A heat health warning system in Thailand including statistic associations and stakeholders consensus exercise

Abstract

Background The aims of this work were to explore the nature of HRIs in Thailand, examine any associations between heat index and heat-related hospitalisations within the general Thai population, and to develop a structure for a HHWS model in Thailand based on the results of statistical analyses and experts’ opinion. Method This study used a mixed methods approach. In phase 1, a daily tally of HRI hospitalisations from the International Classification of Diseases 10 Revision (ICD10) database with diagnosis T67 (effects of heat and light) were obtained between January 2010 and December 2018. Daily temperature and humidity figures from the same period were obtained. The heat index was calculated. Time series and Poisson regression analysis were used to explore the relationship between HRIs and the heat index controlling for day of the week and holiday indicator, with lag times of 0–7 days. Relative Risk (RR) and 95% confidence intervals were calculated based on a Poisson model from each region of Thailand. Next, an e-Delphi exercise with 16 experts and a focus group with key stakeholders and policy makers were applied to select the suitable threshold level.Results Based on expert opinion, the pre-alert level is the level of the heat index below the 75th percentile. The higher level is the level of the heat index from the 75th percentile to the 90th percentile. Lastly, the highest level is the level of the heat index from the 90th percentile. All thresholds were applied depending on the relationship between HRIs and the heat index in each region of Thailand. Conclusion This study found the heat index had positive associations with HRI hospitalisations and vary in each region. Importantly, the results of this study support the view that Thailand should have be spoke HHWS which is different from those that has been operated elsewhere.