Effective disease surveillance is fundamental to public health preparedness and response, helping systems detect threats before they escalate. Nearly 200 countries operate a national disease surveillance system under the WHO International Health Regulations (IHR).
The Challenge
Amid a growing volume of information, traditional approaches relying primarily on manual tracking can be time-consuming and miss critical early signals, many of which first surface in online media and news reports.
Our AI Solution
Media Disease Surveillance (MDS) is an AI-powered media scanning solution that identifies potential disease outbreaks from online news reports in real time, automatically scanning media sources for over 120 health conditions to alert health authorities.
MDS reduces manual monitoring effort by filtering and consolidating relevant information, enables early detection of potential outbreaks through real-time media signals, and helps advance progress on the Integrated Disease Surveillance Program (IDSP) under India’s National Health Mission (NHM), supporting compliance with national and international disease surveillance mandates.
How it Works
01
Continuous Media Scanning
The platform continuously scans online news articles every day across national, regional, and local media sources in multiple languages.
Continuous Media Scanning
The platform continuously scans online news articles every day across national, regional, and local media sources in multiple languages.
02
AI-Driven Analysis
MDS uses an AI pipeline involving multiple NLP models and a large language model (LLM) to identify relevant news articles and public health events.
AI-Driven Analysis
MDS uses an AI pipeline involving multiple NLP models and a large language model (LLM) to identify relevant news articles and public health events.
03
Key Information Extraction
For each health event, the solution extracts entities such as the disease, location, and time of occurrence.
Key Information Extraction
For each health event, the solution extracts entities such as the disease, location, and time of occurrence.
04
Event Clustering and De-Duplication
Related health events are grouped together and duplicates removed, reducing noise for health officials.
Event Clustering and De-Duplication
Related health events are grouped together and duplicates removed, reducing noise for health officials.
05
Human in the Loop
MDS acts as an assistive tool within NCDC’s Media Scanning and Verification Cell (MSVC), with human experts reviewing and validating alerts.
Human in the Loop
MDS acts as an assistive tool within NCDC’s Media Scanning and Verification Cell (MSVC), with human experts reviewing and validating alerts.
Where It's Deployed
Deployed in partnership with the National Centre for Disease Control (NCDC), integrated within the Integrated Health Information Platform (IHIP), supporting the Media Scanning and Verification Cell (MSVC). Supports Hindi, Kannada, Tamil, Bengali, Marathi, English, and seven other Indian languages.
28
States
8
Union Territories
Attribution and Impact
- Increased Event Detection: Compared with human-only media scanning, the solution has increased event detection by identifying 98 percent of outbreak-related news events compared to 2 percent detected manually.
- AI-led Enhanced Surveillance: Over 20,000 health events flagged by the solution were shortlisted by NCDC experts for verification and action.
- Human-in-the-Loop: AI acts as an assistive tool within the MSVC of the NCDC, where human experts review and validate AI-flagged alerts, maintaining a human-in-the-loop system that safeguards accuracy and accountability.
Our Other Solutions
Partners
Looking Ahead
As health information continues to grow in volume, surveillance systems must evolve to keep pace. By continuously learning, adapting, and working closely with public health institutions, solutions like MDS can help translate vast streams of information into timely awareness and action.