Video feedback compared to treatment as usual in families with parent–child interactions problems: a randomized controlled trial

Both groups of families, whether randomized to the VIPI intervention or TAU, improved their observed parent–child interactions, as measured by the EAS scores at T2. This improvement was largest in the VIPI families with the lowest initial emotional availability in the dyad compared with the controls, with significant differences between groups at the post-treatment evaluation. We have not found other RCT studies showing a positive effect of intervention for those parents with the most problematic parenting skills. Parents with less interactional competence often have less capacity for mental representations of their relationships with their children [88,89]. In this study, the video clips visualization of what was taking place in the interactions along with substantial support provided by the therapist, might have strengthened a meta-cognitive capacity or, possibly, an “observational self” in these parents.

VIPI helped families to break unhealthy interaction circles at an earlier stage. In a healthy parent–child relationship, according to Tronick’s Model of Mutual Regulation, the child needs to experience that negative affective states can be changed to positive states through a parental capacity for interactional adjustment and the repair of “mismatches” [90]. The child is believed to develop internal representations of social interactions with parents as positive and repairable, which is fundamental for the child’s development of self-efficacy and efficient coping strategies in the face of adversity. Moreover, neurobiological brain research supports the importance of improved emotional availability between parent and child during the most crucial period for the development of the right prefrontal cortex, an important area for affect and stress regulation and for executive functioning in the child [91].

Consistent with findings from meta-analysis [29], the differences in EAS scores between the VIPI and TAU groups just after treatment might be explained by the use of video, which increased treatment effects, not only on parental sensitivity, but also on parent–child interaction competence as a whole. It is commonly stated that “pictures don’t lie”; moreover, pictures evoke emotions in a different way than words do [52]. With guidance from the VIPI therapists, parents may be able to see for themselves possible mismatches between their children’s signals and their own responses. Further, through VIPI intervention, parents can become observers of their own increased competence, as therapists meticulously points out the beneficial sequences of parent–child interactions. This activity may both reassure parents and boost their self-esteem within their parental roles. Researchers have argued that for parents, the primary source of information about inherent parental competence is through feedback from interactions with their infants [92]. Consistent with a theory of change in addiction research, it is likely that self-observed change (in comparison to receiving the feedback in normal consultations, without video) is more motivating for the maintenance of change in parental behaviour [93]. Possible misinterpretations of what happens in the parent–child interactions might also be resolved when a VIPI therapist articulates his/her own perceptions of the child’s nonverbal cues.

The EAS include a measure of parental hostility observed in the interactions with the child; accordingly, an increased EAS score might also reflect a decreased level of hostility in the relationship. Negative parental emotions related to children or parenting are often not uncovered in usual care [94]. However, the use of video might help parents with negative feelings to gain necessary emotional distance when a situation is more difficult than joyful. Stress caused by negative interactions with a child (for instance, if the child has excessive crying spells) might reduce parent’s reflective functioning, capacity for self-regulation, and therefore capacity to recognize mental states in the child correctly and to stay calm and soothing [95].

As in many intervention studies, one cannot rule out that the early positive intervention effect seen in the VIPI group at T2 was merely due to the overall more intensive treatment exposure experienced by the VIPI families in comparison to the TAU families, since the VIPI group received the VIPI intervention in addition to TAU.

Parental depressive symptoms served as an important moderator of the effect of VIPI compared with TAU through the whole range of EAS scores. Parents with more depressive symptoms, regardless of their initial interactional competence, appeared to benefit most from VIPI, exhibiting increased emotional availability in the parent–child dyad compared with TAU. To our knowledge, video feedback has not previously shown to have a specific beneficial effect on parent–child interactions among parents with depressive symptoms. Since two other effect studies on video intervention found no moderating effect of parental depression [42,46], the effect might be specific to VIPI. Researchers seem to agree that patients’ depressive symptoms are associated with deficits in episodic memory (both visual and verbal) [96]. Since depressed persons also have a tendency both to verbalize about themselves and to recall earlier episodes more negatively than non-depressed persons [97,98] the “empowerment” quality of VIPI might be particularly effective, given the intervention’s almost exclusively focus on positive parent–child interactional sequences. Further, a time limited, structured intervention that targets one issue at a time through the use of both verbal and visual input might be especially beneficial for individuals with impaired concentration, perseverance and executive functioning typical of depressive symptoms. Another advantage of video is its unique potential for studying children’s signals and parental behaviours at a micro-level; slow-speed replay and the opportunity to rewind might facilitate an increase in parental sensitivity.

Feelings of inadequacy and self-blame are frequently present among depressed parents who struggle in their interactions with their children, especially when a child’s temperament is experienced as difficult [52,99]. Parenting problems or having a child with poor interactional competence may also “mask” parental depression or even induce it [100,101], creating a situation that can be easily overlooked during short visits to the well-baby unit [102].

The VIPI intervention did not have any effect on fairly well functioning parent–child dyads. This “ceiling effect” should be of no major concern, since well-functioning parent–child dyads seldom seek help. A similar effect was also found in a previous study with another video feedback method [18]. Thus, VIPI is hardly suitable as a broad preventive measure, especially because we found a tendency, even if just of borderline significance, for parents with good interactional competence and few depressive symptoms to have better outcomes with TAU. This reflects findings from a 2008 study [44], in which the children became less securely attached following video feedback intervention. In fact, competent parents might have their natural interactional flow disturbed if they become more self-conscious and struggle to become “even better”; for instance, by trying to be much more responsive and, thereby, becoming intrusive and leaving the child with less self-reliance [103,104]. “Good enough” parents should instead be reassured that they have inherent competence [105].

The promising VIPI effect compared with TAU seen after intervention (T2) was not replicated at the 6-month follow-up (T3) as presented in Figure 3. However, because both groups experienced considerable improvements in their emotional availability, we might expect that the TAU group had good support from the professionals in the health and welfare systems dealing with the variety of interactional problems within these families. It is also possible that the three home visits by the same supportive research assistants to all the families in both groups represented an intervention in itself. The female research assistants did the filming, had lengthy interviews with the parents, and assisted with completion of the questionnaires, activities that might have instigated increased parental reflection and awareness of their children.

Finally, a longer treatment period or the addition of booster sessions following ended intervention might have contributed to sustained improvement in the parent–child interactions. Egeland and his co-workers have argued that each stage of child development requires different parental skills to match the needs of the developing child [30]. Thus, booster sessions or a longer treatment period might ensure parental responsiveness as children grow older.

Our findings revealed a delayed, but substantial, effect of VIPI on the children’s capacities for self-regulation, compliance, adaptive functioning, autonomy, affect and interaction with others, measured with the total ASQ:SE score at the 6-month follow-up after the completion of treatment (T3). We know that an early capacity for emotion regulation may prevent developmental trajectories for both internalizing and externalizing problems [106-108] and may provide protection against later psychiatric disorders [109,110]. Because this “sleeping effect” was not reflected in a corresponding difference between groups in the EAS score at T3, mechanisms other than a direct link between the observed interactions and actual child development must be involved. Brazelton has argued in his Touchpoints model that child development occurs in bursts followed by regressions and pauses [111]. Because of such “bursts”, it could take some time before the increased emotional availability in the parent–child dyad will affect child development; hence, the differences in emotional availability between groups, which occur just after treatment (T2), first manifest in the child at T3. This finding corresponds to that of a study of video feedback intervention in autistic children, in which increased child social skills, not mediated by intervention effects on parenting, was found in the treatment group compared to controls [112]. Still, one cannot rule out that parental perception of child development evident in the parent-reported ASQ:SE improved more in the VIPI group not because the child had changed much, but because the parents had developed an increased tolerance for minor child problems because of the intervention. This assumption is supported by findings in another Scandinavian study where parents’ assessment of child development with the ASQ:SE was strongly predicted by maternal stress [113].

In a secondary analysis, in which we used parental depressive symptoms, not as a moderator but as an outcome, no differences between groups were found right after treatment, but highly significant differences between groups were seen at the 6-month follow-up in favour of the VIPI group (see Table 5). This finding might merely represent a “sleeping effect” of the increased emotional availability (EA) at T2 at which point the EA was estimated to be significantly higher in VIPI families than in TAU families (see Figure 3). Another scenario is a putative synergetic effect between the reduced concerns for the child’s development and the improved EA, which may have mitigated parental depression symptoms in the VIPI group. How to efficiently treat postpartum depression to achieve a long-term effect on parents’ mood, parent–child interactions, and children’s development has, thus far, been minimally explored [17]. Earlier research on interventions targeting only the depressed parents has shown that such methods might fail to bring about changes in the parent–child relationship or in child development [17,114,115].

VIPI intervention improved the depressive symptoms of parents throughout the study period, which lasted from 9 to 13 months. Among post-natally depressed women, several studies have indicated that without intervention, stability or recurrence of depressive symptoms might be expected in this period [116,117] with one study showing elevated maternal depressive scores at a 3-years follow-up [117].

We found that none of the background variables interfered with the VIPI effect. Our sample is difficult to compare with either the normal or the multi-risk samples presented in the Dutch meta-analyses [36,39] as SES variables in a sample drawn from a less homogeneous population than the Norwegian one could possibly affect treatment outcomes greatly.

To our knowledge, the possible moderating role of deviant parental personality traits on parent–child interactions has not been previously examined; hence, we included such a measure in this study. Within the families with the most compromised parent–child interactions, we found that some parental personality disorder traits had an impact on the VIPI effect in both a positive and a negative direction compared with TAU. Presence of dependent and paranoid personality traits actually served as important moderators in favour of the VIPI intervention, whereas absence of narcissistic traits increased the effect of VIPI compared with TAU. From clinical experience, we know that dependent persons welcome new ideas from others because of their own feelings of inadequacy. However, the effect over time is unclear, because the lack of self-confidence and clinging behaviours are usually profound [118]. Being allocated to the TAU group might also be experienced as more stressful by dependent parents.

One might also suspect that the use of video clips represents a tangible, reassuring element for paranoid people. These individuals’ mistrust in the intentions and actions of others [119] is likely to negatively affect intervention. However, the transparency and clarity afforded by the use of video ensures that there is no hidden agenda, which might be relieving for someone with paranoid traits and help him/her gain trust in both the therapist and the intervention efforts.

For parents with fairly good emotional availability in interactions with their children, the presence of even very few narcissistic traits disfavoured VIPI intervention. The easily violated, slightly narcissistic parent in these high-functioning groups might be disturbed by the fact that they are not perceived as being as exceptional as they thought they were during VIPI intervention, which could possibly lead to a weakening of their interactional competence. Just as likely, when parents with narcissistic traits watch video clips of themselves, their self-preoccupation might increase, and they may become even less aware of their children.

The ICC on the EAS scores in our study was 0.461. In some contexts, this value would be considered low. However, it should be noted that this low ICC is caused by the relatively large residual variance (139.739). The inter-rater variance (22.973) is, by far, the smallest of these variance components, so the contribution to the total variance from inter-rater variance is practically negligible. Even if this variance component were 0 instead of 22.973, the ICC would still be 0.499.

The average Pearson correlation between the raters was 0.63. The Person correlation coefficient and the variance components from the mixed model address different issues. A Pearson correlation coefficient of 0.63 can be considered moderately high, but it does not tell the entire story. For example, if one rater consistently rated scores exactly 20 points higher than another rater, the Pearson correlation between the two would be 1.0. On the other hand, the relative magnitude of the inter-rater variance in the mixed model tells us that there are no large systematic differences between the raters.

As far as we know, there is no potential bias interfering with our analysis.

The history of the seriousness and duration of ongoing depressive symptoms and/or earlier depressive episodes and/or other affective disorders was not available for investigation; neither was information concerning ongoing medication. For parents in our inquiry with moderate depressive symptoms, intervention directed only at the parent–child interactions was effective in decreasing their depressive symptoms significantly. This might not be the case for the most seriously depressed parents, since such were not present in our sample. Therefore, a limitation of our inquiry is that we do not know whether parental cognitive problems related to the most serious depressive states might interfere with the intervention effect.

In 23 families, both parents took part in the VIPI intervention; in most families, the mothers chose to participate in the study. Consequently, no data of the non-participating parents was analysed.

Our follow-up period was limited to 6 months after treatment. A longer follow-up period could provide information about further progress in emotional availability in the parent–child interactions.

Our study indicates that VIPI intervention has an impact on emotional availability in parent–child relationships. Unexpectedly, a specific VIPI effect compared with TAU in families with depressed parents was also found, regardless of existing interactional competence, resulting in increased emotional availability in parent–child interactions. Our findings, together with the long-term main effects on depressive symptoms, support the use of VIPI interventions among depressed mothers with young children. Some clinicians might contend that it is best to treat parental depressive symptoms before working with parent–child interactions; however, our findings give no indications of the necessity of such a sequence.

Even if the study failed to prove long-term effects on emotional availability compared with TAU, the 6-month follow-up evaluation revealed effects on child development. In their clinical work, VIPI therapists experience that one or two “booster sessions” within 6 months of the end of the intervention can enhance the positive effect on the parent–child relationship. However, investigating the effect of such “booster sessions” was not within the scope of our study.