O-42 The process of modelling the number of dengue outbreaks in Jamaica for the period 2000–2014
Author(s):
N Guthrie-Dixon , J Campbell , M Taylor , I Hambleton , G Gordon-Strachan
Type Of Study:
- Analytical Study
- Observational Study
Year of Presentation:
2022
Objective: To develop a statistical model aimed at modelling the number of dengue outbreaks in Jamaica for the
period of 2000–2014 based on the lagged effects of climatic
factors of maximum temperature and rainfall.
Methods: A retrospective study was performed to ascertain the relationship between climatic factors and dengue counts for the 2000–2014 study period. Data concerning the dengue presentations was sourced from the Ministry of Health and Wellness, Jamaica and that concerning the climate parameters from the Climate Studies Group Mona (CSGM) at the Department of Physics, University of the West Indies, Mona Campus, Jamaica. Graphical displays were generated to highlight the patterns for dengue events and the climatic conditions over the period. With dengue being a seasonal epidemic, seasons were defined in 3-month periods of June to August (JJA), September to November (SON), December to February (DJF) and March to May (MAM). To a baseline negative binomial model using nested random effects (Month<Season<Year), varying derivatives of the climate parameters, were added independently and jointly to create additional models.
Results: The dengue incidence rate was greatest in 2012 at 207.6 per 100,000 person-years. Superimposed graphs provided justification for using lagged responses of the climate variables. On adding varying derivatives of the climate variables to the baseline model (AIC = 1463.866), the final model selected (AIC = 1417.415) revealed significant associations with 4-month lagged Lowess smoothed maximum temperature (IRR = 8.96, p < 0.001) and 5-month lagged rainfall (IRR = 1.06, p = 0.001).
Conclusion: Lagged responses of climate parameters can be used as a tool to predict future dengue outbreaks.