Family violence is usually perceived as including mainly child abuse or intimate partner violence. However, child-to-parent aggressive behavior is another understudied facet of domestic violence. Although it has received limited attention until recent years, it is far from being an uncommon phenomenon (1–4). Prevalence rates of child-to-parent violence (CPV), defined by aggression perpetrated at least once in the last year, range from 5 to 21% for physical violence, and 33 to 93% for psychological violence (3, 5). The variety of definitions and measurements used to describe CPV, may explain high discrepancies in reported prevalence (4). Data suggest that adolescents displaying aggressive behavior toward their parents also have violent behavior outside the family (6). Risk factors reported for CPV are similar to those of overall aggressive behavior in adolescents: low socioeconomic status, coercive education, child neglect and abuse and intimate partner violence (2, 7–10).

The present study focuses on a specific type of CPV described as severe tyrannical behavior (STB) (11). STB in children is described as “Oppositional Defiant Disorder with family tyranny” by several clinicians (12, 13) because oppositional defiant disorder (ODD) symptoms in those children are confined to only one setting: the home. Intra-familial ODD (IODD) is considered a mild form of ODD by the DSM-5. STB is a family scenario characterized by an inversion of the family hierarchy, with the child dominating and controlling the family dynamics through child-to-parent psychological and sometimes physical violence. Parents gradually lose control over the situation and accommodate the difficult behavior, by sacrificing their well-being and basic needs, in fear of their own child’s reactions. Any attempt to resist the child’s behavior may trigger violence, tantrums, suicidal threats, or self-harm, leaving the parents helpless and desperate. Children and adolescents with STB show strong resistance to change and poor help-seeking behaviors. The STB family scenario is often disclosed late in the help-seeking process because parents feel ashamed and fear social judgment. Moreover, children with STB preferentially abuse their parents and close family members. Their normal social behavior outside home differentiates them from children with generalized forms of disruptive behavioral disorders such as oppositional defiant and conduct disorders (11). This accentuates the fact that parents rarely feel believed, and suffer isolation and guilt. In addition, while STB is closer to a family pattern than to a formal diagnosis, it is frequently associated with a variety of psychiatric conditions such as attention deficit hyperactivity disorder (ADHD), mood and anxiety disorders, and autism-spectrum disorders (7, 14, 15). Additionally, in the STB scenario, these conditions may be overlooked because of parental accommodation and fear of social stigma in both children and parents. STB shares features with adolescents at an “invisible” risk for psychopathology and suicidal behavior identified by Carli et al. (16). STB has also been associated with failure to successfully go through the developmental stage of emerging adulthood as it fosters dysfunctional dependence and entitlement and has a long-term impact on functioning (17).

Mental health professionals dealing with STB scenarios face various challenges: they have to encourage parents’ communication about family violence, screen for underlying psychiatric conditions in children and build parental skills to change the family dynamics.

Since children with STB usually lack motivation to engage in therapy, we have set up a parent training program based on Non-Violent Resistance (NVR) strategies, with the aim of influencing parental behavior as a key feature in the therapeutic process. NVR has been developed by Haim Omer to help families facing problematic behaviors (18–23) and entrenched dependence in their children (24). NVR parent training focuses on non-escalating coping responses to the child’s violence, de-accommodation, and self-control. Moreover, the program helps parents to step out from secrecy and social isolation by building a support network.

NVR represents an innovative approach in psychotherapy that fosters societal adaptations to parent–child relationships and promotes a new model of parental authority. Specifically, it has been proven effective in supporting parents of children with ADHD (25) and anxiety disorders (26), self-destructive or aggressive behaviors of youth (27, 28) and highly dependent young adults (29). Despite these promising findings, there are no data on NVR therapy involving parents of children with STB, a distressed parent population affected by social stigma from a perceived loss of parental authority and in need of efficient coping strategies (13).

The current study evaluates whether the NVR program, a protocolized 10-session parental-group intervention, was more effective in reducing stress in parents of children aged 6–20 years and displaying STB compared to a treatment as usual (TAU) intervention that provided supportive counseling and psychoeducation. Since this program has not been previously evaluated in this population, the main outcome of this study is efficacy of the NVR program, using the Parenting Stress Index (PSI) (30). Additionally, the study aims to investigate the efficacy of NVR on parental anxiety and depression and children’s behavioral symptoms as secondary outcomes.

This was a single-blinded, randomized, parallel group superiority trial that compared the NVR intervention group with a TAU group during a 4-month period. After this timeframe, parents of the TAU group were also offered to participate in the NVR program. The study was conducted within the context of a clinical treatment aimed at providing a non-pharmacological intervention for parents of children with STB. All screened youth and families were offered services regardless of the study participant agreement. The study was approved by the ethical committee (Comité de Protection des Personnes Ouest III).

The study recruitment occurred from July 2017 to March 2019 with follow-up interviews completed by July 2019.

Eighty two parents of youth aged 6–20 years were enrolled by the Child and Adolescent Psychiatry Department in Saint-Eloi University Hospital in Montpellier (France). Inclusion criteria were: (a) being the parent of a child aged 6–20 years who endorses the DSM-5 criteria of IODD; (b) two or more positive answers to the custom questions routinely used to identify STB in children/adolescents at the University Hospital of Montpellier, related to a 12-month period:

(c) being able to attend at least half of the sessions.

Exclusion criteria were the following: (a) predicted absence from at least half of the sessions; (b) subject deprived of liberty, (c) subject under guardianship or curatorship, and (d) lack of written informed consent from parent.

Participants were recruited at both local and national levels using different methods, including electronic media, print advertising, internal volunteer lists and physician referrals.

Eligible parents were informed by their referent mental health professionals about this study. In other cases, an STB patient organization provided information about the study to their members and these parents directly applied for the study. Once the clinician identified eligible parents and children, they received detailed written and verbal information about the study. After signing informed consent, parents meeting the criteria were invited for the inclusion visit.

The inclusion visit was conducted by a trained research assistant in charge to complete baseline characteristics. The visit also aimed at assessing the child’s medical and psychological history, as well as STB and how it was shaping the family dynamics. The NVR program was explained to participants based on a manual published in 2017 and developed by the principal investigator of this study, Franc and Omer (11). After consent, participants were randomized to the NVR or TAU condition.

Parents in the NVR group received the 10 2-h sessions, over a period of 4 months. The NVR protocol was provided by clinicians trained in the program. TAU received non-pharmacological and pharmacological treatments as usually provided. All participants were interviewed by independent research assistants and completed their assessments at baseline and following their last treatment session (4 months from baseline). Only the participants of NVR were asked for a 4-month post-treatment evaluation (8 months from baseline), thus excluding a follow-up comparison with TAU.

Both parents were allowed to follow the program, but only one of them could complete the questionnaires.

Independent evaluators, blinded to group assignment, administered the self-report measures and conducted the semi-structured interviews listed in Table 1 at pretreatment (inclusion visit, V0), completion (4 months from baseline, V1) and 4-month follow-up (8-month from baseline, V2). All clinician/parent/teacher-rated measures were completed at V0 and V1, unless otherwise noted. Only the NVR group, completed the V2 measures.

The primary endpoint variable was the change from baseline (inclusion visit) to completion of treatment (V1) in the total score of the Parenting Stress Index/Short Form (PSI-SF) (30). We hypothesized that the decrease in PSI-SF total score between baseline and end of treatment would be superior in NVR compared to TAU. Secondary outcomes concern the children’s behavior, parental anxiety and depression by examining pre-post changes and between-group changes at V1. We also examined changes in parental and child variables at 8 months from baseline and 4 months after completion of the NRV program (V2).

Parenting Stress Index/Short Form (PSI-SF) is a short version of the PSI by Abidin in 1995 (30) and is a 36-item, self-report measure of parenting stress. The French version was translated and validated by Bigras et al. (31). The PSI-SF has three subscales: Parental Distress, Parent–Child Dysfunctional Interaction, and Difficult Child. Each subscale consists of 12 items rated from 1 (strongly disagree) to 5 (strongly agree). A total score is calculated by summing the three subscales scores, ranging from 36 to 180. Scores of 90 or above indicate a clinical level of parental stress. Traditionally, the PSI-SF has been used to measure parenting stress in parents from clinical and high-risk populations (32–36), and to measure treatment effectiveness (37, 38). The PSI-SF was administered at baseline, at completion and for the NVR group also at follow-up.

The Hospital Anxiety and Depression Scale (HADS) (39) is a 14-item measure designed to assess anxiety and depression symptoms in adults. Items are rated on a 4-point severity scale. The HADS produces a scale for anxiety (HADS–A) and for depression (HADS–D). To screen for parental anxiety and depressive symptoms, the following interpretation can be proposed for each of the scores: scores of 7 or less are considered no symptomatology; scores between 8 and 10 are considered doubtful symptomatology; scores of 11 and more are considered definite symptomatology. The French version was validated by Duquette (40). The HADS was administered at baseline, at completion and for the NVR group also at follow-up.

The parent/teacher-reported Strengths and Difficulties Questionnaire (SDQ) is a brief behavioral screening questionnaire used in clinical and research settings. The SDQ measures emotional and behavioral symptoms, and their impact on areas of functioning, called the impact score (43). The SDQ symptom part consists of 5 subscales, each containing 5 items. The scales measure emotional symptoms, conduct problems, hyperactivity-inattention, peer relationship problems, and prosocial behaviors. Cut-off scores are: total difficulties: “borderline” = 14–16; “abnormal” ≥ 17; emotional symptoms: “borderline” = 4; “abnormal” ≥ 5; conduct problems: “borderline” = 3; “abnormal” ≥ 4; hyperactivity-inattention: “borderline” = 6; “abnormal” ≥ 7; peer relationship problems: “borderline” = 3; “abnormal” ≥ 4; prosocial behaviors: “borderline” = 5; “abnormal” ≤ 4; impact scores (impact, family, learning, leisure activities, friendship and child’s distress): “borderline” = 1; “abnormal” ≥ 2.

SDQ has become an internationally recognized tool which is extensively used in both French research and clinical settings (44, 45). The parent-rated and teacher-rated SDQ were administered at baseline, at completion and for the NVR group also at follow-up.

The analysis of the primary and the secondary outcomes were performed on the Full Analysis Set (FAS) that will include all subjects who are randomized and have a valid primary efficacy measurement.

The baseline characteristics of children and parents was described with proportions for categorical variables and with means and standard deviations (SD) for quantitative variables. The comparability of the 2 groups was checked for all these baseline characteristics.

The change from pre- to post-treatment on the main and secondary outcomes were compared between groups using a two-sample t-test. Changes in PSI-SF, CBCL and SDQ scores for the NVR group were also analyzed using mixed linear models with a subject-specific random intercept. The fixed effect was the visit or time effect (at baseline, 4- and 8-months). The absolute mean or median difference and its 95% confidence interval were calculated using Hodges-Lehmann Method. The estimated Cohen’s d effect size and its precision 95% confidence interval (CI) were calculated for the main and secondary outcomes. The effect sizes magnitude were classified as: ≥ −0.15 and < 0.15 “negligible,” ≥ 0.15 and < 0.40 “small,” ≥ 0.40 and < 0.75 “medium,” ≥ 0.75 and < 1.10 “large,” ≥ 1.10 and < 1.45 = “very large,” and ≥ 1.45 “huge” (47).

In order to study the persistence of the effect on parental stress, anxiety and depression, the changes of the PSI-SF and HADS between baseline and completion in both groups, and baseline-follow up and completion—follow up in the experimental group was analyzed using paired t-test or Paired Wilcoxon test.

Statistical analyzes were implemented using SAS (Enterprise Guide, version 8.2; SAS Institute; Cary, North Carolina, United States) (48) by the Clinical Research and Epidemiology Unit of the Montpellier University Hospital. A two-sided value of p < 0.05 was considered to indicate statistical significance.

Among the 82 families included nine participants were excluded following randomization and baseline assessment and did not return for treatment sessions or additional research interviews (participants’ flow chart is presented in Figure 1). In both groups the delay of randomization and the initiation of intervention was of 2 months. Six families from NVR (14.3%) and 3 from TAU (7.5%) discontinued the study (p = 0.05) for mainly medical and logistical reasons (NVR: one did not come back after multiple calls, two of them had an improvement/severe deterioration of symptoms, three had organizational or travel-related mobility problems; TAU: two did not reach out after multiple calls, one had organizational issues). A total of 73 participants completed the study and were analyzed (36 NVR and 37 TAU) with a mean (± SD) participation rate in the NVR sessions of 76% (± 28.6) for mothers and 25% (± 35.5) for fathers, considering both parents could be present at the session (Figure 2B). The adhesion rate stayed high throughout the whole intervention period, despite the length of the program (see Figure 2A). The majority of family members participating and completing the assessments were mothers in both groups (Figures 2B, 3). Since father attendance was not sufficient to allow a comparative analysis between groups, we used and compared only mothers’ results (when both questionnaires were available, we used mothers’ scores).

The socio-demographic and clinical results of parents and children are presented in Table 3. There were no socio-demographic differences between the two groups, except for the family financial impact: parents from the NVR group were spending more money on their child, negatively impacting family finances [NVR: 21 (58.3%); TAU: 13 (35.1%)]. Parents of children with STB in the NVR had slightly higher PSI-SF scores than TAU group, with overall high levels of stress in this population (Figure 4). Fathers (mean ± SD: 123.1 ± 18.6) and mothers (120.1 ± 20.0) scores showed similar scores. In the HADS-A, both mothers (11.6 ± 3.8) and fathers (9.0 ± 3.9) showed, respectively, definite and doubtful symptomatology of anxiety. In fact, the largest percentage of mothers had definite symptomatology in each group (NVR: 68.7%; TAU: 45.9%). Although a small number of fathers participating in the study, most of them were in the definite symptomatology in NVR (60.0%) and no symptomatology in TAU (66.7%). HADS-A score in NVR was higher at baseline compared to the control group (Figure 5). Overall, In HADS-D, both mothers (7.9 ± 4.4) and fathers (7.4 ± 4.1) showed no symptomatology of depression (Figure 5). Specifically, the largest percentage of mother had no or doubtful symptomatology in each group (NVR: 60.0%; TAU: 51.3%). Most of fathers were in the no symptomatology in NVR (60.0%) and doubtful symptomatology in TAU (66.7%). There were no significant between-group differences at baseline in the scores of parental questionnaires (Table 4). The teacher-reported SDQ differed significantly at baseline: TAU had higher scores in the total difficulties and hyperactivity/inattention dimensions (Supplementary Table S1).

Between-group comparison of the change in the total score of PSI-SF at completion was not significant: NVR mean (SD) −4.3 (13.9) vs. TAU mean (SD) −7.6 (19.6) two-sample t-test p = 0.42 (Figure 4) with an effect size [95% CI] of −0.19 [−0.67, 0.28]. In the mixed model, the value of p of time effect in NVR group was p < 0.001.

The current study must be considered in light of several limitations. First, NVR was compared to TAU in reducing parental stress, and not to other active and protocolized intervention programs. This leaves the possibility of expectation effects on parental stress reduction in TAU. Moreover, due to ethical reasons, no follow-up assessments were conducted with TAU to confirm that the changes in NVR at follow-up were due to intervention alone. We also consider the short timeframes we used to be a limitation, when the implementation of parental strategies may need some time to unfold and show effects on children’s behavior.

Furthermore, as suggested by the extremely low dropout rate, there might have been a selection bias in our sample, as parents participating in our study were in active search of help, unlike hard-to-reach probably more reluctant to engage in research programs.

In this RCT, contrary to our expectation, we have shown that NVR was not superior to TAU in reducing parental stress at completion of the NVR program for parents of children with STB. However, NVR showed positive outcomes in the follow-up period, pointing to the importance to investigate possible barriers in implementing parental strategies and following this population over longer time periods in future projects. Follow-up assessments are necessary to investigate the possible long-term effects of NVR in the prevention of negative developmental outcomes, although it could be difficult to conduct such studies without adequate control conditions (60). Further studies need to investigate different interventions and treatment outcomes in children with STB, as well as mediators and moderators of treatment effects to understand the mechanisms as this population is understudied despite parental distress and anxiety, and complex symptoms in children. Nowadays, the principal treatment for parent abuse remains family therapy (61) and these data highlight the importance of implementing health care supports and parent therapies like NVR not only for parents but also for child psychological support. At this juncture, it seems essential to better characterize and evaluate the population of children with STB in future studies.