and focus of activities over time based on the participants’ needs will be necessary.

Limitations to knowledge translation and directions for future research

Nearly all interventions were delivered in-person to parents and caregivers while the pediatric patient was in active treatment in a hospital setting. Therefore, it is not known how changes in the setting may change the impact of the interventions. Given many studies did not include caretakers or parents who were non-English speaking, findings may not be generalizable to linguistically and culturally diverse families. More recent studies have demonstrated methods for inclusion (Fortier et al., 2026).

The stressor studied in this review were related to diagnosis of pediatric cancer, which may have some unique attributes as a stressor related to parenting style, developmental factors, and perceived severity. Notably, this review included studies that predominantly measured the interventional effects only on mothers. Given there is a suggestion of differences by age and gender in many studies as previously noted above, this could be an important limitation to translation of findings to the general population. Therefore, future studies should measure and account for these characteristics when evaluating the impact of interventions.

Some but not all of the interventions included in this review had informational components in addition to psychosocial; however, the evidence could not clearly delineate the impact of these aspects, nor which aspect was most beneficial and at what point in the trajectory.

Differences in resources across care settings may pose a threat to intervention fidelity or assessing impact of intervention (e.g. differences in control group resources) for multi-site studies. A salient challenge observed across these studies included striving to find a balance between flexibility for participants and maintaining study rigor. Some studies offered participants a choice between intervention modalities (e.g. in-person or phone-based) to make it convenient for parents and caregivers who are busy and overwhelmed; however, this may unintentionally pose a risk to fidelity and study rigor. Both warrant the need to develop plans to monitor and report fidelity. Additionally, more detailed reporting of protocols, key roles and responsibilities, and intervention fidelity is warranted to support future implementation and dissemination research.

A potential limitation to this study is that the timeframe of the search does not include the period during and after the COVID-19 pandemic. To address this concern, the previously described review papers published after the completion date of our search were examined (Park et al., 2022; Phiri et al., 2023; Prates et al., 2024; Shao et al., 2024), and no additional publications that met our inclusion criteria were included in these reviews that were published after February 2020. We also examined a systematic review specific to digital interventions in this population targeting resiliency published in 2021 after our literature search, and only one of six papers included in that review actually measured resiliency, and the implementation details of this program were minimal, so it would therefore have not changed our findings regarding dissemination and implementation of resiliency interventions (Luo et al., 2021a). Additionally, the studies included were limited to those that specifically included resiliency as the outcome. Although this restriction aligns with the purpose of the current review, future studies may include a broader set of related outcomes such as anxiety, depression, or post-traumatic stress to further the understanding of implementation needs for these kinds of interventions on a wider scale.

In conclusion, while there remain important questions, the compilation of evidence about resiliency interventions suggests that this fills an important need recognized by both participants and providers and that they are beneficial in times of high distress. As the research agenda transitions to understanding implementation of resiliency interventions, it is critical that formative research be conducted to assess implementation context to ensure appropriate adaptations are made based on priority populations and local conditions. This is particularly important given the majority of interventions have been developed and tested in non-Hispanic White and English-speaking populations. For example, for researchers seeking to target racially and ethnically diverse populations, we recommend the use of an adaptation framework (Escoffery et al., 2019) to guide efforts to assess the fit of the existing intervention, decide what parts of the intervention to adapt, make the adaptations, pretest the adaptations, and eventually pilot test the intervention. Delineation and understanding of the mechanistic pathways can also support the study and delivery of resiliency interventions to high-need, vulnerable groups.