Parent-led stepped care for traumatised children: parental factors that predict treatment completion and response

ABSTRACT

Background: Stepped care cognitive behavioural therapy for children after trauma (SC-CBT-CT; aged 7–12 years) can help to increase access to evidence-based trauma treatments for children. SC-CBT-CT consists of a parent-led therapist-assisted component (Step One) with an option to step up to standard therapist-led treatment (Step Two). Studies have shown that SC-CBT-CT is effective; however, less is known about what parent variables are associated with outcome of Step One.

Objective: To examine parent factors and their relationship with completion and response among children receiving Step One.

Method: Children (n = 82) aged 7–12 (M = 9.91) received Step One delivered by their parents (n = 82) under the guidance of SC-CBT-CT therapists. Logistic regression analyses were used to investigate whether the following factors were associated with non-completion or non-response: the parents’ sociodemographic variables, anxiety and depression, stressful life experiences and post-traumatic symptoms, negative emotional reactions to their children's trauma, parenting stress, lower perceived social support, and practical barriers to treatment at baseline.

Results: Lower level of educational achievement among parents was related to non-completion. Higher levels of emotional reactions to their child's trauma and greater perceived social support were related to non-response.

Conclusions: The children seemed to profit from the parent-led Step One despite their parents` mental health challenges, stress, and practical barriers. The association between greater perceived social support and non-response was unexpected and warrants further investigation. To further increase treatment completion and response rates among children, parents with lower education may need more assistance on how to perform the interventions, while parents who are very upset about their child's trauma may need more emotional support and assurance from the therapist.

Trial registration: ClinicalTrials.gov NCT04073862; https://clinicaltrials.gov/ct2/show/NCT04073862. Retrospectively registered 03 June 2019 (first patient recruited May 2019).

HIGHLIGHTS

• The children seemed to profit from Step One despite parents' mental health challenges and practical barriers.

• Response may improve if parents’ emotional reactions to a child's trauma is addressed.

• Parents with lower education may need more assistance in performing Step One.

Antecedentes: la terapia cognitiva conductual de cuidados escalonados para niños después de un trauma (SC-CBT-CT por sus siglas en inglés; de 7 a 12 años) puede ayudar a aumentar el acceso a tratamientos para trauma basados en evidencia para niños. SC-CBT-CT consta de un componente asistido por un terapeuta dirigido por los padres (paso uno) con la opción de pasar al tratamiento estándar dirigido por un terapeuta (paso dos). Los estudios han demostrado que SC-CBT-CT es eficaz; sin embargo, se sabe menos acerca de qué variables principales están asociadas con el resultado del paso uno.

Objetivo: Examinar los factores de los padres y su relación con la finalización y la respuesta entre los niños que reciben el Paso Uno.

Método: Niños (n = 82) de 7 a 12 años (M = 9,91) recibieron el Paso Uno impartido por sus padres (n = 82) bajo la guía de terapeutas de SC-CBT-CT. Se utilizaron análisis de regresión logística para investigar si los siguientes factores estaban asociados con la no finalización o la falta de respuesta: variables sociodemográficas de los padres, ansiedad y depresión, experiencias de vida estresantes y síntomas postraumáticos, reacciones emocionales negativas al trauma de sus hijos, estrés en la parentalidad, menor apoyo social percibido y barreras prácticas para el tratamiento al inicio del estudio.

Resultados: Menor nivel de logro educativo entre los padres se relacionó con la no finalización. Los niveles más altos de reacciones emocionales al trauma de su hijo y un mayor apoyo social percibido se relacionaron con la falta de respuesta.

Conclusiones: Los niños parecieron beneficiarse del Paso Uno dirigido por los padres a pesar de los desafíos de salud mental, el estrés y las barreras prácticas de sus padres. La asociación entre un mayor apoyo social percibido y la falta de respuesta fue inesperada y justifica una mayor investigación. Para aumentar aún más la compleción del tratamiento y las tasas de respuesta entre los niños, los padres con menor nivel educativo pueden necesitar más ayuda en cuanto a cómo realizar las intervenciones, mientras que los padres que están muy perturbados por el trauma de su hijo pueden necesitar más apoyo emocional y seguridad por parte del terapeuta.

ClinicalTrials.gov NCT04073862; https://clinicaltrials.gov/ct2/show/NCT04073862. Registrado retrospectivamente el 3 de junio de 2019 (primer paciente reclutado en mayo de 2019).

背景：创伤后儿童的阶梯式护理认知行为疗法（SC-CBT-CT；7-12 岁）有助于增加儿童获得循证创伤治疗的机会。 SC-CBT-CT 由家长主导、治疗师辅助部分（第一步）和升级到标准治疗师主导治疗（第二步）的选项组成。研究表明 SC-CBT-CT是有效的；然而，对于哪些父母变量与第一步的结果相关联知之甚少。

目的：考查父母因素及其与接受第一步儿童的完成和反应的关系。

方法：7-12 岁的儿童 (n = 82) (M = 9.91) 在 SC-CBT-CT 治疗师的指导下接受父母 (n = 82) 提供的第一步。逻辑回归分析用于考查以下因素是否与未完成或未响应相关：父母的社会人口变量、焦虑和抑郁、压力生活经历和创伤后症状、对孩子创伤的负面情绪反应、养育子女压力、较低的社会支持感知以及基线治疗的现实障碍。

结果：父母较低的教育成就水平与未完成相关。对孩子创伤更高水平的情绪反应和更大的社会支持与未响应有关。

结论：尽管父母面临心理健康挑战、压力和现实障碍，孩子们似乎从父母主导的第一步中获益。更多感知到的社会支持与未响应之间的关联是出乎意料的，值得进一步调查。为了进一步提高儿童的治疗完成率和反应率，受教育程度较低的父母可能需要更多关于如何进行干预的帮助，而对孩子的创伤感到非常不安的父母可能需要治疗师更多的情感支持和保证。

临床试验号 NCT04073862； https://clinicaltrials.gov/ct2/show/NCT04073862。 回顾性注册于 2019 年 6 月 3 日（2019 年 5 月招募了第一位患者）。

KEYWORDS:

• Stepped care
• children
• trauma-focused therapy
• treatment predictors
• attrition

PALABRAS CLAVE:

• Atención escalonada
• niños
• terapia centrada en el trauma
• predictores de tratamiento
• deserción

关键词:

• 阶梯式护理
• 儿童
• 聚焦创伤治疗
• 治疗预测因素
• 流失

1. Introduction

There are several treatments that have been proven to improve post-traumatic stress symptoms (PTSS) in children (International Society for Traumatic Stress Studies, 2018; National Institute for Health and Care Excellence, 2018). Still, there is a gap between the need for and access to evidence-based treatments for traumatised children (Bringewatt & Gershoff, 2010; Holt et al., 2016; Schweer-Collins & Lanier, 2021). This gap may be partially related to a lack of therapist resources, as well as to barriers that prevent parents from seeking treatment for their children (e.g. transportation, work demands, childcare, stigma, and parents who want to help their child themselves; Bringewatt & Gershoff, 2010; Thurston & Phares, 2008).

Stepped care models can help make treatment more available. These models begin with a lower-intensity intervention, followed by higher-intensity treatment if necessary (McDermott & Cobham, 2014). A stepped care model that is specifically for traumatised children has been developed: stepped care cognitive behavioural therapy for children after trauma (SC-CBT-CT; Salloum et al., 2014; aged 7–12 years). The first step in SC-CBT-CT is parent-led with a therapist providing the parent assistance and support (Step One). If the parent-led intervention does not improve the child`s symptoms sufficiently, the treatment can be stepped up to standard therapist-led trauma-focused cognitive behavioural therapy (TF-CBT; Cohen et al., 2017; Step Two). Shifting the task of leading the treatment from therapists to parents makes it possible to address key treatment barriers, such as a lack of therapist resources and challenges in seeking treatment, by facilitating parents in helping their children themselves and by giving parents more control of the time and place of the treatment.

To date, SC-CBT-CT shows comparable results to TF-CBT in reducing PTSS, and cost–benefit analyses have estimated that the cost is reduced by 54%, with a 47% response rate to Step One for intention to treat, and 71% for completers (Salloum et al., 2022). Overall, parents are typically satisfied with the SC-CBT-CT treatment (Salloum et al., 2016), and many have reported that leading their children's treatment in SC-CBT-CT empowered them and improved their relationship with their children (Fagermoen et al., 2023). However, there is currently little data about how parental factors are related to treatment completion and treatment response of the parent-led Step One. This knowledge is critical to inform treatment development and gain a better understanding of which parents this task-shifting model is most suitable for.

Parent-led treatments are in an early phase of development, and only one study has investigated the role that parent factors play in Step One (Salloum et al., 2022; Salloum et al., 2022). That study provides some evidence that certain sociodemographic variables, such as being of a younger age, having a minority background, achieving a lower level of education, and having a lower income are related to poorer treatment outcomes. In addition, caregiver depression was a predictor of non-response (Salloum et al., 2022; Salloum et al., 2022). In support of this finding, caregiver psychopathology has also been found to influence completion and the child`s recovery in therapist-led trauma treatments. For instance, caregiver depression (Danzi & La Greca, 2021; Martin et al., 2019), symptoms related to a parent's own trauma (Canale et al., 2022; Danzi & La Greca, 2021; Martin et al., 2019), distress and feelings of guilt and worry (Cinamon et al., 2021; Holt et al., 2014), and parenting stress (Clark & Nadeem, 2022; Whitson & Kaufman, 2017) have all been found to be related to poorer outcomes in therapist-led trauma treatments.

Another factor that may influence treatment completion and treatment response is perceived social support, as this has been theorised to buffer emotional distress (Thoits, 2011). Several empirical studies have found that the level of a caregiver's distress after their child experiences trauma is negatively related to perceived social support (Deblinger et al., 1993; Hiebert-Murphy, 1998). Furthermore, social support may also help parents, despite their own challenges, support their children during therapy. In Step One, parents are strongly encouraged to mobilise their social support system during the treatment phase, but there is little knowledge of whether this has an actual effect on either treatment completion or how a child responds to the treatment.

Finally, practical barriers may also influence caregivers’ ability to facilitate completion of a treatment (Kazdin et al., 1997). These include the time and cost of transportation to treatment, work demands, childcare, and difficult and time-consuming life transitions like moving to a new home (Bornheimer et al., 2018; Bringewatt & Gershoff, 2010).

1.1. Research aims and hypotheses

The aim of this study is twofold. The first is to investigate baseline parental factors associated with treatment completion of the parent-led Step One. The second is to investigate parental factors associated with treatment response among those children who completed Step One. Based on the currently available research, we expected higher baseline levels of parents’ anxiety and depression, a parent's own stressful life experiences and post-traumatic symptoms, having negative emotional reactions to their child's trauma, parenting stress, and lower levels of perceived social support to be associated with non-completion and non-response. Finally, we expected barriers related to time, resources, and other obligations to be specifically associated with non-completion.

2. Method

2.1. Participants

The participants in this study were children and their caregivers who participated in an open trial that investigated the feasibility of SC-CBT-CT in 11 first-line mental health services that provide short-term treatment for children in Norway. The final sample consisted of 82 child-caregiver dyads. Of that total, 80 had completed the caregiver baseline assessments and were included in the present analysis. Inclusion criteria: 1) age 7–12 years; 2) exposure to one or more potentially traumatic events according to DSM-5 (American Psychological Association, 2013) after age 3; and 3) experience of at least five DSM-5-defined PTSS, including at least one symptom of both re-experiencing and avoidance. Exclusion criteria: 1) the participating caregiver was the perpetrator, or the child was living with a perpetrator; 2) indications of child or parent psychotic symptoms, cognitive disability, active suicidal thoughts, or other conditions that could limit the child's or parent's ability to complete the workbook; 3) the need for an interpreter; 4) parental substance abuse (past three months); or 5) the child was receiving concurrent trauma-focused psychotherapy. See Skjærvø et al., in review, for more details on the study, as well as Tables 1 and 2 for detailed information about the participant children's and caregivers’ characteristics, respectively.

Table 1. Sociodemographic Child Characteristics at Baseline (T1).

Sociodemographics	All participants (n = 82)
	n	%	Mean	SD	Range
Child gender
Female	46	56.1
Male	36	43.9
Child age, mean (SD, range)	82		9.91	1.45	7–12
Child region of birth
Born in Nordic country	77	96.3
Born in non-Nordic country	3	3.8
Child number of trauma	82		3.81	2.11	1–10
Child post-traumatic stress symptoms (CATS 2.0)^a	82		28.96	7.68	13–54
Child index trauma
Violence or witnessing violence in the home	23	28.0
Violence outside the home	14	17.1
Severe and threatening bullying	14	17.1
Illness/sudden death of loved one	14	17.1
Sexual assault	5	6.1
Frightening separation from parent^b	5	6.1
Injury/accident	4	4.9
Natural disaster	3	3.7

Note. N = 82. CATS 2.0 = Child and Adolescent Trauma Screen. Missing data: n = 2 missing on child region of birth. ^aCut off CATS 2.0 = 15. ^bE.g. getting lost in the woods, parent being arrested by police in the home.

Table 2. Sociodemographic Caregiver Characteristics at Baseline (T1).

Sociodemographics	All participants (n = 80)
	n	%	Mean	SD	Range
Caregiver age	80		40.89	6.98	29–66
Caregiver participant in therapy
Biological father	17	21.3
Biological mother	60	75.0
Grandmother/foster mother	3	3.7
Caregiver region of birth
Born in Nordic country	72	93.5
Born in non-Nordic country	5	6.5
Caregiver level of education
High school or lower	36	45.0
University	44	55.0
Caregiver occupational situation
Working/studying	64	80.0
Not working/studying	16	20.0
Caregiver relationship status
Partner	52	65.8
No partner	27	34.2
Caregiver life events/trauma (participants could select more than one type, and thus numbers sum to more than 100%)
Physical violence	39	48.8
Psychological violence	32	40.0
Sexual assault	23	28.7
Illness/sudden death of someone close	29	36.3
Natural disaster/accident	12	15.0
Other	40	50.0
Anxiety/depression (SCL-10S)^a	79		1.64	0.53	1–3.40
Stressful life events (SLESQ)	80		2.85	2.69	0–10
Post-traumatic stress symptoms (PCL-5)^b	58		15.66	14.23	0–57
Parental emotional reactions (PERQ)	79		33.44	10.05	14–65
Parenting stress (PSS)	79		33.18	7.38	18–48
Perceived social support (MSPSS)	79		5.69	0.94	3.17–7.00
Barriers to treatment completion	80		0.89	0.90	0–3.00

Note. N = 80. SCL-10S = Symptom Check-list-10; SLESQ = Stressful Life Events Screening Questionnaire; PCL-5 = PTSD Checklist for DSM-5 (this was only administered to caregivers who indicated exposure to at least one stressful life event in accordance with the SLESQ); PERQ = Parental Emotional Reaction Questionnaire; PSS = Parental Stress Scale; MSPSS = Multidimensional Scale of Perceived Social Support. Missing data: n = 2 had missing caregiver data for the entire baseline assessment, an additional n = 1 had missing data for caregiver relationship status, SCL-10S, PERQ, PSS, and MSPSS; and an additional n = 3 had missing data for caregiver region of birth. ^aCut off SCL-10S = 1.85. ^bCut off PCL5 = 31–33.

2.2. Step one

Step One builds on the treatment principles of standard TF-CBT (Cohen et al., 2017). Step One consists of three therapist-led parent–child therapy sessions, 11 parent-led parent–child meetings over 6–9 weeks, and weekly phone support from the therapist to the parent. In the parent-led parent–child meetings, the child, and the participating parent work together at home on tasks from the Stepping Together workbook (Salloum et al., 2010). The tasks in the workbook focus on building coping skills and having the child complete exposure tasks, including developing a trauma narrative and in vivo exposure to trauma reminders. Families who are unable to complete the workbook are offered an early step up to therapist-led treatment. Children who meet responder criteria after completing the workbook (mid-treatment) continue to a 6-week maintenance phase followed by the final responder assessment (post-treatment). Children who do not meet the responder criteria at mid-treatment and/or post-treatment are stepped up to Step Two, which consists of standard therapist-led TF-CBT treatment. In the original version of SC-CBT-CT, both Step One and Step Two are provided by the same therapist (Salloum et al., 2014). In the current study, we tested an adapted version of SC-CBT-CT where Step One was provided in the first-line mental health services, and Step Two was provided at the specialised child and adolescent mental health services.

Nineteen therapists assisted in administering Step One treatment to the child participants. Skjærvø et al., in review, provides more details about training of the therapists. All therapist-led sessions were audiotaped, and trained supervisors assessed fidelity based on a checklist that Dr Salloum developed. Model fidelity was high (above 95%), the agreement between the ratings exceeded 95%, and all cases were approved.

2.3. Procedures

The Regional Committees for Medical and Health Research Ethics provided ethical approval for this study (ref. 2018/771/REK sør-øst). The 19 therapists pre-screened the children for participation. Prior to the baseline assessment, caregivers provided written informed consent to take part in the study, and the children provided their informed assent to participate. Both children and caregivers completed questionnaires at baseline (pre-treatment), after completing the workbook of Step One (mid-treatment) and after the maintenance phase (post-treatment). All data were collected and stored safely.

2.4. Measures

2.4.1. Demographics

We included information about each parent’s gender, age, region of birth, highest level of educational achievement, current occupational status, biological relationship to their child, and relationship status.

2.4.2. Caregiver measures

Parents’ anxiety and depression symptoms were assessed using the Symptom Checklist-10 (SCL-10S; Strand et al., 2003). Ten items, of which four measure symptoms of anxiety and six measure depressive symptoms, are scored from 1 (not at all) to 4 (extremely), giving a mean score ranging from 1 to 4 (current study α = .88). A score above 1.85 indicated clinically significant symptoms.

Exposure to stressful events was measured with the Stressful Life Events Screening Questionnaire (SLESQ; Goodman et al., 1998). Responses on the list of 15 potentially traumatising events are given as yes or no.

Parents’ PTSS were assessed with the PTSD Checklist for DSM-5 (PCL-5; Weathers et al., 1993). Twenty items are scored 0 (not at all) to 4 (extremely), giving a sum score ranging from 0 to 80 with a cut-off score of ≥ 31 indicating possible PTSD (current study α = .95).

Parents’ emotional reactions to their children's trauma were assessed with the Parental Emotional Reaction Questionnaire (PERQ; Mannarino & Cohen, 1996). We used a modified 14-item version of the PERQ; one item from the standard 15-item version was not applicable to the majority of the caregivers, so it was dropped (‘I feel guilty that I did not know about the trauma sooner’). Responses are given on a scale from 1 (never) to 5 (always), and the summed score ranges from 14 to 70 (current study α = .91).

Parenting stress and rewards were measured using the 18-item Parental Stress Scale (PSS; Berry & Jones, 1995). The items are scored from 1 (strongly disagree) to 5 (strongly agree), with a higher sum score (range 18–90) indicating higher stress (current study α = .78).

Parents’ perceived social support from family, friends, and significant others was measured with the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., 1988). Twelve items are scored from 1 (strongly disagree) to 7 (strongly agree), giving a mean range from 1–7 (current study α = .92).

To assess potential barriers to completing the treatment, we used nine items from the External Demands Subscale of the Barriers to Treatment Participation Scale (illness, economic difficulties, moving, travel, obtaining childcare/a babysitter, caregiver's work schedule, a child's school or leisure activities, disagreements between parents regarding the child's treatment needs, and other; Colonna-Pydyn et al., 2007). Items related to the cost of treatment were removed since the treatment is free of charge in Norway. The number of barriers present for each parent was summed, giving a score on a range from 0–9.

2.4.3. Criteria for non-completion and non-response treatment outcome groups

Non-completion of Step One was defined as not completing the three in-office therapist-led sessions, the workbook, and the mid-treatment assessment.

Treatment non-response to Step One was defined in line with Salloum et al., 2022, by not meeting the following responder-criteria 1) the child reports ≤4 DSM-5 defined symptoms of PTS with an intensity of 2 (often) or higher, using the 20-item validated Child and Adolescent Trauma Screen (CATS 2.0; Sachser et al., 2022) and 2) the therapist rates the child's global improvement (Clinical Global Impression – Improvement [CGI-I]; Guy, 1976) to be improved, much improved or symptom-free on a scale from 1 (symptom-free) to 8 (very much worse).

2.5. Analysis

We investigated the demographic differences between completers and non-completers and responders and non-responders using t-tests and chi-square tests, as well as non-parametric equivalents where indicated. Furthermore, we used logistic regression analyses to investigate the relationship between the dependent variables (non-completion and non-response) and the caregiver factors. Due to low n, all models were unadjusted. To account for the nested nature of the data, with children nested with therapists, we initially attempted to use hierarchical logistic regressions, which allow for random effects of differences between therapists. We entered therapists as the second level of the model. However, the models became unstable, and the width of the confidence intervals rendered the models uninformative. Therefore, we only conducted single-level analyses. All scales were converted into standard measures of 0–10 to simplify the process of interpreting the results of scales with different ranges. In six cases where children met symptom criteria for stepping up at mid-treatment (per-protocol), the therapists based on their clinical judgment and discussions with the families, sent the children to the maintenance phase instead of stepping them up. Of these six cases, four continued to improve and met responder criteria at post-treatment. Being a feasibility trial, we included these six cases as the therapists classified them. For comparison, we have included additional per-protocol analyses in the supplementary files.

The level of statistical significance was set at p < .05. We computed hierarchical logistic regressions using R (Hornik, 2018), while we conducted the remaining analyses using International Business Machines (2021) SPSS, version 28.

3. Results

3.1. Descriptive statistics

Based on predefined criteria, 62 of the 80 dyads that were included in the analyses completed Step One. The other 18 dyads were either stepped up early (n = 9) or dropped out (n = 9) and were all categorised as non-completers. Reasons for early step up and drop out were related to child or parent refusing to continue, parent lack of time/resources, or family conflicts. Of the completers, 51 responded to the Step One treatment, 10 did not respond to Step One treatment and were stepped up, and one dropped out because of therapist absence due to illness. Parents of non-completers tended to have lower levels of educational achievement, x² (1, N = 80) = 13.79, p < .001. There were no significant differences between completers versus non-completers and responders versus non-responders when it came to parents’ genders, ages, regions of birth, occupational situations, relationships to their children, and marital statuses. Consequently, these variables were not included in the tables for logistic regressions.

3.2. Predictors of non-completion

Non-completion was more likely in cases where the participating parent's educational attainment was less than university level. For the remaining variables related to parental mental health and stress, social support, and the number of practical barriers to treatment, the results were inconclusive, showing no statistically significant associations with non-completion (see Table 3).

Table 3. Comparison of Baseline Parental Factors for Treatment Non-Completers and Completers With Unadjusted Odds Ratios (OR) and 95% Confidence Intervals (CI) From Dichotomous Logistic Regressions.

Parental factors	n (%) or mean (SD, range)		OR (95% CI)	p-value	n
	Non-completers (n = 18)	Completers (n = 62)
High school education or less	15 (83.3)	21 (33.9)	0.102 (0.027–0.394)	<.001*	80
Anxiety/depression (SCL-10S)	1.71 (0.64, 1.00–3.40)	1.62 (0.50, 1.00–3.20)	1.102 (0.827–1.467)	.507	79
Stressful life events (SLESQ)	3.33 (2.57, 0–9)	2.71 (2.72, 0–10)	1.133 (0.854–1.502)	.386	80
Post-traumatic stress symptoms (PCL-5)	16.00 (19.16, 0–57)	15.54 (12.13, 0–46)	1.019 (0.736–1.409)	.911	58
Parental emotional reactions (PERQ)	35.39 (12.35, 19–65)	32.87 (9.30, 14–57)	1.159 (0.851–1.578)	.350	79
Parenting stress (PSS)	32.72 (7.21, 19–45)	33.31 (7.49, 18–48)	0.924 (0.551–1.550)	.765	79
Perceived social support (MSPSS)	5.73 (0.86, 4.42–7.00)	5.67 (0.96, 3.17–7.00)	1.042 (0.744–1.467)	.812	79
Barriers to treatment completion	0.89 (0.90, 0–3)	0.92 (0.91, 0–3)	0.966 (0.569–1.642)	.899	80

Note. The sample size for each parameter varies due to missing values in the data set. *p < .05.

3.3. Predictors of non-response

Non-response was associated with parents who showed greater emotional response to their children's trauma and who had higher levels of perceived social support. For the remaining variables related to parental mental health and stress, the results were inconclusive, showing no statistically significant associations with non-response (see Table 4). In the per-protocol analyses, the association between emotional responses to the child's trauma and non-response was not statistically significant (Supplementary Table 1).

Table 4. Comparison of Baseline Parental Factors for Treatment Non-Responders and Responders With Unadjusted Odds Ratios (OR) and 95% Confidence Intervals (CI) From Dichotomous Logistic Regressions.

Parental factors	n (%) or mean (SD, range)		OR (95% CI)	p-value	n
	Non-responders (n = 10)	Responders (n = 51)
High school education or less	3 (30.0)	18 (35.3)	1.273 (0.293–5.532)	.748	61
Anxiety/depression (SCL-10S)	1.58 (0.53, 1.00–2.30)	1.63 (0.51, 1.00–3.20)	0.941 (0.619–1.429)	.774	60
Stressful life events (SLESQ)	1.90 (2.38, 0–7)	2.86 (2.81, 0–10)	0.792 (0.504–1.247)	.315	61
Post-traumatic stress symptoms (PCL-5)	10.45 (8.93, 3–25)	16.21 (12.61, 0–46)	0.698 (0.338–1.441)	.331	41
Parental emotional reactions (PERQ)	38.50 (11.21, 26–57)	31.82 (8.66, 14–55)	1.589 (1.005–2.512)	.047*	60
Parenting stress (PSS)	33.10 (5.07, 26–44)	33.32 (7.99, 18–48)	0.972 (0.505–1.869)	.932	60
Perceived social support (MSPSS)	6.33 (0.75, 5.25–7.00)	5.55 (0.96, 3.17–7.00)	1.947 (1.081–3.507)	.026*	60

Note. The sample size for each parameter varies due to missing values in the data set. *p < .05.

4. Discussion

The parent-led Step One of SC-CBT-CT is developed to address treatment barriers and make treatment more available. However, since this treatment involves shifting the therapeutic intervention tasks from the therapist to the parent, parents may find the treatment to be demanding. Thus, not all parents may be able to lead the treatment in a way that results in positive outcomes for their children. This study has examined whether parent factors are associated with non-completion and non-response to Step One.

In line with results from previous studies of Step One of SC-CBT-CT (Salloum et al., 2022; Salloum et al., 2022), we found that a lower level of parental educational achievement was related to non-completion. Similar results have been found in studies of TF-CBT (Wamser-Nanney & Steinzor, 2017), suggesting that this factor generally leads to challenges with treatment completion. It is possible that in our study, parental education level acted as a proxy for other, unmeasured factors, such as resources, income, or factors related to health literacy and challenges in handling health information or communication with healthcare personnel (Le et al., 2021). Perhaps lower levels of parental education or related factors could have more of an impact on treatment completion in a model that demands significant parental involvement in the treatment process. Clinicians may need to devote greater effort to assessing a parent's individual needs and resources to facilitate completion of the treatment, and future studies should investigate whether there are specific aspects of Step One, for instance the clarity of the workbook, that can be improved to increase completion rates.

In contrast to a previous study of Step One of SC-CBT-CT (Salloum et al., 2022) and studies on TF-CBT (Canale et al., 2022; Danzi & La Greca, 2021; Martin et al., 2019), the results of the current study showed that parents’ symptoms of anxiety, depression, and PTS neither predicted non-completion nor non-response. One reason for this could be that we used the SCL-10S, which includes symptoms of both anxiety and depression, while studies that have found a link between parental depression and treatment outcomes have typically used measures that only assess depression (Martin et al., 2019). Still, our results do not raise immediate concerns that having symptoms of anxiety, depression, or PTS would hinder parents in their participation in Step One. Maybe the structure of Step One with a well-prepared sequence of intervention tasks, a clear expectation to perform them, and weekly follow-up phone meetings with the therapist, have helped parents who lacked initiative and appeared to be disengaged as a result of their depression (Acri et al., 2018) to implement the treatment. It may also be that caregivers who struggle with mental health challenges on their own are even more invested in helping in their child's treatment since they understand the effect that such problems can have (Self-Brown et al., 2016). Also, receiving continuous support from a therapist may assist a parent in focusing on helping their child and setting aside the challenges stemming from their own symptoms.

Another possibility is that the parents learned strategies to cope with their own PTSS when using the workbook with their child. A main focus of Step One is exposure to trauma reminders, and gradual exposure throughout the entire treatment assists in addressing and remedying avoidance. Parents also receive psychoeducation regarding trauma reactions. Therefore, parental behaviours that are related to a parent's own PTSS may be moderated by taking part in the treatment. However, firm conclusions cannot be made since the absence of evidence does not equal evidence of absence (Altman & Bland, 1995). Future studies should thus replicate the analyses with a larger sample.

Although the results showed that the level of parenting stress was not related to any particular outcome, we did find that among parents, greater emotional reactions to their children's trauma was a predictor of non-response. This is in accordance with results from previous studies on TF-CBT (Cinamon et al., 2021; Holt et al., 2014). However, these results were significant only in the naturalistic analyses and not in the per-protocol analyses. This difference could be a real difference between the analyses, but it is also possible that it is due to differences in power. We could not find a clear pattern of characteristics of the families where the protocol to step up was not followed, but we cannot rule out that the therapists judged these parents as better functioning. The results should therefore be interpreted with caution. Although the emotional reactions did not seem to influence the completion of the treatment, there might have been more subtle differences in how the parents performed the tasks that may have limited the efficacy of the treatment. For instance, a parent who is deeply worried about their child's trauma might be afraid to upset their child. Although parents learn about the importance of trauma exposure as part of the Step One process, it can be difficult for parents to initiate trauma-talk with their children and complete the exposure tasks, which would likely cause their children stress. Parents with such concerns may therefore have performed these important tasks during the treatment less thoroughly than parents who were less affected by their children's trauma, perhaps modelling avoidance for their children, which again could maintain their children's post-traumatic reactions. This may have added to their feelings of distress for not being able to help their child.

Thus, clinicians should be aware that parents may need help learning how to cope with their own feelings of distress, guilt, and shame as a response to their children's trauma experiences. Future studies should examine whether these parents could benefit from processing their own reactions. Furthermore, explaining the treatment rationale may be of particular importance for parents who have emotional reactions to their children's trauma to ensure that, despite these reactions, they are able to carry out the treatment with their children. They may also need more therapist support or extra therapist-led sessions.

Contrary to our expectations, we found that a higher level of perceived social support among parents was related with a greater likelihood of non-response. Other studies have found that having social support resources prevents parental distress from having a negative effect on their children's trauma (Deblinger et al., 1993; Hiebert-Murphy, 1998). We can only speculate on what could cause this association between greater parent social support and poorer child outcomes. It may be that unmeasured confounders, for instance a difficult life-situation, increased both availability of support, and barriers to utilising this support, such as fear of not being understood or of taking up too much of others’ time (Charuvastra & Cloitre, 2008; Thoresen et al., 2014). There is also evidence that type of support matters, for instance that emotional support has more of an impact on buffering distress compared to practical support (Thoits, 2011). Future studies would benefit from a larger sample size to include confounders, and a more detailed assessment of types of support to further the understanding of the relationship between parents’ social support and how children respond to parent-led treatment.

4.1. Study strengths and limitations

The current study is the first to examine predictors of non-completion and non-response of the parent-led Step One of SC-CBT-CT conducted by a research group independent of the developers. Furthermore, this is the first study to use a naturalistic sample from first-line mental health services. However, the study has some limitations that warrant mention. First, we may have failed to identify predictors due to the relatively small sample size. Second, ideally, we would have included several predictors in the same models; however, the relatively small groups of non-completers and non-responders limited our analyses to unadjusted models. Third, our attempt to control for therapist effects using hierarchical analyses resulted in unstable models. Therefore, we only used single-level analyses. Fourth, the finding related to parents’ emotional reactions was significant in the naturalistic analyses only and not in the per-protocol analyses. Although we assume the change in sample size is responsible for the discrepancy in findings, this underlines the need for confirming these results in a larger sample. Last, the parents who participated in the study were less diverse in terms of education and ethnic background compared to the general population. Thus, future studies should seek to replicate our findings using other samples of child–parent dyads.

5. Conclusion

The parent-led Step One of SC-CBT-CT addresses the gap between the need for and access to evidence-based treatments. We found that the majority of the children were able to complete the treatment with good results and that their parents` symptoms of anxiety, depression, and PTS, as well as parenting stress, did not seem to affect their children's outcomes. Non-completion was related to parents whose highest level of education was below a university level, while non-response was associated with parents who showed stronger emotional responses to their children's trauma and who had higher levels of perceived social support. Further research on how these parental factors may predict or influence treatment outcomes in parent-led treatment could be used to tailor and select appropriate treatments for trauma-exposed children.

Acknowledgement

EMF had the main responsibility for conducting the analysis and drafting the paper. SMO and IS were the leaders of the project and initiated the data collection. SMO, TKJ, IS and EMF contributed to the conception and design of the study. All the authors contributed to the manuscript revision and approved the final version.

We would like to thank the children, caregivers and therapists participating in the study.

Data availability

The data are not readily available due to the small and potentially identifiable dataset. Data may be made available upon request after the participant list has been deleted by the project leader.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was supported by the Norwegian Directorate of Health and Foundation Dam.