The Role Of Computer-assisted Interventions In The Prevention And Control Of Venous Thromboembolism(VTE)in Hospitalized Patients

Background: Venous thromboembolism(VTE)is a disease of venous return obstruction,which is the abnormal clotting of venous blood,resulting in complete or incomplete obstruction of blood vessels,including deep vein thrombosis(DVT)and pulmonary thromboembolism(PE),which are two clinical manifestations of the same disease in different stages and locations.Among cardiovascular diseases,the incidence of VTE has just risen to the third place after acute myocardial infarction and cerebral infarction,and the mortality rate of hospital-associated VTE is much higher than that of hospital-acquired infections.However,the proportion of hospitalized patients receiving prophylaxis in clinical practice is very low,mainly due to the lack of awareness of VTE risk assessment and concern about the risk of bleeding due to pharmacological prophylaxis,which leads to a worrying situation of low risk assessment rate,low correct assessment rate,and low implementation rate of prophylaxis.Therefore,accurate assessment of VTE risk in hospitalized patients is the key to improving the standardization of VTE prevention and reducing the occurrence of VTE.Computerized intelligent assessment intervention for VTE prevention and control,i.e.,VTE risk assessment of hospitalized patients using artificial intelligencebased clinical-assisted decision support system(AI-CDSS),is expected to improve the current situation of inaccurate VTE risk assessment and less standardized prevention and control in clinical practice.Objectives: To validate the efficacy of computer-assisted intelligent assessment for risk assessment of hospitalized patients.To investigate the impact of using AI-CDSS to assist physicians in VTE risk assessment on the quality of VTE prevention and control and the quality of physician evaluation;to compare the differences between AI-CDSS and physician-performed VTE risk assessment,and to explore the possibility of AICDSS replacing physician-performed VTE risk assessment.Methods: In the first part of the study,a total of 786 patients in Respiratory Medicine Zone 3 in January-February 2021 and all patients in Respiratory Zone 5 in January 2021 were selected using a whole cohort sampling method,and 764 patients in Chest Zone 1in January-February 2021,for a total of 1550 patients,and the AI-CDSS and quality control staff independently assessed VTE risk for all study subjects to evaluate the efficacy of the AI-CDSS.In the second part,a two-way cohort study,a total of 740 medical patients in respiratory area III and respiratory area V in July-August 2020 and784 surgical patients in thoracic area I in July-August 2020,totaling 1524,were selected as the control group by using the whole-group sampling method;a total of 1524 patients in respiratory area III,respiratory area V in July-August 2022 and A total of 685 all medical patients in Respiratory Zone 5 in July 2022 and 719 all surgical patients in Chest Surgery Zone 1 in July-August 2022,a total of 1404 as the control group,to explore the differences in prevention and control quality and physician evaluation quality between the two groups by comparing the quality of VTE prevention and control and physician evaluation quality before and after the AI-CDSS-assisted risk assessment intervention;comparing the AI-CDSS risk assessment results with The results of AICDSS risk assessment were compared with those of physician risk assessment to explore the possibility of AI-CDSS replacing physician risk assessment for VTE.Results:1.The accuracy of AI-CDSS in identifying VTE risk factors in internal medicine patients was 95.58%,the specificity was 97.46%,the sensitivity was 83.80%,the positive predictive value was 83.98%,the negative predictive value was 97.42%,and the results of AI-CDSS in identifying VTE risk factors were in good agreement with the evaluation results of quality control(Kappa value 0.803 p < 0.01).2,The accuracy of AI-CDSS in identifying VTE risk factors in surgical patients was94.70%,the specificity was 96.36%,the sensitivity was 90.17%,the positive predictive value was 90.08%,and the negative predictive value was 96.40%,and the results of AICDSS in identifying VTE risk factors were in good agreement with those of quality control(Kappa value 0.865 p < 0.01).3,After AI-CDSS intervention,the VTE risk assessment rate was significantly improved in medical patients(17.16% vs 100%),all high-risk patients were assessed for bleeding risk in both groups,and the proportion of high and high bleeding risk was statistically different between the two groups(p < 0.01);in surgical patients all inpatients were assessed for VTE risk in both groups,and the VTE high There was a statistical difference between the two groups in the risk of VTE high,medium and low risk(p < 0.01),but there was no statistical difference between the two groups in the proportion of high bleeding risk.4,All VTE prevention quality indicators in medical patients after AI-CDSS intervention were statistically different between the two groups before and after AI-CDSS intervention,and all prevention quality indicators decreased compared with postintervention;VTE prevention implementation rate in surgical patients was not statistically different between the two groups,all patients took preventive measures,physical prevention implementation rate was not statistically different between the two groups,and drug prevention implementation rate after intervention(There was no statistical difference between the two groups in the implementation rate of physical prophylaxis.5.There were no in-hospital VTEs,no adverse events due to prophylaxis implementation,and no hospital-related deaths due to VTEs during the study period.6.After the AI-CDSS intervention,the quality of VTE risk assessment performed by physicians in medical patients improved significantly,with the accuracy of assessment increasing from 89.86% to 94.31%,and the consistency with AI-CDSS enhanced;the quality of VTE risk assessment performed by physicians in surgical patients improved significantly,with the accuracy of assessment increasing from 85.19% to 90.14%,and the consistency with AI-CDSS enhanced.Conclusion1.computer-intelligent-assisted VTE risk assessment,or AI-CDSS,enables effective VTE risk assessment of medical and surgical inpatients;2.AI-CDSS enables comprehensive risk assessment and improves the quality of risk assessment3.AI-CDSS determines risk factors with better efficacy than front-line physicians;4.It is feasible to use AI-CDSS to replace physicians for risk assessment.