Kerala Develops Dengue Early Warning System: District-Level Disease Forecasting Using Machine Learning and Weather Data
Kerala's Institute of Advanced Virology (IAV) has developed a Dengue Early Warning System (DEWS) based on machine learning, disease surveillance, and meteorological data. This system generates weekly risk forecasts for each district based on dengue cases, rainfall, temperature, and average humidity. District-level alerts can assist health authorities in proactively strengthening mosquito control, surveillance, testing, and healthcare preparedness. While this system exemplifies the use of AI/ML in public health, its forecasts are not absolute, as human behavior, urbanization, and other environmental factors also influence the spread of dengue.
Scientists at Kerala's Institute of Advanced Virology (IAV) have developed the Dengue Early Warning System (DEWS). It is designed to forecast potential dengue trends using machine learning, disease surveillance, and weather-related data. The system relies on six years of data—spanning March 2020 to February 2026—to generate weekly forecasts at the district level.
What factors are used in DEWS?
The model primarily incorporates four types of data: recorded dengue cases, rainfall, temperature, and average humidity. Disease-related data is sourced from Kerala's State Surveillance Unit, while meteorological data is obtained from the India Meteorological Department (IMD). Based on this data, the risk of dengue infection in each district is categorized as very high, high, moderate, or low.
The Link Between Weather and Dengue Transmission
The dengue virus is transmitted by infected Aedes mosquitoes, particularly Aedes aegypti. Rainfall creates water-logged areas that facilitate mosquito breeding, while temperature influences mosquito development and the process by which the virus becomes infectious within the mosquito. Humidity can influence mosquito survival; therefore, climatic and meteorological indicators are useful for forecasting the potential seasonal spread of dengue.
Significance of District-Level Forecasting
A key feature of DEWS is its localized or district-specific forecasting. Instead of indicating a uniform risk level for the entire state, it accounts for the specific epidemiological and meteorological conditions of individual districts. This enables health authorities to proactively strengthen measures—such as mosquito control, surveillance, testing, laboratory capacity, and healthcare facility preparedness—in areas identified as high-risk.
Application of Machine Learning in Public Health
Machine learning models assist in predicting potential disease trends by identifying complex relationships across various datasets. A separate study analyzing dengue and climate data from Kerala (2006–2019) also established a seasonal link between dengue cases and monsoon conditions. However, these models serve as forecasting tools rather than definitive predictors. Other factors—such as human behavior, urbanization, water storage practices, mosquito populations, and viral strains—also influence dengue transmission.
Broad Relevance for UPSC
This topic is relevant to Science & Technology, Health, and Biotechnology (GS Paper-III) as well as Climate and Environment (GS Paper-I). It can be understood within the frameworks of the 'One Health' approach, the intersection of climate change and health, AI/ML in healthcare, vector-borne diseases, disease surveillance, and disaster/public health preparedness. Key concepts for the Preliminary examination include DEWS, IAV, Aedes mosquitoes, IMD, the dengue virus, and machine learning-based disease surveillance. For the Main examination, it can be linked to questions regarding the relationship between climate change and vector-borne diseases, as well as the role of AI-based early warning systems.
