Dataset Information

Prediction of central venous catheter-associated deep venous thrombosis in pediatric critical care settings.

ABSTRACT:

Background

An increase in the incidence of central venous catheter (CVC)-associated deep venous thrombosis (CADVT) has been reported in pediatric patients over the past decade. At the same time, current screening guidelines for venous thromboembolism risk have low sensitivity for CADVT in hospitalized children. This study utilized a multimodal deep learning model to predict CADVT before it occurs.

Methods

Children who were admitted to intensive care units (ICUs) between December 2015 and December 2018 and with CVC placement at least 3 days were included. The variables analyzed included demographic characteristics, clinical conditions, laboratory test results, vital signs and medications. A multimodal deep learning (MMDL) model that can handle temporal data using long short-term memory (LSTM) and gated recurrent units (GRUs) was proposed for this prediction task. Four benchmark machine learning models, logistic regression (LR), random forest (RF), gradient boosting decision tree (GBDT) and a published cutting edge MMDL, were used to compare and evaluate the models with a fivefold cross-validation approach. Accuracy, recall, area under the ROC curve (AUC), and average precision (AP) were used to evaluate the discrimination of each model at three time points (24 h, 48 h and 72 h) before CADVT occurred. Brier score and Spiegelhalter's z test were used measure the calibration of these prediction models.

Results

A total of 1830 patients were included in this study, and approximately 15% developed CADVT. In the CADVT prediction task, the model proposed in this paper significantly outperforms both traditional machine learning models and existing multimodal deep learning models at all 3 time points. It achieved 77% accuracy and 90% recall at 24 h before CADVT was discovered. It can be used to accurately predict the occurrence of CADVT 72 h in advance with an accuracy of greater than 75%, a recall of more than 87%, and an AUC value of 0.82.

Conclusion

In this study, a machine learning method was successfully established to predict CADVT in advance. These findings demonstrate that artificial intelligence (AI) could provide measures for thromboprophylaxis in a pediatric intensive care setting.

SUBMITTER: Li H

PROVIDER: S-EPMC8627017 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundAfter deep venous thrombosis (DVT), many patients have impaired quality of life (QOL). We aimed to assess whether pharmacomechanical catheter-directed thrombolysis (PCDT) improves short-term or long-term QOL in patients with proximal DVT and whether QOL is related to extent of DVT.MethodsThe Acute Venous Thrombosis: Thrombus Removal with Adjunctive Catheter-Directed Thrombolysis (ATTRACT) trial was an assessor-blinded randomized trial that compared PCDT with no PCDT in patients with DVT of the femoral, common femoral, or iliac veins. QOL was assessed at baseline and 1 month, 6 months, 12 months, 18 months, and 24 months using the Venous Insufficiency Epidemiological and Economic Study on Quality of Life/Symptoms (VEINES-QOL/Sym) disease-specific QOL measure and the 36-Item Short Form Health Survey (SF-36) physical component summary (PCS) and mental component summary general QOL measures. Change in QOL scores from baseline to assessment time were compared in the PCDT and no PCDT treatment groups overall and in the iliofemoral DVT and femoral-popliteal DVT subgroups.ResultsOf 692 ATTRACT patients, 691 were analyzed (mean age, 53 years; 62% male; 57% iliofemoral DVT). VEINES-QOL change scores were greater (ie, better) in PCDT vs no PCDT from baseline to 1 month (difference, 5.7; P = .0006) and from baseline to 6 months (5.1; P = .0029) but not for other intervals. SF-36 PCS change scores were greater in PCDT vs no PCDT from baseline to 1 month (difference, 2.4; P = .01) but not for other intervals. Among iliofemoral DVT patients, VEINES-QOL change scores from baseline to all assessments were greater in the PCDT vs no PCDT group; this was statistically significant in the intention-to-treat analysis at 1 month (difference, 10.0; P < .0001) and 6 months (8.8; P < .0001) and in the per-protocol analysis at 18 months (difference, 5.8; P = .0086) and 24 months (difference, 6.6; P = .0067). SF-36 PCS change scores were greater in PCDT vs no PCDT from baseline to 1 month (difference, 3.2; P = .0010) but not for other intervals. In contrast, in femoral-popliteal DVT patients, change scores from baseline to all assessments were similar in the PCDT and no PCDT groups.ConclusionsAmong patients with proximal DVT, PCDT leads to greater improvement in disease-specific QOL than no PCDT at 1 month and 6 months but not later. In patients with iliofemoral DVT, PCDT led to greater improvement in disease-specific QOL during 24 months.