Study On The Epidemic Characteristics And Prediction Of Core Pathogen Trend Of Rash And Fever Syndrome In Gansu Province

Objective:To understand the distribution characteristics of the cases and the epidemic law of pathogens of Rash and Fever Syndrome?RFS?in Gansu province,identify high-risk groups and core pathogens,establish a prediction model of the positive rate of core pathogens,and provide scientific evidence for RFS monitoring and early warning and control of related diseases in Gansu province.Methods:Descriptive and predictive research methods were used to conduct statistical description of rash and fever syndrome monitoring cases in Gansu province from May 2010 to October 2018,analyze pathogen detection and related components,and reveal the epidemic characteristics of pathogens.The SARIMA model was constructed based on the prevalence of core pathogens.On this basis,combined with GRNN and NAR neural network models,SARIMA-GRNN and SARIMA-NARNN intelligent combination models were established.MSE,MAE and MPAE values of each model were compared to evaluate the fitting effect and determine the optimal prediction model of RFS core pathogen positive rate in Gansu province.Results:1.From May 2010 to October 2018,a total of 4,720 RFS cases were monitored in Gansu province 4,213 were laboratory tested,of which 2,189 were positive,with a positive rate of51.96%;Except for 2015?82.29%?,the positive rates of other years fluctuated in the range of 40%⁵5%.2.The positive cases was more in men than in women,with no statistically significant difference in the positive rate?P=0.201?.The proportion of children in diaspora and childcare was42.44%and 28.04%respectively,and that of children under 5 years old was 70.26%;there was a significant difference in the positive rate among the age groups?P<0.001?;The seasonal prevalence of case infection was"bimodal",with high incidence in May to June and September to October.The positive rate was different in each region?P<0.001?.3.Among the 9,219 samples,2,206 were positive,of which 2,205 were viral pathogens.The sequence of pathogen spectrum was entero virus?57.10%?,measles virus?20.45%?,varicella-zoster virus?17.87%?,rubella virus?4.40%?,human parvovirus B19?0.13%?and dengue virus?0.05%?;There was significant difference in the positive rate of pathogens in different years?P<0.001?,and the trend of periodic change was obvious,and peaked in 2011?43.79%?and 2015?35.51%?.4.Except for entero virus and measles virus,the positive rate of other viruses showed no significant difference between genders?P>0.05?.Entero virus infection was predominant in the 0^,1^,and 3^years old groups,with the proportion of 56.88%,86.88%and 79.46%,respectively,and the epidemic peak was from May to July and October;Varicella-zoster virus infection mainly occurred in the 5^years old?48.35%?,15^years old?43.12%?and the 55+years old group?74.07%?,with epidemic peak in April to June and November to January of the next year.Measles virus infection was mainly from 35 to 54 years old?56.58%?,and the epidemic peak was from February to April.Rubella virus was high in 5^years old group?13.81%?,epidemic season was from may to August;The composition of positive pathogens was different in different regions.5.The type of pathogen mixed infection was virus mixed infection,with a total of 13 cases,and the infection rate was 0.61%,the main infection mode was measles and rubella virus mixed infection,with a total of 10 cases?76.92%?.In the distribution of subtypes of entero viruses,436cases?47.65%?of EV71 were detected,mainly in the 0³ years old group;406 cases?44.37%?of Cox A16 and mainly in the 5^years old and 15^years old groups;71 cases?7.76%?of other types and mainly in the 35^years old group.The dominant subtypes Cox A16 and EV71 occurred alternately in each year,with Cox A16 dominating in 2010,2013 and 2014,and EV71 dominating in 2011,2015,2017 and 2018.In 2013,2014 and 2015,the proportion of other types was 18.87%,38.60%and 15.29%,respectively.6.The MAE of the SARIMA?2,0,2??0,1,0?₁₂2 and SARIMA-GRNN and SARIMA-NARNN prediction models constructed based on the positive rate of entero virus in the fitting stage was0.111,0.074,0.095,MSE was 0.021,0.010 and 0.016,while MAPE was 36.713%,26.264%and30.648%,respectively.In the verification phase,MAE was 0.187,0.139 and 0.170,MSE was0.041,0.024 and 0.033,and MAPE was 43.736%,34.960%and 42.726%,respectively.The SARIMA-GRNN model was determined as the optimal prediction model by the minimum principle of evaluation index,The predicted results showed that the overall trend of entero virus positive rate from November 2018 to December 2019 was stable,and the epidemic peak was in August?78.57%?and October?78.26%?.Conclusions:1.In Gansu province,RFS positive cases were mainly in children under 5years old,and children in diaspora and kindergartens were the key infected population.2.The overall positive rate of pathogen was at a low level.The top three components of pathogen spectrum were entero virus,measles virus and varicella-zoster virus.EV71 and Cox A16were the main subtypes of entero virus.3.The three models of SARIMA,SARIMA-GRNN and SARIMA-NARNN could be used to predict the positive rate of entero virus,and the combined prediction effect was better than that of the single prediction.Compared with other models,SARIMA-GRNN model had higher prediction accuracy and more stable performance,and could be used as the best predictive model for the positive rate of entero virus in Gansu,currently.