Construction Of Risk Assessment Model Of Upper Venous Thrombosis With Peripherally Inserted Central Catheters Among Patients

Objective:1.Through the Meta-analysis to identify risk factors for upper venous thrombosis with peripherally inserted central catheters.2.Based on the risk factors for upper venous thrombosis with peripherally inserted central catheters which are identified by Meta-analysis,we construct the risk assessment model of venous thrombosis with peripherally inserted central catheters,providing the theoretical basis to clear high-risk patients and prevent the occurrence of thrombosis.Methods:1.We searched domestic and abroad databases such as CNKI(China National Knowledge Infrastructure),Wan fang data,Pubmed,Embase,Science Direct and Cochrane Library to collect the document published of upper venous thrombosis with peripherally inserted central catheters' s influence factors.Then we used the Revman 5.3 software provided by Cochrane Collaboration to analyze the data which was extracted by literature,we used odds ratio and 95% confidence interval to express the effect size of classification data.2.We analyzed the cases of 146 patients with PICC who were hospitalized between July 2015 and February 2016 in a three-level of first-class hospital in Shanxi Province retrospectively,and divided into thrombosis group(n=12),non-thrombosis group(n=134).the occurrence of upper venous thrombosis with peripherally inserted central catheters as the dependent variable;obesity,cancer,history of thrombosis,history of CVC /PICC insertion,puncture times more than two,arm swelling of tube side limb,long-term bedridden(>72h)as independent variables to establish logistic regression model.Results:1.Based on the process of literature screening and criterion,10 articles were chosen for Meta-analysis.Including 5 cohort studies and 5 case-control studies.The total of casesare 3471 patients,including thrombosis group and non-thrombosis group were 662 and2809 respectively.The seven factors of the merged result of meta analysis have statistical significance,they were obesity(OR=2.20,95%CI(1.41,3.44));cancer(OR=1.68,95%CI(1.12,2.53));history of thrombosis(OR=1.77,95%CI(1.36,2.32));history of CVC/PICC insertion(OR=3.83,95%CI(1.51,9.66));puncture times more than two(OR=1.77,95%CI(1.09,2.87));arm swelling of tube side limb(OR=15.44,95%CI(6.44,37.05));long-term bedridden(>72h)(OR=9.44,95%CI(3.87,23.02)).These factors were risk factors for upper venous thrombosis with peripherally inserted central catheters.2.We constructed the logistic regression model according to the result of logistic regression analysis,we got the logistic regression model as following: Logit(P)=-1.681+2.184 × obesity + 2.283 × cancer + 2.086 × history of thrombosis + 2.177 × history of CVC /PICC insertion+ 3.518 × puncture times more than two + 3.027 × arm swelling of tube side + 1.153 × long-term bedridden(>72h).The prediction accuracy was 95.9% of the model,the area under the receiver operating characteristic curve is 0.908,the prediction ability of the model to the occurrence of upper venous thrombosis with peripherally inserted central catheters was good in statistical meaning.And we got the best threshold is4.632,sensitively and specificity were 81.2% and 75.5% respectively.Conclusion:The risk assessment model we have constructed were convenient and practical,nurses can use it to evaluate patients with peripherally inserted central catheter individually,clear high-risk patients to prevent the occurrence of thrombosis.