Unlocking the transformative potential of data science in improving maternal, newborn and child health in Africa: a scoping review protocol Authors: Joseph Akuze, Bancy Ngatia, Samson Yohannes Amare, Phillip Wanduru, Grieven P Otieno, Rornald M Kananura, Kirakoya-Samadoulougou Fati, Agbessi Amouzou, Abiy Seifu Estifanos, Eric Ohuma Published: 2024 ABSTRACT ABSTRACTIntroductionApplication of data science in maternal, newborn, and child health (MNCH) across Africa is variable with limited documentation. Despite efforts to reduce preventable MNCH morbidity and mortality, progress remains slow. Accurate data are crucial for holding countries accountable for tracking progress towards achieving the Sustainable Development Goal 3 targets on MNCH. Data science can improve data availability, quality, healthcare provision and decision-making for MNCH programmes. We aim to map and synthesise data science use cases in MNCH across Africa.Methods and analysisWe will develop a conceptual framework encompassing seven domains: (1) infrastructure and systemic challenges, (2) data quality, (3) data governance, regulatory dynamics and policy, (4) technological innovations and digital health, (5) capacity development, human capital and opportunity, (6) collaborative and strategic frameworks and (7) recommendations for implementation and scaling.We will use a scoping review methodology involving literature searches in seven databases, grey literature sources and data extraction from the Digital Health Atlas. Three reviewers will screen articles and extract data. We will synthesise and present data narratively and use tables, figures and maps. Our structured search strategy across academic databases and grey literature sources will find relevant studies on data science in MNCH in Africa.Ethics and disseminationThis scoping review does not require formal ethical review and approval because it will not involve collecting primary data. The findings will showcase gaps, opportunities, advances, innovations, implementation and areas needing additional research. They will also propose next steps for integrating data science in MNCH programmes in Africa. The implications of our findings will be examined in relation to possible methods for enhancing data science in MNCH, such as community and clinical settings, monitoring and evaluation. This study will illuminate data science applications in addressing MNCH issues and provide a holistic view of areas where gaps exist and where there are opportunities to leverage and tap into what already exists. The work will be relevant for stakeholders, policymakers and researchers in the MNCH field to inform planning. Findings will be disseminated through peer-reviewed journals, conferences, policy briefs, blogs and social media platforms. Strengths and limitations of this study This scoping review is the first to examine the role and potential of data science applications in maternal, newborn, and child health (MNCH) in Africa, with assessments on healthcare infrastructure, data quality improvement, innovative data collection and analyses, policy formulation, data-driven interventions, technologies for healthcare delivery and capacity building. This scoping review is the first to examine the role and potential of data science applications in maternal, newborn, and child health (MNCH) in Africa, with assessments on healthcare infrastructure, data quality improvement, innovative data collection and analyses, policy formulation, data-driven interventions, technologies for healthcare delivery and capacity building. We will conduct systematic searches across multiple databases (PubMed, Scopus, Web of Science, Google Scholar, CINAHL, EMBASE and Ovid) and grey literature. We will conduct systematic searches across multiple databases (PubMed, Scopus, Web of Science, Google Scholar, CINAHL, EMBASE and Ovid) and grey literature. Focusing on studies that have used data science, we will synthesise our findings with detailed explanations, informative charts, graphs and tables. Focusing on studies that have used data science, we will synthesise our findings with detailed explanations, informative charts, graphs and tables. The study will deliver actionable recommendations for stakeholders engaged in MNCH policy formulation, strategic planning, academia, funders and donors and clinicians aimed at improving MNCH outcomes in Africa. The study will deliver actionable recommendations for stakeholders engaged in MNCH policy formulation, strategic planning, academia, funders and donors and clinicians aimed at improving MNCH outcomes in Africa. Our scoping review will primarily rely on published literature in English and French, therefore, will omit valuable insights that may have been published for non-anglophone and francophone regions of Africa. Our scoping review will primarily rely on published literature in English and French, therefore, will omit valuable insights that may have been published for non-anglophone and francophone regions of Africa. Background The 2015 Sustainable Development Goals (SDGs) require countries to commit to ending preventable maternal, newborn and child deaths by 2030. These goals include reducing under-five mortality in each country to at least 25 per 1000 live births, neonatal mortality to at least 12 per 1000 live births, and the global maternal mortality ratio to at least 70 per 100 000 live births.1 2 There needs to be high-quality data on all pertinent indicators for nations to be held responsible for their promises.3 Although countries have done significant work towards achieving these targets, progress is still sluggish, particularly regarding birth and death registration, and survival and well-being of the most vulnerable.4 The United Nations International Children’s Emergency Fund monitors child well-being through 48 indicators5 for example, birth registration, education, protection from harm, clean and safe surroundings and poverty eradication in life, but data gaps remain a significant challenge in many countries.6 Similarly, over 190 countries worldwide offer data to different United Nations agencies on various indicators like mortality, stillbirth, morbidity, preterm birth and low birth weight, among other variables.7 However, it is remarkable that the world’s regions with the highest burdens, like Africa, face a Data Rich Information Poor syndrome.810 Accurate data on maternal newborn and child health (MNCH) morbidity and mortality are essential for holding leadership in African countries accountable and achieving the goals of reducing preventable deaths and improved well-being.11 In addition to accountability, accurate data would guide better-targeted interventions for the most vulnerable to obtain the most significant impact despite limited resources. Countries need to set priority agendas for MNCH—there is a critical need for data-driven approaches to identify and prioritise key areas.1218 During the recent Coronavirus pandemic, there were new advancements using data science in tracking and reporting global mortality such as rapid mortality surveillance and epidemic response within the WHO’s Toolkit for Routine Health Information Systems Data.19 However, the pandemic also significantly disrupted MNCH indicators tracking, as well as essential services like immunisation and broader healthcare provision, including widespread school closures.52023 Data science, an interdisciplinary field amalgamating scientific methodologies, algorithms and systems, is instrumental in deriving insights and knowledge from structured and unstructured data.24 It synthesises elements from mathematics, statistics, computer science and specific domain knowledge such as MNCH to facilitate data engineering, organisation and management, analysis and interpretation and dissemination of large volumes of data for informed decision-making.25 Over the last two decades, data science has been applied in several sectors, including healthcare, finance and marketing, to solve complex challenges, forecast outcomes and inform strategic approaches.26 Data science applications have also enabled tracking and progress of several SDG targets with improvements in interoperability and improved routine information systems.27 28 In MNCH, especially in Africa, various applications have been documented, for example, predicting infant mortality in Rwanda using Machine Learning,29 prediction of poliovirus, vaccine coverage and immunisation programmes in Africa30 31 and using data science approaches for decision-making and utilisation of maternal healthcare services in rural Ethiopia.32 33 With data science applications, there is potential to integrate African MNCH data with other planetary datasets like temperatures, seasonality, rainfall, air pollution, flooding and National Aeronautics and Space Administration’s space data.24 Data science initiatives elsewhere (Europe, the USA and Asian settings) have been found useful in improving data quality, care and service provision, planning and priority setting as well as implementing targeted interventions for MNCH, for example, digitisation of health information systems.3437 Utilising data science in MNCH, mortality and morbidity tracking, data collection and reporting has the potential to improve the quality, comprehensiveness and utilisation of data in Africa. Fortunately, many efforts to harness data science in Africa are underway, such as District Health Information Software 2 (DHIS2), digital payments, mobile money, eHealth and mHealth platforms.38 However, many of these efforts act in isolation, are poorly coordinated and are not implemented as part of national efforts that can guide policy decisions. In this study, we conduct a scoping review that aims to review the use, role and contribution of data science to improved MNCH in Africa in recent years. Specifically, the study aims to provide a landscape of data science in MNCH by mapping where, how and for what MNCH outcomes data science approaches have been applied to in the African region. Study rationale The WHO has developed a 5-year global strategy for digital health.39 It has also established the Digital Health Atlas (https://digitalhealthatlas.org/) to track planning, development, implementation and scale-up of digital health initiatives.38 The Digital Health Atlas shows projects at different stages of development and their potential impact. Over 600 initiatives exist in Africa, with 316 focusing on the MNCH thematic areas. However, a significant number (n=284) lack documented data science involvement, therefore, creating a blind spot in our understanding of how these initiatives effectively leverage data science and have little information about their stages of implementation.38 Consequently, the extent to which data science has been used to improve MNCH remains largely speculative. This limited knowledge about data science utilisation in MNCH hinders effective resource allocation and stalls progress towards reducing maternal and child mortality rates in Africa. In addition, it is impossible to know whether current initiatives are missing vital tools for improving MNCH, for example, utilising predictive modelling for identifying high-risk pregnancies and optimising resource allocation. By identifying data science practices and experiences within the African continent, we can better estimate their potential impact on reducing adverse MNCH outcomes and improving data quality and use for evidence-based decision-making. This in turn could lead to improved healthcare delivery and more targeted interventions for the already-vulnerable mother and neonatal group. This scoping review aims to address this critical knowledge gap by discussing the existing gaps and opportunities for data science in MNCH, illustrate the future trajectory of this emerging field in Africa, and highlight why it is a key pillar for accelerating progress towards national and regional targets. The findings of our review will stimulate further research into this field, as well as highlight case studies and impact made in different African countries. This will also provide a valuable reference for decision-makers to identify potential areas for investment, policy intervention and collaboration in the field. Study objectives To map and synthesise use cases of data science applications in MNCH in Africa. We will also highlight case studies and identify existing gaps and opportunities and next steps for further improvement and integration of data science in MNCH in Africa. Methods and analysis Conceptual model This section presents the proposed conceptual approach which will be adapted for data extraction and synthesis. Conceptualising data science review for MNCH in the African context In our scoping review, we shall use a conceptual approach to structure the investigation of how data science has been applied in the field of MNCH in Africa notably at clinical, community and district or regional levels. Our conceptual framework (figure 1 and box 1) draws inspiration from the Measure Evaluation’s Health Information System Strengthening (HISS) model designed to assess, plan and improve health information systems in low- and middle-income countries.40 Our adapted model goes beyond the HISS model by incorporating additional domains specific to data science (notably technological innovations, capacity development for data science skills, policy and regulatory considerations and recommendations for implementing and scaling). This approach shall define the scoping review into interconnected domains that shall cover existing gaps, opportunities and strategic imperatives for integrating data science into MNCH initiatives across the continent. The conceptual framework is designed to guide a comprehensive exploration of the current state, identify gaps and barriers and spotlight opportunities for harnessing data science in the MNCH context more effectively in Africa. Conceptual framework for harnessing data science role in improving MNCH in Africa. The conceptual model represents the process of inquiry for data science’s role in MNCH in Africa. In step 1, the main domains of inquiry represented with blue boxes are defined. Step 2 is the problem and gap analysis represented by orange boxes. Step 3 is the potential opportunities analysis represented by yellow boxes. Finally, step 4 will include sustainability, institutionalisation and scale-up potential for each domain. Conceptual framework for harnessing data science to improve maternal, newborn and child healthcare (MNCH) in Africa Domain I. Infrastructure and systemic challenges Assessing the existing data science initiatives for data infrastructure and data use: evaluation of the accessibility, interoperability and robustness of current health data systems and data use ecosystems for integrated care and personalised medicine. Examining the multistakeholder approaches that impact the utility of MNCH data. Domain II. Data quality Standardisation and integration: addressing the challenges related to data standardisation and quality (accuracy, completeness and reliability of data) to facilitate accurate and reliable health decision-making utilising data science tools and techniques. Data analysis for quality assurance: methodologies and frameworks for analysing data quality, to ensure data accuracy, completeness and reliability. Domain III. Data governance, regulatory dynamics and policy Data stewardship practices: investigating the capacity for effective data governance, focusing on the ethical management and use of health data. Regulatory environment analysis: examination of the regulatory landscape affecting the application of data science in MNCH, identifying barriers and enablers within government policies and their impact on moving it forward or hindering innovation. Compliance and ethical guidelines: ensuring that data science practices comply with local and international regulations and ethical standards. Policy impact assessment: assessing how changes in policy can either promote or hinder the use of data science in MNCH. Domain IV. Technological innovations and digital health Leveraging digital technologies: exploring the potential of mobile health, telemedicine and artificial intelligence to revolutionise MNCH data collection, analysis (including advanced data analytics) and application. Innovation in MNCH data management: assessing the impact of data science tools and methodologies in transforming integrated healthcare delivery, personalised medicine and MNCH outcomes. Domain V. Capacity development, human capital and opportunity Human resource potential: strategies for building data science capabilities for MNCH to maximise the opportunities. Training and education: prioritising the development and expansion of data science knowledge and skills for data analysis among MNCH health professionals and implementing partners. Domain VI. Collaborative and strategic frameworks Fostering strategic partnerships: identifying and leveraging partnerships and collaborations across sectors to advance data science applications in MNCH. Engaging stakeholders to promote data-driven decision-making in MNCH healthcare. Domain VII. Recommendations for implementation and scaling Operationalising data use in health: developing actionable strategies for the implementation of data science initiatives in maternal, newborn and child healthcare. The role of data analysis in operationalising data use in MNCH for data-driven decision-making. Scaling and sustainability: recommendations for scaling successful interventions and ensuring their sustainability through continuous learning and adaptation. Scoping review protocol design We propose to use a scoping review design guided by the methodological framework developed by Arksey H and O'Malley L,41 because the implementation of data science within MNCH in Africa at regional, national and subnational levels is an emerging field with poor documentation currently. The Arksey H and O'Malley L framework proposes four stages to identify, collate and summarise the existing literature with scientific rigour. These stages include: Stage 1: identifying the research question Stage 1: identifying the research question Stage 2: identifying relevant studies and data sources Stage 2: identifying relevant studies and data sources Stage 3: selection Stage 3: selection Stage 4: collating, summarising of data and synthesis Stage 4: collating, summarising of data and synthesis The scoping review shall springboard our synthesis of the seven domains listed in our conceptual framework. We shall present the screening process using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews flow diagram.42 Stage 1: identifying the research questions We have six research questions that we intend to answer notably: Where have the data science projects and initiatives in MNCH been implemented in Africa? Where have the data science projects and initiatives in MNCH been implemented in Africa? What are the experiences and practices of data science in MNCH within Africa to improve MNCH outcomes? What are the experiences and practices of data science in MNCH within Africa to improve MNCH outcomes? What are the existing gaps and opportunities in using data science in MNCH in Africa to improve MNCH outcomes? What are the existing gaps and opportunities in using data science in MNCH in Africa to improve MNCH outcomes? How can we leverage the identified gaps and opportunities to advance data science in MNCH in Africa to improve MNCH outcomes? How can we leverage the identified gaps and opportunities to advance data science in MNCH in Africa to improve MNCH outcomes? What is the future trajectory of data science in MNCH in Africa? What is the future trajectory of data science in MNCH in Africa? What are the opportunities and challenges for effective use and scale-up of data science to optimise decision-making and improve MNCH outcomes in Africa? What are the opportunities and challenges for effective use and scale-up of data science to optimise decision-making and improve MNCH outcomes in Africa? Stage 2: identifying relevant studies and data sources Searching strategy A comprehensive literature search will be conducted in the following databases: PubMed, Scopus, Web of Science, Google Scholar, CINAHL, EMBASE and Ovid (table 1 and online supplemental table S1). In addition, reference lists and peer-reviewed studies will be searched. Other sources of unpublished work, such as websites and blogs by both organisations and individuals, will also be searched through searching for relevant reports obtained from within our existing networks, relevant organisations and government policies published before December 2023. All global multicountry and African data from the WHO’s Digital Health Atlas shall be downloaded and reviewed with a focus on projects implemented with components of the MNCH thematic area. Table 1Summary of proposed search termsNo.Proposed search components and keywords*1Data science AND maternal health OR newborn health OR neonatal health OR child health OR perinatal health AND Africa2Data science AND maternal health OR newborn health OR neonatal health OR child health OR perinatal health AND country (where country represents one of each country in Africa)*Detailed search terms by database are in the appendixsupplemental table S1. Summary of proposed search terms Detailed search terms by database are in the appendixsupplemental table S1. Inclusion and exclusion criteria Articles and any relevant literature published in English and French discussing the status and progress of data science approaches in MNCH in Africa will be included. We will exclude articles without a specific focus on Africa and those not highlighting the application of data science in MNCH, as well as systematic and scoping reviews. Stage 3: selection of articles Document screening and selection The screening and selection process will include removing duplicates, reviewing the titles to exclude irrelevant articles (peer-reviewed and grey literature) and assessing the executive summaries or abstracts to further filter documents. Three reviewers will be involved: two reviewers will assess the peer-reviewed articles and grey literature, while a third reviewer will review the data from the Digital Health Atlas to identify projects in MNCH thematic area. Data extraction and collection Once the three reviewers reach a consensus on the final documents for the scoping review, one individual will carry out the data-extraction process of the peer-reviewed articles by entering the data into Rayyan for data management.43 Subsequently, the two remaining reviewers will each focus on the grey literature review and the synthesis and analysis of data from the Digital Health Atlas. Due to limitations of scope, time and budget, we shall use only one reviewer for each review and data extraction (peer-reviewed articles, grey literature and the data from the Digital Health Atlas). The data-extraction process will focus on obtaining information about the author of the document, year, organisation, population, country, donor, study aims and objectives, study design or methods and approaches used (including sample size), context, intervention (type, duration, recipients), key concepts, comparisons made and main outcomes. The extracted data will further be collated, synthesised and analysed. Similar attributes will be extracted in addition to the project name, software and challenges for the Digital Health Initiatives Atlas data. Stage 4: collating, summarising of data and synthesis This process will involve reviewing the data extracted and summarised using descriptive analysis. The purpose of this is to synthesise and present an overview of all articles reviewed. Where possible, maps, graphs and tables will be used to visualise the different attributes of data science use in MNCH in Africa based on our proposed conceptual framework (figure 1 and box 1). Patient and public involvement Since this is a protocol for a scoping review, patients and the public were not involved in the design or proposed research. Proposed timeline Conceptualising of the scoping review including the framework began in December 2022. From January 2023 to December 2023, we began drafting a concept note with consultations with African researchers to develop concepts on ‘The State of Data Science for Health in Africa’ on MNCH and other thematic areas of global health (notably environmental and planetary health, value for data, infectious diseases, policy, etc). These efforts are coordinated by the African Population and Health Research Center (APHRC), and the National Institute of Health (NIH) Fogarty International Center and have been funded using pooled resources from NIH, Wellcome Trust and the Bill and Melinda Gates Foundation. We also formulated the African MNCH Data Science writing team for the scoping review. Database searches, screening and data extraction began at the end of June 2024, and we anticipate that the scoping review will be complete by the end of the fourth quarter of 2024. Ethics and dissemination The scoping review of already-published literature on data science in MNCH within the African context does not require ethical approval because only published and publicly accessible data and literature shall be used. For dissemination, the findings of this scoping review aim to be shared broadly with stakeholders in the MNCH thematic area, including healthcare professionals, policymakers and academic researchers. The strategy includes publications in peer-reviewed journals, presentations at conferences related to health informatics and maternal health and summaries for non-academic audiences to ensure broad accessibility. Additionally, the review will leverage digital platforms and social media to enhance reach and engagement, facilitating the translation of research findings into practical interventions that can be implemented to improve MNCH care and outcomes in Africa. This approach ensures that the insights derived from the application of data science in MNCH are disseminated effectively, fostering collaboration and innovation in the field. We shall also disseminate findings through Nature Africa, policy briefs, blogs, webinar series and workshops. Discussion This study shall contribute to a publication series on the State of Data Science for Health in Africa by APHRC and NIH, showcasing the expected transformations and advancements in fields of biomedical and behavioural research with data science applications.44 Our study has several strengths; notably, data science, being an emerging field in the African continent, has limited documentation available. Therefore, our scoping review is the first to review its applications in the MNCH context. It is timely and shall address issues concerning data science on the African continent in MNCH thematic areas, including but not limited to service, data governance, capacity, data quality and policies. Our study shall use a rigorous methodology including reviewing literature and data on data science initiatives. Finally, our study shall formulate recommendations for various stakeholders on the African continent. Although we shall ably review and assess literature and data from multiple databases and grey literature, our study has one major limitation, notably, mainly publications in English shall be included, therefore limiting our understanding of the application and impact of data science in MNCH from other non-anglophone or francophone countries in Africa. Although efforts to include French articles shall be done. Data availability No underlying data is associated with this article. Extended data Open Science Framework (OSF): Unlocking the transformative potential of data science in improving maternal, newborn and child health in Africa: A scoping review protocol https://doi.org/10.17605/OSF.IO/8PYNC supplementary material