Home‐Based Learning Opportunities, Responsive Caregiving and the Development of Preschool‐Aged Children in Low‐ and Middle‐Income Countries of the East Asia and Pacific Region: A Systematic Review

ABSTRACT

Background

Optimal early childhood development predicts lifelong health and well‐being. A child's immediate environment, especially the home, shapes cognitive, physical, language, motor, social and emotional development. Contextually relevant data on the proximal settings that support preschool‐aged children are lacking in low‐ and middle‐income countries (LMICs) across the East Asia and Pacific region. This systematic review synthesises evidence on associations between home‐based learning opportunities and caregiving experiences and the cognitive and overall development of children aged 2–5 years in LMICs of the East Asia and Pacific region.

Methods

We searched Medline, Embase, PsycINFO, Emcare and Global Health for studies published between January 1990 and January 2025. The search strategy used subject headings and free text terms using the condition, context and population (CoCoPop) framework to capture studies on cognitive or overall development (condition), home‐based learning opportunities and responsive caregiving (context) and children aged 24–59 months in LMICs in the East Asia and Pacific region (population). Data were independently extracted and appraised using the Joanna Briggs Institute Critical Appraisal Tools risk of bias and quality assessment checklists. The findings were summarised using narrative syntheses.

Results

Studies were from lower and upper middle‐income countries in East Asia, and the study quality ranged from medium to high. Eighteen of the 19 included studies reported significant positive associations between home‐based learning opportunities and caregiving and higher child development. The strongest effects were linked with psychosocial stimulation, academic‐focused caregiving and enriched home learning environments. Measures of learning opportunities and caregiving varied widely.

Conclusions

The review findings suggest that home‐based learning opportunities and responsive caregiving experiences may be associated with improved developmental outcomes among 2‐ to 5‐year‐old children. Although these patterns are consistent with broader evidence, the heterogeneity of measures and limited representation mean that conclusions should be interpreted with caution. Future research could prioritise underrepresented settings, including low‐income countries and Pacific Island nations, to ensure policy reflects regional diversity. The heterogeneity in measures highlights the need for standardised tools to better capture associations between child development and the home environment.

Summary

• Significant positive associations between home‐based learning opportunities, caregiving and higher child development were reported in 18 of 19 papers.

• Findings point to modifiable factors, such as caregiving behaviours, parental knowledge and access to learning materials, that may support child development.

• Future research could prioritise underrepresented settings, including low‐income countries and Pacific Island nations, to inform inclusive policy.

• Wide variability in measures highlights the need for standardised tools to better assess associations between child development and the home environment.

1. Introduction

Optimal early childhood development (ECD) is an important predictor of future adult health and well‐being (Black et al. 2017; Richter et al. 2017). Research and programmatic attention have traditionally focused on the window from conception to age 2 (Black et al. 2017; Britto et al. 2017; Richter et al. 2017; Walker et al. 2011). However, the 2‐ to 5‐year age range—also referred to as the next 1000 days—is now recognised as a significant life phase. This is due to the potential for cumulative and continuous development among preschool aged children who experience appropriate enabling environments (Draper et al. 2023; Draper et al. 2024; Nores et al. 2024; Tomlinson et al. 2021). The next 1000 days presents opportunities to build on the developmental gains made during infancy, setting children on an optimal trajectory into middle childhood, adolescence and beyond (Draper et al. 2024). Significant language, literacy, cognitive, numeracy, social and self‐regulation development occurs during this time (Draper et al. 2024), and both physical and cognitive catch‐up growth are possible (Crookston et al. 2013; Fink and Rockers 2014; Sudfeld et al. 2015).

Cognitive development refers to the biological and experiential processes by which children acquire more mature ways of thinking about the world, including memory, attention, problem solving, reasoning, organising and creating new ideas (American Psychological Association 2023). It is associated with changes in the brain that result from biological maturation, experience and the environment (American Psychological Association 2023). Brain development occurs fastest in the first 5 years of life, and poverty, together with poor caregiving environments, can adversely affect this process (Shonkoff and Garner 2012; Walker et al. 2011).

The quality of a child's immediate environment influences their cognitive, physical, language, motor, social and emotional development. The proximal setting has the most direct influence on child development (Bronfenbrenner and Morris 2006). It includes a child's family, peers and school and is a core area of focus articulated in the World Health Organization's Nurturing Care Framework (World Health Organization 2018a). Nurturing care refers to a stable environment that supports five essential components for optimal ECD: health, nutrition, safety, early learning and responsive caregiving. Currently, an estimated 182 million children aged 3–4 years (75%) in low‐ and middle‐income countries (LMICs) lack minimally adequate nurturing care, across at least one component each of the five dimensions of nurturing care (McCoy et al. 2022). Due to many factors such as poverty and lack of access to health and education, children may, for example, experience low weight for age, inadequate water, sanitation and hygiene, physical punishment and a lack of parental stimulation or learning materials such as books in the home (McCoy et al. 2022). Evidence from a meta‐analysis shows that whereas nutrition interventions alone have only small effects on cognitive and language outcomes, stimulation interventions yield substantially larger benefits, highlighting the need to strengthen early caregiving (Aboud and Yousafzai 2015).

Because young children tend to spend most of their time with parents or primary caregivers, these relationships influence a child's development (Bronfenbrenner and Morris 2006). Early experiences shape neural architecture, with chronic stress, deprivation and limited stimulation disrupting the development of executive function and self‐regulation skills (Nelson et al. 2023). A study of caregiver‐reported developmental measures in 35 LMICs estimated that 80.8 million (32.9%) children aged 3 and 4 years experienced low cognitive and/or social emotional development (McCoy et al. 2016). Low cognitive development was defined as an inability to follow simple directions and work independently, and low social–emotional development was defined as an inability to control aggression, avoid distraction and/or get along with other children (McCoy et al. 2016).

1.1. Home‐Based Responsive Caregiving and Learning Opportunities

Caregiver–child interactions, including engagement in stimulation, play or providing learning opportunities are a crucial aspect of nurturing care supporting cognitive and social–emotional development (Bornstein and Putnick 2012; Engle et al. 2011). Responsive caregiving is defined as behaviours in which the caregiver promptly follows the child's lead, offers emotional support, adjusts their actions in direct response to the child's cues and engages in ways that are conceptually linked to the child's behaviour, ensuring interactions are developmentally appropriate rather than intrusive or controlling (Black and Aboud 2011). Parenting quality, in terms of the extent to which a child receives responsive caregiving, is linked with positive impacts on child development (Britto et al. 2017; Jeong et al. 2017). Studies report higher developmental scores for children experiencing more parental psychosocial stimulation and caregiving practices (Ma et al. 2023; Rocha et al. 2022; Russell et al. 2022; Urke et al. 2018; Watanabe et al. 2005).

Early learning refers to any opportunity for children to interact with people, places or objects, recognising that every interaction, or absence of interaction, contributes to brain development (World Health Organization 2018b). Some examples of home‐based learning opportunities include talking, singing or telling stories with a caregiver, playing independently with toys or household objects, looking at picture books or interacting with siblings or peers through shared play. Parental engagement in learning opportunities activities, for example, reading, naming objects and having books in the home, was associated with better child development (Hasan et al. 2023; Mohammed et al. 2023; Rey‐Guerra et al. 2022; Walker et al. 2000). Studies suggest associations between early learning and responsive caregiving and higher cognitive and developmental outcomes (Jeong, Franchett, et al. 2021; Prime et al. 2023). Furthermore, nurturing early home environments during the preschool years can mitigate the negative effects of early cumulative adversities on cognitive development (Trude et al. 2021).

1.2. East Asia and Pacific Region

The East Asia and Pacific region is economically, geographically and culturally diverse and is home to one‐quarter of the global child population (around 580 million) (World Bank Group 2024). The region faces socio‐economic and urban–rural disparities that impact access to and quality of early childhood education and care, including in the home settings (UNICEF 2024). Socio‐economic disparities occur within and between countries in the region and are affected by factors such as parental education, occupation and family wealth (Sun et al. 2018). Urban–rural disparities are common in the East Asia and Pacific countries, with rural children often experiencing lower development than urban children (Rao et al. 2019). The Sustainable Development Goals (SDGs) emphasise the importance of ECD and the need for research to meet these targets, particularly SDG 4.2, which aims to ensure all children have access to quality ECD, care and pre‐primary education (United Nations 2015).

Although progress has been made in the East Asia and Pacific region, inequities persist. UNICEF estimates that at the height of the COVID‐19 pandemic lockdowns, over 150 million children under 5 years old experienced disruptions to health, nutrition, early learning and caregiving environments (UNICEF 2021). Indeed, some countries were already off track for SDG 4.2 before the pandemic, with COVID‐19 further slowing progress across child health, education and social protection systems (Economic and Social Commission for Asia and the Pacific (ESCAP) 2020). Even prior to the pandemic, available national estimates of children aged 36–59 months who were developmentally on track ranged from one‐third to over four‐fifths, with countries including Cambodia, Lao PDR and Myanmar showing substantially lower ECDI performance than Mongolia, Vanuatu and Kiribati (UNICEF 2020). Despite gains in pre‐primary participation in the region, approximately one in four children do not attend preschool, and these children tend to be from the most disadvantaged families (United Nations Children's Fund (UNICEF) 2019). Together, these patterns highlight gaps in children's early environments and highlight the need for stronger evidence on the factors shaping early development.

Yet, despite this clear need, there remains limited observational research examining how home environments relate to the development of children aged 2–5 years in LMICs in the East Asia and Pacific region. Reviews of observational studies are an important component of public health research to capture naturally occurring variations between the exposures and outcomes of interest. Although existing systematic reviews and meta‐analyses have explored parenting and child development in LMICs more broadly, they are limited in scope and relevance to this context. Specifically, these reviews have predominantly synthesised data from interventions studies rather than observational studies; they have focussed on broader age ranges rather than the preschool age groups; and they encompass regions beyond East Asia and Pacific (Jeong, Franchett, et al. 2021; Jeong, Pitchik, and Fink 2021; Rao et al. 2017; Yang et al. 2021; Zhang et al. 2021).

1.3. The Current Study and Aim

Nuanced, context‐specific data on how features of the home environment relate to young children's development are needed to inform locally relevant policy in the East Asia and Pacific region (Draper et al. 2024). A systematic review focused specifically on observational studies from the East Asia and Pacific LMICs can fill this evidence gap and provide a stronger empirical foundation for resource allocation and planning of ECD.

Our aims were to (1) identify observational studies examining home‐based learning opportunities and responsive caregiving experiences among children aged 2–5 years in LMICs of the East Asia and Pacific region; (2) evaluate the quality of the studies; and (3) synthesise the reported associations between home‐based learning opportunities, responsive caregiving experiences and children's cognitive and overall development, including any moderators or mediators assessed within the studies.

2. Methods

This review followed the methods and reporting requirements for systematic reviews and meta‐analysis set out in the Preferred Reporting Items for Systematic Reviews and Meta‐analysis (PRISMA) and Conducting Systematic reviews and Meta‐analyses of Observational Studies of Etiology (COSMOS‐E) guidelines (Dekkers et al. 2019; Moher, Liberati, Tetzlaff, Altman, and PRISMA Group 2009). The protocol was registered on the International prospective register of systematic reviews, PROSPERO, database (ID CRD42023443851) on 20 July 2023.

The review pertains to the LMICs in the East Asia and Pacific Region, as classified by the World Bank, which include Cambodia, China, Indonesia, Lao PDR, Malaysia, Mongolia, Myanmar, Papua New Guinea, the Philippines, Thailand, Timor‐Leste, Vietnam, Fiji, Kiribati, Marshall Islands, Federated States of Micronesia, Samoa, Solomon Islands, Tonga, Tuvalu and Vanuatu (World Bank Group 2024).

2.1. Search Strategy

An initial search was conducted in Google Scholar and Medline to explore relevant text words and phrases contained within titles and abstracts and to determine how indexers used subject terms. A search plan containing identified key words and index terms was subsequently developed and carried out across five databases (Ovid Medline, Ovid Embase, Ovid PsycINFO, Ovid Emcare and Ovid Global Health), retrieving papers from 1990 to 2025. The search strategies used a combination of subject headings (MeSH) and free‐text terms using the CoCoPop framework to include three concepts (Hosseini et al. 2024). Condition included cognitive development and overall child development; context included home‐based learning opportunities and/or responsive caregiving; and the population included children aged 2–5 years in LMICs of the East Asia and Pacific region. Searches were adapted to the specifications of each of the five databases and conducted on 16 October 2023 and again 31 January 2025. A full Ovid Medline search strategy is available in the Supporting Information (see Data S1).

2.2. Selection Criteria

Studies were included if they satisfied all the following inclusion criteria:

• Children aged 24–59 months or with a mean age under 59 months in LMICs in the East Asia and Pacific Region, as classified by the World Bank.

• Home‐based learning opportunities and/or responsive caregiving experiences (e.g., maternal or paternal engagement, parenting style and home learning activities such as reading, singing, playing, telling stories, going outside or having books or toys in the home).

• Child cognitive development, or overall index of child development, numeracy–literacy measure, numeracy measure or executive function.

• Observational studies including cross‐sectional, case–control, cohort studies and secondary analyses of data that quantified the associations between cognitive development or proxy measures and home‐based responsive caregiving and/or learning opportunities.

• Peer‐reviewed papers published in English between 1 January 1990 and 31 January 2025.

Studies were excluded if they reported any one or more of the following criteria:

• Children with special needs in institutions or humanitarian settings.

• Child outcomes focussing only on nutrition (e.g., height for age and weight for age outcomes) or health (e.g., child immunisation) or security and safety observations (e.g., exposure to violence in community—witness or victim).

• Early learning and caregiving environments outside the home (e.g., early childhood centre setting or community‐based settings such as toy and book libraries).

• Motor development or social–emotional development.

• Countries outside the World Bank East Asia and Pacific region.

• High‐income countries in World Bank East Asia and Pacific region (Hong Kong excluded as high‐income area).

• Intervention studies, studies of pooled data across multiple countries without individual country‐level data.

2.3. Study Selection

This systematic review was managed using the Covidence Systematic Review Management System, a web‐based collaboration software platform that streamlines the production of systematic reviews (Veritas Health Innovation 2024). The papers were extracted and uploaded to Covidence and screened for duplicates. With oversight from senior team members (J.F. and T.T.), the titles and abstracts of papers were screened by one reviewer (S.P.) and those meeting the inclusion criteria progressed to the full‐text review stage. Any conflicts arising during the title and abstract screening were raised and resolved with the team. Full‐text review papers were then retrieved and assessed by two reviewers (S.P. and T.H.) for potential eligibility in the final selection.

2.4. Data Extraction

Data extraction and quality assessment of the papers meeting the inclusion criteria were independently undertaken by two reviewers (S.P. and T.H.). Data from the included papers were extracted by S.P. and T.H. into a summary form that included author, year, location, study design, sample size, measure of exposure (responsive caregiving and/or learning opportunities measure) and outcomes (child cognitive or overall development measures), timing of assessment and main findings. We extracted only those associations that included responsive caregiving and/or learning opportunities as experiences. Other associations reported in the same studies were not systematically captured and are not included in our counts. The study team reviewed and extracted data collaboratively and resolved any differences in interpretation through discussion.

2.5. Quality Appraisal

The risk of bias and quality assessment was appraised independently by S.P. and T.H. using the Joanna Briggs Institute Critical Appraisal Tools checklists (Joanna Briggs Institute 2017). The eight criteria were each scored 0 = no, 1 = unclear or 2 = yes. The total score for each paper was calculated by dividing the numerator (summed score of the eight questions) by the denominator (maximum possible score of the eight questions; 16) and then multiplying by 100 to obtain the percentage score. Given the absence of standard guidelines to categorise studies based on their quality, we followed Burger et al.'s model, which involves a recommendation of high quality for studies scoring more than 85%, medium quality for studies scoring from 50% to 84% and low quality for studies scoring less than 50% (Burger et al. 2020). The intraclass correlation coefficient (ICC) evaluated the level of agreement of the study quality scores between the two assessors. ICC ranges from 0 to 1. ICC values less than 0.5, between 0.5 and 0.75, between 0.75 and 0.9 and greater than 0.9, indicates poor, moderate, good and excellent agreement, respectively (Koo and Li 2016). The ICC was calculated using SPSS Version 29. Following independent scoring, we reached consensus on the final scores for the included papers by resolving any discrepancies through discussion.

2.6. Synthesis of the Results

We conducted narrative syntheses of the findings of associations between cognitive and overall child developmental outcomes and home‐based learning opportunities and responsive caregiving experiences. Additionally, we reported measures of responsive caregiving, learning opportunities and early childhood developmental outcomes and modifiers of these relationships.

3. Results

3.1. Study Characteristics

The study selection process is presented in the PRISMA flow diagram (Figure 1). A total of 11 571 initial studies were identified in the five databases, after which 4088 duplicates were removed. The remaining 7483 studies were screened according to the eligibility criteria, producing a short list of 90 studies. Multi‐country papers were eligible only when at least one country met the inclusion criteria and individual country‐level data were available. Studies were excluded if they were intervention trials, unpublished, conducted outside the home setting, focused on children younger than or older than 5 years, involved children with special needs or those in institutional or humanitarian settings or were set in high income or non‐regional countries. Although formally part of China, studies from Hong Kong were excluded due to their high‐income setting. Eligible outcomes included cognitive development, overall developmental indices, numeracy and executive function measured using continuous or categorical scale. Finally, 19 studies were included in the systematic review covering a combined study population of 44 717 parent and child pairs. Due to the heterogeneity of measures of child development, home‐based learning opportunities and responsive caregiving among studies, meta‐analysis was not possible. All manuscripts were published between 2012 and 2024 (Table 1). Ten studies were secondary analyses, seven were cross‐sectional, and two were longitudinal studies. Some of the secondary analyses used country‐level data from multiple countries. Within the East Asia and Pacific region, 16 studies were conducted in upper‐middle income countries (China: 8, Indonesia: 2, Mongolia: 1, Thailand: 5), and nine were conducted in lower‐middle income countries (Cambodia: 2, Lao PDR: 1, Papua New Guinea: 1, Timor‐Leste: 1, Vanuatu: 1 and Vietnam: 3) (Figure 2). Sample sizes varied substantially; the smallest was 58 mother and child pairs in a cross‐sectional study in Indonesia (Warsito et al. 2012), and the largest sample size was 6557 mother and child pairs in a secondary analysis in Thailand (Topothai et al. 2024). The most common sampling methods were multistage stratified cluster sampling following the standard UNICEF Multiple Indicator Cluster Surveys (MISCs) method (Duc 2016; Manu et al. 2019; Palakawong‐Na‐Ayudhya and Rukumnuaykit 2019; Topothai et al. 2024; Topothai et al. 2022) and convenience sampling, recruiting participants from specific locations like childcare centres and preschools (Liu et al. 2020; Wang et al. 2022; Wei et al. 2023; Wu et al. 2020; Yang et al. 2023).

FIGURE 1.

PRISMA flow diagram of included studies.

TABLE 1.

Summary of included studies.

Author and yearCountry	Study design and sample size	Responsive caregiving/early learning measure. Timing of assessment	Child cognitive development measure. Timing of assessment	Main findings
Berkes et al. 2019 Cambodia	Secondary analysis of data from a cohort study of children in rural Cambodia 6457 caregiver and child pairs	Measure: Study specific survey measuring cognitive parenting (activities), social emotional parenting (parent–child bonding) and negative parenting (punitive behaviour) Timing: 3–5 years postpartum	Measure: Executive function, language and early numeracy based on MELQO toolkit Timing: 3–5 years	Cognitive parenting has the strongest association with cognitive scores. A one‐SD increase in cognitive parenting associated with a 0.092–0.098 SD increase in language and early numeracy scores The association of cognitive parenting and early numeracy is 54% weaker for stunted children The wealth gap in cognitive competencies widens with age (Early numeracy age 3:0.12 SD—age 5: 0.88 SD)
Duc 2016 Vietnam	Secondary analysis of MICS 2011 1459 mother and child pairs	Measure: Adult engagement in learning, violent discipline using MICS based on modified CTSPC Timing: 36–59 months postpartum	Measure: Child development, literacy‐numeracy domain, learning domain using MICS ECDI module Timing: 36–59 months	Children whose parents engaged in 4+ learning activities is significantly associated with ECDI (AOR, 1.55; 95% CI 1.13–2.14) and literacy–numeracy domain (AOR, 2.01; 95% CI 1.39–2.91). Children who experienced physical punishment were 0.69‐fold (95% CI 0.51–0.95) less likely to be on track Children whose mother believes in physical punishment were less likely to be on track for literacy–numeracy (AOR, 0.48; 95% CI 0.32–0.73) but more likely for the learning domain (AOR, 2.06; 95% CI 1.13–3.76)
Ernawati et al. 2019 Indonesia	Secondary analysis of a longitudinal study 150 mother and child pairs	Measure: Psychosocial aspects of caregiving using HOME Timing: 36 months postpartum	Measure: Cognitive development using BSID‐III Timing: 36 months	Higher cognitive development is significantly associated with psychosocial caregiving (OR: 3.54, 95% CI 1.43–8.76, p = 0.003); learning stimulation (OR: 4.21, 95% CI 1.70–10.42, p = 0.001); language stimulation (OR: 2.68, 95% CI 1.03–6.98, p = 0.031); and academic stimulation (OR: 5.99, 95% CI 1.73–20.8, p = 0.001)
Likhitweerawong et al. 2023 Thailand	Cross‐sectional study 1540 parent and child pairs	Measure: Parenting style using Short Form of PSDQ‐Thai Timing: 2–5 years postpartum	Measure: Executive function using BRIEF‐P‐Thai Timing: 2–5 years	Permissive parenting style is significantly associated with increased risk of poor executive function (OR: 2.75, 95% CI: 1.24–6.13, p < 0.006)
Lin et al. 2020 China	Cross‐sectional study 163 child and father pairs	Measure: Play beliefs using CPPBS Timing: About 3 years postpartum	Measure: Overall child development using CDSC Timing: Mean age 38.73 months	Pragmatic Fathers' (who value pre‐academic behaviours over free play) predicted better overall development (B = 0.17, p 0.05) compared with ‘Hedonistic Fathers’ (who value free play over pre‐academic behaviours)
Liu et al. 2020 China	Cross‐sectional study 748 parent and child pairs	Measure: Parenting self‐efficacy using SEPTI‐TS and parental involvement using IT‐HOME Timing: 25–36 months postpartum	Measure: Cognitive competence using BSID‐II Timing: 25–36 months	Cognitive competence is directly associated with parental involvement (B = 0.18 p < 0.01) and parental self‐efficacy (B = 0.51, p < 0.01) Parental education predicted both parental self‐efficacy (B = 0.63, p < 0.01) and parental involvement (B = 0.42, p < 0.01) Parental education (B = 0.598, p < 0.01) and household income (B = 0.25, p < 0.01) directly affected children's cognitive competence
Manu et al. 2019 Lao PDR, Vietnam	Secondary analysis of MICS Parent and child pairs: 4457 Lao PDR, 1185 Vietnam	Measure: Availability of children's books using ECDI modules in MICS Timing: 36–59 months postpartum	Measure: Literacy‐numeracy using LNI derived MICS ECDI Timing: 36–59 months	Lao PDR: Children with at least one children's book is more likely to be on track for LNI (ARR: 1.87, 95% CI: 1.59–2.19) Vietnam: Children with at least one children's book is more likely to be on track for LNI (adjusted RR: 1.72, 95% CI: 1.37–2.15).
Nguyen et al. 2018 Vietnam	Secondary analysis of data from a PRECONCEPT RCT 1458 mother and child pairs	Measure: Home learning environment using IT‐HOME to assess social, emotional and cognitive support available to child Timing: During the first 24 months postpartum	Measure: BSID‐III including cognitive development Timing: 2 years old	Significant positive associations between HOME and cognitive development (β = 0.13, 95% CI 0.01–0.25, p = 0.04) Children living in more stimulating environments (i.e., higher HOME scores) had 0.23 higher z‐scores in cognitive development The negative associations between stunting on development were modified by HOME; the associations were strong among children living in homes with a poor learning environment
Palakawong‐Na‐Ayudhya and Rukumnuaykit 2019 Thailand	Secondary analysis of MICS‐4 4362 mother and child pairs	Measure: Adult‐child activities using MICS Timing: 36–59 months postpartum	Measure: Cognitive development assessed from EDCI MICS module Timing: 36–59 months	For children who live with both parents, cognitive development significantly associated with adult child activities singing, outings and playing have positive statistical effects For children who do not live with both parents or live with others, the positive effect remains only for singing
Rao et al. 2022 China	Secondary analysis of EAP‐ECDS validation study 1598 parent and child pairs	Measure: Home learning environment using EAP‐ECDS Timing: 3–5 years postpartum	Measure: Child development using EAP‐ECDS Timing: 3–5 years	Urban children scored 0.25 SD higher on the EAP‐ECDS than rural children (t = −5.32, p < 0.001) Maternal education modestly associated with ECD (β = 0.23, p = 0.003) There is no association between home learning environment and ECD (β = −0.01, p < 0.58)
Sun et al. 2018 Cambodia, China, Mongolia, Papua New Guinea, Timor‐Leste and Vanuatu	Secondary analysis of EAP‐ECDS Parent and child pairs: Cambodia 1189, China 1169, Mongolia 1230, Papua New Guinea 1639, Timor‐Leste 1176 and Vanuatu 730	Measure: Parental engagement (cognitive caregiving and socio‐emotional caregiving) adopted from MICS Timing: 36–50 months postpartum	Measure: Child development using EAP‐ECDS Timing: 36–50 months	Significant moderating effects of maternal engagement were found in the relationship between SES and child performance on approaches to learning (estimate = −0.09, SE = 0.03, p < 0.05) and on cognitive development (estimate = 0.05, SE = 0.03, p < 0.059) in Timor‐Leste Significant moderating effects of parental engagement found between family SES and child performance on cognitive development in Papua New Guinea (estimate = −0.07, SE = 0.02, p < 0.01) and on approaches to learning in Timor‐Leste (estimate = −0.06, SE = 0.03, p < 0.05) and Vanuatu (estimate = −0.15, SE = 0.06, p < 0.05)
Supanitayanon et al. 2020 Thailand	Longitudinal study 291 parent and child pairs	Measure: Positive parenting using PSDQ‐short version Timing: 3 and 4 years postpartum	Measure: Cognitive development using Mullen Scales of Early Learning Timing: 3 and 4 years	Positive parenting behaviour at 4 years of age is correlated with duration of verbal interaction between caregivers and children during screen time (r = 0.22, p < 0.01) Positive parenting behaviours at 3 years of age is correlated with ELC at age 3 (r = 0.22, p < 0.01) and age 4 (r = 0.27, p < 0.01) Positive parenting behaviours at 3 years of age is correlated with ELC at age 4 (r = 0.27, p < 0.01)
Topothai et al. 2024 Thailand	Secondary analysis MICS data 6557 child and mother pairs	Measure: Books in home using MICS Timing: 24–59 months postpartum	Measure: Overall development using MICS ECDI2030 Timing: 24–59 months	More books at home (AOR = 1.59 for 3–9 books; AOR = 2.40 for 10+ books) positively associated with being developmentally on track
Topothai et al. 2022 Thailand	Secondary analysis MICS data 5787 child and mother pairs	Measure: Parental interactions using MICS Timing: 36–59 months postpartum	Measure: Overall development using MICS ECDI Timing: 36–59 months	Significantly higher chance of being developmentally on track for children who had appropriate parental interactions (more than four out of six interactions) than children with less than four interactions (OR = 1.52, 95% CI 1.14–2.04, p < 0.001)
Wang et al. 2022 China	Cross‐sectional study 122 child and mother pairs	Measure: Maternal nurturing and stress using PBR; Parent‐grandparent co‐parenting quality using the Parents' Perceptions of the Coparenting Relationship Scale Timing: 3 years postpartum	Measure: EF assessed using tasks adopted in Chinese preschooler studies; EF using ECBQ Timing: Children aged 3 years	Coparenting has no direct relationship with EF, but there was an indirect relationship between co‐parenting and cognitive flexibility through the mediation of maternal nurturing and stress Maternal nurturing positively predicted children's cognitive flexibility (B = 0.18, p < 0.05), but maternal stress is more likely to lower cognitive flexibility (B = −0.24, p < 0.01)
Warsito et al. 2012 Indonesia	Cross‐sectional study 58 mother and child pairs	Measure: Psychosocial stimulation using HOME Timing: 3–5 years postpartum	Measure: Cognitive development using Indonesia Department of National Education instrument Timing: Children aged 3–5 years	Children with high psychosocial stimulation scored the highest on cognitive development (p < 0.001) Psychosocial stimulation had a significant effect on cognitive development (β 0.407, p = < 0.001)
Wei et al. 2023 China	Cross‐sectional study 126 mother and child pairs	Measure: Observational maternal parenting using the Parental Warmth and Control Scale; mother rated maternal parenting using PSDQ Timing: 42–54 months postpartum	Measure: EF using the Backward Digit Span, Corsi Block Tapping, Day‐Night Stroop and HTKS. Mother Rated EF using the Childhood Executive Functioning Inventory Timing: Children aged 42–54 months	The latent performance‐based EF was uniquely predicted by maternal positive control and negative control in mother–child interaction Children's EF difficulties reported by mothers were predicted by mother‐reported warmth and support and autonomy granting Positive control is correlated with HKTS (r = 0.31, p < 0.05), warmth and support is correlated with working memory (r = 0.40, p < 0.05) and planning (r = 0.32, p < 0.05) and autonomy granting is correlated with working memory (r = 0.39, p 0.05) and planning (r = 0.35, p < 0.05)
Wu et al. 2020 China	Longitudinal study 730 parent and child pairs	Measure: Parenting quality assessed using parenting assessment tool Timing: 18 months postpartum	Measure: Child development using ASQ‐3 Timing: 36, 42 and 48 months	Child communication scores more likely to be higher if parents have knowledge of parenting concept (β a  = −0.58, p < 0.01) and parent–child relationship (β a  = −0.26, p < 0.01) language stimulation (β a  = −1.48, p < 0.001) and safety (β a  = −0.48, p < 0.001) Children receiving poor quality parenting more likely of suspected development delay (OR: 2.74, 95% CI 1.17–6.40) compared with children receiving high‐quality parenting
Yang et al. 2023 China	Cross‐sectional study 90 parent and child pairs	Measure: Parent–child engagement in math‐related home activities using questionnaire adapted from LeFevre et al. Timing: 4–5 years postpartum	Measure: Numeracy skills measured using WOC, NLET SCT and arithmetic task; Executive function measured using DCCS Timing: Children aged 4–5 years	Parent–child engagement in formal numeracy activities was significantly and positively associated with child's knowledge of cardinality—the number of elements in a given mathematical set (β = 0.19, p < 0.05) Diversity of parent number talk was significantly and positively associated with child's symbolic number comparison (β = 0.36, p < 0.05)

Abbreviations: ASQ‐3 = Ages and Stages Questionnaire, BRIEF‐P = Behaviour Rating Inventory of Executive Functioning‐Preschool, BSID‐III = Bayley Scales of Infant and Toddler Development‐Third Edition, CDSC = China Developmental Scale for Children, CPPBS: Chinese Parent Play Beliefs Scale, CTSPC = Parent to Child Conflict Tactics Scale, DCCS = Dimensional Change Card Sort, EAP‐ECDS = East Asia–Pacific Early Child Development Scales, ECBQ = Early Child Behaviour Questionnaire, ECD = early childhood development, ECDI = Early Childhood Development Index, EF = executive function, ELC = early learning composite, HOME = Home Observation for Measurement of Environment Inventory, HTKS = Head‐Toes‐Knees‐Shoulders, IT‐HOME = Infant/Toddler Home Observation for Measurement of the Environment, LN = literacy–numeracy index, MELQO = Measuring Early Learning Quality and Outcomes, MICS = Multiple Indicator Cluster Survey, NLET = number line estimation task, PBR = Parent Behaviour Report, PSDQ = Parenting Styles and Dimensions Questionnaires, RCT = randomised controlled trial, SCT = symbolic comparison task, SEPTI‐TS = Self‐Efficacy for Parenting Task Index‐Toddlers Scale, SES = socio‐economic status, WOC = What's‐on‐this‐card.

FIGURE 2.

Distribution of studies in LMICs of the East Asia and Pacific region.

3.2. Study Quality

Methodological quality was rated high for six studies (> 85%) (Lin et al. 2020; Liu et al. 2020; Nguyen et al. 2018; Sun et al. 2018; Warsito et al. 2012; Wei et al. 2023) and medium (50%–85%) for 13 studies (Berkes et al. 2019; Duc 2016; Ernawati et al. 2019; Likhitweerawong et al. 2023; Manu et al. 2019; Palakawong‐Na‐Ayudhya and Rukumnuaykit 2019; Rao et al. 2022; Supanitayanon et al. 2020; Topothai et al. 2024; Topothai et al. 2022; Wang et al. 2022; Wu et al. 2020; Yang et al. 2023) (Table 2). The inter‐rater reliability was ICC 0.79, indicating good reliability.

TABLE 2.

Quality assessment using Joanna Briggs Institute Critical Appraisal Tools checklists.

Author and year	Q1. Were the criteria for inclusion in the sample clearly defined?	Q2. Were the study subjects and the setting described in detail?	Q3. Was the exposure measured in a valid and reliable way?	Q4. Were objective, standard criteria used for measurement of the condition?	Q5. Were confounding factors identified?	Q6. Were strategies to deal with confounding factors stated?	Q7. Were the outcomes measured in a valid and reliable way?	Q8. Was appropriate statistical analysis used?	Score (yes = 2, unclear = 1, no = 0)	Quality (low < 50%; medium 50%–85%; high > 85%)
Berkes et al. 2019	Yes	Unclear	No	Yes	Yes	Yes	No	Yes	11/16 = 0.69	Medium
Duc 2016	Yes	Yes	No	Yes	Yes	Yes	No	Yes	12/16 = 0.75	Medium
Ernawati et al. 2019	Yes	Unclear	Yes	Yes	Unclear	Unclear	Yes	Unclear	12/16 = 0.75	Medium
Likhitweerawong et al. 2023	Yes	Yes	No	Yes	Yes	Yes	No	Yes	12/16 = 0.75	Medium
Lin et al. 2020	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	14/16 = 0.88	High
Liu et al. 2020	Yes	Yes	Yes	Unclear	Yes	Yes	Yes	Yes	15/16 = 0.94	High
Manu et al. 2019	Yes	No	No	Unclear	Yes	Yes	No	Yes	9/16 = 0.56	Medium
Nguyen et al. 2018	Yes	Yes	Yes	Unclear	Yes	Yes	Yes	Yes	15/16 = 0.94	High
Palakawong‐Na‐Ayudhya and Rukumnuaykit 2019	Yes	No	No	Unclear	Yes	Yes	No	Yes	9/16 = 0.56	Medium
Rao et al. 2022	Yes	Unclear	No	Yes	Yes	Yes	Yes	Yes	13/16 = 0.81	Medium
Sun et al. 2018	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	14/16 = 0.88	High
Supanitayanon et al. 2020	Yes	Yes	No	Yes	Yes	Yes	Unclear	Yes	13/16 = 0.81	Medium
Topothai et al. 2024	Yes	Yes	No	Unclear	Yes	Yes	No	Yes	11/16 = 0.69	Medium
Topothai et al. 2022	Yes	Yes	No	Unclear	Yes	Yes	No	Yes	11/16 = 0.69	Medium
Wang et al. 2022	Yes	Unclear	No	Yes	Yes	Yes	Unclear	Yes	12/16 = 0.75	Medium
Warsito et al. 2012	Yes	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	15/16 = 0.94	High
Wei et al. 2023	Yes	Unclear	Yes	Yes	Yes	Yes	Unclear	Yes	14/16 = 0.88	High
Wu et al. 2020	Yes	Yes	Unclear	Yes	Yes	Yes	No	Yes	13/16 = 0.81	Medium
Yang et al. 2023	Yes	Unclear	Unclear	Yes	Yes	Yes	Unclear	Yes	13/16 = 0.81	Medium

All studies described the criteria for inclusion in the sample, and around half described the study participants in detail. Five studies measured responsive caregiving and learning opportunities in valid and reliable ways, for example, through observer‐reported measures such as the Home Observation for Measurement of Environment (HOME) Inventory (Ernawati et al. 2019; Liu et al. 2020; Nguyen et al. 2018; Warsito et al. 2012; Wei et al. 2023). Four studies measured the outcomes in replicable ways through direct assessment of children using the Bayley Scale of Infant and Toddler Development‐Third Edition (BSID‐III) (Ernawati et al. 2019; Lin et al. 2020; Liu et al. 2020; Nguyen et al. 2018), and two studies used the East Asia‐Pacific Early Child Development Scales (EAP‐ECDS) (Rao et al. 2022; Sun et al. 2018). The MICS is a parent report measure and is not considered a valid and reliable measure of home‐based responsive caregiving, learning opportunities or child development outcomes. Overall, 26% of studies employed direct assessments for home environment measures, and 31% of studies used direct assessments of child development measures.

Most studies conducted multiple regression modelling to determine associations between child development or cognitive development outcomes and home‐based learning opportunities and responsive caregiving experiences while adjusting for covariates. Only one study (Ernawati et al. 2019) did not adjust for any covariate. The most common covariates controlled for in the studies were maternal education (n = 14), household wealth (n = 13), child sex (n = 12), child age (n = 11), participation in an early childhood education program (n = 6), urban/rural (n = 5) and maternal age (n = 4). Four papers specifically articulated potential confounding factors: Likhitweerawong et al. (nutritional deficiency, adverse childhood experiences and limited accessibility receiving healthcare resources) (Likhitweerawong et al. 2023), Manu et al. (wealth index quintile, maternal education, rural or urban residence and age of the child) (Manu et al. 2019), Nguyen et al. (socio‐economic position (SEP), maternal ethnicity, occupation, education, IQ, child age, gender, birth weight centile and group assignment) (Nguyen et al. 2018) and Yang et al. (children's executive function) (Yang et al. 2023).

3.3. Home‐Based Learning Opportunities and Responsive Caregiving Measures and Indicators

Among the 19 studies, the most used measures of home‐based learning opportunities and responsive caregiving were the ECD module of the MICS (n = 6), HOME (n = 4), Parenting Styles and Dimensions Questionnaires (PSDQ) (n = 3) and the EAP‐ECDS (n = 1). Three studies used specifically developed measures to quantify parenting and learning opportunities (Table 1). Validity refers to how accurately a tool measures the outcome of interest. Across the studies, validity testing ranged from basic assessments such as face validity to more rigorous psychometric analyses, which reported internal consistency estimates (Cronbach's alpha). Basic assessments include face validity that is conducted for the MICS, whereby the questionnaires are reviewed and adapted to ensure local context appropriateness in the country‐specific setting. Face validity is important, but it is not a rigorous type of validity. Seven studies specifically reported strong reliability with high internal consistency estimates (Cronbach's alpha) (Berkes et al. 2019; Likhitweerawong et al. 2023; Liu et al. 2020; Sun et al. 2018; Wang et al. 2022; Wei et al. 2023; Yang et al. 2023). Other papers reported locally validated measures. For example, Lin et al. described that the measures had been developed, revised and validated in collaboration with parents and early childhood education experts in China, whereas Nguyen et al. described construct validity verification of the instruments during pretesting sessions (Lin et al. 2020; Nguyen et al. 2018).

Responsive caregiving was measured using 16 distinct indicators across the studies, highlighting substantial variation in how this dimension is gauged. Some examples of the 16 indicators are parenting style, parental self‐efficacy, parental involvement, parental interactions, caregiver non‐verbal reasoning, cognitive parenting, positive parenting, negative parenting and maternal and paternal engagement. Parenting style, for instance, was assessed using the 32‐item Short Form Parenting Styles and Dimensions Questionnaire (PSDQ‐Thai), which rates authoritative, authoritarian and permissive behaviours on a 5‐point Likert scale (Likhitweerawong et al. 2023). Although parenting self‐efficacy was measured using the Self‐Efficacy for Parenting Task Index‐Toddler Scales (SEPTI‐TS), a 53‐item measure of domain‐specific parenting self‐efficacy that assesses seven dimensions of caregiving: emotional availability, nurturance, safety, discipline, play, teaching and routines, which captures parents' confidence in providing responsive structured and developmentally supportive care to toddlers (Liu et al. 2020). There was less variation in the indicators employed for learning opportunities with six in total used across the studies, including home learning environment, activities (reading, storytelling, singing, playing and outings), having children's books in the house, playing with handmade toys, playing with toys from shops and adult engaged in learning. Not all measures assessing responsive caregiving aligned completely with established definitions. Some overlapped with aspects of the learning opportunities environment, making distinguishing between the two dimensions at times challenging.

3.4. Cognitive or Early Childhood Developmental Measures

Although cognitive development was our primary focus, we also considered overall child development as a secondary outcome. This was assessed in studies where overall child development was investigated, instead of cognitive development. Of the 19 studies, the main child development outcome assessed was overall child development (n = 7), followed by cognitive development (n = 6), executive function (n = 4), literacy‐numeracy (n = 2), language and early numeracy (n = 1), learning (n = 1) and numeracy (n = 1). Cognitive development and related outcomes were assessed using a diverse range of measures across the studies—19 measures in total—with a few studies employing more than one measure to capture the desired information. The most used measures were the MICS Early Childhood Development Index (ECDI) (n = 5), the BSID‐III (n = 3) and the EAP‐ECDS (n = 2). Two studies used measures developed specifically for the research (Table 2).

3.5. Associations Between Cognitive Developmental Outcomes and Home‐Based Learning Opportunities and Responsive Caregiving Experiences

Overall, 18 of the 19 studies reported a significant positive association between quality of home‐based learning opportunities and/or responsive caregiving experiences and higher child developmental outcomes (Table 2). In each of the included studies, between one and four exposures of responsive caregiving and/or learning opportunities were assessed for their potential positive association with a child development measure. Only one exposure was not significantly associated with higher child development scores.

Being developmentally on track means a child is progressing through developmental milestones at a typical rate and is achieving the skills expected for their age. For example, the ECDI is calculated by obtaining the percentage of children who are developmentally on track in at least three of four domains. Generally, the coding and scoring of the home‐based learning opportunities and responsive caregiving variables followed the definitions and indicators set out in the MICS and are indicated below. Studies measuring overall child developmental scores found that children had significantly higher chances of being developmentally on track when they received parental engagement (Sun et al. 2018; Wu et al. 2020), maternal engagement (Sun et al. 2018) and paternal engagement, particularly from ‘pragmatic fathers’ who value pre‐academic behaviours over free play (Lin et al. 2020), if they engaged in four or more stimulating activities with a relative (Duc 2016; Topothai et al. 2022), and had multiple books at home (Topothai et al. 2024). Duc (2016) found children whose mothers believed in physical punishment were less likely to be developmentally on track for literacy–numeracy (measured by three indicators: name at least 10 letters of the alphabet; read at least four simple, common words; knowing names and recognising symbols of all numbers from 1 to 10), but more likely to be on track for the learning domain (measured by two indicators: following simple directions on how to do something correctly; able to do something independently). One secondary analysis found no association between the home learning environment (measured by a six‐item scale whether caregiver provided home learning activities to child such as reading, telling stories and singing songs) and child development, instead revealing that maternal education and provincial economic status were the strongest predictors (Rao et al. 2022).

Higher cognitive development was significantly associated with psychosocial caregiving (assessed by how parents engage in learning, language and academic stimulation, modelling, warmth and acceptance, caregiving variation, punishment and acceptance and physical environment) (Ernawati et al. 2019), parental involvement (Liu et al. 2020), positive parenting (Supanitayanon et al. 2020) and high psychosocial stimulation (defined as satisfying 46 or more of the 55 item observer reported HOME) (Warsito et al. 2012). Higher cognitive development was significantly associated with learning, language and academic stimulation (Ernawati et al. 2019), stimulating home environments (Nguyen et al. 2018) and adult child activities (i.e., singing, playing and outings) (Palakawong‐Na‐Ayudhya and Rukumnuaykit 2019).

In one secondary analysis, executive function was significantly associated with cognitive parenting, which measured how parents actively interact with their child such as playing games, reading books and playing with toys or objects (Berkes et al. 2019). Executive function was significantly associated with mother–child interaction in one cross‐sectional study (Wei et al. 2023). Permissive parenting style was significantly associated with lower executive function in one cross‐sectional study (Likhitweerawong et al. 2023), whereas another cross‐sectional study investigated grandparent involvement and found co‐parenting had no direct relationship with executive function (Wang et al. 2022).

A secondary analysis investigating the association of children's books found that one or more books at home was associated with a higher likelihood of being on track for literacy and numeracy (Manu et al. 2019). A cross‐sectional study found that parent–child engagement in formal numeracy activities was significantly and positively associated with child numeracy (Yang et al. 2023).

3.6. Modification and Mediation Analyses

Three of the 19 studies tested the potential modifiers of the association. Sun et al. (2018) demonstrated that parental engagement in learning activities moderated the relationship between socio‐economic position and child development, particularly in the cognitive domain. Berkes et al. (2019) revealed that the positive effects of how parents engage with their child (playing games, reading books and playing with toys or objects) on language and early numeracy were stronger for non‐stunted children, indicating that nutritional status modifies the impact of parenting. Nguyen et al. (2018) showed that the negative associations between stunting and development were modified by HOME. The Liu et al. (2020) was the only paper testing potential mediations. The study found that higher socio‐economic position was associated with better cognitive competence in young children, mediated by increased parental involvement and parental self‐efficacy.

4. Discussion

Our aims were to identify observational studies examining home‐based learning opportunities and responsive caregiving experiences among children aged 2–5 years in LMICs of the East Asia and Pacific region; evaluate the quality of the studies; and synthesise the reported associations between these experiences and children's cognitive and overall development. Included studies were from lower and upper middle‐income countries in East Asia. The study quality ranged from medium to high. Eighteen of the 19 studies reported significant positive associations between home‐based learning opportunities and responsive caregiving experiences and higher child development. Across the included studies, between one and four experiences of responsive caregiving and/or learning opportunities were assessed; only one experience was not significantly associated with higher child development scores. Our findings suggest that home‐based learning opportunities and responsive caregiving experiences may be associated with improved developmental outcomes among 2‐ to 5‐year‐old children. The strongest effects were linked with psychosocial stimulation, such as emotional support, language and academic stimulation; academic‐focused caregiving, for example, structured numeracy activities; and enriched home learning environments, including the availability of children's books and adult–child interactions.

The findings in this review are consistent with global evidence demonstrating that responsive caregiving, early stimulation and enriched home learning environments are important predictors of early cognitive development across LMIC settings (Black et al. 2017; Jeong, Franchett, et al. 2021). Similar to studies from sub‐Saharan Africa, South Asia and Latin America, the associations observed in this review reinforce that caregiver–child interactions, access to learning materials and parental engagement may be linked to better developmental outcomes (Britto et al. 2017; Aboud and Yousafzai 2015, 134). However, the review also highlights substantial evidence gaps across the East Asia and Pacific region. Most studies originated from a small number of middle‐income countries, with no data from low‐income countries, nor many Pacific island nations including Fiji, Kiribati, Samoa, the Solomon Islands, Tonga and Tuvalu. Generalisability therefore remains uncertain given wide variation in structural and cultural conditions, alongside the limited representation of data in many countries in the East Asian Pacific region.

The caregiver–child relationship plays a central role throughout life, including during the 2‐ to 5‐year development stage, with responsive caregiving shown to support secure attachment, cognitive development and the development of executive functioning (Bornstein and Putnick 2012; Engle et al. 2011). The data in this review align with existing evidence of positive associations between responsive caregiving and child development (Ma et al. 2023; Rocha et al. 2022; Russell et al. 2022; Urke et al. 2018; Watanabe et al. 2005) and with the broader literature on the influence of the proximal environment on early child development (Britto et al. 2017; Bronfenbrenner and Morris 2006).

Home‐based learning opportunities, such as reading, storytelling, singing and engaging in play, were positively associated with child development in most studies reviewed, consistent with research highlighting the importance of these activities for cognitive and social–emotional outcomes (Hasan et al. 2023; Rey‐Guerra et al. 2022). However, one secondary analysis in our review found no significant association between these activities and child development (Rao et al. 2022). The study involved regression analyses of data about 1598 children aged 3–5 years across five Chinese provinces, finding that maternal education and provincial economic status were the strongest predictors of child development. These findings highlight how socio‐economic and urban–rural disparities may shape child development, pointing to the need to consider broader structural influences such as economic policy, education systems and regional infrastructure. The WHO social model of health emphasises that development is shaped by the conditions in which people live and grow and calls for multisectoral action to address inequities in these environments (Solar and Irwin 2010).

There was considerable heterogeneity in how studies measured home‐based learning opportunities and responsive caregiving. Learning opportunities were measured more consistently than responsive caregiving, with six indicators used across the studies. Responsive caregiving was assessed using 16 distinct indicators, reflecting wide variation in how this dimension was defined and captured. Many papers in fact assessed other important constructs, for example, parenting styles or parent involvement in play. This highlights the substantial variation in how this dimension is characterised and assessed and reveals a broader pattern of blurred conceptualisation of responsive caregiving in the literature. This aligns with findings from a scoping review of Nurturing Care Framework measurement tools, which highlighted a lack of consistency and validation, particularly for the responsive caregiving and learning opportunities components (Jeong et al. 2022). The inconsistency in terminology and operationalisation is a recognised challenge in the field (Black et al. 2024). Responsive caregiving, though nearly universal across cultures, is characterised by diverse expressions across daily activities such as talking, playing, feeding and bathing (Black et al. 2024). The absence of universal indicators and the broad variability in operational definitions currently limit comparisons among studies. The responsive caregiving component of the Nurturing Care Framework requires the most attention, particularly in establishing concrete measures and indicators (Black et al. 2024; Richter et al. 2020).

Similarly, our review also revealed much variation in child development measures, with a wide range of tools and indicators used. The measurement of child development is challenging as there is a need to find a balance between population‐level coverage and culturally appropriate testing (Rao et al. 2020). The objective is to produce globally comparable data that accurately reflect local cultures and contexts while also providing relevant policy and implementation evidence (Gove and Black 2016). The 2023 launch of UNICEF's globally standardised Early Childhood Development Index 2030 (ECDI2030) provides a tool to measure developmental progress in the 24‐ to 59‐month age range and to support national monitoring of SDG 4.2.1 (UNICEF 2020). The ECDI2030 provides population‐level assessment of child development, enabling valuable cross‐country comparability, though it should be noted that some of its constructs may not fully capture the breadth of culturally normative caregiving practices across all country contexts. In addition to population‐level tools, individual developmental assessments are being explored. Studies are underway to determine whether newly developed tools, for example, the Early Childhood Development Assessment Scale‐Direct Assessment (ECDAS‐DA) can add value to the measurement of SDG Target 4.2.1 by complementing the caregiver reported ECDI2030 (Richards et al. 2023). Multiple measures could assist in providing a clearer picture of child development, especially in LMICs. Further, emerging instruments such as the Early Learning (EL) tool, which captures the levels of stimulation available to children in resource constrained settings, are being developed to help assess learning opportunities on child development, though validation studies in different settings are required (Hentschel et al. 2024).

Most studies in our review relied on caregiver reports. Concerns have been raised about the reliance on parent reports for child development data collection at the population level, due to potential biases (Fernald et al. 2017). For example, parent reporting may increase the likelihood of social desirability bias where answers are embellished to be more socially acceptable and/or recall bias where past events are incorrectly remembered (Althubaiti 2016). Although direct observation of child and parent behavior is preferred for obtaining child development data, this approach is not feasible in many settings such as resource‐constrained areas.

A confounder is a variable that influences both the independent and dependent variables, potentially distorting the true association between them. Controlling for confounders is necessary to isolate the effects of the independent variables of interest and is typically achieved by including covariates in regression models. In this review, most studies incorporated covariates in regression analyses, thereby reducing bias and enhancing the internal validity of their findings. Frequent inclusion of maternal education, household wealth, child demographic characteristics and early childhood programme attendance suggest an awareness of socio‐demographic influences on child development outcomes.

However, only a minority of studies explicitly identified and justified their selection of confounding variables. This raises the possibility that important influences embedded in children's proximal and distal environments were not accounted for. Poor parental mental health, for example, depression, anxiety and chronic stress, plus the pressures of poverty, food insecurity and gender inequality, can substantially reduce caregivers' capacity to provide nurturing care (Lin et al. 2025; Marlow et al. 2022). Parental mental health problems in LMICs are consistently associated with poorer social–emotional development among children aged 24–59 months, with affected children showing nearly twice the odds of developmental difficulties (Honda et al. 2023).

Few studies controlled for preschool quality or attendance, despite recruiting children from early learning centres. In LMIC settings, preschool provision varies widely in quality, affordability and daily duration. Potential confounding by preschool characteristics cannot be ruled out, particularly where centre‐based learning may partially compensate for or amplify home‐based experiences. Not controlling for such factors limits explanatory depth and may obscure how broader contextual factors interact with home learning experiences to influence developmental outcomes.

Mediation and effect modification were conducted in several studies in our review, highlighting that the influence of the home environment on ECD may be shaped by broader socio‐demographic factors. Mediation occurs when the effect of an independent variable on a developmental outcome is transmitted through a third variable, known as a mediator. The mediation analysis by Liu et al. (2020) illustrates a clear pathway through which SEP affects cognitive development: Families with greater resources tend to demonstrate higher parental involvement and self‐efficacy, which in turn supports children's cognitive competence (Liu et al. 2020). These findings reflect broader evidence showing that poverty constrains access to stimulation, early education and responsive caregiving (Cuartas et al. 2020; Frongillo et al. 2017; Lu et al. 2020). Moderation refers to a situation where the strength or direction of the relationship between two variables depends on a third variable, which is the moderator variable. Often moderation analyses include testing potential moderator variables such as child age, child sex, urbanicity, maternal education and household wealth on the associations between the exposure of interest and child developmental outcomes. These variables are commonly used in moderation analyses because they represent key socio‐economic, demographic and environmental factors that can influence the relationship between child outcomes and exposures, for instance, learning opportunities/caregiving factors. In the case of Sun et al. (2018), this study showed that parental engagement can buffer or amplify the effects of SEP and child developmental outcomes. Specifically, the impact of SEP on cognitive performance in Papua New Guinea, and on approaches to learning in Timor‐Leste and Vanuatu, varied depending on levels of parental engagement. These patterns suggest that even within resource‐constrained settings, caregiver behaviours can play a protective role. Overall, these mediation and moderation findings outline the layering of socio‐economic factors that can influence the relationship between the home caregiving and learning environment and child development.

A limitation of this study is that the search was restricted to English; hence, relevant papers in languages other than English might have been missed. Additionally, most of the studies in this review were observational, limiting causal inferences. The substantial heterogeneity in the measures and indicators used for responsive caregiving, home learning experiences and child development made it inappropriate to conduct a meta‐analysis. Additionally, the variation in validity, particularly lower validity of measures such as MICS, reduces the rigour of some studies included in this review. Despite these limitations, we believe this systematic review provides a solid overview of the positive associations between the home‐based learning opportunities and responsive caregiving environment and the development of preschool‐aged children in the region. It highlights the importance of strengthening caregiver education and home stimulation, particularly by addressing modifiable factors, such as caregiving behaviours, parental knowledge and access to learning materials, to support optimal child development.

To inform ECD planning, future research could help to identify the needs of preschool‐aged children in low‐income contexts and the Pacific Islands where data are scarce. Greater insight across key subgroups including region, socio‐economic position and sex may also support ECD policy and programming. This review reinforces calls from researchers for greater conceptual clarity and standardised measurement of components of the Nurturing Care Framework to support cross‐cultural comparison and national monitoring systems (Black et al. 2024; Richter et al. 2020). This includes the development and validation of concrete indicators for responsive caregiving and the harmonisation of learning opportunities indicators. The review also aligns with recommendations for the refinement of child development assessments to ensure culturally grounded yet globally comparable measures to accurately capture child development in population‐level surveys (Gove and Black 2016; Munoz‐Chereau et al. 2021; Rao et al. 2020).

5. Conclusion

The review findings suggest that home‐based learning opportunities and responsive caregiving experiences may be associated with improved developmental outcomes among 2‐ to 5‐year‐old children. Although these patterns are consistent with broader evidence, the heterogeneity of measures and limited representation mean that conclusions should be interpreted with caution. Future research could prioritise under‐represented settings such as low‐income countries and Pacific Island nations. Addressing these gaps is essential to ensuring that policy and programming efforts reflect the diversity of cultural, geographic and socio‐economic realities across the region. The substantial heterogeneity in measures and indicators related to caregiving, learning opportunities and child development revealed in the review highlights the need for standardised measures to appropriately capture the associations between child development and the home environment.

Author Contributions

Sally Popplestone: conceptualization, methodology, data curation, formal analysis, writing – original draft, writing – review and editing. Tomoko Honda: methodology, data curation, formal analysis, writing – review and editing. Thach Tran: conceptualization, methodology, supervision, writing – review and editing. Yeji Baek: methodology, writing – review and editing, supervision. Lorena Romero: methodology, data curation, writing – review and editing. Jane Fisher: conceptualization, methodology, supervision, writing – review and editing.

Funding

Jane Fisher is supported by the Finkel Professorial Fellowship, which is funded by the Finkel Family Foundation.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supporting Information.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.