Abstract

Towards the development of an early-warning system on disease outbreak
Track: Health and Human Services
Authors: Poh Lai, Ho-Ting Wong, Kim-hung Kwong

Disease modeling using GIS is primarily retrospective and lacks forward warning capacity to prevent/notify disease spread in real-time. We developed an early warning system to detect disease outbreaks by geographic localities. Our grid-based model is founded on known patterns of SARS and H5N1 outbreaks in Hong Kong. We implemented the SEIR (Susceptible-Exposed-Infected-Removed) epidemiological approach and modeled for within-cell and cross-cell transmissions using multi-criteria spatial overlay to incorporate environmental/social variables that interact in space and time to affect the patterns of disease transmission. Results of geo-statistical and spatial-scan measures show that our GIS model can explain spatial variance of disease outbreak up to 64% but the model is rather sensitive to changes in the modeling parameters, such as suitable cell size (granularity) and not fewer than 8-10 total cases (sensitivity). Our spatial model was able to predict potential locations of outbreak to forewarn communities needing early intervention or control measures.