Study Design
This randomized controlled trial was part of a pilot study in which the KIPA-intervention was developed (Schepper et al. 2024) and evaluated in terms of adherence, acceptance and feasibility (Herrmann et al. (submitted)) . To obtain initial information on the effectiveness of KIPA, we conducted the intervention in a RCT design (waitlist control group) shown in figure 1. The CONSORT Statement (Schulz et al. 2010) was applied in this study.
--- PLEASE INSERT FIGURE 1 HERE ---
Participants and Procedure
Families were consecutively recruited from August 2021 to July 2023. Participants in acute treatment were recruited on the ward for pediatric oncology at the University Hospital Leipzig, Germany. Participants in follow-up care were recruited in an outpatient cancer counseling center in Leipzig (Elternhilfe für krebskranke Kinder Leipzig e.V.). Both institutions cooperate closely in terms of psychosocial care of pediatric cancer patients and their families.
Inclusion criteria were (1) oncology patients aged 0-17 years with the diagnosis of any cancer entity within the last ten years and a minimum time of four weeks since cancer diagnosis and their families, (2) minimum moderate FoP level of at least one family member (FoP-Q-SF/P ≥ 26; FoP-Q-SF/C ≥ 19). Exclusion criteria were (1) insufficient German language skills, (2) inability to follow the intervention due to cognitive impairments assessed by the treating psychosocial staff, (3) palliative care.
After receiving written informed consent, parents and/or children completed measurements of FoP and secondary outcomes. Additionally, information on sociodemographic characteristics, diagnosis and treatment of the child was obtained from one parent. After this, the participants were randomly assigned to either the intervention group or a waitlist control group, who received usual care as described in the German national guideline for psychosocial care in pediatric oncology (Schröder et al. 2019) and participate in the intervention after a four months waiting period.
As figure 1 shows, participants in the intervention group completed questionnaires at three measurement points: before the intervention (BL), after the intervention (T1) and four months later (T2). Waitlist controls had an additional measurement point before their 4-month waiting period (pre-baseline, PBL), subsequently received the intervention at the end of their waiting period and, like the intervention group, completed questionnaires at BL, T1 and T2.
Randomization was performed separately in acute treatment and follow-up care. Unrestricted randomization was not possible due to the small number of cases and the need for approximately equal group sizes (Lachin 1988) . A replacement randomization (Schulz and Grimes 2007) was carried out. The participants were consecutively placed on the randomization lists and assigned according to the date of their baseline measurement by study coordinator (JH).
--- PLEASE INSERT FIGURE 2 HERE ---
As the flow chart for participation and drop-out (figure 2) shows, N=180 families were eligible for inclusion and informed about the study. Of those n=52 families (response rate 29%) wished to participate, n=41 were randomized and n=29 finished the intervention. Of n=12 families who did not finish the intervention (drop-out rate 29%) n=5 died, n=1 had no further interest. In addition, we were unable to start the intervention with n=4 families on the acute ward because no psychosocial professional was available in time due to the increased number of patients, and n=2 families left the acute ward. In view of the high number of drop-outs for various reasons we decided in April 2023 to include all further participants on the acute ward in the intervention group. In addition, we carried out a per-protocol analysis and analyzed the feasibility of KIPA with a focus group in which the participating psychosocial professionals reflected among others on the conditions for successful implementation of KIPA in different settings (e.g. acute treatment); (Herrmann et al. (submitted)) .
N=29 families with a child with cancer were included in our analysis. Either one parent (n=20), a parent-child (older than six years) dyad (n=6) or – at the request of three families – both parents (they were not analyzed as dyads, therefore n=6) participated. In total, N=32 data sets were analyzed, including n=14 participants in the control group and n=18 in the intervention group.
--- PLEASE INSERT TABLE 1 HERE ---
Intervention
The intervention (table 2) is based on cognitive behavioral and solution-focused therapeutic approaches and was developed under consideration of a recent theory of FoP described above and existing therapeutic manuals, for example psychotherapy of dysfunctional FoP in adult oncology patients (Dinkel 2018; Waadt et al. 2011) as well as programs for the prevention and management of separation anxiety (Blatter-Meunier and Schneider 2011) , social anxiety, and depression in children (Aune and Stiles 2009) .
--- PLEASE INSERT TABLE 2 HERE ---
KIPA is designed for family dyads to participate together (e.g., parent-child dyads, both parents) as well as for a parent or a child individually. Individual participation is possible, if the child cannot take part due to their young age (0-6 years) or low FoP, but the parent shows elevated FoP or vice versa. In this case the total number of sessions remains the same, but a special focus is placed on attachment and the child’s perspective is included using circular conversation techniques.
The intervention was carried out by psychosocial professionals (psychologists, social workers) from the participating institutions usually at one to three-week intervals between sessions. Therapists documented their adherence to the manual by written descriptions of every module, which were analyzed in a separate paper (Herrmann et al. (submitted)) .
Measures
Sociodemographic and medical data
Parents provided sociodemographic and medical information about themselves and their ill child via a questionnaire before intervention.
Fear of Progression Questionnaire Short-Form for Children with Cancer and their Parents (FoP-Q-SF/C and P)
The FoP-Q-SF contains the subscales affective reactions, partnership/family, occupation and loss of autonomy (Hinz et al. 2015) . The adaption for parents of children with cancer shows good reliability (Cronbach´s α=0.89) and validity ( Clever et al. 2018 ). This also applies to the adaption of the questionnaire for children with cancer (Cronbach´s α=0.86 ; Luz et al. 2020) . Children and parents indicate their burden from FoP in twelve items on a five-point Likert scale from (1) never to (5) very often. The total score ranges from 12 to 60, with higher scores indicating higher levels of FoP. To evaluate KIPA in families with at least moderate FoP levels, we used a cut-off for participation (total score FoP ≥26 for parents and ≥19 for children).
Secondary Outcomes
We assessed parental anxiety and depression with the German version of the Hospital Anxiety and Depression Scale (HADS; Herrmann-Lingen, C. et al. 2011) , HRQoL was measured with the Short Form-36 Health Survey (Morfeld et al. 2011) and posttraumatic stress symptoms with the German version of the PCL-5 (Krüger-Gottschalk et al. 2017) .
To assess children’s anxiety we used the subscales “animal phobias”, “medical fears” and “separation anxiety” from the German Version of the Fear Survey Schedule for Children – Revised (PHOKI; Döpfner et al. 2006) . Further we measured depressive symptoms with the DTK-II short form ( Rossmann 2014 ), HRQoL with the KIDSCREEN-10 (Ravens-Sieberer et al. 2010) and posttraumatic stress symptoms with the CATS-D for children and adolescents (Sachser et al. 2017) .
Statistical analysis
Analyzes were performed using R Version 4.2.2. The significance level was set to α≤.05 for all analyses. Scores of main and secondary outcomes were calculated only if no more than 1% of the items were missing.
To analyse group differences in the main and secondary outcomes (RQ1), we first had to control for between-group differences before intervention/waiting period (BL/PBL). Despite randomization, such differences may occur in small samples and mask correlations after intervention/ waiting period (T1/BL). We used regression analyses to control for variable levels via residualization. After z-standardization and proof of normal distribution with Shapiro-Wilk test, we used the Mann-Whitney-U-test to estimate group-differences as well as Hedge’s g effect sizes after intervention/ waiting period (T1/BL). Due to the small number of participating children, we only ran this analysis for parents.
To evaluate the pre-post and long-term intervention effect (RQ2), we calculated the differences in FoP between three measurement points (BL, T1, T2) for the full sample, using the Mann-Whitney-U-test and Hedge’s g effect size. Additionally, we calculated differences in FoP within the waitlist control group from PBL to BL via Mann-Whitney-U-test.
To explore the influence of medical and sociodemographic variables on the pre-post intervention effect (RQ3) an exploratory multiple regression analysis was conducted with FoP post-intervention (T1) as outcome and FoP pre-intervention (BL), time since diagnosis and age of the child as predictors. We used z-standardization before interpretation and examined assumptions for multiple regression analysis.