Fine-grained dengue forecasting using telephone triage services

Area of Science:

• Public Health
• Epidemiology
• Health Informatics

Background:

• Lack of real-time disease surveillance data leads to delayed epidemic detection in developing countries, causing significant loss of life.
• Dengue fever poses a substantial public health threat, particularly in regions with limited surveillance infrastructure.

Purpose of the Study:

• To evaluate the efficacy of a telephone triage service for early dengue forecasting in Pakistan.
• To demonstrate the feasibility of building an accurate, location-specific disease forecasting system using public health hotline call data.

Main Methods:

• Statistical analysis of dengue-related phone calls to a large-scale telephone triage service.
• Forecasting suspected dengue cases at a subcity level with a lead time of 2 to 3 weeks.
• Operational deployment and data collection over three years, receiving over 300,000 calls.

Main Results:

• Accurate forecasting of suspected dengue cases with a correlation of up to 0.93.
• Successful 3-year operational scale deployment in Pakistan.
• Dissemination of predictions to public health officials, informing government dengue containment strategies.

Conclusions:

• Analysis of public health hotline call volume provides a viable method for accurate, location-specific disease forecasting.
• Telephone triage services can serve as a critical component of real-time disease surveillance and epidemic preparedness.
• This approach offers a scalable and cost-effective solution for early warning systems in resource-limited settings.