Mapping Data Sources for Local Decision-Making on Maternal and Child Health in Tribal Primary Health Centre Settings of Andhra Pradesh, India

Background: Health systems in low- and middle-income countries are frequently described as "data rich, information poor", collecting substantial amounts of data that rarely inform local decision-making. In tribal settings, this challenge is compounded by geographic isolation, fragmented governance, sectoral silos, and the absence of disaggregated tribal health data within routine health information systems. We conducted a systematic mapping of data sources available for maternal and child health (MCH) decision-making at tribal Primary Health Centres (PHCs) in Andhra Pradesh, India. Methods: Using a participatory data discovery approach embedded within an action research project, we mapped data sources across three PHCs under the Integrated Tribal Development Agency (ITDA) - Rampachodavaram, Alluri Sitarama Raju District of Andhra Pradesh, India. Data discovery proceeded through three phases: document review, key informant interviews with Medical Officers and frontline health workers, and stakeholder validation. Sources were classified using the HEALTHY framework (Healthcare, Education, Access, Labour, Transportation, Housing, Income) and the Keller's data discovery typology (Designed, Administrative, Opportunity, Procedural). Accessibility was assessed based on whether Medical Officers could retrieve data for local planning and decision-making. Results: We identified 28 distinct data sources relevant to MCH decision-making. Healthcare dominated (57.1%), while determinant domains remained underrepresented: Housing (10.7%), Income (10.7%), Education (7.1%), Labour (7.1%), Transportation (3.6%), and Access to healthy choices (3.6%). By data origin, Administrative sources predominated (46.4%), followed by Opportunity (21.4%), Procedural (17.9%), and Designed (14.3%). Despite 67.9% of sources having digital components, only 32.1% were fully accessible to Medical Officers, with 10.7% partially accessible and 57.1% inaccessible at the PHC level. Accessibility barriers were consistent across data categories, ranging from 50.0% to 66.7% inaccessibility. Conclusions: The tribal PHC data ecosystem exhibits a fundamental mismatch between data generation and local utility. Data is predominantly collected for administrative reporting rather than local decision-making. Addressing MCH outcomes in tribal populations requires reorienting health information systems toward local needs.