Abstract Number: PB2290
Meeting: ISTH 2020 Congress
Theme: Venous Thromboembolism and Cardioembolism » VTE Diagnosis
Background: Statistical tools can be useful to determine the main factors of venous thromboembolism recurrence (VTR) since it extracts patterns among selected factors.
Aims: This work aimed to use techniques of fractional factorial design to obtain the main predictive factors of VTR, using the neural model for prediction of VTR, developed by Martins [Unicamp, Thesis (2018)].
Methods: Thirty-nine clinical and laboratory factors were considered. We considered 235 patients with a history of pulmonary, lower limbs and central nervous system thrombosis. Fractional factorial design techniques of five types were used. The statistical significance of the main factors was determined using t-student test. All calculations were performed using the neural model and the Statistica software.
Results: For all designs, the leukocyte level was indicated as statistical significant, with negative effect. The higher it’s level, the lower the risk of a thrombotic recurrence episode. Also were founded as significant factors: tabagism and hormone therapy, with positive effects. A total of fifteen factors were found as statistical significant: time of use of anticoagulants, age, residual thrombus, proximal or distal thrombus, diabetes mellitus, renal, protein S, red blood cells, C-reactive protein, dyslipidemia, hemoglobin and platelets (and also the first three factors). In this study it was found that the longer the anticoagulation time, the greater the chance of recurrence of thrombosis. The literature shows that this effect is opposite. This result may be due to the fact that, the greater the chance of a recurrence, the longer the anticoagulation time to avoid it. This fact is inherent in the collected data.
Conclusions: The results of this work showed that it’s possible to determine the main factors of VTR using a neural model, combined with a design of experiments technique. These factors can be used in the formulation of new (and more accurate) predictive models/scores.
To cite this abstract in AMA style:
Ottaiano GY, Annichino-Bizzacchi JM, Maciel-Filho R, Martins TD. Determining Predictive Factors of Venous Thromboembolism Using Design of Experiments [abstract]. Res Pract Thromb Haemost. 2020; 4 (Suppl 1). https://abstracts.isth.org/abstract/determining-predictive-factors-of-venous-thromboembolism-using-design-of-experiments/. Accessed October 2, 2023.« Back to ISTH 2020 Congress
ISTH Congress Abstracts - https://abstracts.isth.org/abstract/determining-predictive-factors-of-venous-thromboembolism-using-design-of-experiments/