Integration Standards & Maturity Model
For Ugandan Health Supply Chain Systems Version 2.0 - Post-Validation Revision
Introduction
This document provides a realistic framework for data exchange across Uganda's health supply chain systems. It is grounded in the empirical findings of the 2025 Baseline Assessment, which revealed that 89% of facilities face unreliable internet, and that system fragmentation and parallel paper-based workflows are the norm, not the exception.
Objective
The objective is not to mandate an unattainable real-time integration, but to define a pragmatic maturity model that guides facilities from their current state of digital disconnection toward incremental levels of interoperability.
Foundational Principles
Grounding in Baseline Reality
The baseline assessment found:
System Fragmentation
Facilities operate multiple disconnected digital systems (DHIS2, eAFYA, CSSP, eLMIS) alongside paper records
Infrastructure Gaps
Unreliable power and internet prevent consistent use of API-dependent systems
Conclusion: Current integration efforts must prioritize practical, incremental data exchange over complex, real-time API orchestration.
Integration Maturity Model
This model defines a graduated pathway. The current project focus is Level 2 (Opportunistic Sync) as the primary target.
Level 0: Paper-Centric
All records on paper registers. No digital entry.
Physical transport of paper reports to district for manual entry
~40% of HCIIs. Focus: Digitization at source
Level 1: Single-System
One digital system used (e.g., DHIS2 mobile), reducing duplicate entry but isolated
Weekly export of CSV/Excel files, emailed or physically transported
~35% of facilities. Focus: Standardizing on offline-first app
Level 2: Opportunistic Sync
PRIMARY TARGET. Offline-first app used. Data synced when connectivity allows
App syncs JSON packets to central server when online. Pulls reports and forecasts
80% of facilities by project end
Level 3: Real-Time
Full API-based, real-time data exchange between national systems
RESTful APIs, FHIR resources. Assumes stable connectivity
Regional Referral Hospitals & District HQs only. Future state
System-Specific Integration Standards
DHIS2 Integration
Level 2 (Opportunistic Sync):
Central server periodically pulls aggregated consumption and stock data from DHIS2's API
Pushes back: Tier 2/3 forecast recommendations, stockout alerts, override logs as analytics outputs
Level 3 (Real-Time):
RESTful API and FHIR standards for bidirectional data exchange with sub-minute latency
e-LMIS Integration
Level 2 (Opportunistic Sync):
Facility-level orders (Tier 1 rule-based + human overrides) and stock counts collected via offline-first app
During sync, data pushed to central queue → processed in batch jobs (every 4-6 hours) to update e-LMIS
Eliminates need for facility staff to use multiple systems
Delivery information pulled from e-LMIS and pushed to facilities during next sync
Level 3 (Real-Time):
Direct RESTful API integration for real-time stock visibility and automated order submission
eAFYA Integration
Level 2 (Opportunistic Sync):
Patient attendance and service utilization data used at district level as Tier 3 forecasting feature
District IT officers export CSV files weekly using built-in export functionality
CSV files uploaded to district server via secure file transfer (SFTP or web interface)
Data informs allocation but does not trigger real-time deductions at facility level
Level 3 (Real-Time):
Direct integration for automatic deduction of dispensed commodities using FHIR MedicationRequest resources
WFP LESS Integration
Level 2 (Opportunistic Sync):
Delivery status updates pulled by central server via API calls (every 6-12 hours)
Pushed to relevant facilities during next sync
Provides near-real-time visibility into last-mile logistics
Supports proactive redistribution planning when delivery delays detected
Level 3 (Real-Time):
API endpoints for live shipment tracking and automated receipt confirmations
Nutrition Appointment Platform (WHO/WFP/UNICEF/UNHCR)
This platform supports nutrition service delivery for vulnerable populations including refugees and climate-affected communities.
Level 2 (Opportunistic Sync):
Central server pulls nutrition program enrollment data and appointment schedules (weekly)
Data feeds into Tier 3 district-level forecasting models as demand signals for nutrition commodities (RUTF, micronutrient supplements, etc.)
Supply adequacy indicators pushed back to Nutrition Appointment Platform to inform program planning
Level 3 (Real-Time):
Real-time bidirectional integration where appointment confirmations automatically trigger commodity allocation adjustments
Data Governance, Security, and Compliance
All data exchange must adhere to Uganda's Data Protection and Privacy Act (2019).
Security Standards
Encryption (Transit)
TLS 1.3 for all transmitted data
Encryption (Rest)
AES-256 for mobile devices via Android Keystore System
Authentication
Role-based access control (Principle of Least Privilege)
Roles: Facility Staff, Store Manager, District Supply Officer, National Administrator
Conflict Resolution Protocol
CRITICAL DESIGN PRINCIPLE
In case of conflicting data edits from multiple sources, the system applies a 'local user edits always win' policy.
Rationale: The health worker physically present at the facility has ground truth. If a facility worker manually adjusts an order quantity after counting physical stock, and a district officer or automated system also modified it, the facility worker's version is preserved.
Implementation:
All conflicts logged to override audit trail for district officer review
District officer can contact facility to understand reasoning
Cannot retroactively override without facility consent
Exception: During declared emergencies (disease outbreak, humanitarian crisis), district officers can flag orders for mandatory review, but facility's data remains system of record until facility explicitly approves changes.
Data Validation Rules
The system enforces these plausibility checks at data entry and upon central receipt:
Positive integers
Order quantities must be positive integers
Historical cap
Orders cannot exceed 200% of 6-month historical maximum without override justification
Facility type
Orders must align with facility type capacity (e.g., HCII cannot order surgical equipment)
Non-negative stock
Stock levels cannot be negative
Duplicate detection
Check for identical facility ID, commodity code, and timestamp within 24-hour window
Audit Trails
All transactions must be logged with:
User ID and role
Timestamp (UTC)
Action type (data entry, sync, override, forecast generation)
Forecast tier that generated the recommendation (Tier 1, 2, or 3)
Manual override justification (free text)
Tier 2 Forecast Data Management
Architectural Note
Tier 2 hierarchical statistical forecasts are pushed to facilities as informational guidance only. They are stored separately from the operational SQLite database to prevent concurrent write conflicts during background sync.
Facility staff can view Tier 2 forecasts in a read-only dashboard section but the primary Tier 1 rule-based forecast remains the default for order generation.
Realistic Integration Scenarios
Based on baseline findings and stakeholder feedback, the following scenarios must be supported at Level 2:
Disease Outbreaks
Facility uses offline app to mark Outbreak Mode
Justifies large manual order override (e.g., 300% increase in antimalarials)
Override synced to district, alerting them to potential redistribution need
Override audit log captures justification
System flags for district review without blocking the order
Delivery Delays
District officer sees LESS update about road washout affecting Karamoja shipments
Uses central dashboard to manually reassign shipments to accessible facility
Reassignment pushed to facilities during next sync
Affected facility receives notification explaining delay and alternative pickup arrangements
Budget Adjustments
Mid-year budget change entered into central system
New budget ceilings pushed to facilities during next sync
Facilities see updated, budget-aware Tier 2/3 forecast recommendations
Tier 1 rule-based forecast continues generating orders within new budget constraints automatically
Power/Network Downtime
Core workflow continues uninterrupted on offline app
All data stored locally in encrypted SQLite database (WAL mode)
When connectivity restored (automatically detected or manually triggered), data synchronized
Automatic conflict resolution favors facility's local edits
No operational delay even after weeks of offline operation
Nutrition Program Expansion
WFP scales up nutrition program in refugee settlements (e.g., +500 beneficiaries)
Nutrition Appointment Platform updates enrollment data
Central server pulls data during weekly scheduled sync
Tier 3 models automatically adjust predicted demand for RUTF and micronutrients
District officers receive alerts and can proactively reallocate stock before shortages occur
Monitoring, Evaluation, and Compliance
Key Performance Indicators
Integration Maturity
Facilities operating at Level 2+
80% by project end
Technical Reliability
Sync success rate
>95%
Technical Reliability
Data packet completeness
Monitored
Technical Reliability
Average time since last sync
Monitored per facility
Forecast Accuracy
MAPE Tier 1
<30%
Forecast Accuracy
MAPE Tier 2
<25%
Forecast Accuracy
MAPE Tier 3
<20%
User Engagement
Daily active users
Monitored
User Engagement
Override frequency
<15% of orders
Operational Impact
Stockout frequency/duration
Reduced from baseline
Operational Impact
Time spent on supply chain tasks
50% reduction from baseline
Compliance Audits (Semi-Annual)
Technical Integration
Data flows between systems, API endpoint functionality, error logs
Data Quality
Completeness, accuracy, timeliness of synced data, conflict resolution logs
Usability
Frontline worker interviews confirming workload reduction
Privacy Compliance
Audit trails, encryption implementation, role-based access controls
Post-Grant Implementation Roadmap
Phase 1
Months 1-3
Deploy offline-first mobile app with Tier 1 forecasting. No external system integration. Build user trust
Phase 2
Months 4-6
Activate Level 2 integration with e-LMIS and DHIS2. Deploy district servers with Tier 2/3 forecasting
Phase 3
Months 7-9
Add eAFYA, WFP LESS, Nutrition Appointment Platform integration at Level 2. Weekly data flows
Phase 4
Months 10-12
Pilot Level 3 real-time integration at 2-3 regional referral hospitals. Evaluate before broader rollout
Governance Structure
District Health Officers
Primary system owners, core partners throughout implementation
MoH Technical Working Groups
Policy-level review, national digital health strategy alignment
Quarterly Review Committee
Cross-stakeholder review of integration performance, override patterns, refinements
Conclusion
This integration framework is designed to meet Uganda's health supply chain where it currently operates, not where we wish it to be.
By prioritizing Level 2 (Opportunistic Sync) integration, we provide a pragmatic path that respects infrastructure constraints while delivering tangible benefits:
Reduced duplicate data entry
Enhanced supply visibility
Better-informed decision-making
The framework explicitly rejects integration approaches that would increase frontline worker burden or require always-online connectivity. Instead, it builds incrementally from the offline-first foundation, adding integration touchpoints only where they provide clear value without introducing new points of failure.
This is integration designed for resilience, not fragility.
Document Information
Version
2.0 (Post-Technical Validation)
Date
November 2025
Authors
Gideon Abako, Timothy Kavuma (IFRAD)
Validator
Ojok Ivan, Kyambogo University
License
CC BY 4.0
Last updated