Effectiveness of a group intervention to improve mental health in siblings of children with chronic disorders: a cluster randomized controlled trial

Abstract

Objective

We evaluated the effectiveness of SIBS, a preventive intervention for siblings and parents of children with chronic disorders (CDs).

Methods

This two-arm, unmasked cluster randomized controlled trial registered on ClinicalTrials.gov (NCT04056884), included 288 siblings (M age = 10.4 years, SD = 1.9) and parents of children with CDs (mainly neurodevelopmental disorders) randomly assigned to intervention (k = 34, n = 137) or waitlist control (k = 35, n = 151) groups. Eligible siblings were aged 8–16 years and had a sibling diagnosed with a CD. SIBS is manual-based and was delivered as five sessions over 2 weeks in primary care and hospital settings across Norway. Three sessions are separate for siblings and parents, and two are integrated sibling–parent dialogues. The primary outcome was sibling mental health, rated by siblings, parents, and teachers. The secondary outcome was parent–child communication, rated by siblings and parents. Analyses included intention-to-treat (ITT) and complier average causal effects (CACE).

Results

Although not statistically significant, at 3-month follow-up, the intervention group showed fewer mental health problems (ITT: sibling-rated d = −0.16, 95% CI [−0.49, 0.17]; parent-rated d = −0.10, 95% CI [−0.48, 0.12]; teacher-rated d = −0.18, 95% CI [−0.50, 0.29]) and higher-quality parent–child communication (ITT: sibling-rated d = 0.21, 95% CI [−0.10, 0.52]; parent-rated d = 0.24, 95% CI [−0.07, 0.55]) compared to waitlist.

Conclusions

The SIBS intervention demonstrated small, consistent positive effects on sibling mental health and parent–child communication. This suggests SIBS is a promising preventive program for siblings of children with CDs.

siblings, parents, intervention, prevention, randomized controlled trial

Mental health problems are a leading cause of health-related disability among children and adolescents, representing a considerable burden on society in terms of healthcare costs and productivity losses (Plana-Ripoll et al., 2019). Consequently, establishing effective prevention for children at risk of mental health problems is of high importance, and calls have been made for research that informs policymaking about preventive efforts (Patel et al., 2023). The current study focuses on an understudied group of children at risk of developing mental health problems, i.e., siblings of children with chronic disorders (CDs) (Pinquart, 2023; Wolff et al., 2023a). CDs include a range of congenital or acquired physical, developmental, and/or psychological conditions such as neuromuscular diseases, intellectual disability, and autism spectrum disorder, which despite their diversity share enduring effects on health, development, and quality of life for the child with disorder and also their families. Siblings of children with various CDs are at elevated risk of psychosocial problems. A meta-analysis of outcomes for siblings of children with chronic physical health conditions identified higher levels of internalizing, externalizing, and overall behavioral problems (effect sizes ranging from g = 0.13 to 0.22; Pinquart, 2023). Similarly, reviews of siblings of children with neurodevelopmental disorders have shown that siblings face increased risks of stress, internalizing and externalizing challenges, and reduced overall well-being (Wolff et al., 2022, 2023a). Therefore, the current study adopts a dimensional approach to sibling adjustment, focusing on a wide range of CDs across somatic, intellectual, and other neurodevelopmental functioning domains despite differences in causes and progression across the CDs.

Family members influence one another reciprocally at both dyadic levels and systemic levels according to family systems theories (Minuchin, 2018; Watson, 2012). For families with childhood CDs, multiple risk factors influence how siblings are affected. One risk factor is that families with pediatric CDs display lower levels of open, warm, and structured communication and higher levels of hostile/intrusive and withdrawn communication compared to families without CDs (Murphy et al., 2017). Importantly, open and warm communication and information about the disorder can buffer against the mental health risks (Haukeland et al., 2020). Experiencing better communication with parents may facilitate better understanding of the CD, and siblings may seek more emotional support from their parents (Haukeland et al., 2015; Incledon et al., 2015).

Another risk factor for siblings is parent mental health problems, which is common among parents of children with CDs due to increased caregiving responsibilities (Cohn et al., 2020). Emotional and behavioral difficulties in siblings may arise due to shifts in family roles and responsibilities (Meadan et al., 2015), with differences in parent mental health further influencing sibling adjustment through interconnected family subsystems (Fredriksen et al., 2023; Zahl et al., 2024). Finally, the extent to which the CDs affect daily family life (i.e., severity) is consistently associated with lowered sibling mental health, both for chronic somatic and neurodevelopmental disorders (Pinquart, 2023; Wolff et al., 2022). In this study, severity refers to the degree of functional impairment and the need for support in everyday life, including difficulties with age-appropriate self-care, behavioral and emotional challenges such as mood disturbances (e.g., depression, anxiety, delusions), and disruptive behaviors (e.g., aggression, rages, or age-inappropriate actions) as operationalized by the Developmental Behavior Checklist Primary Carer Version (DBC-P; Einfeld et al., 2002).

The family system perspective highlights the interplay of factors such as disorder severity, family resources, and individual perceptions in shaping family experiences with childhood CDs (Tudor et al., 2018). For siblings, these dynamics can result in both challenges and strengths. While siblings of children with CDs face heightened risks for mental health difficulties, they also exhibit notable strengths. Research highlights positive outcomes, such as increased prosocial behavior, mentalizing skills, and adaptive caregiving roles, which can enhance quality of life, self-esteem, and social relationships (Orm et al., 2022b; Pinquart, 2023).

Interventions have been developed to mitigate risks for siblings. A recent review of 24 intervention studies for siblings of youth with neurodevelopmental disorders showed promising improvements in sibling self-reported emotional and behavioral adjustment across the reviewed studies (Wolff et al., 2023b). However, only nine of the 24 studies were quantitative, of which only six were randomized controlled trials (RCTs). These six RCTs showed statistically significant improvements in one or more of their target outcomes, which ranged from mental health problems to stress and self-esteem with a median effect size of d = 0.62. However, there were about as many nonsignificant as significant findings (specifically, 76 nonsignificant and 68 significant findings), suggesting that even the best of available evidence from sibling interventions is mixed and varies across outcome and informant type.

Multiple approaches have been taken in these past sibling intervention programs. These include psychoeducational formats with information about the CDs and teaching siblings problem-solving/coping skills, noneducational support groups, and programs focusing on stress-reducing activities; most programs focus on providing direct support to siblings with limited parent involvement (Wolff et al., 2023b). In the three previous RCTs that included parent involvement, one program was delivered via telephone support (Giallo & Gavidia-Payne, 2008), one comprised an information sheet with parent tips (Roberts et al., 2015), and the third included one parent session (Williams et al., 2003). To the best of our knowledge, no previous program involved direct training of parental communication or parent–sibling dialogues. Given the vast documentation of communication difficulties in families of children with CDs, the lack of addressing parent communication with siblings is a gap in the field (Murphy et al., 2017). Recognizing both the risks and strengths of sibling experiences is crucial for developing balanced interventions. Supporting parent–sibling communication and providing social support for siblings may help foster adaptive potentials while addressing mental health concerns (Haukeland et al., 2020; Incledon et al., 2015; Kirchhofer et al., 2022).

Here, we examine a manual-based preventative intervention for siblings and parents of children with CDs called SIBS. SIBS was developed based on empirical evidence about risk factors affecting children’s adaptation to illness and on qualitative studies of siblings’ diagnostic knowledge, emotional experiences, and coping (Haukeland et al., 2020). The family-oriented approach is based on the broad literature documenting the family dynamics involved in the risk for siblings (Wolff et al., 2022). The focus on parent–sibling communication specifically is based on review findings of less open/warm communication in families with childhood CDs (Murphy et al., 2017). During the development of SIBS, multiple rounds of user feedback with families and care providers were conducted. Patient user organizations for autism, cerebral palsy, congenital heart disease, and Down syndrome, as well as parent unions for children with a range of disabilities and mental health disorders, were involved. The overarching aim of SIBS is to prevent sibling mental health problems and enhance parent–sibling communication about two empirically documented sibling challenges, i.e., siblings’ disorder knowledge and emotional experiences (Haukeland et al., 2020).

Calls have been made to address the mental health needs of siblings and parents of children with CDs and to provide standards of care (Salley et al., 2024; Wolff et al., 2023b). Approaches to overcome barriers of implementing evidence-based interventions in general include shortening treatment protocols and developing shorter, intensive, and concentrated treatment models (Stoll et al., 2020). SIBS specifically aims to overcome implementation barriers by being adjusted to fit real-world health service settings, making the intervention more accessible, as well as adjusted to the families’ expressed needs and capacities. SIBS is delivered as part of regular primary care and specialist health care centers, and not in university or extraordinary camp settings, enhancing accessibility and relevance for ordinary practice. SIBS comprises five sessions; three of the sessions (1, 2, and 4) are separate groups for siblings and parents that focus on introduction, the CD, and siblings’ emotional experiences, respectively. The sibling sessions have a support group format focusing on sharing and reflecting on CD understanding and emotional experiences. The parent sessions have an educational format focusing on increasing parental communication skills, focused on the principles listen, explore, and validate. In parent sessions 2 and 4, parents are shown animated psychoeducational videos in which these principles (listen, explore, and validate) are explained and demonstrated. In addition, example videos with parent–child dialogues about illness in which the parents show various levels of listening, exploring, and validating behaviors are shown. The task for parents is to identify the behaviors, followed by discussions and role plays about how to enhance their own use of these communication behaviors. Two sessions (3 and 5) are parent–sibling dialogues where siblings talk privately with their parent with brief group leader supervision to enhance parental communication competence. Group sessions 1 and 2 prepare siblings and parents for the dialogue in session 3, and group session 4 prepares siblings and parents for the dialogue in session 5. To enhance age adaptation, the maximum age range within sibling groups is set to 4 years from the oldest to the youngest participant. The SIBS group leader training and manual emphasizes age adjustment, e.g., using more frequent breaks, more play-based tasks, and simpler language with younger participants. See Figure 1 for an intervention overview.

Figure 1.

Overview of the SIBS intervention.

SIBS was initially evaluated in an uncontrolled trial with 99 families of children with CDs (Haukeland et al., 2020). The initial study showed significantly improved sibling mental health, family communication, and disorder knowledge up to 6 months post-intervention (mean effect size d = 0.43). The study showed promising acceptability and feasibility, with high user satisfaction and good manual adherence. Now, we examine the effects of SIBS in a cluster RCT with a larger sample size based on a priori power analysis with appropriate accounts for similarities within and between clusters (intracluster correlation coefficient [ICC]). We use a multi-informant approach to capture siblings’ mental health from their own perspectives, as well as from their parents and teachers. The main outcome, siblings’ mental health, is compared between the two intervention arms 3 months post-intervention/waitlist (T3m). We include sibling and parent reports on family communication as a secondary outcome. Based on prior findings on the impact of parent mental health and severity of the CD on sibling risk (Macias et al., 2007; Pinquart, 2018), we include these as potential moderators. We hypothesize that siblings randomized to intervention groups will have more improved parent, teacher, and self-reported mental health and parent–sibling communication at 3 months post-intervention compared to siblings randomized to waitlist groups. There is little knowledge on potential outcome moderators in the sibling field, and hence limited basis for hypotheses (Wolff et al., 2022). However, due to the risk that parental mental health and disorder severity represent for siblings (Cohn et al., 2020; Pinquart, 2018), we hypothesize that the effect of the intervention on sibling mental health outcomes and communication quality will depend on these two factors. We expect that higher parent mental health problems will negatively affect intervention outcomes, since parents with poor mental health may struggle more to implement communication changes. We expect that higher perceived disorder severity will positively affect intervention outcomes, since the intervention may appeal more and be more relevant to families who experience high disorder severity (Pinquart, 2023).

Methods

This two-arm, unmasked cluster RCT was preregistered in August 2019 (www.ClinicalTrials.gov; NCT04056884). The sample size was determined based on power analysis of change in the main outcome variable in the uncontrolled trial adjusted for the inflation factor due to group clustering (Donner et al., 1981; Fjermestad et al., 2020). The recruitment stopped when the sample size reached 288, which was the determined size based on this a priori power analysis. We followed the Consolidated Standards of Reporting Trials (CONSORT) checklist, and the final checklist is available as Supplementary File S1.

Recruitment

The participants were recruited from multiple sources, including health services, patient user associations, and advertisements, between September 2019 and September 2022. The health services (primary care and hospitals) recruited families known for their services (e.g., disability services, child and adolescent mental health clinics) by verbal invitation, mail, and/or online information. The patient user organizations informed members about SIBS on their websites, social media, and at user events. One-page advertisements presenting SIBS as an intervention for families of children with CDs with a QR code for more information and consent forms were placed at health centers across the participating sites. All groups were delivered across seven public primary or specialist health care sites.

Eligibility criteria

The eligibility criteria for siblings of a child with a CD were aged 8–16 years. The initial preregistered eligibility CDs were neurodevelopmental disorders, including rare disorders (defined as affecting fewer than 1:2,000 people). This criterion was expanded with the following CDs 3 months into the trial: cancer, congenital heart disease, diabetes, and eating disorders. The decision to expand the criterion was made because initial recruitment was slow, patient user groups and health professionals involved with patients beyond the neurodevelopmental disorder spectrum requested to be included in the trial, and the risk for siblings is documented beyond the neurodevelopmental disorder spectrum (e.g., Pinquart, 2023). The child with a CD had to be aged 0–18 years and receive health or special education services, intended as indicator of clinical impairment. See Table 1 for final sample characteristics. For siblings, being diagnosed with any of the inclusion CDs was an exclusion criterion. This was for ethical reasons, as other services would be indicated for these children, and to avoid conflicts and dilemmas in group discussions about the CD. Parents needed to be able to understand the Norwegian language. There were no other parent-related exclusion criteria.

Table 1.

Demographic characteristics of participating families.

	Intervention, N (%)	Waitlist controls, N (%)
Participating child
Sexa
Female	79 (58.1)	80 (53.3)
Male	57 (41.9)	70 (46.7)
Age, M (SD)	10.2 (2.1)	10.6 (1.8)
Sibling order
Younger	58 (52.7)	60 (49.2)
Older	46 (41.8)	54 (44.3)
Twin	6 (5.5)	8 (6.6)
Ethnicity
European	100 (90.9)	112 (91.8)
Asian	0	5 (4.1)
African	3 (2.7)	2 (1.6)
Mixed	7 (6.4)	3 (2.5)
Family
Participating child living with
Both parents	86 (78.2)	101 (82.8)
Mostly father	3 (2.7)	0
Mostly mother	21 (19.1)	20 (16.4)
Other carers	0	1 (0.8)
Age of father, M (SD)	44.3 (7)	45 (6.2)
Fathers’ highest completed education
Secondary school	4 (3.6)	10 (8.2)
High school	38 (34.5)	33 (27.0)
University ≤4 years	24 (21.8)	36 (29.5)
University >4 years	44 (40)	43 (35.2)
Fathers’ employment status
Public sector	28 (25.5)	38 (31.1)
Private sector	66 (60)	75 (61.5)
Student	0	2 (1.6)
Unemployed	4 (3.6)	2 (1.6)
Disability benefit	2 (1.8)	1 (0.8)
Retired	0	0
Other	10 (9.1)	4 (3.3)
Age of mother, M (SD)	42 (5.3)	42.4 (5.4)
Mothers’ highest completed education
Secondary school	3 (2.7)	5 (4.1)
High school	23 (20.9)	17 (13.9)
University ≤4 years	33 (30)	40 (32.8)
University >4 years	51 (46.4)	60 (49.2)
Mothers’ employment status
Public sector	51 (46.4)	62 (50.8)
Private sector	37 (33.6)	38 (31.1)
Student	0	0
Unemployed	3 (2.7)	3 (2.5)
Disability benefit	11 (10)	7 (5.7)
Retired	0	0
Other	8 (7.3)	12 (9.8)
Self-reported family financial situation
Poor/very poor	10 (9.1)	4 (3.3)
Medium	31 (28.2)	38 (31.1)
Good/very good	68 (61.8)	80 (65.6)
Child with diagnosis
Age, M (SD)	10.6 (3.4)	10.5 (3.3)
Female	44 (40.0)	41 (33.6)
Diagnosisb
ADHD	50	47
Anxiety disorders	7	4
Asperger syndrome	19	16
Autism	18	27
Cerebral palsy	7	4
Down syndrome	10	19
Eating disorders	9	6
Intellectual disability	16	15
OCD	2	7
ODD/CD	8	7
Other diagnosesc	8	4
Rare disordersd	22	29
Somatic disorders	20	9
Specific developmental disorder	6	4
Tourette syndrome	11	17

Note: The table is based on the total randomized sample (ITT). All demographic variables are parent-reported; the total does not equal N = 288 due to missing responses on demographics. ADHD = attention-deficit/hyperactivity disorder; ODD/CD = oppositional defiant disorder/conduct disorder; OCD = obsessive–compulsive disorder; SD = standard deviation; M = mean; ITT = intention-to-treat.

a

The Norwegian common term for sex/gender does not distinguish between sex and gender; the option of “other” was provided in the form.

b

The number of diagnoses in the table is higher than 288 because many children had more than one diagnosis (up to 3). Therefore, percentages are not provided for the diagnoses. All diagnoses were arranged into primary, secondary, and tertiary diagnoses by two experienced clinicians according to the International Classification of Diseases, 10th Revision (ICD-10 manual) and cross-checked with clinic registry.

c

Other diagnoses, for example, bipolar disorder, depression, adjustment disorder, and unspecified/mixed emotional disorders.

d

Rare disorders are defined as affecting fewer than 1 out of 2,000 people.

Randomization and masking

Site coordinators enrolled and assigned participants to age-appropriate groups (maximum 4 years between the youngest and the oldest participant). Each time a site had recruited a minimum of three eligible siblings, the group was given a sequence number. The randomization status was determined based on a computerized randomization list (randomizer.org) administered by an independent research coordinator (Urbaniak & Plous, 2013). Participants and providers were masked for randomization status until 2 weeks before intervention. The protocol plan was to randomize groups of six participants (Fjermestad et al., 2020); however, we lowered the number to at least 3 due to recruitment delays and COVID-19 restrictions that hindered gatherings of larger groups. Hence, whereas the plan was to conduct 48 groups with six participants each, the result was 69 groups with between two (due to no-show) and seven participants (mean participant number = 3.5). The trial was unmasked as the outcomes are sibling- and parent-reported. We decided that masked clinical evaluators were not suitable for a prevention trial.

The SIBS intervention and group leaders

The SIBS sessions were delivered over 2 days (sessions 1, 2, and 3 on the first day, and sessions 4 and 5 one week after), 1 week apart. For practical reasons, one parent per sibling was invited to participate. The SIBS sessions were led by a group leader and a facilitator who were employed in health services (e.g., psychologist, social worker, nurse) or were advanced clinical psychology students. All group leaders and facilitators in this study had completed a standardized training comprising a 1-hr e-learning course and 2-day practical training. Group leader adherence and competence were measured based on video observations of 93% of all group leaders (26 of 28) (Bjaastad et al., 2016; Fjermestad et al., 2020). The two group leaders who were not observed delivered only one group each during the trial, and videos were missing from these two groups. The mean adherence score was 87.5% (with 100% representing all manual elements), and the mean competence score was 3.8 (SD = 0.6) on a 1 (minimum) to 5 (maximum) scale.

Outcomes

All measures were completed within 14 days prior to intervention (Tbaseline) for both groups before randomization status was revealed and 3 months after completing intervention for the outcome measures (T3m). The end date for follow-up was December 2022.

Primary

The Strengths and Difficulties Questionnaire (SDQ; Goodman, 2001) was used to measure sibling mental health at Tbaseline and T3m. The 25 items are rated on a 3-point Likert scale from 0 (not true) to 2 (certainly true) and were completed by participating parents and siblings. To preserve confidentiality, teachers were invited to complete the SDQ via parents. Internal consistency (α) was 0.80 for siblings, 0.81 for mothers, 0.82 for fathers, and 0.81 for teachers.

Secondary

The Parent–Child Communication Scale (PCCS; Orm et al., 2022a) was used to measure parent-sibling communication at Tbaseline and T3m. The items (six for parents; eight for siblings) are rated on a 5-point Likert scale from 1 (never) to 5 (always) and were completed by participating parents and siblings. Internal consistency (α) was 0.84 for siblings, 0.72 for mothers, and 0.63 for fathers.

Moderators

The Symptom Checklist-90-Revised (SCL-90-R; Derogatis & Unger, 2010) was used to measure parent mental health. The 90 items are rated on a 5-point Likert scale and were completed by participating parents at Tbaseline. Internal consistency was (α) 0.97.

The DBC-P (Einfeld et al., 2002) was used to measure perceived disorder severity. The 96 items are rated on a 3-point Likert scale and were completed by parents at Tbaseline. Internal consistency was (α) 0.94.

Treatment satisfaction

An overall satisfaction item ranging from 0 (not satisfied) to 4 (very satisfied) was rated by participating parents immediately post-SIBS. Siblings were asked, “What did you think of being in the SIBS group?” and the scale ranged from 0 (really bad) to 4 (really good).

Ethical considerations

The study was approved by the Regional Committee for Ethics in Medical and Health Research (#2018/2461). Both parents and children 16 years and older provided informed consent. Parents consented on behalf of children <16 years old. There were no interim analyses and/or stopping rules. If deterioration in siblings was detected (i.e., poorer mental health for the participating sibling), families could contact the site coordinator or project management. There were no observed or reported adverse events (e.g., injury, psychological harm, trauma, or death) in or following any of the groups.

Statistical analysis

ITT and CACE

We considered two different analytical paradigms to estimate the effects of the intervention on primary (sibling, participating parent, and teacher-rated SDQ) and secondary outcomes (participating parent- and sibling-rated PCCS): intention-to-treat (ITT) and complier average causal effect (CACE). Under ITT, the traditional paradigm used to analyze effectiveness, the difference between groups, does not take into account the extent to which the intervention group adhered to treatment completeness (i.e., some families were present over the 2 days of the intervention, as the protocol was designed, and some were only present a single day). CACE involves a mixture modeling methodology (Imbens & Rubin, 1997), which allowed us to robustly estimate the effect of SIBS exposure among those who received the intervention (i.e., compliers) versus those who would be potential compliers but were not exposed to the intervention (i.e., control group). We measured adherence using an attendance list, and we opted for a partial compliance analysis, where at least 1 day of attendance was considered as adherence because we are assuming participating for a single day could impact outcomes. For further information regarding the CACE paradigm and its assumptions, see Jo et al. (2008). We conducted all analyses using maximum likelihood with robust standard errors (MLR), which accounts for the nonindependence of observations (i.e., families nested in groups). We then computed robust standard errors considering multilevel structure by the command in Mplus called (TYPE = Complex) using a sandwich estimator. For all analyses, we conducted multiple imputations of the missing measures; hence, all randomized subjects were analyzed.

Linear regressions and adjustments

For both analytical paradigms (ITT and CACE), we conducted four linear regressions (i.e., one per outcome) using baseline values of the outcome variables as adjustments, i.e., residualized change score (Mun et al., 2009).

Interactions

We tested if the effects of the intervention on the five outcomes were moderated by baseline parent mental health and baseline participating parent-rated CD severity. We chose these moderators as they are potentially modifiable. We did not include the nonmodifiable moderators’ sex and ethnicity.

Missing data and effects estimates

For both paradigms of the ITT and CACE analyses, we carried out multiple imputations of missing data using the Bayes estimation of an unrestricted variance–covariance model. We included the following variables in the imputation at baseline and follow-up of the SDQ (sibling, mother, father, and teacher), baseline and follow-up of PCCS (sibling and parent), age, sex, randomization status, SCL-90-R, and DBC-P scores. We generated and used 20 imputation data sets in the subsequent analysis using the Rubin method (Rubin, 1987), via an MLR estimator. Adopted statistical significance was 0.05, and standardized effect sizes were Cohen’s d. We carried out all analyses via structural equation modeling approach in Mplus (Muthén & Muthén, 2015), Version 8.0. The deidentified data set generated and analyzed during the current study will be available from the corresponding author upon reasonable request.

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

In total, 288 families were allocated into 69 groups and group-randomized to intervention or waitlist. The two groups had a proximate demographic distribution. Female participants constituted 58.1% and 53.3% of the total intervention and waitlist groups, respectively. The mean sibling participant age was 10.2 years (SD = 2.1) and 10.6 years (SD = 1.8), and participants of White-European ethnicity constituted 90.0% and 91.8%. In the final sample, the most frequent CDs were ADHD (attention-deficit/hyperactivity disorder), rare disorders (i.e., prevalence <1 out of 2,000), autism, Asperger syndrome, and Down syndrome (Table 1; Figure 2). Descriptive statistics of all baseline measures are presented in Table 2. The mean satisfaction score for siblings was 3.4 (SD = 0.8), and 3.2 (SD = 0.8) for parents, with a minimum score of 0 and a maximum positive score of 4.

Figure 2.

CONSORT flowchart of participants and groups.

Table 2.

Overview of Baseline Measures.

	Wait-list groups				Intervention groups
Informant, scale	M	SD	Max	Min	M	SD	Max	Min
Sibling
SDQ	9.43	5.62	30	0.00	10.21	5.19	26	0.00
PCCS	31.43	5.35	40	20	30.77	6.00	40	16.0
Parent
SDQ	9.60	5.80	23	0.00	9.7.8	6.36	23	0.00
PCCS	23.99	3.13	30	16.0	23.71	3.22	30	15.0
SCL-90-R	0.54	0.44	1.94	0.00	0.66	0.48	2.26	0.02
DBC-P	48.00	23.18	112	8.00	47.09	25.29	115	2.00
Teacher
SDQ	6.63	5.26	24	0.00	6.38	4.50	22	0.00

Note. M = mean; SD = standard deviation; SDQ = Strengths and Difficulties Questionnaire; PCCS = The Parent–Child Communication Scale; DBC-P = The Developmental Behavior Checklist Primary Carer Version; SCL-90-R = Symptom Checklist-90-Revised.

The baseline mean levels of sibling mental health were elevated compared to the normative means for Danish 10- to 12-year-olds (girls = 5.27, boys = 6.02) (Youthinmind, n.d.), suggesting that siblings in our study have more mental health challenges compared to the general level for Scandinavian children. The mean baseline level of sibling-reported communication quality with parents was proximate to the mean level of a similar sample of siblings of children with CDs (Orm et al., 2022b). The baseline mean level of parent mental health problems was elevated compared to the normative mean in a nonclinical Norwegian sample (0.34) (Derogatis & Unger, 2010), suggesting an elevated mental health burden for parents in our study compared to the general level in Norwegian men and women. Teachers rated girls but not boys as having raised difficulties for total difficulties on the SDQ (Prentice et al., 2024). The mean baseline level of perceived diagnosis severity was slightly above the clinical cut-off for emotional and behavioral disturbances (46) (Einfeld et al., 2002).

See Table 3 for the ITT and CACE standardized effect sizes, their robust standard errors, and p values after adjusting for the baseline score of the outcome. The effect sizes for ITT and CACE were small, had the expected direction (i.e., better mental health and better parent–sibling communication quality), and were not statistically significant (all p values were >0.05). Note that the point estimates for CACE were stronger than ITT, as expected because we compare those families attending at least one session versus the potential adherent families in the control arm. Due to cluster effects, we calculated the ICC for the five outcomes at baseline. Those ICCs are fundamental for guiding sample size calculation of future cluster ICCs. The ICC from SDQ scores were 0.12 (95% CI [00.0, 0.39]) for teachers’ ratings, 0.15 (95% CI [0.00, 0.33]) for parent ratings, and 0.00 (95% CI [0.00, 0.13]) for sibling ratings. As for the moderators, interaction effects for parent mental health and disorder severity were not significant for any outcome.

Table 3.

Results under both analytical paradigms for the two outcome measures.

	ITT					CACE
Scale	Informant	Cohen’s d	RSE	95% CI	p	Cohen’s d	RSE	95% CI	p
SDQ	Sibling	−0.16	0.17	[−0.49, 0.17]	0.34	−0.23	0.25	[−0.72, 0.25]	0.35
	Teacher	−0.10	0.20	[−0.50, 0.29]	0.61	−0.14	0.29	[−0.70, 0.42]	0.63
	Parent	−0.18	0.15	[−0.48, 0.12]	0.23	−0.25	0.22	[−0.68, 0.19]	0.27
PCCS	Sibling	0.21	0.16	[−0.10, 0.52]	0.19	0.28	0.23	[−0.17, 0.72]	0.22
	Parent	0.24	0.16	[−0.07, 0.55]	0.13	0.32	0.21	[−0.09, 0.73]	0.13

Note. ITT = intention-to-treat; CACE = complier average causal effect; RSE = robust standard error; SDQ = Strengths and Difficulties Questionnaire; PCCS = The Parent–Child Communication Scale; CI = confidence interval.

Discussion

In this cluster RCT of an intervention for siblings of children with CDs, we found small, positive effect sizes using both ITT and CACE paradigms. These effects were found for all three informant sources (siblings, parents, teachers) for the main outcome, sibling mental health, and from both informant sources (siblings, parents) for the secondary outcome, parent–sibling communication. The magnitude of the effects is on par with effect sizes typically found across prevention studies (Rothwell et al., 2018). Preventative interventions typically have smaller effect sizes than treatment interventions (Rapee, 2008). The small positive effect found in our study could ultimately have substantial importance at a larger population level and yield promise for SIBS as a preventative intervention.

We did not, however, find sufficient evidence to statistically prove evidence for SIBS as more effective than waitlist. Factors associated with the intervention’s content, duration, or delivery may account for the lack of significant findings. Focusing on parent–sibling communication, disorder information, and sharing emotional experiences may not be the most effective preventive approach. Alternative strategies like support groups, psychoeducation, or recreational activities could be more effective. A recent review of 24 psychosocial interventions and support groups for siblings of youth with neurodevelopmental disorders, including six RCTs, reported moderate improvements in mental health, stress, and self-esteem, with a median effect size of d = 0.62 (Wolff et al., 2023b); however, methodological variations, including differences in intervention format, duration, and delivery settings, as well as underpowered samples and reliance on single informants, limit the robustness of these findings. The content and format of current sibling interventions vary widely, ranging from individual psychoeducation to group-based social support, with or without parental involvement (Wolff et al., 2022). Moreover, interventions delivered in controlled settings, such as university clinics or camps, often face barriers to real-world implementation due to their intensive nature and limited accessibility. Efforts to overcome these challenges include the development of shorter, more intensive treatment models designed for regular health service settings (Stoll et al., 2020). The SIBS intervention seeks to address these gaps by integrating into primary and specialist healthcare contexts, enhancing its accessibility and alignment with families’ needs. Unlike many previous interventions, SIBS emphasizes feasibility and relevance for ordinary practice while aiming to support siblings’ mental health and social well-being. Alternatively, extending SIBS, either through additional sessions, booster sessions, or homework between sessions, may enhance its effectiveness.

The statistically nonsignificant findings may be partly attributed to reduced statistical power allied to the ICCs, which were different from zero. Appropriate accounting for similarities among subjects in and between groups (clusters) normally results in a net loss of power. In primary care research in general, clustered designs are frequently used; however, studies often fail to recognize the implications of clusters for the analytical method and results. Although sibling interventions are often group-based, of the current reviewed RCTs for siblings of youth with neurodevelopmental disorders (Wolff et al., 2023b), none of those studies described as group-based report accounting for similarities among subjects in and between groups (i.e., the cluster effects, ICCs), increasing risk of Type I errors substantially.

Moreover, in contrast to previous studies of sibling interventions, we included analysis based on both compliance status (CACE) and the traditional approach of ITT where adherence is not considered. We were interested in assessing the size of the benefit of receipt (CACE) of the SIBS intervention, as well as the offer (ITT) of it (Peugh et al., 2017). Effect sizes under CACE were larger than for ITT, meaning that for those who complied with their randomization status and received intervention, when compared with estimated compliers in the control group, the outcomes were more positive than what was found under ITT, where all participants randomized to the intervention and waitlist were compared, regardless of participant compliance. The choice of design and analysis of our study can therefore be considered a methodological step forward for research on sibling-directed group interventions, as we have used two different yet complementing analytical approaches as well as accounted for similarities within and between groups in the reporting of the results.

Sample characteristics may also help explain the lack of statistically significant findings, particularly the fact that not all siblings had clinical levels of mental health problems. A stepped-care model approach targeting siblings who exceed thresholds for emotional and behavioral problems may be more useful than delivering SIBS independent of siblings’ risk profiles (Bower & Gilbody, 2005; Richards, 2012). A stepped-care approach would supersede the current preventative approach, and conducting additional moderator analysis would surpass the protocol plan (Fjermestad et al., 2020). Future studies are encouraged to further investigate who would benefit the most from intervention and support.

Our study highlights the importance of follow-up timing. The timepoint for the effect comparison was 3 months post-intervention, avoiding inflated short-term effects but limiting our understanding of the long-term adaptability and effect of the intervention. In four of the six previous RCTs for siblings of youth with neurodevelopmental disorders, the initial outcomes were reported within 2 weeks after intervention and may therefore only capture immediate post-intervention effects compared to a 3-month follow-up design (Wolff et al., 2023b). A related issue is the time schedule for delivering the five SIBS sessions. In the current study, the five sessions were delivered as two half-day sessions 1 week apart. This timeline may be too short to elicit family communication change. Spreading the sessions over multiple weeks would allow for more homework practice and longer reflection time while potentially making attendance more complicated for families. Further implementation research of the SIBS intervention is indicated to assess optimal delivery timing.

The effects of SIBS were not moderated by parent mental health or perceived disorder severity. This was surprising and positive, as both parent mental health and disorder severity have been suggested as key risk factors for sibling mental health (Tudor et al., 2018; Wolff et al., 2022, 2023a). To the best of our knowledge, the only previous RCT that included disorder severity as a moderator for intervention outcomes focused on siblings of children with autism spectrum disorders (Jones et al., 2020), in contrast to our study that includes a range of CDs. Our results call into question whether parental mental health and disorder severity function as moderators for intervention outcomes for siblings of children with a range of CDs, or if these moderators are more relevant for specific subgroups of CDs. Although inclusion criteria were initially based on a cross-disorder approach to sibling risk, the final sample comprised primarily of siblings and parents of children with neurodevelopmental disorders, which may be considered less acute compared to medical disorders such as cancer or heart diseases and may explain difficulties to detect moderating effects of disorder severity. Alternatively, other aspects related to the CD altogether may impact intervention outcomes more, such as to what extent the disorder impacts everyday life in terms of the load of treatment management. Cousino and Hazen (2013) found that higher levels of general parenting stress were related to increased parental responsibility for managing treatment, regardless of the illness’s duration or severity, reflecting that there are important nuances to disorder severity and its impact on the family system. The measure of disorder severity used in this study (DBC) was initially developed to capture emotional and behavioral problems in children with intellectual and developmental disabilities (Einfeld et al., 2002) and may not capture all relevant aspects of severity for these families and may also explain difficulties in capturing potential moderating effects on intervention outcomes. We encourage further investigation of these and other potential moderators as well as mechanisms of efficacy in future studies (Silverman et al., 2024).

While general mental health measures are valuable as outcome measures, sibling-specific tools can capture nuanced aspects of their experiences, beyond mental health symptoms (Veerman et al., 2025). Our results may underscore the importance of applying a multidimensional approach that also incorporates adaptive outcomes in investigating effects of a family-oriented sibling intervention. Strengths-based outcomes like quality of life, self-esteem, and social relationships remain understudied, and we encourage future interventions to carefully select outcome measures that can capture the unique sibling experience (Wolff et al., 2022).

Limitations

Response bias may be present due to the open recruitment method. Including a broad range of CDs provided the advantage of reaching a large sample size, making the results applicable to a wide variety of conditions, supported by a cross-disorder understanding of sibling risk (Pinquart, 2023). However, diagnostic groups within this CD sample were too small and unevenly distributed to allow for subgroup analysis beyond clinical severity and posed challenges for generalizability of findings within disorder subgroups.

We excluded siblings who were diagnosed with any of the included CDs (neurodevelopmental or somatic). This may represent bias in the type of siblings that were selected for the study. A practical and ethical dilemma in sibling intervention research is what the optimal inclusion criteria for siblings should be, and to what extent interventions should be differentiated for CDs with documented genetic risk (e.g., autism) relative to noncongenital disorders (e.g., acquired physical disabilities following accidents). Larger studies that can differentiate between sibling subgroups are warranted to enhance the field’s knowledge regarding such dilemmas.

Another recurring problem in sibling research is that siblings of families with lower socio-economic status and from ethnic minorities are not well represented, which was also the case for this study (Wolff et al., 2023b). Potential response bias also applies to the recruitment method for including teacher ratings, as parents were asked to forward questionnaires to the sibling’s teacher.

This RCT was conducted during the COVID-19 pandemic. Although we had few cancellations of entire groups and were able to adjust to the temporary health regulations, many groups were smaller than originally planned, and recruitment was delayed. The power calculation for the current trial was based on a plan of conducting 48 groups, which was extended to 69 groups to retain power. Due to the change in group sizes, we were not able to analyze group leader effects. Although we can only speculate how this study would have proceeded in the absence of the pandemic, or how the pandemic may have influenced outcomes, we find it important to address these extraordinary circumstances.

Conclusion

The SIBS-RCT represents important progress toward establishing evidence-based interventions for siblings of children with CDs, a group at risk of developing mental health problems. We identified small, positive, albeit statistically nonsignificant, effects for sibling mental health for the intervention group using a pre-evaluated manual-based intervention with measures of adherence, multi-informant reports, and outcomes measured 3 months after intervention. The initial findings from the SIBS-RCT provide the basis for refining and advancing targeted, evidence-based preventive interventions for siblings of children with CDs and their families. The SIBS intervention may be enhanced by incorporating additional sessions that emphasize parent–sibling communication and the inclusion of booster sessions to reinforce therapeutic gains over time. Targeting siblings with elevated mental health symptoms, in line with an indicated prevention model, may optimize intervention outcomes by focusing on those most likely to benefit. Future intervention planners are encouraged to employ multiple informants to capture a comprehensive understanding of family dynamics, which is essential for addressing the complex needs of families managing childhood CDs, to carefully consider implications for analysis when interventions are provided in groups (clusters), and to include potential moderators based on current sibling risk factors. This will help further set the stage for rigorously evaluated interventions to support families of children in pediatric care.

Acknowledgments

We thank the participating study sites, the municipalities of Asker, Lillestrøm, and Rælingen, and the hospitals Telemark, St. Olav, Lovisenberg, Østfold, and Innlandet for their significant contributions to this trial. We also thank our dedicated research assistants Jakob Marentius Andreassen, Live Amalie Tynes Anderson, Malene Brattvåg Bøe, Kaia Fløtten, Elina Alexandra Løken Garras, Mari Trageton Ileby, Mikkel Stormyr, Thea Sundal, Mariam Tashakori, and Rebecca Furuholmen Møller Warmedal. Finally, we thank the participating families.

Supplementary material

Supplementary material is available online at Journal of Pediatric Psychology (https://academic.oup.com/jpepsy/).

Author contributions

Solveig Kirchhofer (Writing original draft [lead], Writing – review and editing [lead], Data curation [equal], Project administration [equal]), Trude Fredriksen (Writing – review and editing [support], Data curation [equal], Project administration [equal]), Stian Orm (Writing – review and editing [support], Data curation [equal], Project administration [lead]), Matteo Botta (Writing – review and editing [support], Data curation [lead]), Erica Zahl (Writing – review and editing [support], Data curation [equal]), Hugo Cogo-Moreira (Writing – review and editing [support], Formal analysis [lead], Methodology [lead]), Caitlin Prentice (Writing – review and editing [support], Data curation [equal]), Torun Marie Vatne (Writing – review and editing [support], Conceptualization [equal], Project administration [equal], Visualization [equal]), Yngvild Bjartveit Haukeland (Writing – review and editing [support], Conceptualization [supporting], Project administration [supporting]), Wendy K. Silverman (Writing – review and editing [support], Conceptualization [equal], Supervision [equal]), and Krister Fjermestad (Writing – review and editing [support], Conceptualization [lead], Data curation [equal], Funding acquisition [lead], Methodology [equal], Project administration [lead], Supervision [equal])

Funding

This research was funded by the Norwegian Research Council [#321027].

Conflicts of interest: We have no conflicts of interest to disclose.