Evaluating the impact of the Mellow Babies group-based parenting programme for supporting at-risk mothers in Tajikistan

Early intervention parenting programmes

From a public health perspective the above evidence highlights the importance of targeting interventions towards mothers in low-income families (14, 15) and that interventions should aim to promote positive childhood experiences (16). Factors including timing and severity of adversity may also be important methodological considerations with evidence suggesting that severe maternal deprivation experienced between 0 – 2 months had a significant impact on functioning in middle childhood and these effects were greater that at earlier developmental stages (17). The study provides evidence and a rationale for engaging in early intervention and prevention of childhood trauma.

Meta-analytic studies indicated that parenting interventions (such as Triple P, Incredible Years and Mellow Parenting (MP)) can have a positive impact on children’s behaviour alongside improving parenting skills and parental wellbeing (18, 19). However, there are challenges with the existing evidence base for these interventions including recruitment biases towards including middle-class families in most of the research, raising concerns about the exclusion of families from low socioeconomic backgrounds, over-reliance in the evidence base on data from high income countries (19)and challenges in establishing the independence of evaluation frameworks from intervention delivery (20).

Although the evidence base for delivery of parenting programmes in low and middle income countries is growing (21) these generally focus on the impact on parenting approaches and child outcomes, with relatively less focus on the impact on caregiver mental health. There are also challenges remaining around how these evaluations work with contextual factors such as poverty, adversity and cultural sensitivity.

There is a growing body of evidence around the application of parenting programmes that use combined targeting of parenting, child behaviour and parental health outcomes. A UK evaluation of MP reported improvements in maternal mental health, parenting confidence, and child conduct problems after participation in the programme(22). Further, most participants in these programmes self-identified as unemployed and presenting with other population level vulnerabilities. A subsequent evaluation of MP reported similar outcomes in terms of wellbeing and confidence for both mothers and fathers who participated in the intervention (23). However, there is limited recent evaluation evidence of the effectiveness of Mellow Parenting in LMIC settings(24).

Tajikistan context

The current study evaluated the delivery of Mellow Babies (a version of Mellow Parenting adapted for infants) in the Central Asia nation of Tajikistan. Mellow Parenting has been adapted to be delivered in non-high resource settings such as Central and Eastern Europe and the Commonwealth of Independent States (CEE/CIS), where health policy has moved to cease the practice of end institutional care rearing for children under the age of 3 years, replacing this with national social protection systems supporting children to live within a family environment. Mellow Parenting groups were introduced to Tajikistan in 2010 in the context of a gatekeeping and reintegration approach to develop family support alternatives and prevent babies at risk entering institutional care settings such as Baby Homes (25). The Putting Families First Project also built on previous work to develop family support and early intervention alternatives to institutional care and included Mellow Parenting Programmes as part of a package of support for families in the community (26).

Tajikistan has seen a steady decrease in poverty rates over the past decade but continues to face economic challenges following COVID-19 and the recent conflict in Ukraine (27). High underemployment within Tajikistan has been linked to high levels of Migration to neighbouring countries such as Russia, which in turn leads to increased vulnerability for “left-behind” women and children (28).

Research suggests high prevalence rates for depression (26%), anxiety (17%), and PTSD (17%) amongst Tajik women (29). Further, the high levels of migration mean that Tajik wives are vulnerable to being separated from their husbands, having little-to-no financial or social support, and often left dependent on their in-laws (30). Evidence suggests that Tajik women are at high risk of domestic violence at the hands of their husband/in-laws and at least 23% of married women have experienced domestic violence (31). Violence against women and girls (VAWG) is a major problem in Tajikistan, driven by conservative gender norms, the culturally ascribed position of young women, and poverty, although there is evidence for the effectiveness of violence reduction programmes (30). Tajik women also have elevated levels of stress and poor living conditions, with suicide rates increasing by 176% from 2008 – 2010, and suicide completion and attempt statistics for Tajikistan were reported as three suicides for every one attempt, in contrast to the global statistics one suicide for every 10 – 20 attempts (32). As childhood experiences of poverty, poor maternal health and limited access to healthcare all increase the risk of poor child outcomes, targeted early interventions for at-risk families in Tajikistan may be one way to mitigate the impact of these social and demographic risk factors.

Current Project

The current study analysed pre– and post-data collected from mothers who participated in the Mellow Babies intervention, targeting families of babies aged 0 – 18 months. We sought to evaluate associations between participation in a Mellow Babies group and maternal mental health outcomes for Tajik mothers. The secondary aim was to evaluate evidence for associated improvements in parenting stress, quality of life, and child behaviour following participation in a Mellow Babies group. Finally, we aimed to identify baseline predictors of outcomes in the sample. It was hypothesised that participation in a group would be associated with improved maternal mental health, and reduced parenting stress but that sociodemographic factors would affect the degree of improvement on outcomes.

Materials and Methods

Participants and eligibility

To be eligible for participation in the group mothers had to have at least one child under the age of 18 months with whom they had frequent contact. It was a requirement that their children participated in Mellow Babies groups.

The total sample was n=195. The mean age of participants was 33.20 years of age (SD=7.52), and the mean age of babies was 5.52 months (SD=3.09). All participants attended 12 or more sessions of the programme. There were 194 participants who completed at least one outcome measure at baseline and post intervention. Figure 1 details the flow of participant data through the study.

• Download figure
• Open in new tab

Figure 1: Flow diagram of participant data in study

Data analysis and approvals

Ethical approval for the data analysis was received from the University of Edinburgh Health in Social Science Ethics Committee. Associations between participation in MB and outcomes were analysed using paired-samples t-tests. Where data were non-parametric the Wilcoxon signed rank test was used. Pearson’s correlations were conducted to test for associations between baseline demographics and scores on pre-intervention measures. Finally, one-way ANCOVAs were carried out to measure the effect of categorical demographics on outcomes, assessing interactions between pre and post scores controlling for the following baseline demographics: residence, education, employment, child disability, and marital status. Effect sizes were reported as Cohen’s d for pre-post comparisons, and η ² for ANCOVAs.

Results

Mental health, parenting and child outcomes

Pre-post intervention comparisons (Table 2) reported significant pre-post intervention reductions on maternal DASS total scores (t(193) = 13.69, p<.001), Depression (t(193) = 13.29, p<.001), Anxiety (Z = –9.39, p<.001) and Stress (Z = –9.72, p<0.001) subscales. For parenting hassles, PDHS Intensity and Frequency significantly reduced from pre to post intervention (t(181) = 14.30, p<.001; Z = –9.41, p<0.001), as did the Parenting Tasks and Challenging Behaviour subscales (t(181) = 11.38, p<.001; t(181) = 12.88, p<.001). Finally, QoLESS scores indicated a significant improvement in self-reported quality of life across the intervention (t(173) = –13.63, p<.001, d = 1.03). There was also a significant decrease in child behaviour (RBCL Z = –8.98 p<.001). Cohen’s d indicated large effect sizes across all tests. Due to the number of t-tests carried out, the p value was adjusted using an alpha correction by calculating p value/number of t-tests resulting in a corrected p-value of p<.005. All tests remained significant after correction.

Relationships between baseline demographics and outcomes

There were significant correlations between the child’s age and both baseline PDHS Intensity (r(127) = –.208 p = .018) and Challenging Behaviour (r(129) = –.295 p = .001). This indicated that parents of younger children reported increased difficulties on these scales. Further, there were negative correlations between child age and PDHS Frequency (r(129) = –.243, p = 0.005) and RBCL Child behaviour scores (r(129) = –.308, p<.001. This indicated that parents of younger children also reported increased difficulties on these scales. Child age was not significantly correlated with maternal DASS score or Quality of life. Participant age was not significantly correlated with baseline measures.

Interactions between baseline demographics and outcomes

ANCOVA analyses were carried out to assess interactions between baseline scores and the baseline demographics (residence, education, employment, child disability, and marital status) on outcome (see Table 3).

Discussion

Our results represent the largest evaluation to date of a parent and infant mental health interventions in Tajikistan to date. Our findings suggest participation in Mellow Parenting was associated with significant post-intervention improvements in self-reported parental mental health capacity to manage parenting stress, reduced problematic child behaviour, and improved quality of life at the end of the intervention; with all changes consistent with large effect sizes. These findings are consistent with existing research into the effectiveness of MP (22, 23), but extend the evidence base into middle resource settings, and with consistently larger effect sizes. The mean age of children in the sample was around 5, suggesting that MP is a potentially effective intervention in the perinatal period. Further, MP was acceptable and feasible to deliver in the Tajik context, with over 80% of participants attending all 14 group sessions, representing higher engagement than UK samples(38). Pre-intervention data suggested mean levels of depression and anxiety in the sample consistent with mild common mental disorder, which had dropped to minimal levels post-intervention.

The current study also reported improvements in child behaviour and parenting hassles, suggesting that participation in MP is associated with broad changes in both maternal mental health and parenting outcomes and may ultimately improve child-related outcomes. This is consistent with network mapping evidence from Tajikistan, indicating that availability of psychological support was linked to more positive relationships between caregivers and children(39).

Further, demographic data indicated a broad range of markers indicative of increased vulnerability within the sample including high levels of unemployment, child disability and education ending at secondary school level. Our demographic data also suggested that vulnerability and risk in mother-infant dyads may be nested within demographic vulnerabilities, consistent with the observation that maternal mental health chronicity emerges within the context of social determinants (6). For instance, those participants who were not currently married showed higher levels of mental health symptoms post-intervention. Participants in rural areas reported higher levels of depression and parenting challenges at post-intervention than those in urban areas, although rural mothers also reported lower anxiety. In addition, employment was associated with higher levels of distress and parenting challenges, perhaps signalling stress emerging from pressure to maintain livelihoods. These differential findings for specific demographic factors highlight the challenges of matching intervention delivery to need.

Given the Tajik context, these outcomes could be interpreted in terms of levels of support available to mothers. Tajikistan has moved towards developing family support and child centres which may be accessed by parents of disabled children (40). Within Tajik culture, it is also common for husbands to migrate for employment to provide and for women to live with extended family (41). Therefore, mothers who identified as “unemployed” may be receiving support from in-laws. Consequently, future evaluations should take into account the role of cultural context as both a baseline factor and a process factor in the delivery of the intervention.

In addition, pre-post intervention change on most of our secondary outcomes (parenting, child behaviour, and quality of life) were subject to significant interactions with baseline demographic characteristics. More positive parenting outcomes were associated with demographics such as having a child with a disability, living in an urban area, being unemployed and being married, and lower levels of post-intervention child behaviour problems were associated with baseline unemployment, and marriage. With the exception of unemployment, each of these interactions could represent the impact of access to wider or more organised support mechanisms (e.g. community, health and social care availability). Further, rural location was also associated with lower child behaviour problems, which could also represent availability of supports in the wider family and community. For disabled children, parents reported less challenging behaviour on the PDHS but increased behaviour problems on the RBCL. Over recent years there have been increases in the level of support available to Tajik mothers, particularly those with additional needs, including the family support centres and living with in laws (40, 41). However, more research is needed to unpick the nature of interactions between intervention and demographic factors.

In addition, we identified significant correlations between child age and parenting stress suggesting that mothers of younger children identified elevated levels of parenting stress and higher levels of behaviour problems. This could fit with the relevance of sensitive periods in child development . For instance, mothers adjusting to the parental role, particularly in the context of additional needs or existing vulnerabilities could experience increased levels of stress in the early months post-partum. Equally, mothers may be more aware of challenging/problem behaviours and accordingly more likely to report them.

Strengths of the current research

The present study builds on existing research on the effectiveness of Mellow Parenting (18, 23, 38) in improving infant and parental mental health, extending this work to a global setting, in a middle resource nation. This adds to the cross-cultural evidence base for MP. Further, the outcome measures used to collect data were all standardised questionnaires, providing high levels of reliability and enabling comparison with other MP programmes and other parenting programmes in general. In addition, the high level of engagement indicates that MP is acceptable to, and well-tolerated by Tajik participants, indicating that there is a basis for scaling up both intervention and evaluation of MP in the country. Furthermore, delivery of the intervention occurred in different sites within Tajikistan, Khuajand, Panjakent and Dushanbe, within community settings, and offered an example of positive NGO government partnership working.

That said, we acknowledge several limitations. Due to the “real world” design of the research, there was no control group, limiting comparisons with families who did not receive an intervention. Therefore, it was not clear if improvements entirely were due to the programme or non-specific factors such as time. Further, given that the data was collected as part of routine delivery of MP in Tajikistan, demographic data was missing for many participants. Although data was collected using standardised measures, two measures (RBCL and PDHS) were administered to children outwith the intended age-group. In addition, only total scores of each measure were recorded. Therefore, it was not possible to record item-level validity. In addition, given that participants’ children in the current study were 18 months and younger, the findings for parenting stress and child outcomes should be interpreted with caution. Future research should aim to use outcome measures which have been standardised for the age groups participating in the study (22). Finally, all outcomes used in the present research relied on self-report measures, reported by mothers who participated in the Mellow Babies group. Future research should therefore aim to collect observational data to accompany self-reported outcomes.

In summary, our evaluation identified that, for Tajik mothers, participation in the Mellow Babies programme was associated with improvements in parental mental health, parenting, child, and quality-of-life outcomes. Future research is required to evaluate MP using controlled trials or implementation designs, as well as examining the long-term impact of participating in Mellow Parenting. Further, our findings suggest that there is sufficient infrastructure in Tajikistan to support this work (42).