Study On Influenza Surveillance Based On Hospital Respiratory Syndrome Data

ObjectiveThis study,taking the second affiliated hospital of Nanjing medical university as an example,analyzed the diagnosis and treatment information data of The health information platform of Jiangsu province(passed by the hospital via the front machine)to understand its data characteristics;In this paper,the case data of respiratory syndrome was analyzed to evaluate its effect on the outbreak and early warning of influenza(influenza).Thus explore the utilization value and utilization mode of medical and health data information.Data and Methods1 The data sourceThe clinical diagnosis and treatment information of the second affiliated hospital of Nanjing medical university is from The health information platform of Jiangsu province;The data on the monitoring of influenza-like and influenza cases in Nanjing are all from National Disease Supervision Information Management System.2 Data sortingThe inclusion and exclusion criteria were established according to the case definition of respiratory symptoms.According to this standard,the clinical diagnosis and treatment data of the second affiliated hospital of Nanjing Medical University were selected to obtain the information of respiratory syndrome,and eventually forming respiratory syndrome cases information database.Preliminary classification and summarization were taken to respiratory syndrome,influenza and influenza-like cases,to form a statistical chart for easy analysis and utilization.3 Analysis methodsAccording to the arrangement form of statistical data,diagrams and statistics to describe the distribution of respiratory syndrome cases,influenza-like cases and influenza cases,such as the distribution of clinic time,department,etc.;Using SPSS 17.0 software to develop normality test of respiratory symptoms,flu cases and influenza-like cases time distribution sequence.The correlation between respiratory symptoms,influenza cases and influenza-like cases was described by means of dislocation correlation.(Pearson correlation analyze normal distribution data,Spearman correlation analyze abnormal distribution data);Finally,use the time series analysis method of ARIMA,build a model in week,to predict the influenza and of influenza-like cases and other cases of respiratory symptoms,and verify the predictive effect.Results 1 Distribution of respiratory syndromeRespiratory tract infection,tracheitis and pneumonia are the first three contributory cases of respiratory symptoms,who mainly visited pediatrics,emergency department and internal medicine.2 The correlation between respiratory symptoms and influenza and influenza-like cases.The correlation between the incidence of influenza cases and the incidence of influenza in 4 weeks was the best,with the correlation coefficient of 0.598(p < 0.01).The correlation between respiratory tract infection and the time distribution of influenza samples in the same week was the best,with correlation coefficient of 0.342(p < 0.01).The incidence of respiratory syndrome was not well correlated with the time distribution of influenza cases in Nanjing.The correlation between respiratory infection and lung infection was good.Among them,the correlation between respiratory tract infection and flu case 6 weeks in advance was the best,with correlation coefficient of 0.369(p < 0.01);The correlation between pulmonary infection and influenza cases 1 week in advance was the best,with correlation coefficient of 0.357(p < 0.01).3 ARIMA modelThe ARIMA model was established by using the weekly statistical data of "respiratory tract infection","influenza" and "influenza-like case" of 98 weeks from the 1st week of 2014 to the 46 th week of 2015.The best models of them were ARIMA(1,1,1),ARIMA(1,0,0)and ARIMA(1,0,0).Using the data from week 47 to 53 of 2015,the actual value is within the 95% confidence interval of the predicted value,and the coincidence rate is 100%.Among them,the short-term prediction accuracy of "flu-like case" was good,and the errors in the 47 th and 48 th week were 1.31% and 5.17% respectively.Conclusions1?Respiratory symptoms cases of the second affiliated hospital of Nanjing medical university are mainly in pediatric department,emergency department and medical treatment.So the above departments should be the focal monitoring work of the respiratory symptoms.2?Respiratory symptoms can prompt flu outbreaks or epidemics several weeks in advance,with respiratory infections and influenza-like cases showing better results.3?ARIMA model can be used to predict influenza cases,influenza-like cases,respiratory tract infections and other respiratory symptoms.