Snow integrated communicable disease prediction service

Abstract

Objective: This thesis mainly focused on construction of an integrated infectious disease prediction service that predicts and visualizes prediction results in time and space. Methods: We have used weekly aggregated laboratory confirmed cases of various diseases collected from the Snow system, which is an infectious disease surveillance system that covers Troms and Finnmark counties of north Norway. Influenza A dataset is applied for modeling SIR(S) model and various diseases datasets applied to a Bayesian model. The infectious disease prediction service prototype was constructed following an iterative and incremental approach where the entire development process was composed of four activities. Results: The prediction service framework facilitates the process of integrating various models and allows their evaluation. Currently, the system contains two mathematical models that demonstrate the effectiveness of the architecture in integrating new models. Conclusion: The framework can significantly improve the status of disease prediction systems, investment and time of development. It also speeds up mathematical modeling through its integrated environment for testing and evaluating different mathematical models against other existing models. Thus, the project contributes to improve the overall disease prediction accuracy and increase the benefits from prediction. Keywords: Infectious disease, Influenza, Mathematical model, Prediction, Mathematical model evaluation, Spatiotemporal Epidemiological Modeler, Visualization, Integrated infectious disease prediction.