Mid-term effects of a parenting programme on early childhood development and maternal health in Vietnam: protocol for a follow-up of a cluster-randomised controlled trial

Abstract

Introduction

Parents are crucial as providers of nurturing care for their children to grow and thrive. A community-based multicomponent intervention for women’s health and early childhood health and development was conducted in rural Vietnam, spanning from mid-pregnancy through the first postpartum year. A cluster-randomised controlled trial of the programme conducted between 2018 and 2021 found that children aged 2 years whose parents were in the intervention group and had received the programme had higher cognitive, language and motor development than those in the control group who had received the usual standard of care. The intervention was also found to be cost-effective and improve equity. The aim of this follow-up study is to examine the mid-term effects of the intervention and describe the views of caregivers, service providers and policymakers on the programme, the evidence of its impact and its translation into policies and practice.

Methods and analysis

This is a follow-up study of participants in a cluster-randomised controlled trial to examine the mid-term effects of the intervention when children are 7 and 9 years old in Vietnam. We will re-enrol mother–child dyads from the original trial (622 women and their children in the intervention; 546 dyads in the control group). The primary outcome is children’s intellectual ability. Secondary outcomes include children’s emotional and behavioural development, quality of life, anthropometric indices, health conditions, health behaviour and education and parenting, maternal mental health, maternal quality of life, cardiovascular health and biomarkers of ageing together with women’s and children’s experiences of family violence.

Outcomes will be measured using culturally and linguistically adapted standardised tools and study-specific structured questions. Differences in outcomes between trial groups will be assessed using mediation analyses and mixed-effects models, including a fixed effect for the trial group and adjusting for cluster effects as a random effect based on the intention-to-treat principle. An economic evaluation will estimate the cost-effectiveness of the intervention in improving child development at ages 7 and 9 years from healthcare and societal perspectives. A qualitative study will be conducted alongside the follow-up study using purposive sampling to describe the perspectives of caregivers, service providers and policymakers on the intervention, the evidence about its impact and its integration into national policies and services.

Ethics and dissemination

Ethics approval was obtained from the Monash University Human Research Ethics Committee (2025–46034-125952). Study findings will be disseminated through multiple channels, including peer-reviewed publications, national and international conference presentations, seminars, policy briefings, lay language summaries to communes and media reports.

Trial registration number

ACTRN12625001113448.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Using a mixed-methods approach, this follow-up study of a cluster-randomised controlled trial will examine the mid-term effects and cost-effectiveness of the intervention over time, and the perspectives of caregivers, service providers and policymakers.

• A multidisciplinary team with expertise in global health, child health and development, psychology, biostatistics, molecular and perinatal epidemiology, epigenomics and health economics from Vietnam, Australia and the United States will undertake the research using culturally and linguistically adapted standardised tools and study-specific structured questions to ensure contextual relevance and validity.

• The Wechsler Intelligence Scale for Children Fourth Edition, the measure of the primary outcome, has been formally translated, adapted and validated for use among children in Vietnam.

• Loss to follow-up and reliance on some recall-based data may introduce bias into the findings.

Introduction

The early years of life lay the foundation for lifelong health and well-being.¹ However, analyses of household survey data indicate that approximately 25% of children in low-income and middle-income countries are not developmentally on track in at least three of the four early childhood development domains: literacy–numeracy, physical, social–emotional and learning.^2–4 Parents, as primary caregivers, are fundamentally important to optimal child development.¹ Parents can provide nurturing care to ensure that children have good health; are well nourished, safe and secure, given opportunities for early learning; and receive responsive care, with support from enabling policies, programmes and services.¹ A meta-analysis shows that parenting interventions aimed at improving parents’ interactions, behaviours, knowledge, beliefs, attitudes or practices with their children can improve child cognitive development (standardised mean difference 0.32, 95% CIs 0.23 to 0.40), globally.⁵ Further benefits include language development, motor development, socio-emotional development, parenting knowledge, parenting practices and parent–child interactions.⁵

A community-based multicomponent intervention ‘Learning Clubs’, now known as Hanh Trinh Dâu Đời, Early Journey of Life, was conducted in a rural province in Vietnam to improve women’s health and the early health and development of their children.^{6 7} The intervention aimed to address eight modifiable risk factors to early childhood development, including intrauterine growth restriction, stunting, iron deficiency anaemia, iodine deficiency, unresponsive caregiving, insufficient cognitive stimulation, maternal mental health problems and family violence. The intervention consisted of 19 locally facilitated community-based parenting group sessions, delivered from mid-pregnancy to the first postpartum year and one home visit in the first two postpartum weeks, in addition to the usual standard of maternal and child healthcare. The control group received only the usual standard of maternal and child healthcare. A cluster randomised controlled trial showed that children in the intervention group had higher cognitive development than those in the control group at age 2.⁷ Additionally, the intervention improved language and motor development, was found to be cost-effective and improved equity.^7–9 However, its effects beyond the short term have not yet been examined. Most existing research has focused on short-term outcomes (1–3 years after programme completion), whereas evidence on mid-term (4–9 years) or long-term (10+years) effects remains limited in low-income and middle-income countries.¹⁰

The primary objective of this follow-up study is to examine the mid-term effects of the Learning Clubs on children’s intellectual ability at ages 7 and 9 years. The secondary objectives include examining (1) the mid-term effects on children’s health and socio-emotional development and on women’s health, (2) the economic value of the intervention in mid-childhood and (3) the perspectives of parents, service providers and policymakers on the intervention, evidence about its impact and its integration into policies and services.

Methods and analysis

Study design

This is a follow-up study of a two-arm, parallel-group cluster-randomised controlled trial to examine the mid-term effects of the multicomponent intervention in Vietnam. The original trial assessed the effects of the intervention on child cognitive, language, motor and social–emotional development at age 2 (Trial registration number ACTRN12617000442303).^{6 7} This follow-up study will measure child development outcomes at ages 7 and 9 to examine the mid-term effects of the intervention (Trial registration number ACTRN12625001113448) (figure 1).

Figure 1. Study design.

Learning clubs trial

In the original trial,^{6 7} women in the intervention group were invited to attend eight group sessions during pregnancy and 11 sessions in the first postpartum year and to have one home visit, in addition to receiving the usual standard of care (antenatal health checks, birth in a medical facility and participation in national immunisation and growth monitoring programmes). Women in the control group received only the usual standard of care.

The contents of the programme included education to optimise maternal nutrition, mental health, stage-specific parenting knowledge, infant care skills, infant health and development, first aid and home hygiene and gender-based violence and empowerment. The programme was participatory, incorporating demonstrations of caregiving and hands-on practice, scenario-based discussions, role play, short videos and the sharing of experiences. Each group included 10–15 women at the same life stage, meeting every 2–4 weeks in local community centres for 60–90 min. Fathers and grandparents were encouraged to participate. Group sessions were facilitated by trained local members of the Vietnam Women’s Union with support from a commune health worker and a kindergarten teacher to some content. All facilitators received supportive supervision. Take-home posters and a family booklet summarising the content were provided for discussion and ongoing reference.

In the trial, a sample size of 1008 pregnant women from 84 clusters was needed to detect a difference in the primary outcome (cognitive score below −1 SD) of 15% in the control group and 8% in the intervention group (80% statistical power, a significance level of 0.05 and an intracluster correlation coefficient of 0.03) at trial inception.^{6 7} An independent statistician randomly selected 84 communes (cluster units) from a list of 112 communes in Ha Nam province. These 84 communes were then randomly allocated to either the intervention or control groups, with 42 communes each. Women who were pregnant and up to 20 weeks of gestation living in the randomly selected communes were invited to participate in the study. Data at the end of the original trial were contributed by 616 women and their children in the intervention group and 544 women and their children in the control group.

A cluster-randomised controlled trial of the programme found fewer children with Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III), cognitive scores below −1 SD in the intervention group (3%) than in the control group (6%) at age 2 years; however, the difference was not statistically significant (OR 0.55, 95% CI 0.26 to 1.17).⁷ The trial also found that children in the intervention group had higher Bayley-III cognitive development scores than those in the control group (mean difference 4.00, 95% CI 2.56 to 5.43), as well as improved language development (mean difference 2.48, 95% CI 0.55 to 4.41) and motor development (mean difference 2.73, 95% CI 1.23 to 4.22). The programme was cost-effective to improve child cognitive development, with incremental cost-effectiveness ratios (ICER) of approximately 0.5% of Vietnam gross domestic product per capita, alongside benefits in language and motor development.⁸ Further, greater cost-effectiveness was found among disadvantaged groups, indicating improved equity.⁹ The study protocols and results are available elsewhere.^{6–9 11–13}

Study setting

Vietnam is a lower-middle income country in Southeast Asia with a population of approximately 100 million people. Over 60% of the population lives in rural areas. The cluster-randomised controlled trial was conducted in Ha Nam,^{6 7} a typical rural Red River delta province in Northern Vietnam. Ha Nam province has been merged into Ninh Binh province, following government restructuring in June 2025.

According to the census in 2021, Ha Nam had a population of 875 200, crude birth rate of 14.8 per 1000 people, under-5 mortality of 18 per 1000 livebirths and average age of first marriage of 26 years.¹⁴ The average household size was 3.4 members.¹⁵ The monthly average income per capita was 4372 thousand Vietnamese dong (around US$160), and the main source of income was salary or wage (52.4%), followed by self-employment in non-agriculture, forestry and fishery sectors (30.6%) in 2021.^{14 15}

Participants

Mother–child dyads who completed the baseline survey of the original trial are eligible to participate in this follow-up study. To maximise the re-recruitment rate, the successful strategies used in our previous follow-up studies^{16 17} conducted in the same setting will be employed, which achieved re-recruitment rates of 87%–92% among women and their infants. Commune health centre staff will contact eligible participants by telephone to inform them about the follow-up study and invite them to participate.

In Vietnam, local commune authorities maintain records of residents and their addresses, and citizens are required to update these records when they relocate. Further, people in rural Vietnam often live close to their extended families and maintain strong relationships with their neighbours. Consequently, when a participant has moved outside the community, updated contact details can often be obtained through local authorities, family members residing in the area or former neighbours. Ha Nam Province is located approximately 60 km south of Hanoi, the country’s capital and a major destination for rural-to-urban migration. As a result, most migration from the study area is to Hanoi. We will therefore attempt to contact and follow-up all participants who have migrated to Hanoi, where the project office is based.

During the original 3-year trial, the migration rate was 3.6%, and the refusal rate was 9.4%.⁷ The national mortality rate among children aged 1–4 years is approximately 5.33 per 1000 live births.¹⁸ No deaths occurred among children in our cohort from birth to 2 years of age, despite a national infant mortality rate of approximately 12.01 per 1000 live births.¹⁸ Overall, we anticipate achieving a re-recruitment rate of 80%, corresponding to approximately 900 of the 1168 children in the original cohort.¹⁸

Measures

Primary outcome

The primary outcome is child intellectual ability, measured using the Wechsler Intelligence Scale for Children, Fourth Edition (WISC-IV)¹⁹. A Full-Scale Intellectual Quotient (FS-IQ), ranging from 40 to 160 points will be estimated based on four indexes: (1) verbal comprehension index, (2) perceptual reasoning index, (3) working memory index and (4) processing speed index.^{19 20} A description of the measure is provided in table 1, with further details provided in online supplemental table 1.

Table 1. Measures.

Variables	Tools
Child at ages 7 and 9
Intellectual ability (primary outcome)	Wechsler Intelligence Scale for Children, Fourth Edition19
Emotional and behavioural development	Strengths and Difficulties Questionnaire29 30
Quality of life	EuroQol 5-Dimension Youth version 3-Level21
Anthropometric indices	Stadiometer and scale
Health conditions	Study-specific structured questions
Health behaviour: screentime, sleep time, physical activity	Study-specific structured questions
Education	Study-specific structured questions
Education expenditure	Study-specific structured questions
Healthcare expenditure	Study-specific structured questions
Mothers when their child is seven and 9 years old
Mental health	Self-Reporting Questionnaire 2031
Quality of life	EuroQol 5-Dimension 5-Level32
Parenting practices	Longitudinal Study of Australian Children,33 Brief Parental Self Efficacy Scale,34 UNICEF Multiple Cluster Index Surveys35
Intimate partner violence	WHO Multi-country Study on Women’s Health and Domestic Violence against Women36
Sociodemographic factors	Study-specific structured questions
Maternal cardiovascular disease risk factors and obstetric outcomes*	Study-specific structured questions
Maternal ageing biomarkers*	DNA methylation-derived estimates of biological age
Maternal cardiovascular health*	Questionnaires, blood pressure and lipid levels

Further details for each variable are provided in online supplemental tables 1 and 2.

^*Maternal health assessment when the child is 7 years old.

Secondary outcome

Secondary outcome measures include children’s emotional and behavioural development, quality of life, anthropometric indices, health conditions, health behaviour, educational practices and parenting. Additionally, maternal outcomes including mental health, quality of life, experiences of intimate partner violence, cardiovascular health and ageing biomarkers will be measured (table 1 and online supplemental tables 1 and 2).

Sociodemographic factors

On re-enrolment, sociodemographic factors will be measured to assess changes since the original trial, including marital status, births of other children, number of household members, whether a parent is working away, household wealth, main caregiver for children, women’s and men’s education level and occupation and household relocation through study-specific structured questions (table 1 and online supplemental table 1).

Costs

Costs will include intervention implementation and household participation costs and child healthcare and education expenditures from both healthcare and societal perspectives. Intervention and household participation costs will be obtained from a published cost and cost-effectiveness analysis, conducted at age 2.⁸

Expenditure on children’s healthcare and education in the past 12 months will be collected through caregiver surveys at ages 7 and 9. All costs in Vietnamese dong will be converted to US dollars.

Procedures

Individual consent will be obtained from women who will participate in the study. To measure child development, consent will be obtained from the parent or main caregiver to interview and assess their children.

Data collection will be conducted when children are aged 7 and again at aged 9 through direct assessments and surveys in private rooms at local commune health centres. All data collectors will be blinded to the trial arm allocation of the woman and her child.

The primary outcome, child intellectual ability, will be assessed by psychologists trained to use WISC-IV Vietnam validation. Child anthropometry data will be collected by health workers from the commune health centres. These data will be recorded on paper and then entered online using KoboToolbox.

Child-related questions will be reported by caregivers, except for the quality of life at age 9, as the EuroQol 5-Dimension Youth version 3-Level self-complete version is recommended for children aged 8–11 years.²¹ Based on participants’ preferences, the quality of life at age 9 will be completed by either the child or the caregiver. All other data will be self-reported by caregivers. Research staff will provide instructions for completing the questionnaire using KoboToolbox on a tablet. They will also be available to answer any questions.

Height, weight and blood pressure will be measured by trained research staff. Blood draws will be conducted by laboratory staff trained in phlebotomy. Four 2 mL tubes (two heparin tubes and two K2-EDTA tubes) will be collected. One heparin tube and one K2-EDTA tube will be transported to Ha Nam General Hospital for analysis within 24 hours. The tests include glycated haemoglobin and low-density lipoprotein/high-density lipoprotein/total cholesterol.

The two remaining tubes will be stored at Ha Nam General Hospital at –65 °C. On completion of the fieldwork, all stored samples will be transferred to –80 °C storage at Bach Mai General Hospital (Hanoi) and kept frozen until analysis of ageing biomarkers.

Several strategies will be implemented to minimise loss to follow-up and missing data. First, barriers to participation will be reduced by keeping data collection procedures simple, efficient and appropriate for participants, providing flexibility in data collection methods (self-completion or interview) and scheduling data collection at times convenient for participants and maintaining privacy by not using local data collectors. Second, an active participant-tracking system will be established and maintained, including follow-up through telephone contacts. Good relationships with participants will also be fostered through regular communication, including sending thank-you messages. Third, strong relationships with local authorities will be maintained to secure the support needed for participant tracking. Fourth, data collection procedures will be standardised to ensure consistency and quality. Data collectors will complete a 2-day enumerator training and a 9-day WISC training before commencing fieldwork. Fifth, missing data will be monitored in real time using KoboToolbox, and data collectors will be promptly notified of any missing or incomplete responses. Finally, each data collection team will include a supervisor responsible for ensuring data quality, including reviewing completed questionnaires in the field each day for completeness and logical consistency. These strategies have been successfully implemented in the main trial and in previous longitudinal studies conducted in the same setting, resulting in very low rates of loss to follow-up and missing data (<5%).^{7 16 17 22 23}

Data analysis

Linear mixed-effects models will be used to assess differences in primary outcomes between trial groups at ages 7 and 9 years, following the intention-to-treat principle. For secondary outcomes, mixed-effects logistic regression models will be used for binary outcomes, and linear mixed-effects models will be used for continuous outcomes. In all models, trial group (intervention vs control) will be included as a fixed effect, and commune will be included as a random intercept to account for clustering (Equation 1). Intervention effects will be estimated as mean differences for continuous outcomes and ORs for binary outcomes, with corresponding 95% CIs. Consistent with the analytical approach used in the original trial, no adjustment will be made for baseline covariates. Models will be fitted separately for the 7-year and 9-year follow-up assessments. For continuous outcomes, statistically significant mean differences will be converted to Cohen’s d to facilitate comparisons of effect sizes across outcomes and time points.

$${\displaystyle {\displaystyle Y_{ij}={\beta }_0+{\beta }_{1}Tria{l}_{Arm}+u_j+{\epsilon }_{ij}}}$$ (Equation 1)

Where:

• Y_ij = outcome

• β₀ = overall intercept

• β₁ = fixed effect of trial arm

• uj = random intercept for commune j

• ε_ij = individual-level residual error

Exploratory moderation analyses will be conducted for the primary outcomes. Potential moderators include child sex, household wealth at baseline and maternal education level at baseline. Each moderator will be examined separately for each primary outcome by including the moderator and its interaction with trial group in the linear mixed-effects models described above. Moderation will be assessed based on the direction, magnitude and statistical significance of the interaction coefficient, with a statistically significant interaction providing evidence that the intervention effect varies across levels of the moderator.

The mechanism of missing data will be examined by comparing available data between groups with and without missing outcome data, as well as by conducting Little’s missing completely at random (MCAR) test. If the data are found to be MCAR, full information maximum likelihood will be used to handle the missing data. If MCAR is not supported, we will adjust for covariates associated with loss to follow-up or missing data in the analyses. Data analysis will be conducted using Stata V.19.5 (StataCorp. 2025, TX).

Economic evaluation

An economic evaluation will estimate the cost-effectiveness of the intervention in improving child development at ages 7 and 9 from both healthcare and societal perspectives. The ICER will be calculated per FS-IQ point gained and per SD improvements at ages 7 and 9, respectively, in accordance with common reporting practices and the cost-effectiveness study of the original trial.^{8 24 25} The ICERs will be estimated by dividing the mean difference in costs by the mean difference in outcomes between the trial groups. Costs will include intervention cost obtained from the original trial⁸ and child healthcare and education expenditures at ages 7 and 9. Mean difference will be estimated using generalised linear mixed models, adjusting for cluster effects. Following the overall data analysis methods, trial group (intervention vs control) will be included as a fixed effect, and commune will be included as a random intercept to account for clustering. Additionally, non-parametric cluster bootstrapping will be performed to examine uncertainty around the ICERs.

The time horizons for the cost-effectiveness analysis will be 7.5 years at age 7 (from mid-pregnancy, 6 months before birth, to child age 7) and 9.5 years at age 9 (from mid-pregnancy to child age 9). An annual discount rate of 3% will be applied, consistent with previous studies,^{8 9 24} whereas alternative discount rates of 0% and 5% will be applied in sensitivity analyses.

The economic evaluations will follow the Consolidated Health Economic Evaluation Reporting Standards 2022 and best practices for trial-based economic evaluations.^{26 27}

Qualitative study

Study participants and data collection

A qualitative study will be conducted using purposive sampling to include caregivers, service providers and policymakers to understand the perspectives of community representatives. First, approximately 10 caregivers from the intervention group will be invited to participate in semi-structured interviews or focus group discussions to share their experiences with the intervention and their perspectives on its mid-term effects.

Second, semi-structured interviews will be conducted with approximately five service providers, including health workers, Women’s Union members and kindergarten teachers to gather their experiences delivering the intervention, lessons learnt, challenges and suggestions for sustaining and improving the programme. Potential participants will be identified and contacted through professional networks.

Lastly, approximately two policymakers will be invited to participate in semi-structured interviews to explore national early childhood development policies and plans, the use of intervention evidence in policymaking and the opportunities and challenges related to programme integration or scale-up at the national level. Government representatives from the Ministry of Health and the Provincial Department of Health will be identified and contacted through professional networks.

Each focus group discussion or semi-structured interview will last approximately 40 min and will be conducted face to face in private rooms, preferably at commune health centres or government offices. Interpretation from English to Vietnamese will be provided by a trained interpreter during interviews and discussions.

Data analysis

With participants’ consent, data will be audio-recorded and transcribed in English. Descriptive coding will identify common themes, and thematic analysis²⁸ will be performed using NVivo V.15.

Data management

Data will be downloaded directly to a secure Monash University server and be accessible only to approved research staff. The follow-up data will be linked to the original trial data using the unique identification number assigned to each participant. Survey responses and personal information will be stored separately, and each participant will be identified only by a unique identification number for analysis.

Ethics

Ethics approval to conduct the original randomised controlled trial was obtained from the Monash University Human Research Ethics Committee, Melbourne, Australia (Certificate Number 20160683), and the Institutional Review Board of the Hanoi School of Public Health, Hanoi, Vietnam (Certificate Number 017–017-377IDD-YTCC).

For this follow-up study, ethics approval was obtained from the Monash University Human Research Ethics Committee, Australia (Certificate Number 2025–46034-125952), with additional approval by the Ha Nam local government.

All data will be password-protected, and participants will be identified by a unique identification number, with all personally identifying information removed to maintain confidentiality. For informed consent, parents or main caregivers will be provided with a consent form that outlines the study objective, procedures, potential risks and benefits, information on confidentiality and data use. The form will emphasise the voluntary nature of participation and the option to withdraw at any time. Additionally, interviewers will be available to answer any questions parents may have, providing further clarification to help them make an informed decision.

Regarding potential harms and benefits, discussions of personal experiences such as parenting, intimate partner relationships and domestic violence may be distressing or uncomfortable. To minimise this discomfort, all participants will be fully informed about the aim of the research, their right to skip or refrain from answering any questions, take a break or withdraw from the survey at any time. If a participant discloses an experience of domestic violence, they will be provided with information about local support services and referred to the Women’s Union, a socio-political organisation representing the legal and legitimate rights and interests of Vietnamese women. As potential benefits, participants will have the option to receive a summary of the research findings, which could enhance their parenting knowledge and support their children’s development. In cases where children are found to have ‘extremely low’ intelligence scores, parents will be informed and provided with a referral to an appropriate child disability service. Further information on appropriate specialists and facilities for follow-up visits will be provided.

Dissemination

We will widely disseminate our results through multiple channels, including open-access peer-reviewed publications, national and international conference presentations, seminars and technical reports. Additionally, we will engage with print, broadcast and social media both in Vietnam and internationally. We will conduct national policy briefings in Vietnam and provide local-language reports and lay summaries specifically designed for participants, communities and policymakers.