Psychosocial interventions for rehabilitation and reintegration into daily life of pediatric cancer survivors and their families: A systematic review

Background The survival rate of childhood cancer patients increased over the past decades. However, even after successful treatment the transition back to normalcy is often a major challenge for the whole family. Therefore, this study aims to provide an overview of psychosocial interventions for childhood cancer survivors and their families in the first years after the end of cancer treatment. Methods We conducted a systematic review following the PRISMA Checklist (Preferred Reporting Items for Systematic Reviews and Meta-Analyses; PROSPERO registration number: CRD42017059782). In November 2016 and September 2017, we searched the databases CINAHL, MEDLINE, PSYNDEX, and Web of Science. We included studies investigating psychosocial interventions for childhood cancer survivors diagnosed under the age of 21, their family members or the family as a whole. Further, we summarized the study characteristics and conducted a narrative synthesis of the results. Finally, we assessed the study quality with the Effective Public Health Practice Project Quality Assessment Tool. Results We identified a total of 8215 records based on our database searches and 17 additional records through hand searches. We included 33 articles in the qualitative synthesis. Most of the studies described interventions for the cancer survivor (n = 15). Nine studies investigated interventions for the whole family, and two studies interventions for siblings. The interventions mainly take place in an outpatient group setting (n = 15). Overall, most of the studies reported a significant psychosocial benefit of the interventions. However, the quality of the included studies was limited. Conclusion In summary, we identified a broad range of different interventions and thus could give a comprehensive overview of existing interventions for childhood cancer survivors and their families. However, there is a necessity for high quality studies. The results may help to optimize health care services that support families with the re-entry into daily life.


Introduction
A childhood cancer diagnosis turns the lives of patients and family members upside-down.The incidence rates range from approximately 120 per million to more than 170 per million in developed countries [1].Diagnostic and treatment methods improved a lot in the past decades.The five-year survival rate for childhood cancer in developed countries increased from 65% in the 1980s to approximately 80% nowadays [1], which leads to a growing population of childhood cancer survivors.
However, even despite a complete remission of childhood cancer, the transition back to normalcy is often long and difficult [2].After the completion of treatment, childhood cancer patients report significantly lower physical well-being than their healthy peers [3].The psychological well-being of cancer patients seems to increase after diagnosis [4][5][6].Nevertheless, pediatric cancer survivors are at risk for developing stress-related mental disorders and are highly distressed [7][8][9].Even more than seven years after diagnosis, a relevant proportion of adult survivors of childhood cancer report clinically relevant psychological distress symptoms whereof some fulfil diagnostic criteria of a post-traumatic stress disorder (PTSD) [10].
Notably, also the patient's family as a system and the family members individually can suffer from negative psychosocial consequences of childhood cancer [11,12].After the first shock caused by the cancer diagnosis, parents have to adapt to the new situation, their new roles and responsibilities [13].Especially an active treatment status is associated with low parental quality of life and high psychological distress [6,11,14].Still, post-traumatic stress symptoms or PTSD can also occur in parents after successful treatment of childhood cancer [15] and siblings of childhood cancer patients are also affected [16].Siblings have to fulfil new roles and responsibilities within their families [17].Further, siblings often get less attention by their parents than patients and in some families the extended family cares for the siblings [12].Furthermore, they are emotionally burdened, report a poor quality of life and school difficulties can occur as a consequence of the cancer disease of their sibling [17].Some siblings also report cancer-related traumatic stress reactions or fulfill the criteria for PTSD [18,19].During the first two years after diagnosis, the burden of siblings seems to be highest [17].In summary, childhood cancer seems to be a family challenge that goes far beyond cancer treatment.
In addition to medical follow-up, evidence-based interventions are required to alleviate negative effects in childhood cancer patients, their parents and siblings [20,21].In the first years after the long and intensive cancer treatment, the physical and mental rehabilitation and reintegration into daily life of survivors and their family members (e.g.school life, working life, social life) are particularly important.Various interventions for childhood cancer survivors and their family members were evaluated and summarized in prior reviews.Some of these reviews included interventions on psychosocial and neurocognitive outcomes as well as health promotion solely for childhood cancer survivors [22][23][24][25] and a further review refers to psychosocial interventions solely for siblings [26].Other reviews included studies on interventions for families of patients and survivors [27][28][29][30][31][32][33].However, previous reviews often included a wide range of studies with regard to the aim of the intervention (psychosocial to neurocognitive interventions, e.g.[22]), the included diseases (not only cancer, e.g.[28]), the time since diagnosis (newly diagnosed patients to survivors, e.g.[31,33]) or the age of the childhood cancer survivors (children and adults up to 56 years, e.g.[22]).
The burden and needs of affected families change over time [12] and depending on their current age, patients face different challenges and may have specific developmental needs (e.g.keeping up in school, development of autonomy).In this systematic review, we investigate interventions for childhood cancer survivors and their family members with a narrower focus in terms of type of intervention, diagnoses, time since cancer treatment, and age of the survivor.The review aims to provide an overview of psychosocial interventions for childhood cancer survivors diagnosed before the age of 21 and their family members in the first years after the end of acute cancer treatment.

Methods
We developed our research questions in concordance with the PICO criteria [34].Thus, we specified the participants (P), interventions (I), and outcomes (O) in our research questions.We included all studies reporting interventions independent of the presence or absence of a comparison group.Therefore, we did not specified the comparison (C) in our research questions.This systematic review addresses the following two research questions: 1. Which psychosocial interventions for rehabilitation and reintegration into daily life of pediatric cancer patients and their families after the end of acute cancer treatment were evaluated and published?
2. What are the effects of these interventions on psychosocial outcomes in the family members?
The reporting of this systematic review follows the PRISMA Checklist (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) to conduce to transparency and completeness (S1 Table) [35,36].We published a review protocol in the International prospective register of systematic reviews (PROSPERO; www.crd.york.ac.uk/PROSPERO/;PROSPERO registration number: CRD42017059782; S1 File).

Eligibility criteria
The eligibility criteria address three aspects: Study characteristics, participants and intervention.The inclusion criteria for study characteristics were: (1) Language English or German, (2) full text accessible, (3) no conference proceedings, (4) article published in a peer-reviewed journal, (5) primary research (no study protocols or intervention descriptions) and ( 6) not only qualitative research.The participants in the studies had to meet the following inclusion criteria: (1) Cancer patients and/or their family members, (2) the patient was diagnosed with cancer before the age of 21 and (3) no primary focus on palliative cancer patients.Finally, the inclusion criteria for the evaluated interventions were: (1) Psychosocial intervention (not primary pharmacological, neurocognitive or educational interventions), (2) cancer specific intervention, (3) child-, family-or parent-focused intervention and (4) intervention directed primarily at patients or family members after the end of cancer treatment to five years later (no primary focus on long-term survivors or patients receiving acute cancer treatment).

Study selection
As a first step, duplicates were removed and the first author (MLP) screened the titles.Second, the first author screened the abstracts of the eligible studies.Third, the first author and an independent member of the research team (LI) screened the full texts.Disagreements between the raters were solved by discussion.

Data collection and quality assessment
After the full text screening, two independent members of the research team (MLP and LI) extracted data from the included studies using a data extraction form.The data extraction form included information on study characteristics (authors, year, title, language, country), methods (study design, sample size, patient's age at diagnosis, patient's age at intervention, time since diagnosis, time since end of cancer treatment) and investigated interventions (name of intervention, target group, setting, psychosocial outcomes, effects of the intervention).
Additionally, both members of the research team assessed the quality of the included studies independently with the Effective Public Health Practice Project Quality Assessment Tool (EPHPP Quality Assessment Tool) [43,44].The EPHPP Quality Assessment Tool assesses six components: Selection bias, design, confounders, blinding, data collection methods, and withdrawals and drop-outs.Ratings (strong, moderate, weak) are applied on all six components.Additionally, studies receive a final grade.Studies with no weak ratings and at least four strong ratings on the six components receive the final grade strong [43].Studies with two or more weak ratings receive the final grade weak.All other studies are considered moderate.Disagreements between raters were solved by discussion.

Data synthesis
As anticipated, a quantitative data synthesis was not feasible due to the clinical, methodological and statistical heterogeneity of the studies.Therefore, we conducted a narrative synthesis as predefined in our study protocol [45].We referred to the recommendations of the Cochrane Consumers and Communication Review Group on data synthesis and analysis [46].

Study selection
We identified a total of 8215 potentially relevant records through our database searches in November 2016.An additional 17 records were identified through hand searches (e.g.reference lists of relevant papers and previous reviews).After the removal of duplicates, 5888 titles and in a second step, 251 abstracts were screened.After this screening process, we excluded 5751 records.The remaining 137 full-texts were assessed for eligibility.First, 30 articles were included in the qualitative synthesis.The main reasons for the exclusion of 107 full-text articles were: Intervention not primarily conducted after the end of cancer treatment to five years later (n = 30), no primary research (n = 19), only qualitative research (n = 17), no psychosocial intervention (n = 15), and conference proceeding (n = 12).Three studies were identified through our search update in September 2017 and therefore we included a total of 33 studies in the qualitative synthesis.The study selection process is illustrated in the PRISMA flow diagram in Fig 1.

Study characteristics and synthesis of results
In the 33 included papers, 24 interventions were evaluated.Four interventions were each investigated in two or more papers (Group Social Skills Intervention Program [47][48][49][50], Surviving Cancer Competently Intervention Program [51,52], Quality of Life in Motion [53,54], Family oriented rehabilitation [55][56][57][58][59]).However, the samples or the focuses of the articles differ and some interventions have been revised over time.Therefore, we decided to include all papers independently in the descriptive analyses of the study characteristics and results.Table 1 provides an overview of the relevant study information.
Synthesis of the psychosocial effects of the interventions.In the following, we summarize the study results of the 28 papers describing effects on psychosocial outcome parameters   [79], psycho-education [73]).The studies reported a positive impact of these interventions on social skills [75], anxiety scores [79], quality of life [70], and other psychosocial variables [67,73], partly with mixed results [47][48][49][50].We also identified outpatient group interventions for families, parents and siblings.Studies on interventions for survivors and parents or the family as a whole reported significant positive changes in various psychosocial outcome parameters (e.g.social skills, posttraumatic stress) in children and their family members [51,52,77].We identified one outpatient group intervention for mothers only [69].The maternal hope score increased and the depression score decreased significantly.Further, a paper evaluating an outpatient group intervention for siblings found a reduction of anxiety and depression symptoms post intervention, in particular in the less resilient group [76].
Four different cancer camps were evaluated in the included studies [61,62,71,72].Cancer camps had a positive impact on social support in childhood cancer patients and survivors [71] as well as on the independence, social skills, and self-esteem of survivors and siblings [61].The evaluation of a camp for adolescents with cancer could not reveal any changes in loneliness scores from pre to post [72].
Five studies evaluated a family oriented rehabilitation program [55][56][57][58][59]. A positive impact of the rehabilitation program on the physical and psychological situation of patients, siblings and parents as well as improvements in partner relationship could be found from pre to post and in comparison to a comparison group [55,56,58,59].At 6.5 years follow up, the psychosocial situation of families participating in the rehabilitation program did not significantly differ from the situation of families from the general public [57].
Three studies evaluated computer-based interventions and generally found mixed evidence [65,66,74].Whereas a computer-based support group intervention for parents led to a significant decrease of maternal depression and paternal anxiety and stress [65], an e-mental health intervention for parents did not improve parental quality of life, psychological functioning, and family functioning [66].
Four studies described outpatient individual interventions [53,54,63,64].A physical and psychosocial intervention for children and their parents did not lead to an improvement of psychosocial outcomes with some exceptions for parent-reported subscales [54].In contrast, interventions for brain tumor survivors had a positive impact on psychological outcome parameters [63] as well as social skills [64].
Only one study assessed psychosocial outcomes in participants of a home-based intervention [78].The authors evaluated an aerobic intervention for children and adolescents with acute lymphoblastic leukemia and could not reveal any significant intervention or time effects for any fatigue subscale and measurement time points in the intent-to-treat analysis [78].However, the per-protocol analysis revealed significantly lower general fatigue in the intervention group than in the comparison group at 1-month follow up.

Risk of bias
The quality of all studies was assessed with the EPHPP Quality Assessment Tool [43,44].Most of the studies received the final grade weak (n = 19).Another 13 studies were considered moderate and only one study received the final grade strong.The most frequent weak ratings were assigned in the categories selection bias (n = 19) and data collection methods (n = 22), whereas only five studies received the rating weak for their study design (Table 2).In the EPHPP Quality Assessment Tool, the selection bias is determined by the representativeness and the participation rate.Most of the studies reported a participation rate below 60% (n = 12 studies).In nine studies, participation rates were not reported at all.In six studies, 60% to 79% of the patients or family members participated and in six studies, the reported participation rates were between 80% and 100%.In eight studies, the authors compared nonparticipants with participants.They found no significant differences regarding demographic characteristics (e.g.age, ethnicity, gender).However, there were significant differences in the diagnoses of the children [48].One study additionally found differences in the fitness level and internalizing behavioral problems [54].Participants reported a lower fitness level and more internalizing behavioral problems than nonparticipants did.

Discussion
We identified 33 studies evaluating psychosocial interventions for childhood cancer survivors and/or their family members during the first five years after the end of cancer treatment.The investigated interventions show a heterogeneity of aims, target persons and settings.
Most of the investigated interventions aim to improve psychological burden.Others aim to improve social skills or provide social support and psychoeducation.Besides, nine interventions aim to improve both the physical and psychological situation [54-59, 70, 78, 79].These results are consistent with a review analyzing the association between physical exercise and well-being in several subgroups (e.g.adult cancer patients), which found a positive impact of exercise and physical interventions on mental well-being in adult cancer patients [80].Another review investigated physical exercise training interventions for children and young adults during and after cancer treatment [81].The authors could not reveal any positive impact of exercise on well-being [81].In this review, we found mixed results on the impact of physical activity interventions or combined physical and psychosocial interventions on mental health [54-59, 70, 78, 79].However, we only included studies with a psychosocial focus and other studies on physical interventions for childhood cancer may have been omitted.
In our systematic review, we analyzed studies evaluating interventions for survivors, siblings, parents, and for the whole family.Numerous studies have shown that some siblings experience high levels of psychosocial burden [17].Even though siblings of long-term survivors do not report elevated distress levels per se, a childhood cancer disease can be a risk factor for long-term distress in some siblings [82].Social support seems to be a protective factor in siblings of childhood cancer patients and thus psychosocial interventions focusing on social support could prevent long-term negative psychosocial consequences.However, only two of the included studies investigated interventions addressing siblings [61,76].Furthermore, the family as a whole has to adapt to the illness related changes.A recent systematic review revealed a significant association between family functioning and child adjustment [83].Therefore, psychosocial support for the family as a whole is highly recommended.Only 9 of the 33 included studies investigated interventions for the family as a whole [51, 52, 55-59, 62, 68].Most of the studies focused on children from approximately 8 to 18 years, while younger children and their family members received less attention.After treatment, also preschool children have a lower health-related quality of life than their healthy peers [84].Their scores seem to normalize after two to three years after treatment.However, the time from the end of cancer treatment to normalcy can be very challenging and thus psychosocial support is nevertheless recommended [84].
The included studies investigate interventions in various unconventional child-and familyoriented settings.Among others, we identified studies investigating camps, computer-based interventions and half of the included studies evaluated group interventions.This indicates that different attempts have been made to improve the psychosocial situation of affected families.Even though clinical efficacy could not be confirmed in all of the included studies, for all settings at least some studies revealed a statistically significant benefit and therefore offer starting points for further research.It is important to mention, that the high heterogeneity of different interventions and settings might also be a result of differences in health care systems in the different countries.For instance, the family oriented rehabilitation program is a special feature of the German health care system.
Overall, the methodological quality of the included studies assessed with the EPHPP Quality Assessment Tool was rather poor.Only one study showed a strong quality and the majority of the studies received the final grade weak.Especially high selection bias and poorly described data collection methods caused unsatisfactory final grades.Only five studies received the grade weak for their study design and thus many of the included studies applied appropriate designs and methods but lacked adequate reporting.Some studies used validated and reliable measurement scales but did not describe their psychometric properties.Since the EPHPP Quality Assessment Tool Dictionary does not provide unambiguous recommendations on this topic [85], we decided to rate studies as strong in this component if they described the reliability and validity of the instruments or referred to their study protocol in which they had described the psychometric properties.Other studies did not describe the blinding nor did they report drop-out rates.However, it is very likely that small sample sizes and a high selection bias are consequences of this particular population.The recruitment of childhood cancer survivors and their families can be difficult.First, childhood cancer is a rare disease [1].Second, even after successful treatment of the cancer disease, survivors as well as parents and siblings can suffer from high levels of distress, which can reduce the willingness to participate [3,11,12,17].Only few studies provided information on nonparticipants.Participants and nonparticipants were compared with respect to demographic and medical characteristics (e.g.age, ethnicity, gender, diagnosis).Only one study compared the groups regarding their physical and psychological situation [54].Due to the lack of information on nonparticipants, a self-selection bias cannot be ruled out and therefore the generalizability of the study results is limited [86].Furthermore, many studies assessed multiple outcome parameters and did not adjust for multiple testing.Multiple testing leads to an type 1 error inflation [87].Thus, the number of falsely rejected null hypotheses increases with the number of test hypotheses.With regard to the included evaluation studies, this means that the probability of finding false-positive effects of the investigated interventions increases with the inclusion of multiple outcome parameters and the omission of adjustment.Therefore, it is difficult to draw a conclusion on the efficacy of the investigated interventions based on the included studies.Additionally, the assessment of multiple outcomes within and across the studies contributes to difficulties in the aggregation of study results.Four studies had a cross-sectional design [60,62,71,74].More than half of the included studies did not have a control or comparison group.Only 8 of the 33 included studies are randomized controlled trials [48,52,54,[66][67][68][69][70].Taking together these methodological limitations, the results of the studies can only be interpreted with caution.The effects of the interventions might have been overestimated in the included low quality studies.More high quality studies are necessary to validate prior findings.
There are several implications for research and clinical practice.From a health care research perspective, more high quality studies investigating the efficacy of psychosocial interventions for childhood cancer survivors and their family members are necessary.Only reliable studies can affect the psychosocial aftercare of survivors and their families substantially.Therefore, more high quality randomized controlled trials should be conducted.Additionally, future studies should counteract the low reporting quality by following reporting guidelines.Furthermore, siblings and the family as a whole should also be addressed in psychosocial interventions after the successful treatment of the patient.Overall, the investigated interventions helped families to improve their mental well-being and enhance social skills.These results can be used to optimize health care services that help families with the re-entry into daily life.

Strengths and limitations
This systematic review has some limitations on study level and at review level.Firstly, most of the studies were conducted in North America and Europe.Therefore, the results cannot be generalized to other parts and cultures of the world.Moreover, we included only studies in English or German.Consequently, we may have overlooked studies in other languages.Due to the methodological heterogeneity of the included studies, we could not conduct a quantitative synthesis of the study results.Lastly, we conducted the systematic database search in four databases that are relevant for this field of research and conducted additional hand searches.We had to add only few studies to the records identified through our database search.Nevertheless, relevant studies published in peer-reviewed journals that are not covered by these databases, might have been missed.
This systematic review has also several strengths.First of all, we published a review protocol in the International prospective register of systematic reviews (PROSPERO) to conduce to transparency.Further, this review was conducted and reported according to the PRISMA checklist [35].Two independent raters assessed the inclusion and exclusion criteria to increase objectivity.Finally, we conducted a quality assessment of the included studies to ensure the reliability of the results.

Conclusions
In conclusion, this review provides a comprehensive overview of existing interventions for childhood cancer patients and their family members.However, the high heterogeneity of addressees, methods and outcomes calls for further systematic research.To our knowledge, this is the first systematic review focusing on the vulnerable time after cancer treatment to five years later.Our results suggest that various interventions can help to improve the well-being of affected families.Still, more high quality studies are necessary to validate previous findings and to develop future comprehensive interventions.Finally, this systematic review gives implications for future research and may help to optimize health care services for childhood cancer survivors and their families.

Table 1 . Overview of the included studies (N = 33). Name of the intervention 1. Study; 2. Country 1. Study design (groups; measurement time points); 2. Sample size 1. Age at diagnosis; 2. Time since diagnosis; 3. Time since end of treatment Target group Setting Psychosocial effects of the intervention
3-month follow-up); 2. n = 29 children 1. N/A; 2. N/A; 3. N/A 8 to 18 year-olds, cancer diagnosis, post traditional medical treatment; Their siblings Three-day camp for the oldest camper group precedes week-long camp for all campers Increased independence, social skills and self-esteem from pre to post, only significant for self-esteem Mothers' depression decreased significantly from the second measurement time point to post; Fathers' anxiety decreased significantly from pre to post; Fathers' stress decreased significantly from the second measurement time point to post; Mothers and fathers perceived mutual support to some extent from participating in the CMSG

Study design (groups; measurement time points); 2. Sample size 1. Age at diagnosis; 2. Time since diagnosis; 3. Time since end of treatment Target group Setting Psychosocial effects of the intervention
[53,60,62,68,74]71/journal.pone.0196151.t001brieflyandsynthesizethefindingsaccordingtothe setting of the interventions.In five studies, the authors only assessed the participants' satisfaction and the feasibility of the interventions and no psychosocial outcome parameters[53,60,62,68,74].A comprehensive overview of the interventions and study characteristics is provided in Table1.15studiesdescribeinterventions in an outpatient group setting[47-52, 60, 67, 69, 70, 73, 75- 77, 79].Most of these interventions target childhood cancer survivors and cover a wide range of different programs (e.g.yoga program https