Background:
Seasonal influenza epidemics occur each year and has a major toll on populations across the globe. About 20% of the children and 5% of the adults worldwide develop symptomatic influenza each year. Although influenza is mostly a self-limiting disease, there are few other diseases that cause as much absenteeism from work and schools, suffering, visits to outpatient clinics, hospitalizations, mortality and economic loss.
Objectives:
Our goal is to study seasonal influenza in Israel through the long-term highly resolved datasets held at the ICDC. Statistical models will be used to gain insights into the seasonal dynamics, progression, the spatial spread across the country. It will also yield important details (for modeling also) of flu natural history parameters: reproductive number R0, infectiousness of influenza in Israeli households, identify and target groups for vaccination. The analysis will be used to examine appropriate vaccination policies suitable for the Israeli context.
Hypothesis:
We will test the following hypothesis concerning influenza dynamics in Israel: 1) There is a generic spatial pattern in the evolution of the spread of influenza in Israel. 2) Annual influenza epidemics start first at pre- and school aged children and then spread to older age groups in a time lag of about 2 weeks.
Methods:
A novel and comprehensive statistical and time series analysis will be conducted on influenza data collected at the Israel Center for Disease Control. The data includes patient visits to Maccabi Health Services with influenza like illness between 1998-2007, and has high resolution both in space (clinic locations) and time (daily). Apart from classical time series analysis (correlograms, variograms, spectral analysis) we will use likelihood approaches for parameter fits and markov random field models to capture spatio-temporal dynamics.
|