High-quality data serves as the cornerstone for robust planning and decision-making in public health. From designing health programs grounded in the needs of communities to developing advanced prediction and surveillance platforms to stay ahead of health threats, data plays an instrumental role in maximizing health equity in a measurable, transparent, and accountable way. Additionally, the benefits extend beyond the health sector. Research by Dalberg and the Global Partnership for Sustainable Development Data found that a one-dollar investment in data across multiple sectors including health can yield an average economic return of $32, multiplying the opportunities for social impact. 

Unlocking these benefits is contingent on the effective use of data for decision-making, a challenge often faced by low- and middle-income countries (LMICs). While health officials collect a variety of health data, it often does not translate to informing day-to-day operational decisions. Challenges manifest along the data value chain—including data collection, processing, and use—and relate to the following: 

• People – Low data literacy among public health workers affects the quality of data collection, analysis, and use, ultimately leading to distrust in the validity of data-driven insights and their potential to inform decision-making. 
• Technology – Limited awareness, affordability, access, and adoption of tools (software and hardware) for data collection and analysis lead to low automation of routine processes and inefficiencies in producing insights. 
• Institutional systems and processes – Weak governance mechanisms for data access, sharing and privacy—compounded by limited human and financial resources across the data value chain—impede the ability of Ministries of Health (MoHs) to leverage data effectively and ethically. 

These limiting factors contribute to poor decision-making for resource allocation, disease prevention, and healthcare delivery. Three responses to these data constraints will be critical to advancing the potential of health systems. 

1. Equip MoH workers with more advanced decision-intelligence skills

Health workers at all levels of the public health system need to understand what data can and cannot do, how to ask good questions, and how to interpret data to inform decisions. As stated by the former Chief Data Scientist at Google Cassie Kozyrkov, “With more interest in data-driven everything… and more complex organizational decision-making, better decision skills that take technology into account are needed urgently.” Meeting this need requires targeted trainings and toolkits on data for decision-intelligence, as part of MoH officials’ core training modules.

For example, the African Health Initiative (AHI) was launched in 2007 by the Doris Duke Foundation to strengthen health systems in sub-Saharan Africa. The initiative supports partnerships and large-scale models of care that link implementation research and workforce training directly to the delivery of integrated primary healthcare. Three AHI partnership projects in Ethiopia, Ghana, and Mozambique found that mentorship, joint analysis of data, and results-sharing bolsters sustainable engagement in data use for decision-making. Ensuring that training and capacity-building are adaptable to context, feasible to trial, and easy to introduce and maintain will be essential to the success of such programs. 

2. Design decision-support tools that cater to different levels of data acumen

Wide variation in data acumen among personnel in public health departments leads to suboptimal interaction with data and application of insights. The emergence of generative artificial intelligence (GenAI) applications for decision-support tools presents a promising opportunity for low data acumen users in the public health sector and beyond. These advanced AI tools function as assistants, particularly for users with limited data literacy. By simplifying the adoption of decision support mechanisms, they empower users to effectively leverage data-driven insights.  

GenAI tools hold strong potential to democratize access to sophisticated data analysis and decision-making tools in the healthcare sector. Importantly, such decision-support tools should not just target the highest-level decision-makers. They should be made available to all health workers at the facility, district, subnational, and community levels to truly capacitate health workforces.

Examples of GenAI-enabled decision-support tools:

The Bill & Melinda Gates Foundation Grand Challenges Initiative, focused on catalyzing equitable use of AI in LMICs, awarded Dalberg Data Insights grant funding to develop an AI-enabled health analyst. The tool helps public health officials “talk” to the data in the world’s largest health information management system – DHIS2 – using a chatbot. Questions could include: 

“What was the vaccination coverage rate for the BCG vaccine in Malawi in 2023?”

“How has the vaccination coverage for measles change in Malawi over the past five years?”

“Can you share the historical data on report completeness for last year and highlight any areas where completeness was an issue?” 

The chatbot functionality allows MoHs to move away from traditional business intelligence techniques and dramatically reduces the time required to generate insights. 

As part of the Rockefeller Foundation’s Intelligent Community Health Systems (iCoHS) initiative, UNICEF partnered with Dalberg Data Insights (DDI) to develop the Health Insights and Visualization for Essential Services (HIVES) application in Uganda. The tool provides access to data on the continuity of health services at facility and community levels. HIVES improves data handling by fixing time delays and data cleanliness issues by linking datasets to Uganda’s existing national health data system (DHIS2). It offers easy-to-understand data visualizations, automated data emails and monthly reports on over 150 health indicators. 

3. Strengthen data regulations and governance

Strong data governance is fundamental in the realm of public health, where patient consent and privacy are of paramount importance. Yet, the absence of mature privacy laws inhibits responsible management of sensitive data in many LMICs. This can preclude MoHs from processing even basic personal information needed to uncover population health needs and design targeted interventions, without running the risk of potential legal ramifications.  

Enforcing national (and as appropriate, sub-national) laws and guidelines to facilitate data sharing with the appropriate consent or legal basis, together with technical and organizational support for MoHs to protect against accidental or unauthorized disclosure, alteration, or loss of data, are prerequisites for effective and compliant data-driven decision-making. This becomes even more relevant as governments recognize and apply AI in the public health sector, which demands much larger datasets and further regulation for transparency, accuracy, and fairness of algorithms. 

The growing significance of data in enhancing healthcare delivery in LMICs and beyond is evident. We now need coordinated action to bolster data ecosystems for healthcare. Bilateral and philanthropic donors have a key role in covering the short-term costs of data system development and training health workers to more effectively utilize those for decision-making. MoHs need to institutionalize these measures by allocating ongoing resources for training and evolution of data processes in line with improvements in technology. Policymakers must prioritize the development and enforcement of guidelines and regulations to ensure that technology developers and government officials know what to build and how to implement it safely and equitably. Only by bringing these activities forward across sectors can we truly leverage the full potential of data for healthcare.