Improving health and development through play - Evaluation of a comprehensive ECCE programme at scale in Ghana
Lead Research Organisation:
Institute for Fiscal Studies
Department Name: Grants Administration Department
Abstract
Children who receive quality ECCE are proven to be healthier, do better and stay longer in school, and thereby have better economic trajectories in adult life. Improving ECCE is therefore critical, especially in remote and vulnerable areas, such as northern Ghana, where most families live on less than US$2 per day, 20% of children under the age of 5 are stunted and 39% 3-4 years olds are off-track cognitively.
Previous research by members of this research team has shown that an intervention by the NGO Lively Minds (LMNGO), designed to combat the twin problem of low kindergarten quality and poor parenting practices/mindsets, significantly improved children's health, cognitive skills and socio-emotional development in remote, poor, rural areas of northern Ghana. These promising findings were a key driver of the decision by the Ministry of Education (MoE) to work with LMNGO to scale up the programme in 60 of Ghana's 228 districts to reach 4,000 pre-schools and up to 1.3 million children over the next four years.
However, from previous studies we know that programmes that are impactful at smaller scale may not maintain their effectiveness once adapted for and implemented at the scale necessary to reach those in need of them. In this project we, therefore, aim to find out whether the promising effect of the Lively Minds programme can be sustained as coverage is expanded and ownership of the programme is handed from LMNGO to the MoE. We will also study the causal pathway between the programme and child developmental outcomes. Analysis of the efficacy trial results conducted by this research team revealed some of these pathways (especially for impacts on socio-emotional development), drivers of the substantial health effects that were identified remain unexplained.
To achieve these research aims, we have brought together a multidisciplinary research team, consisting of in-country and international economists, epidemiologists, WASH experts and developmental psychologists. Working in close partnership with LMNGO and the three ministries which oversee ECCE provision in Ghana (MoE, Ministry of Health and Ministry of Gender, Children and Social Protection) this team will embed a randomised controlled trial in the scale-up, collecting data on child development and pathways through which the Lively Minds programme could affect it after 4 terms of exposure. We will also collect routine data in schools to assess how well the programme is being implemented.
The Lively Minds programme is unusual. ECCE programs are often one-dimensional, targeting either parents or pre-schools, and/or focusing on one area of child development. LMNGO developed a hybrid ECCE program, designed with scalability in mind, that simultaneously improves the home and pre-school environments, as well as children's health alongside cognitive and socio-emotional development at a projected cost of $7 per child per year when scaled, using only locally available resources and infrastructure and empowering marginalised and isolated mothers in rural communities. From starting in a few villages in Ghana and Uganda ten years ago it will now be taken over by the government of Ghana and will reach over a million children in the next 5 years. We will evaluate whether it maintains its effectiveness through this process of substantial scaling and transmission of ownership and whether there are ways in which it could be improved. This will first and foremost provide important evidence for the MoE on effectiveness of its policies. However, it will also help the wider community of stakeholders in LMICs seeking cost-effective, innovative, cross-sectoral and scalable ways to intervene in order to meet Sustainable Development Goals 3 and 4 ensuring that all children, including those in the most marginalised and vulnerable populations, receive nurturing care and opportunities for early learning.
Previous research by members of this research team has shown that an intervention by the NGO Lively Minds (LMNGO), designed to combat the twin problem of low kindergarten quality and poor parenting practices/mindsets, significantly improved children's health, cognitive skills and socio-emotional development in remote, poor, rural areas of northern Ghana. These promising findings were a key driver of the decision by the Ministry of Education (MoE) to work with LMNGO to scale up the programme in 60 of Ghana's 228 districts to reach 4,000 pre-schools and up to 1.3 million children over the next four years.
However, from previous studies we know that programmes that are impactful at smaller scale may not maintain their effectiveness once adapted for and implemented at the scale necessary to reach those in need of them. In this project we, therefore, aim to find out whether the promising effect of the Lively Minds programme can be sustained as coverage is expanded and ownership of the programme is handed from LMNGO to the MoE. We will also study the causal pathway between the programme and child developmental outcomes. Analysis of the efficacy trial results conducted by this research team revealed some of these pathways (especially for impacts on socio-emotional development), drivers of the substantial health effects that were identified remain unexplained.
To achieve these research aims, we have brought together a multidisciplinary research team, consisting of in-country and international economists, epidemiologists, WASH experts and developmental psychologists. Working in close partnership with LMNGO and the three ministries which oversee ECCE provision in Ghana (MoE, Ministry of Health and Ministry of Gender, Children and Social Protection) this team will embed a randomised controlled trial in the scale-up, collecting data on child development and pathways through which the Lively Minds programme could affect it after 4 terms of exposure. We will also collect routine data in schools to assess how well the programme is being implemented.
The Lively Minds programme is unusual. ECCE programs are often one-dimensional, targeting either parents or pre-schools, and/or focusing on one area of child development. LMNGO developed a hybrid ECCE program, designed with scalability in mind, that simultaneously improves the home and pre-school environments, as well as children's health alongside cognitive and socio-emotional development at a projected cost of $7 per child per year when scaled, using only locally available resources and infrastructure and empowering marginalised and isolated mothers in rural communities. From starting in a few villages in Ghana and Uganda ten years ago it will now be taken over by the government of Ghana and will reach over a million children in the next 5 years. We will evaluate whether it maintains its effectiveness through this process of substantial scaling and transmission of ownership and whether there are ways in which it could be improved. This will first and foremost provide important evidence for the MoE on effectiveness of its policies. However, it will also help the wider community of stakeholders in LMICs seeking cost-effective, innovative, cross-sectoral and scalable ways to intervene in order to meet Sustainable Development Goals 3 and 4 ensuring that all children, including those in the most marginalised and vulnerable populations, receive nurturing care and opportunities for early learning.
Technical Summary
A critical knowledge gap for development of effective ECCE policy is what intervention strategies are effective at scale. We aim to fill this knowledge gap through the evaluation of a holistic ECCE programme Lively Minds, implemented by the Government of Ghana at scale. The programme, shown to be effective in an efficacy trial, will be scaled to all pre-schools in 60 most deprived districts in Ghana to reach over 1 million children over 10 phases from September 2021 to May 2024. We propose to integrate a stepped-wedge cluster randomised controlled trial into this scale-up focusing on 4 phases (Sept.'21 - Sept. 23), 48 districts, randomly sampled 144 schools and 1,440 children.
We will estimate the average treatment effect of the scaled Lively Minds program on child development (health, socio-emotional and cognitive) using an ANCOVA strategy. We will establish how impacts vary by child, parent, teacher and district characteristics and identify, through machine learning methods, which characteristics are associated with largest treatment effects.
We will further assess heterogeneity in impacts to understand the role that degree of intervention fidelity and compliance play, to be able to advise on how the programme design can be adapted to maximise cost-effectiveness. In order to establish causal links, we will construct a counterfactual group among control areas by using baseline data to identifying those where intervention fidelity/compliance would likely be high/low.
Using descriptive mediation analysis followed by structural estimation we will shed light on whether observed impacts are achieved through changes within the school and/or the children's home environments, including teacher and caregiver behaviour. We will explicitly model children's outcomes as a function of key hypothesised mediators, using an instrumental variable approach to address endogeneity of these mediators.
We will estimate the average treatment effect of the scaled Lively Minds program on child development (health, socio-emotional and cognitive) using an ANCOVA strategy. We will establish how impacts vary by child, parent, teacher and district characteristics and identify, through machine learning methods, which characteristics are associated with largest treatment effects.
We will further assess heterogeneity in impacts to understand the role that degree of intervention fidelity and compliance play, to be able to advise on how the programme design can be adapted to maximise cost-effectiveness. In order to establish causal links, we will construct a counterfactual group among control areas by using baseline data to identifying those where intervention fidelity/compliance would likely be high/low.
Using descriptive mediation analysis followed by structural estimation we will shed light on whether observed impacts are achieved through changes within the school and/or the children's home environments, including teacher and caregiver behaviour. We will explicitly model children's outcomes as a function of key hypothesised mediators, using an instrumental variable approach to address endogeneity of these mediators.
