Quantifying the relationship between climatic indicators and leptospirosis incidence in Fiji: A modelling study

Leptospirosis, a global zoonotic disease, is prevalent in tropical and subtropical regions, including Fiji where it's endemic with year-round cases and sporadic outbreaks coinciding with heavy rainfall. However, the relationship between climate and leptospirosis has not yet been well characterised in the South Pacific. In this study, we quantify the effects of different climatic indicators on leptospirosis incidence in Fiji, using a time series of weekly case data between 2006 and 2017. We used a Bayesian hierarchical mixed-model framework to explore the impact of different precipitation, temperature, and El Niño Southern Oscillation (ENSO) indicators on leptospirosis cases over a 12-year period. We found that total precipitation from the previous six weeks (lagged by one week) was the best precipitation indicator, with increased total precipitation leading to increased leptospirosis incidence (0.24 [95% CrI 0.15-0.33]). Negative values of the Niño 3.4 index (indicative of La Niña conditions) lagged by four weeks were associated with increased leptospirosis risk (-0.2 [95% CrI -0.29 --0.11]). Finally, minimum temperature (lagged by one week) when included with the other variables was positively associated with leptospirosis risk (0.15 [95% CrI 0.01-0.30]). We found that the final model was better able to capture the outbreak peaks compared with the baseline model (which included seasonal and inter-annual random effects), particularly in the Western and Northern division, with climate indicators improving predictions 58.1% of the time. This study identified key climatic factors influencing leptospirosis risk in Fiji. Combining these results with demographic and spatial factors can support a precision public health framework allowing for more effective public health preparedness and response which targets interventions to the right population, place, and time. This study further highlights the need for enhanced surveillance data and is a necessary first step towards the development of a climate-based early warning system.