Tanzania's Ministry of Health deployed an AI triage layer in its DHIS2 disease-surveillance platform, clearing 85% of a 15,000-alert backlog and reducing processing time from 36–48 hours to near-instant, routing 23% of alerts for human epidemiological investigation.
26
Regions covered by e-IDSR
>15,000
Unreviewed alert backlog (2021-2023)
~23%
Alerts routed to human epidemiologists
36-48 hours
Processing time per alert batch, before
85%
Backlog cleared following deployment
5
Initial rollout regions
21
Remaining regions targeted for expansion
95%
Target auto-triage rate
>70
Countries using DHIS2
Details
Maturity
Scaling
Promoter
Tanzania Ministry of Health
Period
2021–present
Keywords
health, disease surveillance, public health
Context
Tanzania's Electronic Integrated Disease Surveillance and Response (e-IDSR) system collects disease-alert reports from health facilities across all 26 regions via the DHIS2 platform; as facility reporting improved over 2021-2023, a backlog of over 15,000 unreviewed alerts accumulated, hampering timely outbreak detection.
Objectives
Clear the alert backlog and speed up triage to prevent recurrence, at a time when outbreak-detection speed is critical.
Activities
The Tanzania Ministry of Health, the UDSM DHIS2 Lab, and the U.S. CDC developed a rule-based AI alert-triage tool embedded in DHIS2 that automatically classifies incoming alerts, routing approximately 23% to human epidemiologists and filtering the remainder, with machine-learning enhancements planned; initial rollout covered five high-priority regions, with expansion underway to the remaining 21.
Results
Processing that previously took 36-48 hours per alert batch now occurs in near-real time, and the system successfully cleared 85% of the 15,000-alert backlog following deployment.
Conclusions
The Ministry of Health is expanding the tool nationwide with a target of 95% auto-triage; DHIS2's use in over 70 countries suggests the approach could be adaptable to other low- and middle-income country health systems.
Implementation
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Commonly funded by
National / regional programmes
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Do you run this practice?
Claim it —
verified implementers get a public contact pathway and can propose corrections.
Data sources
Where this practice's information was retrieved from, and when.
WFP's AI-assisted HungerMap Live platform, built with Google.org and covering 50+ countries, helped identify worsening hunger in Somalia's Buur Hakaba …
Sierra Leone's National Medical Supplies Agency and University of Pennsylvania researchers built a decision-aware ML tool that raised essential-medicine consumption …