Predictors For Abundance Of Small Mammals In Households Of Rodent Borne Diseases Foci In Western Of Yunnan Province

[Objectives]To know the distribution and proportion of small mammals in the households ofrodent-borne diseases foci in Western of Yunnan Province and to analyze theecological environment, geographical landscape, social economy, and humanintervention factors affecting the abundance of small mammals for establishing aepidemiological model to predict small mammals abundance and finding out a newmethod to control the abundance of small mammals.[Methods]1．Study samples and objects: Eight hundred households (20households eachvillage) were randomly selected in40natural villages of10counties from rodent-borne diseases (plague, epidemic hemorrhagic fever, bartonella and anaplasmosis)foci in Western of Yunnan Province as study unit. Small mammals and keyenvironment factors were collected via placing live-cages and observation andquestionaires.2．Collecting small mammals: Small mammals in households were capturedusing live-cages and fried ham sausage bait. Five cages each household were placedfor3continuous nights. The cages were checked every morning. When capturingsmall mammals, a new cage with same bait was placed in the same place. Thecaptured small mammals were brought laboratory for identification and counting. 3．Collecting potential factors affecting small mammals abundance: Self-madequestionairs (village and household) and face to face interview were applied tocollect potential factors (including ecological environment, geographical landscape,social economy, and human intervention) affecting small mammals abundance.4．Data management and analysis: EpiData3.02was used to establish the database and stata format was exported. Data were analyzed using R3.0.3and statisticalpackages (epicalc, MASS, pscl, and lmtest). Chisq test, logistic regression, negativebinomial regression, and hurdle negative binomial regression were applied to explorethe factors affecting small mammals abundance. The number of small mammals inthe villages, the number of small mammals in the households, the number of Rattustanezumi in the households, and seeing small mammals in the households wereregarded as dependent variabls and all possible factors in the questionairs wereregarded as independent variabls. Univerate analysis for each variable wasconducted and the variable with P value less than0.2was chosed to enter theprototype multiple regression model. The final models were refined using backwardelimination and stepwise to reduce independent variable (α<0.05as statisticalsignificance level). Coefficients and95%CI of the logistic regression componentwere exponentiated to obtain prevalence odds ratios (OR) and the negative binomialcomponent were exponentiated to get abundance ratios (AR).[Results]1． Four hundred and twenty-one small mammals were captured in800households of40natural villages in10counties and were divided into9species,6genera,2families,2orders. Of421small mammals, Rattus tanezumi (278), Rattusnorvegicus (30), Mus musculus (76), Rattus nitidus (23), Suncus murinus (7), Apodemus chevrieri (2), Anourosorex squamipes (2), Crocidura attenuate (2), and Muscaroli (1). Rattus tanezumi was the dorminat species. The range of the number ofsmall mammals for10counties was16to100and the average of captures for eachcounty was42small mammals. The highest number of small mammals (100) wascaptured in Mangshi city. 2．The negative binomial regression model for the number of small mammals of40villages was established. The following factors were confirmed: The number ofsmall mammal captures in the village with Yi2.08times higher than that of thevillage with Han ethinic group; The number of small mammal captures in the villageplanting fruits nearby was increased1.71times; The number of small mammalcaptures in the village with more than80%households with chickens was increased1.58times; The number of small mammal captures in the village with killing smallmammal behavior within6months was increased2.42times; The number of smallmammal captures in the village with average humidity more than55%was increased1.82times.3．The hurdle negative binomial regression model for the number of smallmammals of800households was established. The factors were confirmed as follows:The probability of capturing small mammals in the household with family memberreciving high education was55%less than that of the household with family memberreciving middle education and below; The probability of capturing small mammals inthe Dai and Yi households increased2.35-3.01times compared with the Hanhouseholds; The probability of capturing small mammals in the households whichgarbage put in the outdoor (within20meters and more than20meters) decreased67-72%; The probability of capturing small mammals in the households of seeing smallmammals and vegetables grown near house (within50meters) increased1.54-1.70times; In the households which thought there were many small mammals, theprobability and number of capturing small mammals increased1.6and1.74times,respectively; The number of capturing small mammals in the households with thehouse of brick and wood and soil floor decreased45-71%.4．The hurdle negative binomial regression model for the number of Rattustanezumi of800households was established. The factors were confirmed as follows:The probability of capturing Rattus tanezumi in the household with family memberreciving high education was65%less than that of the household with family memberreciving middle education and below; The probability of capturing Rattus tanezumi inthe Dai households increased2.74times compared with the Han households; The probability of capturing Rattus tanezumi in the households keeping pigs wasdecreased47%; The probability of capturing Rattus tanezumi in the households withseeing small mammals and in the households which thought there were many smallmammals, having other house around the house, and vegetables grown near house(within50meters) increased1.80-2.42times; In the households with measures toprevent small mammals, the probability of capturing Rattus tanezumi decreased34%;The number of capturing Rattus tanezumi in the households with soil and wood houseand cement floor increased2.30-4.16times.5．The logistic regression model for wether or not sighting of small mammals inliving area was established. The factors were confirmed as follows: The probability ofsighting of small mammals in living area in the Naxi and Zang households decreased60-63%compared with the Han households, but the probability of sighting of smallmammals in the other ethnic group households was1.91times higher than Hanhouseholds; the probability of sighting of small mammals in the households with thewater storage container with a cover and keeping cats decreased39-41%; In thehouseholds and fields where there were many small mammals, and the householdsseeing faeces of small mammals, the probability of sighting of small mammals inliving area increased1.83-6.48times.[Conclusions]1．Small mammals are divided into9species,6genera,2families,2orders inthe households of Western of Yunnan Province. Rattus tanezumi is the dorminatspecies. The small mammals density had exceed the5%of warning value in Mangshiand Yongren county, the risk of rodent borne diseases occurrence is still exists. Themeasures killing small mammals and the surveillance of rodent borne diseases shouldbe conducted and strengthened in order to reduce the risk of rodent borne diseasesoutbreaks and epidemics.2．The number of small mammal captures will increase in the village withaverage tempreture and average humidity more than20℃and55%, ethnic group,more than80%households with chickens, the village planting fruits and wheatnearby. The Centers for Disease Control and Prevention should pay more attention to the village with above characteristics.3．The probability or number of capturing small mammals will increase in thethe Dai and Yi households and have a lot of family members, seeing small mammalsin the households and vegetables grown near house. The households with the abovecharacteristics need to be concerned. However, the probability or number of capturingsmall mammals will decrease in the households of family member reciving higheducation, garbage put in the outdoor, the house of brick and wood and soil floor. Itcontributes to control the abundance of small mammals and occurrnce of rodentborne diseases in the households having the above characteristics.4．The probability or number of capturing Rattus tanezumi will increase in theDai households, which thought there were many small mammals, having other housesurrounding the house, vegetables grown near house, and the house of wood andcement floor. The households having the above characteristics in plague foci shouldbe paied more attention to. The probability or number of capturing Rattus tanezumiwill decrease in the households with measures to prevent small mammals and familymember reciving high education. It contributes to prevent the occurrence of plagueand other rodent borne diseases.5． In the western of Yunnan Province, lots of ethnic groups, undevelopedeconomic and culture, poor living environment, and favorable ecologicalenvironment and climate for small mammals living and reproduction. are mainreasons for higher small mammals abundance and higher occurrence of rodent bornediseases. So the local government should take measures to develop the economy andimprove living conditions and sanitary condition, protecting the ecologicalenvironment in order to control and reduce the occurrence and epidemic of rodentborne diseases.