Construction And Applications Of Standardized Network Information Platform For Acute Pancreatitis

Diagnosis and treatment of acute pancreatitis(AP)is facing many bottlenecks,such as lack of multi-disciplinary collaboration(MDT)concept for treatment;difficulty in timely access to rapidly developing new technologies in remote areas;large amounts of data generated by diagnosis and treatment affect the accuracy of diagnosis and treatment;lack of information follow-up for patients and so on.From the macro policy point of view,the state also emphasizes the necessity of information construction,so it is urgent to establish a platform with wide coverage and embedded multiple functional modules.However,at present,the construction of information technology for AP diagnosis and treatment at home and abroad is obviously inadequate.The purpose of this study is to establish AP multi-disciplinary diagnosis and treatment information platform(APnet)based on the Internet,and to develop remote consultation,online education,AP shared database and assistant diagnosis and treatment decision-making system,so as to provide a new model for AP clinical diagnosis and treatment guidance,give full play to the role of high-quality medical resources,and then improve the quality and efficiency of diagnosis and treatment.First,we built APnet.The system application layer adopts three-tier architecture design of C/S/S system,MVC hierarchical framework model,componentized algorithm,multi-tier fine-grained authority control,and embedded data interface settings.Gigabit Ethernet and other technologies are used in the supporting platform;dual cluster system with hot backup is used in the host platform;IP-SAN storage system is used in the storage platform;virtual disk/NDMP and LAN mode are used in the backup platform;and network firewall/PPP protocol/static routing system is used in the security platform.APnet is deployed in star structure,centered on the pre-switching server,and has a variety of adapters.The application development framework is oriented to PC terminal,APP terminal and micro-messaging terminal.Audio,graphic and text transmission uses WebRTC technology to achieve multi-point bidirectional transmission.APnet provides remote consultation,two-way referral,distance education,remote ward rounds,academic exchanges and other functions.APnet has 23015 public fans and 12074 registered users of APP,covering 3140 hospitals.Based on APnet platform,more than 220 teleconsultation and two-way referral have been carried out,most of which are critically ill patients.It has realized online education functions including remote training,remote difficult case discussion and remote ward rounds.It has carried out 212 courses,covering more than 100,000 medical staff.Secondly,establish AP business database.Importing application data according to data,format and protocol,completing metadata classification,maintenance and expansion,standardizing dictionary content,formulating database table structure according to variables,implementing definition and association,carrying out data storage structure,index strategy,storage allocation,programming,data loading,establishing standard database,and delivering and delivering test in APnet.The database includes nine sub-databases,including basic information database,symptom and sign database,diagnosis and evaluation database.Married patients accounted for 75.4%;patients with type A,B,O and AB accounted for 22.4%,33.4%,13.5%and 30.7%respectively;patients with mild/moderate AP(MAP/MSAP)and severe AP(SAP)accounted for 85.8%and 14.2%respectively.Biliary pancreatitis,alcoholic pancreatitis and hyperlipidemic pancreatitis accounted for 86.7%,7.7%and 3.7%respectively.Nanjing,Suzhou and Yangzhou ranked the top three in the number of patients.6%of MAP/MSAP patients were unmarried,which was lower than 16%of SAP patients.Compared with ordinary residents,the proportion of AB blood group patients was significantly higher(13.5%vs 8.8%).SAP was more significant(18.7%).Cluster analysis showed that the three attributes of AP,including etiology,severity and ABO blood group,played a key role in the clustering process.Samples are divided into five categories by self-organizing mapping network,each with its own epidemiological and clinical characteristics.Thirdly,the data of AP diagnosis and treatment are deeply excavated,and an assistant diagnosis and treatment decision-making system is established.We used BP network,RBF network and logistic regression to predict the occurrence of AP with portal vein thrombosis(PVT).After pretreatment,353 cases were randomly divided into training group,validation group and test group according to the ratio of 5:1:1.Age,sex,hematocrit(Hct),prothrombin time,fasting blood glucose,D-dimer,serum calcium concentration,triglyceride and serum amylase(AMY)were included in the input layer.Programming prediction is carried out by using the toolbox of MATLAB2016b.BP network model test showed that the sensitivity(SEN)was 80.0%,the specificity(SPE)was 85.7%,and the accuracy was 83.3%.RBF network test results were SEN 78.2%,SPE 92.0%,accuracy 88.1%.Logistic regression model showed that SEN was 65.3%,SPE was 83.6%,and accuracy was 75.3%.The results of three models show that the RBF network is the most accurate and the area under ROC curve is the largest,followed by BP network,and the logistic regression is the lowest.There were significant differences among the three groups(P<0.05).The validity of neural network model is higher than that of regression analysis model,and RBF network is better than BP network.AMY,D-dimer and Hct are important factors affecting AP-associated PVT,and their relative importance are 100%,92.3%and 68.9%respectively,Based on the established model algorithm,we developed an assistant diagnosis and treatment decision-making system,which was pushed and implemented on APnet.In summary,the APnet,which includes remote consultation,online education,AP database and auxiliary decision-making system function modules,provides a new way of thinking,theoretical and practical basis for AP multi-disciplinary standardized diagnosis and treatment,and provides a strong support for the sharing and accessibility of high-quality medical resources.