Effect and prediction of long-term weather and pollutant exposure on hemorrhagic fever with renal syndrome: based on statistical models.

Abstract

Background: Previous studies have typically explored daily lagged relationships between hemorrhagic fever with renal syndrome (HFRS) and meteorology, with a limited seasonal exploration of monthly lagged relationships, interactions, and the role of pollutants in multiple predictions of hemorrhagic fever.

Methods: Our researchers collected data on HFRS cases from 2005 to 2018 and meteorological and contaminative factors from 2015 to 2018 for the northeastern region. First, we applied the moving epidemic method (MEM) to estimate the epidemic threshold and intensity level. Then, we used a distributed lag non-linear model (DLNM) and a generalized additive model (GAM) with a maximum lag of 6 months to evaluate the lagged and interaction effects of meteorological and pollution factors on the HFRS cases. Multiple machine learning models were then applied after Spearman's rank correlation coefficient analysis was performed to screen for environmental factors in the Northeastern region.

Results: There was a yearly downward trend in the incidence of HFRS in the northeastern region. High prevalence threshold years occurred from 2005 to 2007 and from 2012 to 2014, and the epidemic months were mainly concentrated in November. During the low prevalence threshold period, the main lag factor was low wind direction. In addition, the meteorological lag effect was pronounced during the high prevalence threshold period, where the main lag factors were cold air and hot dew point. Low levels of the AQI and PM₁₀ and high levels of PM_2.5 showed a dangerous lag effect on the onset of HFRS, while extremely high levels of PM_2.5 appeared to have a protective effect. High levels of the AQI and PM₁₀, as well as low levels of PM_2.5, showed a protective lag effect. The model of PM_2.5 and the AQI interaction pollution is better. The support vector machine (SVM)-radial algorithm outperformed other algorithms when pollutants are used as predictor variables.

Conclusion: This is the first mathematically based study of the seasonal threshold of HFRS in northeastern China, allowing for accurate estimation of the epidemic level. Our findings suggest that long-term exposure to air pollution is a risk factor for HFRS. Therefore, we should focus on monitoring pollutants in cold conditions and developing HFRS prediction models.