Predicting town-level cases of Lyme disease in Southern Maine
Abstract
Maine has some of the highest rates of Lyme disease in the United States. Lyme disease is treatable when detected, but can have severe health effects when left untreated. Much scientific work on Lyme disease prevention has explored environmental covariates with Lyme disease rates or tick populations, to gain either mechanistic understanding of ticks as vectors or predictive understanding of how Lyme disease cases correlate with environmental variables. However, our current predictive understanding is at national, regional, or county-scale resolution; in contrast to evidence that the majority of infections happen locally. To predict and prevent Lyme disease spread we need an understanding of how the environment affects infection rates at the scale of infection.
We fit competing Bayesian models of Center for Disease Control annual observations of Lyme disease incidence for each town in Cumberland County, ME for 2009-2016, with 2018 and 2017 withheld for validation. We compared null models to models driven by in-situ and remotely-sensed environmental parameters: temperature, humidity, and precipitation. We also compared null models dynamic in time to null models dynamic in space. We evaluated models using DIC and hold-out cross validation. Even at the scale of county, we found evidence of spatial heterogeneity. We found that across towns, underlying Lyme disease detection rates were high. However, there were large town-to-town differences in detection rates and order-of-magnitude differences in probabilities of infection. Annual-scale humidity proved to be a poor predictor of Lyme disease incidence. This work helps test assumptions about underlying reporting rates and informs tailored monitoring and prevention efforts. In addition, it is a novel example of testing predictive models of disease at the local scale.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMGH41B1206M
- Keywords:
-
- 0230 Impacts of climate change: human health;
- GEOHEALTH;
- 0240 Public health;
- GEOHEALTH;
- 0245 Vector born diseases;
- GEOHEALTH;
- 1813 Eco-hydrology;
- HYDROLOGY