Insights from disease ecology: focus on hand, foot and mouth disease in China

Abstract

Outbreaks of hand, foot and mouth disease (HFMD) have become increasingly regular in the Asia-Pacific region, and China has experienced annual epidemics each year since 2007. This project studied large-scale environmental drivers of infectious disease, with a focus on understanding HFMD epidemics in China between 2008 and 2011. First, I assessed the potential of using associations with landscape pattern to discriminate diseases having a wild animal reservoir (wild-zoonoses) from those that do not, and tested the hypothesis that landscape ecology, as measured by select land cover pattern metrics, is more strongly associated with the incidence of wild-zoonoses than with the incidence of diseases lacking a wild animal reservoir. Quasi-Poisson regression models were used to estimate county-level associations between land cover pattern metrics and the incidence of three wild-zoonoses and eight diseases lacking a wild animal reservoir, for each county in the contiguous United States. The absolute strengths of the associations between each pattern metric and each disease were compared to determine whether the strongest associations were observed with wild-zoonoses. When sorted by absolute strength of association, wild-zoonoses had the strongest associations with six of the ten pattern metrics (p=0.008), and with all four of the metrics that measure land cover shape (p=0.002), suggesting that associations with land cover pattern metrics may offer insight into the existence of a wild animal reservoir for an emerging or otherwise poorly understood infectious agent Second, I sought to fill important gaps in our understanding of the ecology of HFMD by looking at land cover and land cover pattern, and their associations with HFMD incidence in China. Univariate and multivariate associations were estimated between each of twelve landscape ecology variables, plus population density, and HFMD incidence using quasi-Poisson regression models. Decreased elevation and vegetation density were significantly associated with increased rates of HFMD; and increased division, disaggregation, and diversity of land cover types were associated with increased rates of HFMD. The results suggest connections between landscape ecology and HFMD incidence that warrant further investigation, and support previous studies that have found local transmission to be more important than distant transmission. Finally, several studies have found associations between weather and HFMD, suggesting that climate change could have a role in the recent growth of HFMD in China. I sought to determine if climate change could underlie the recent emergence and growth of HFMD in China by developing a weather-based predictive model of HFMD and applying that model to historical climate data. When monthly climate-based HFMD predictions were regressed against calendar time, I found evidence of a significant increasing secular trend, with predicted rates for 2011 being 94% higher than those for 1982 (Incidence rate ratio (IRR): 1.937; 95% confidence interval (CI): 1.933, 1.940). Most of the increase in the predicted HFMD incidence occurred between 2002 and 2011, with predicted rates for 2011 being 49% higher than those for 2001 (IRR = 1.490; 95% CI: 1.488,1.493). Our climate-based retrospective predictions suggest that changing climate should have made weather increasingly favorable to HFMD during our thirty-year study period and we find that the data are compatible with climate change playing a role in the recent growth of HFMD in China.