Participatory Data Science for Maternal & Child Health in Tribal Andhra Pradesh

Background

Tribal primary health centres (PHCs) in India operate within data-rich routine health information systems (RHIS), yet local decision-making for maternal and child health (MCH) remains constrained: data is fragmented, misaligned with local needs, and rarely available in usable, actionable formats. The result is a wide gap between centrally oriented reporting and the ground-level realities of MCH service delivery. This PhD asks how a participatory data-science approach can re-orient routine systems to support PHC-level decisions, transforming under-used data into meaningful, context-specific, and actionable insight.

Approach

The study uses Action Research with a data-science approach, structured around a “Three Co’s Framework” (co-design and co-creation) as iterative cycles of co-design, implementation, and reflection with frontline staff. Fieldwork is sited at the tribal PHCs of Boduluru, Vadapalli, and Gangavaram in ITDA-Rampachodavaram, Alluri Sitarama Raju district, following IEC approval in June 2023 and field cycles through 2023-2024 (colloquium January 2026). Methods combine R-based spatial data science, integration of HMIS/RCH data sources, participatory design workshops, and stakeholder interviews with medical officers and ANMs to ground tool design in real clinical workflows.

What we found

• PHC medical officers and ANMs face genuine usability barriers in existing portals (for example the RCH portal), including excessive nesting, no aggregated PHC-level view, and rigid “pending” logic that diverges from clinical reality.
• The most valued capabilities are simple and action-oriented: at-a-glance PHC summaries, drill-down by ANM, search by indicator or location, and actionable lists (for example identifying severe-anaemia cases at Hb < 7 g/dL for immediate follow-up).
• Low-bandwidth and offline tolerance are non-negotiable design constraints in this setting, not optional features.
• Co-design shifts ownership: when frontline workers help define indicators and visualisations, routine data begins to function as a decision trigger rather than an administrative obligation.

Outputs & impact

The PhD colloquium was completed in January 2026. Three of five planned thesis papers are available as medRxiv preprints (March 2026); the remaining two — a Conceptual Framework paper and a Dashboard Assemblage paper — are in preparation:

1. Mapping Data Sources for Local Decision-Making on Maternal and Child Health in Tribal Primary Health Centre Settings of Andhra Pradesh, India
2. Data use practices and challenges for maternal and child health decision-making in tribal primary health centres in Andhra Pradesh, India
3. Co-creating data science solutions for maternal and child health decision-making in tribal primary health centres: an action research using the Three Co’s Framework

Tangible outputs include co-designed PHC dashboards, reproducible data-integration algorithms, and anonymised research-level datasets. The intended impact is a transferable model for participatory data science in low-resource health systems, demonstrating how co-designed tools can shift institutional data-use culture toward locally grounded, evidence-based MCH decisions.