Title: Time series models for tracking and forecasting epidemics
Authors: Andrew Harvey - University of Cambridge (United Kingdom)
Paul Kattuman - University of Cambridge (United Kingdom) [presenting]
Abstract: As an epidemic takes hold, projections of its trajectory enable health care providers to plan and organize clinical and human resources to meet treatment requirements. A new class of time series models is developed that reflect epidemic trajectories and are able to make good forecasts even before new cases and/or deaths reach their peak. The models are relatively simple and transparent, and their specifications can be assessed by standard statistical test procedures. The nature of epidemic trajectories leads us to formulate a class of models in which the logarithm of the growth rate of the cumulative series follows a downward time trend. Allowing this trend to be time-varying introduces flexibility and enables the effects of changes in policy to be tracked and evaluated. The models are able to adapt as the response of the population changes over time. The framework can be extended to modelling the relationship between two or more series. When there is balanced growth, simple regression models can be used to forecast using leading indicators. The models are applicable in a wide range of disciplines.