| Schistosomiasis is an important public health issue in China. However, prospects for the control of schistosomiasis have been less optimistic in recent years, and there was an upward trend of the incidence rate of schistosomiasis in some areas. Therefore, the study the control of snails and integrated control strategies for schistosomiasis were important at this stage. In this paper, mixed linear models, time series analysis models, generalized additive models and generalized estimating equations were used to analyse the characteristics of oncomelania distribution and time trend among the three types (diked area, bottomland area, and hills area), establish the prediction models of oncomelania distribution for three types, explore the influencing factors of oncomelanla distribution in different scales and the complex nonlinear relationship between oncomelania distribution and factors in Hubei Province, and assess the effect of a comprehensive strategy to control transmission of Schistosoma japonicum in Gong’an County. The paper included four sections.Part1Analysis of the characteristics of oncomelania distribution in Hubei ProvinceObjectivet To investigate the difference of oncomelania distribution and time trend among three types of oncomelania habitat in Hubei Province. Methods:A retrospective survey methods was used to collect the oncomelania data among30counties in Hubei Province from1980to2009every five years (the last time was a four-year interval). Two indicators, including the area and density of oncomelania, were extracted from the raw data. According to the main type of environment of oncomelania habitat, the30counties were divided into three types, which were diked area, bottomland area, and hills area. A mixed linear model was used to analyse the difference of oncomelania’s area and density and time trend among the three types.Results:There were higher oncomelania’s area both in bottomland area and diked area than in hills area (P=0.0179and P<0.0001.respectively).And the whole area has a downward trend about oncomelania’s area (P=0.0139). There were higher oncomelania’s density in diked area than in hills area (P=0.0098). And the diked area has a downward trend about oncomelania’s density (P=0.0005). Conclusion:There were differece of oncomelania distribution and time trend among three types of oncomelania habitat in Hubei Province.Part2Study of the time characteristics and prediction model of oncomelania distribution in Hubei ProvinceObjective:Analyse the time characteristics of oncomelania distribution among three types of oncomelania habitat in Hubei Province, and establish the prediction models of oncomelania distribution for three types.Methods:A retrospective survey methods was used to collect the oncomelania data among3counties, which represent three types of oncomelania habitat, in Hubei Province from1980to2009every years. The autoregressive integrated moving average (ARIMA) model was used to predict oncomelania density for three types. Meanwhile, the climate data about the3counties from1980to2009every years were collected. The multivariate ARIMA (ARIMAX) model was used to predict oncomelania density in considering the influence of climate factors for three types. Results:The ARIMA models of oncomelania density for diked area, bottomland area, and hills area were(1-B)(1-0.867B6)d1=ε1,(1-B)(1+1.014B+0.776B2)d,=ε1,(1-B)d,=(1-0.475B)ε1,respectively. And the ARIMAX models of oncomelania density and climatic factors for diked area and hills area were d1=16.363+(-0.501-0.380B)tem1+(1+0.535B)v1,,respectively. In ARIMA and ARIMAX models, d1, tem1and sun1represents the oncomelania density, the daily average temperature and sunshine duration, respectively;ε1and v1represents the white noise series; B represents the delay operator. The results showed that ARIMA and ARIMAX models have good predictive effect, most of the measured value were in95%confidence interval of predictive value. Conclusion:The ARIMA and ARIMAX models can be used to predict oncomelania density in practical work, however, established prediction models needs consider the characteristics of the environment.Part3Study of the influencing factors of oncomelania distribution in different scales in Hubei ProvinceObjective:To explore the influencing factors of oncomelania distribution in different scales in Hubei Province, specially focusing on the complex nonlinear relationship between oncomelania distribution in small scale and factors. Methods:In a larger scale (county level scale), a retrospective survey methods was used to collect the oncomelania density data among30counties in Hubei Province in2000,2005and2009. Meanwhile, the climate data about the30counties in the3time point and their last years were collected. A linear mixed model was used to analyse the relationship between oncomelania density and climate factors. In a small scale (sampling area level scale), a cross-sectional study about oncomelania distribution was conducted in5counties in2010. Two indicators, including the density of oncomelania and the number of positive oncomelania, were collected in the survey. The3S technology was used to collect the information of environment and terrain about the sampling point,and the last year hydrological data about the sampling point was also collected. A generalized additive model was used to analyse the the complex nonlinear relationship between oncomelania distribution in small scale and factors. Results:In the larger scale, the average temperature has positive effect on oncomelania desity in the whole area (P=0.0150). The rainfall has positive effect on oncomelania density both in diked area and bottomland area (both P<0.0001), however, the rainfall has no effect on oncomelania density in hills area. In the small scale, the gradient of sampling areas has a negative impact on oncomelania density. There were complex nonlinear relationship between oncomelania density and factors,which included last year’s watered out days, elevation and surface temperature of sampling areas (P<0.0001, P=0.0462and P=0.0031,respectively). If a sampling area has the characteristic about last year’s watered out days of90days, an elevation of29meters and the surface temperature of26degrees, it would have a highest oncomelania density in all areas.In addition. the elevation of sampling areas has a negative impact on the number of positive oncomelania. There were complex nonlinear relationship between the number of positive oncomelania and factors,which included last year’s watered out days, gradient and surface temperature of sampling areas (P<0.0001,P=0.0026and P<0.0001,respectively). Conclusion:In the large scale, the climate factors have different effect on oncomelania density in three type areas. In the small scale, There were complex nonlinear relationship between hydrological, environmental and topographical factors and oncomelania distribution.Part4Study of the control strategy for schistosomiasis japonica in marshland and lake regions of Hubei ProvinceObjective:To assess the effect of a comprehensive strategy to control transmission of Schistosoma japonicum in marshland and lake regions. And to explore the specific measures of controlling schistosomiasis japonica in Hubei Province. Methods:In a cluster randomized controlled trial, we implemented an integrated control strategy in twelve villages from2008through2011in Gong’an County, Hubei Province. The routine interventions included praziquantel chemotherapy and controlling snails, and were implemented in all villages. New interventions, mainly consisting of building fences to limit the grazing area for bovines, building safe pastures for grazing, improving the specialized schistosomiasis clinics at the village level and strengthening health education, improving the residents’health conditions and facilities, were only implemented in six intervention villages. Five outcome measures, including the primary one (the prevalence of S. japonicum in humans) and four secondary ones (the rate of S. japonicum infection in bovines, cow dung, snails and mice), were determined annually and used to assess the strategy’s effect. A generalized linear model with a logit link and a binomial error distribution was used to analyze the risk of S. japonicum infection in humans. Generalized estimating equations of parameters with an unstructured variance-covariance matrix were used to account for repeated measures on individuals during the study period. Results:The rate of S. japonicum infection in humans, bovines, snails, cow dung and mice in the intervention group decreased from3.41%in2008to0.81%in2011.3.3%to none,11of6,219to none,3.9%to none and31.7%to1.7%, respectively (P<0.01for all comparisons). In contrast, there were no significant statistically reductions of S. japonicum infection in humans, bovines and snails from2008to2011in the control group (P>0.05for all comparisons). Moreover, the generalized estimating equations showed that there was a higher infection risk in humans in the control group than in the intervention group (OR=1.250, P=0.001) and an overall significant downward trend in infection risk during the study period (P<0.001). Conclusions:The integrated control strategy, designed to reduce the role of bovines and humans as sources of S. japonicum infection, was highly effective in controlling the transmission of S. japonicum in marshland and lake regions of Hubei Province. The specific measures, including building fences to limit the grazing area for bovines and building safe pastures for grazing, met the actual situation in Hubei Province, and they can be used to control schistosomiasis japonica in Hubei Province. |
PDF Full Text Request