Significant health risks arise in Thailand from dengue but little work has been conducted to quantify the extremities of dengue outbreaks - where health systems are likely to be most stretched. In this paper, we detail the utility of tools derived from extreme value theory (EVT) in modelling the extremes in dengue case counts observed during outbreaks using 25 years of province level dengue case count data in Thailand from 1993 to 2018. We assess the validity of the EVT toolkit by comparing them against 8 competing benchmarks. The inhomogeneous point process representation (IPP) was found to perform best on 5 in and out of sample criterion such as parameter stability, distributional characteristics and out of sample coverage. Lastly, by using the IPP to infer future extreme dengue events, IPP found stark differences at the province level in the mean level of dengue case counts that is expected to be exceeded over the next 10 years. The IPP model also found that high probability that dengue extreme events will reach levels above and beyond the observed historical maximums. EVT shows considerable potential in aiding health planners for the risk management of dengue. The results in this paper can be easily translatable to any infectious disease observed over a long period.
Keywords: Dengue; Extreme Value Theory; Extremes; Outbreaks; Thailand.
Copyright © 2020 The Author(s). Published by Elsevier B.V. All rights reserved.