Abstracts: ISEE 22nd Annual Conference, Seoul, Korea, 28 August-1 September 2010: Statistical Methods in Environmental Health Research
1Queensland University of Technology, Brisbane, Australia; 2Umeå Universitet, Umeå, Sweden; and 3London School of Hygiene and Tropical Medicine, London, United Kingdom.
Abstracts published in Epidemiology have been reviewed by the societies at whose meetings the abstracts have been accepted for presentation. These abstracts have not undergone review by the Editorial Board of Epidemiology.
Researchers have theorized about the existence of a susceptible pool who are at greater risk of death from acute events such as extreme temperatures. This theory was developed because of highly changeable patterns in the risk of death which pointed towards a change over time in the at-risk population. We aimed to use mathematical modeling to estimate the number of people in the susceptible pool over time.
We modeled monthly data using a state space model with 3 states: Healthy, Susceptible, and Dead. We tried a range of different assumptions for the transition rates between states, including a seasonal risk of death and susceptibility. We aimed to create a seasonal pattern in deaths that peaked in January. We modeled acute increases in the transition rate to death to mimic the effects of cold spells and heat waves.
The size of the susceptible pool was highest after summer and lowest after winter, a pattern explained by a build-up during periods of low risk, and depletion during periods of high risk. The peak in deaths due to a cold spell was followed by mortality displacement which was much stronger when the susceptible pool size was 1% compared with 10%. The median months of life lost due a cold spell was just 3 for a 1% pool size, and 22 for a 10% pool. Depletion in the susceptible pool following a heat wave early in summer meant that an identical heat wave later in summer caused fewer deaths.
Changes in the susceptible pool can explain some of the highly variable associations between exposure to extreme temperatures and the risk of death. Our approach has the potential to improve the early preparedness of heat intervention strategies by estimating expected burdens for the coming summer.