Validation Of Risk Assessment Tools For Venous Thromboembolism In Medical Inpatients With Cancer

Objective:To estimate the incidence of venous thromboembolism(VTE)and its risk factors in medical inpatients with cancer,and provide evidence for targeted prevention of VTE;and to evaluate the validity of the Khorana score and Padua risk assessment model in identifying VTE risk patients among medical inpatients with cancer,and compare the validity between the Khorana score and Padua risk assessment model in predicting VTE risk,and provide evidence for VTE evaluation and prevention in medical inpatients with cancer,ultimately reduce incidence of VTE in medical inpatients with cancer.Methods(1)A total of 3730 medical records of patients who met the oncology department and met the inclusion and exclusion criteria were selected by convenient sampling from January 1,2015 to December 31,2017.We collected general patient information by self-made general data sheet,and assessed 3730 medical inpatients with cancer by Khorana score.(2)Based on the first part,we used the Padua risk assessment model to assess 3730 medical inpatients with cancer,and compared the validitybetween the Padua risk assessment model and Khorana score in identifying VTE risk patients among medical inpatients with cancer.(3)All data were analyzed by SPSS18.0.The patient general information,Khorana scores and Padua scores were calculated by descriptive statistics.Binary Logstic regression analysis was used to analyze the risk factors related to VTE and the risk stratification effect of the two risk assessment models.The prediction effect was analyzed by ROC curve,and the area under the ROC curve of the two models was compared using the non-parametric test of MedCalc.Results(1)The incidence of VTE in medical inpatients with cancer was 4.21%.Increasing age,indwelling PICC,bed rest,and trauma or surgery within the past month were all risk factors related to VTE(OR> 1,P <0.05).The risk of VTE in men was 0.650 times comparing that in women(P <0.05).The type of cancer was also an influencing factor of VTE,with gastric and pancreatic cancer patients having the highest risk of VTE.(2)The area under the ROC curve of the Khorana score in medical inpatients with cancer was 0.741 [95% CI: 0.698-0.786].The risk prediction ability of Khorana score was moderate,and its cut-off value was 2 points.For each score increase in Khorana score,the VTE risk became 2.457 times as much as before.The patients identified as high risk and moderate risk by the Khorana score had a VTE risk of 15.131 times and 2.145 times that oflow-risk patients,respectively.(3)The area under the ROC curve of the Padua risk assessment model in medical inpatients with cancer was 0.527 [95% CI: 0.489-0.585].The risk prediction ability of Padua risk assessment model was poor.The risk of VTE occurrence in patients identified as high-risk by the Padua risk assessment model was 1.336 times that of low-risk patients,but this VTE risk stratification was not statistically significant(P = 0.092).(4)The evaluation results of the Khorana and Padua risk assessment models for medical inpatients with cancer were inconsistent(Kappa coefficient was 0.025).The two areas under the ROC curve were obviously different(Z = 6.005,P < 0.01),and the area under the ROC curve of the Khorana score was larger,which indicated that Khorana score had a better predictive power.Conclusions(1)The incidence of VTE in medical inpatients with cancer was 4.21%,and the influencing factors were age increase,indwelling PICC,bed rest,trauma or surgery within the past month,women and the type of cancer.(2)The Khorana score can be used to predict the VTE risk of medical inpatients with cancer,and the predictive ability is medium.(3)Padua risk assessment model has a weaker VTE risk prediction capability for medical inpatients with cancer than Khorana score,and cannot be used for medical inpatients with cancer.