The Role of Paternal Accommodation of Paediatric OCD Symptoms: Patterns and Implications for Treatment Outcomes

Fathers’ contribution to a child’s development and coping with mental disorders is a largely neglected area (Panter-Brick et al., 2014). Accommodation of children’s obsessive-compulsive disorder (OCD) symptoms by mothers is suggested to have a negative impact on treatment outcomes (Strauss et al., 2015); yet, little is known about the role of fathers in OCD. In this study, we examine the pattern of paternal accommodation of children’s OCD and how it affects treatment outcomes.

OCD is common in children and young people with prevalence estimates ranging between 1% - 4% in epidemiological studies (Heyman et al., 2001, Valleni-Basile et al., 1996). The disorder is associated with marked functional impairment in the young person as well as significant family burden and distress (Amir et al., 2000, Cooper et al., 1996, Calvocoressi et al., 1995). The term family accommodation (FA) has been formally used in the OCD literature to refer to the involvement and participation of family members in an individual’s OCD rituals (e.g. providing reassurance, providing items, assisting in avoidance, modifying routines).

To date, a number of studies have examined FA in children with OCD (Lebowitz et al., 2016). Most studies have reported high rates of accommodation among families of children with OCD, suggesting that up to 60–96% of relatives assist or modify their behaviour to accommodate their child’s OCD symptoms (Bipeta et al., 2013, Caporino et al., 2012, Flessner et al., 2011, Futh et al., 2012, Garcia et al., 2010, Lebowitz et al., 2014, Merlo et al., 2009, Peris et al., 2008, Shafran et al., 1995, Stewart et al., 2008, Storch et al., 2007). In addition to being common, various child- and parent-level factors have been associated with parental accommodation, including OCD symptom severity, functional impairment, child’s internalizing and externalizing symptoms, and parent psychopathology (i.e. maternal anxiety and depression) (Caporino et al., 2012, Flessner et al., 2011, Lebowitz et al., 2014, Merlo et al., 2009, Peris et al. 2008, Stewart et al., 2008, Storch et al., 2007, Strauss et al., 2015, Wu et al., 2016, Wu et al., 2019). A handful of studies across both the pediatric and adult OCD literature have also observed an association between FA and treatment outcomes (Cherian et al. a., 2014, Ferrao et al., 2006, Garcia et al., 2010, Merlo et al., 2009; Rudy et al., 2014), albeit not consistently (e.g. Torp et al., 2015). Specifically, lower FA scores at pre-treatment and greater reduction in FA at post-treatment were associated with better outcomes as indicated by lower OCD severity scores (Ferrao et al., 2006; Merlo et al., 2009; Garcia et al., 2010; Rudy et al., 2014). According to the theoretical cognitive-behavioural therapy (CBT) framework, FA hinders CBT effectiveness by reinforcing OCD fears and avoidance (Peris et al., 2008; Storch et al., 2007). Nevertheless, studies have differed in the extent to which their treatments targeted FA, which might explain the inconsistent findings regarding FA as predictor of treatment outcomes (e.g. Torp et al., 2015).

Despite the increased interest in understanding factors that promote FA, the extent to which patterns of accommodation vary between family members - and specifically between mothers and fathers of young people with OCD - remains unclear. Evidence suggests that mothers and fathers interact differently with their children, with fathers contributing uniquely to their child’s behavioral and psychological development across various mental health diagnoses (Bögels and Phares, 2008; Lewis and Lamb, 2003; Ramchandani et al., 2005, 2013). Yet, the role or response of fathers to their child’ OCD symptoms and treatment outcomes is a largely neglected area of research. Indeed, the available OCD literature have either only involved one parent (most commonly the mother) or clustered together different kinship of family members within the same study. It is possible that there might be a differential response between family relatives to a child’s OCD symptoms. For instance, in the adult OCD literature, it has been found that spouses/partners endorse significantly higher FA scores than other family members, such as parents, siblings, children, and cousins (Gomes et al., 2014), and that the burden of caring for an individual with OCD tends to fall on one family member (Cooper et al., 1996). Of relevance to the present study (albeit only 3 participants in their sample had a diagnosis of OCD), Thompson-Hollands et al. (2014) examined parental accommodation among mothers (n = 68) and fathers (n = 51) of children with anxiety disorder. The authors found a medium association between mother and fathers reports of accommodation (r = 0.27, p = 0.06) and that the vast majority of mothers (97%) as well as fathers (88%) engaged in accommodating behaviours. Thus far, the only study that has directly examined differences between mothers (N = 41) and fathers (N = 29) in response to children’s (N = 43) OCD symptoms (Futh et al., 2012) failed to find differences between parents in understanding, narrative, coping, and distress associated with family accommodation. However, the generalizability of these findings is significantly limited by the small sample size, the qualitative nature of the study, and the self-reported diagnostic status of the child (Futh et al., 2012). To our knowledge, no study has empirically examined maternal and paternal accommodation of symptoms separately in a pediatric sample with OCD. Therefore, the question as to how FA differs between mothers and fathers, and how these differences impact or interact with the child’ symptoms and treatment remains to be addressed.

The current study sought to extend the existing work by examining maternal and paternal FA of child OCD symptoms. The first aim of the current study was to examine whether the frequency, type of accommodating behaviours, and correlates of FA of OCD symptoms differed between mothers and fathers. Based on the limited research (Futh et al., 2012; Gomes et al., 2014), we expected to find no differences in the frequency and type of accommodation between mothers and fathers. Based on previous research examining correlates of FA (Flessner et al., 2011, Gomes et al., 14, Peris et al., 2008, Stewart et al., 2008), various child- and parent-related variables were chosen as candidate correlates, including the child’s age and gender, OCD symptoms severity, internalising and externalising difficulties, depressive symptoms, general functioning, and parents’ psychopathology. OCD symptom severity and parent psychopathology were hypothesised to be associated with parental accommodation based on previous findings (e.g. Flessner et al., 2011, Peris et al., 2008, Storch et al., 2007, Wu et al., 2016). However, due to the lack of studies, no specific hypotheses were advanced with respect to how correlates of FA would differ between mothers and fathers. The second aim of the study was to examine whether the association between FA and CBT outcomes differed between mother- and father-reported FA. Since family accommodation has been linked to reduced treatment response irrespective of kinship of family members, both maternal and paternal accommodation of OCD symptoms were hypothesized to predict poorer treatment outcomes.

Demographic and clinical characteristics of the sample are summarized in Table 1. Parental accommodation was common; 80% of mothers and 57% of fathers reported accommodating their child’s OCD symptoms daily in at least one item. In addition, there was a moderate intra-class correlation (ICC) between mother and father FAS-PR total score (ICC = 0.68, p < 0.0001). As shown in Table 2 and Fig. 1, similar patterns of accommodating behaviours were observed for both parents, with provision of reassurance, participation in rituals, and facilitation of avoidance being endorsed most frequently. Relative to fathers, mothers reported significantly higher rates of accommodation on the 12-item FAS-PR Total score (t(208) = 7.59, p < 0.0001) and on the subscales of Avoidance of triggers (t(208) = 5.14, p < 0.0001) and Involvement in compulsions (t(208) = 8.32, p < 0.0001) .

Configural and metric invariance held, the latter suggesting that the same latent factors were being measured in each rater. However, scalar invariance was not achieved; examination of the modification indices suggested a point of localized strain in the intercept of item 1; accordingly, a partial scalar invariance model was estimated in which the intercept of item 1 (“How often do you reassure your child?”) was allowed to differ between groups, resulting in a good-fitting model. This suggests that, at the same level of the latent trait of family accommodation, mothers tend to score higher in item 1 than fathers (Fig. 1, Table 2). A residual variance invariance model held well. Since the threshold for non-invariance in the scalar model was just over the limit (limit: a change of ≤ −0.010 in CFI, supplemented by a change of ≥0.015 in RMSEA or a change of ≥0.010 in SRMR; Scalar invariance model comparison: ΔCFI = − 0.011, ΔSRMR = 0.010), and the model itself had an acceptable fit, we conducted the analyses including item 1. However, a sensitivity analysis was conducted excluding this item. Full description and results of measurement invariance can be found in the Supplemental material (Methods, Results, and Table S2).

Based on the number of mothers and fathers who reported daily FA in at least one of the FAS-PR items, there were 114 participants (55%) for whom both parents reported daily FA, 59 participants (28%) for whom only one parent reported daily FA (specifically, 52 mothers and 6 fathers), and 36 participants (17%) for whom neither parent reported daily FA. Bonferroni-corrected one-way ANOVA revealed that those with both parents accommodating daily, compared with those for whom neither parent accommodated daily, scored higher in CY-BOCS (p = 0.002), CGAS (p < 0.001), BDI (p = 0.021), SDQ total score (p = 0.003), SDQ emotional (p = 0.006), peer problems (p = 0.005) and prosocial subscales (reversed) (p = 0.008), mother DASS total score (p = 0.012), and the depression (p = 0.010) and stress subscales (p = 0.011), father DASS total score (p = 0.002), and anxiety (p = 0.006), depression (p = 0.014) and stress subscales (p = 0.001), and mother (p < 0.001) and father FAS-PR (p < 0.001). In addition, compared with young people for whom only one parent accommodated daily, those with both parents accommodating daily showed higher scores in SDQ prosocial subscale (reversed) (p = 0.003), mother DASS stress subscale (p = 0.001), father DASS total score (p = 0.047) and stress subscale (p = 0.039), and mother (p < 0.001) and father FAS-PR (p < 0.001).

With regard to correlates of maternal versus paternal FA, no significant correlations were found between parents’ FAS scores and child demographic variables (age, gender, duration of illness) (all p > 0.05). In contrast, the child’s OCD symptom severity (CY-BOCS), general functioning (CGAS), depressive symptoms (BDI-Y), child’s emotional and behavioural difficulties (SDQ total score and all subscales expect SDQ hyperactivity) as well as mothers’ and fathers’ psychopathology (DASS total score and their subscales) were significantly correlated with mother and father FA (Table S3). These variables were therefore retained as independent variables and entered into regression models.

The model examining correlates of maternal FAS accounted for 40% of the variance, with OCD symptom severity (β = 0.28, p < 0.001), mother DASS (β = 0.34, p < 0.001), father DASS (β = 0.15, p = 0.045), and SDQ total score (β = 0.16, p = 0.028) being significantly associated with maternal FA. With regard to paternal accommodation, the model accounted for 34% of the variance, with OCD symptom severity (β = 0.21, p = 0.008), paternal DASS (β = 0.32, p < .001), and SDQ total score (β = 0.16, p = 0.045) making a significant contribution. BDI and CGAS were not associated with maternal or paternal accommodation of OCD symptoms. Additional exploratory post-hoc analyses were carried out using SDQ and DASS subscales. For maternal FAS, the model accounted for 42% of the variance, however only CY-BOCS total score was significantly associated with FAS (β = 0.26, p = 0.001). For paternal FAS, the model accounted for 37% of the variance, with CY-BOCS total score (β = 0.20, p = 0.013), SDQ conduct problems (β = 0.21, p = 0.012) and CGAS (β = −0.17, p = 0.036) being significantly associated with FAS (Table S4).

In terms of treatment outcomes, higher maternal FA significantly predicted higher scores on the CY-BOCS at post-treatment (β = 0.27, p < 0.001) after adjusting for pre-treatment severity (Table 3). Similarly, father’s accommodation significantly predicted higher post-treatment OCD severity (β = 0.23, p = 0.003). Ninety-two participants (74%) met criteria for treatment response. Both mothers’ (OR = 0.96, p = 0.037) and fathers’ (OR = 0.95, p = 0.004) FA significantly predicted treatment response status, with higher levels of accommodation predicting lower likelihood of response at post-treatment. Similarly, rates of treatment response in participants with two parents accommodating daily (n = 63) were lower than in participants with one parent accommodating daily (n = 37) or no parent accommodating daily (n = 24) (60% vs 84% vs 96%, χ² = 13.98, p = 0.001).

Levels of FA decreased after treatment in both mothers and fathers (B = -22.12, p < 0.001), with a rater by time interaction (B = 2.79, p = 0.039), in which FA decreased slightly more in mothers. In models adjusted for changes in CY-BOCS scores, the decrease in FA was still evident (B = -9.36, p < 0.001), however the interaction effect disappeared (B = 1.70, p = 0.162). At post-treatment, mothers scored higher than fathers in FA in both unadjusted (B = 1.99, p = 0.038) and adjusted models (B = 3.08, p = 0.001).

Fathers’ response to their child’ OCD symptoms remains a largely neglected area of research. This is the first and largest quantitative study to date to examine levels and correlates of parental accommodation of OCD symptoms in young people and its association with treatment outcomes with a specific emphasis on investigating and comparing maternal and paternal accommodation.

The results of the current study indicate that both parents engage in high and similar patterns of accommodation of their child’s OCD symptoms. Specifically, both mothers and fathers reported provision of reassurance and facilitation of avoidance as the two most frequent types of accommodation provided. Furthermore, for most of our sample (n = 114, 55%) both parents accommodated their child’s OCD daily. This subgroup of young people scored higher on measures of OCD severity, functional impairment, child emotional and behavioural difficulties, and parent psychopathology compared to those where only one (n = 59, 28%) or no (n = 36, 17%) parent accommodated daily. Notably, a moderately high correlation between mothers’ and fathers’ FAS-PR scores was also observed, suggesting that if one parent accommodates a child’s symptoms the other seems more likely to accommodate as well. These findings suggest that parents might take a similar approach to a child’s OCD. Whilst both parents give accounts of being similarly drawn into rituals, parents differed in the levels of symptom accommodation, with mothers accommodating more than fathers. Furthermore, our results indicate that in cases where only one parents accommodates, this tends to be the mother, which might simply relate to the amount of time a caregiver spends with a child with OCD. Although this study did not collect information regarding how much time each parent spent with their child, future studies may benefit from addressing this limitation. Nonetheless, both parents highly accommodate their child’s OCD symptoms, highlighting the importance of considering both parents in the assessment and treatment of OCD to ensure that they are both able to withdraw successfully from OCD symptoms rather than having one parent inadvertently maintain a cycle of rituals and avoidance. Our finding that mothers and fathers report a highly similar pattern of accommodation is in line with previous investigations on this topic (Futh et al., 2012; Gomes et al., 2014, Thompson-Hollands et al., 2014). Indeed, these results extend the findings of a recent study on parents of young people with anxiety disorders, whereby the authors found 97% of mothers and 88% of fathers engaged in FA as well as a medium correlation between mother and father reports of accommodation (r = 0.27, p = 0.06) (Thompson-Hollands et al., 2014). However, our findings also indicate differences in the extent of FA between parents, with mothers accommodating their child’s OCD symptoms to a far greater extent than fathers. The lack of research comparing mothers versus fathers’ FA and the small number of fathers participating in prior studies in OCD may have provided limited power to detect differences in the extent of FA between parent dyads and supports the scope for further research to address questions concerning paternal role in OCD.

The current study also sought to examine correlates of parental accommodation for mothers and fathers. The finding that OCD symptom severity predicted both maternal and paternal involvement in rituals is not surprising. Indeed, this result emerges consistently as one of the factors most relevant to understanding FA (Storch et al., 2007; Peris et al., 2008) and, furthermore, has been supported by two recent meta-analyses that reported a medium-sized correlation (r = 0.35–0.42) between FA and OCD symptom severity (Strauss et al., 2015, Wu et al., 2016). Moreover, post-hoc analyses in our study indicated that symptom severity remained a significant correlate of both maternal and paternal FA even after taking into account additional measures in the model. Alongside OCD severity, our findings confirmed a relationship between child’s emotional and behavioural difficulties (as measured using the SDQ) and parental psychopathology (as measured by the DASS) with FA. This was found to be equally true for mothers and fathers. In addition, fathers’ distress (DASS) was also associated with mothers FA, though the reverse was not observed. Unfortunately, the causal direction of this association cannot be established. Children with heightened emotional/behavioural difficulties might facilitate parent’s accommodation. Similarly, distressed parents could be more likely to accommodate in response to their child’s difficulties. Yet, the reverse pattern however could also be true; that is, FA could lead to increased child’emotional/behavioural difficulties and/or parental distress. These two alternatives would have different implications for treatment and remain an important question to be addressed in future longitudinal research. The finding that symptom severity exerts an influence on mothers’ as much as on fathers’ likelihood to accommodate strongly supports the need for education for all family members regarding this coercive cycle and management strategies for such behaviour. In addition, our results highlight the potential value of screening and providing additional support for children who present with generalized heightened emotional/behavioural difficulties and/or whose parents present with psychopathology. Our findings also reinforce the need to consider fathers’ perspective as potentially indirectly influencing maternal accommodation.

Clinical practice guidelines currently recommend the involvement of parents according to the needs of the child (NICE, 2005). Certainly, there is a growing body of literature demonstrating good clinical outcomes for family-based CBT, with emerging evidence supporting the additive effect of parental involvement in OCD treatment (e.g. Lebowitz et al., 2020, Storch et al., 2007; Freeman et al., 2014; Rudy et al., 2014). As hypothesized, the results in this study support the association between family accommodation and treatment outcomes. Similar to maternal FA, we found that greater levels of fathers’ accommodation of symptoms predicted more severe OCD symptoms at post-treatment, even after controlling for pre-treatment severity. Moreover, in this study, paternal accommodation was found to be a significant predictor of response status. This is a novel finding that encourages more consideration and research on fathers’ perspective in pediatric OCD. Since fathers are usually less involved in treatment, they may continue to engage in accommodation of OCD symptoms inadvertently maintaining or reinforcing OCD behaviours.

Treatment response was poorer for those participants with both parents accommodating daily (60% response rate) than those with only one (84% response rate) or no (96% response rate) parent accommodating, providing preliminary evidence to support the inclusion of both parents in family-based CBT for OCD. Emerging evidence supports the benefits of family-based CBT, incorporating treatment modules that address FA (e.g. Grunes et al., 2001; Rudy et al., 2014). In a small trial, greater OCD reductions were observed for patients assigned to a behaviour therapy plus family intervention group, compared to those in the behaviour therapy only condition (Grunes et al., 2001). Future research is warranted to examine the additional benefit of involving both parents, as opposed to one relative, in family-based interventions for pediatric OCD. Indeed, whilst common clinical practice tends to involve the participation of one parent in treatment, usually mothers (Iversen et al., 2012), our results provide preliminary support in favor of further consideration of fathers’ role in their child’s OCD treatment. For instance, research has shown that the involvement of fathers in treatment of disruptive behaviors in adolescence can lead to improved treatment outcomes (e.g. Bagner and Eyberg, 2013; Lundahl et al., 2008). In contrast to the literature on externalizing behaviors (Iversen et al., 2012), this issue remains to be examined in anxiety disorders and/or OCD. However, the current findings point to the potential benefit of examining this issue further and investigating the impact of the involvement of fathers in treatment.

A number of shortcomings ought to be considered when interpreting the findings of this study. First, as with previous research examining parental accommodation in paediatric OCD, information on mothers, fathers, and families was limited minimizing our ability to fully understand how parents’ role and involvement might vary based on certain clinical or demographic information (e.g. age, SES, married status, ethnic status). Future research would benefit from collecting further demographic and clinical information on parents. Second, only families where both parents completed the FAS were included in the study. Although the majority of young people in this study (91.5%) were cohabitating with both their mother and father, differential patterns of FA may be observed or be influenced by parents’ living arrangement. Third, whilst data on ethnicity and family composition were not available for all participants, anecdotally the sample consisted of largely White, British families. Therefore, our findings may not be generalizable to a range of families from differing ethnic backgrounds. Fourth, the study relied on parents as central informants and no measures were in place to ensure mothers’ and fathers’ responses on the FAS-PR were collected independently from the other parent. Therefore, these findings should be replicated using observational and clinician-administered measures of FA. Fifth, we did not include a measure of the time that parents spent with their child. It is possible that the difference in the extent of FA between mothers and fathers is simply related to the time spent with their child as opposed to the type of the caregiver, specifically. This requires investigation in future studies. Sixth, only a sub-group of our sample had treatment outcome data available. Although those with and without outcome data were comparable with respect to most variables of interest (with the exclusion of father-reported FA), they may have differed with respect to other clinical characteristics that were not measured in this study. Seventh, the lack of measures of maternal and paternal involvement in the OCD treatment is another limitation in that it does not allow one to assess their respective impact on treatment outcomes. It is possible that the extent to which FA is directly addressed in treatment influences results across studies on FA in OCD; as such, future research examining FA would benefit from careful measuring, monitoring and tracking of parental involvement in treatment. Eighth, post-treatment CY-BOCS was not routinely collected by an independent clinician in all cases. However, it is unlikely that clinicians’ ratings were directly influenced by parent reports of family accommodation. Finally, causality between FA and treatment response cannot be established from this study. More research is needed to confirm the direction of causation in order to inform intervention strategies. This could be addressed in randomized controlled trials which are beginning to emerge (e.g. Grunes et al., 2001; Freeman et al., 2014).

Notwithstanding these limitations, the inclusions of both parents in the examination of FA of OCD is an important contribution to the pediatric OCD literature. The present study found that mothers and fathers are more similar in their propensity to accommodate child OCD than previously thought. The findings also highlight the value of appropriate screening and targeting of both maternal and paternal symptom accommodation in pediatric OCD. Further research is needed to investigate the contribution and involvement of fathers in OCD treatment.