Using massive, multi-regional EHR data to estimate the impacts of climate change on fungal disease epidemiology in the U.S.

点击翻译按钮获取中文摘要

ABSTRACT In the past 40 years, fungal diseases have emerged as a pressing health concern, as incidence rates have increased markedly, novel pathogens have emerged, resistance to antifungal drugs has risen and prevalence of immunosuppressive conditions has increased. Changes in climate may exacerbate fungal disease risks by shifting the suitable environmental habitat for pathogens, lengthening the transmission season of spores, increasing the frequency of extreme climate events that mobilize spores and further straining adaptive capacity of vulnerable populations. A mechanistic, quantitative, and causal understanding of the effects of climate on the distribution and severity of fungal diseases will be critical to protecting the health of high-risk groups in the coming decades, and to addressing factors leading to health disparities across axes of social vulnerability. To date, epidemiologic studies of fungal diseases in the U.S. have been limited in spatiotemporal scope and sample size, precluding robust characterization of the impact of climate and extreme events on fungal disease risk. While largely untapped as a resource for investigating fungal disease, electronic health record (EHR) and infectious disease surveillance systems generate massive health datasets that can be used to estimate risk factors for fungal infections, including candidiasis, cryptococcosis, aspergillosis, blastomycosis, histoplasmosis, coccidioidomycosis and dermatomycosis. Via partnership with Cerner (Kansas City, MO), we will analyze de- identified EHR data for over 96 million patients, 1.4 billion visits, and 4.7 billion clinical events. A subset of the databases is available with geographic locations of patient 3-digit zip codes. We will also analyze surveillance data on all reported cases (>95,000) of coccidioidomycosis in California since 2000, geolocated to patient address. After addressing misclassification, selection, and missing data biases in the EHR, our team will estimate regional trends in incidence, hospitalization and mortality rates for fungal diseases in the U.S. We will apply modern time series approaches to understand the effects of climate variability, including in temperature, precipitation, and humidity, on fungal disease incidence and geographic emergence. We will investigate the impacts of extreme events such as heat waves, dust storms, tropical cyclones and flooding on incidence and examine whether investment in infrastructure can mitigate effects. We will determine whether exposure- response relationships between climate and fungal infection are modified by socio-economic status and race, and identify individual- and community-level factors that mediate the relationship between social disadvantage and fungal infections and severe outcomes, such as preexisting comorbidities and housing quality. We will apply our estimated exposure-disease relationships to quantify the proportion of fungal disease incidence since 2000 that is attributable to anthropogenic climate change, and project mid- and end-of-century estimates of the distribution and burden of key diseases. The project will yield robust understanding of the climate epidemiology of major mycoses in the U.S., and the role of social disadvantage and resilience in moderating future impacts.