Success/Failure Prediction of Noninvasive Mechanical Ventilation in Intensive Care Units^[*]

Summary

Objectives: This paper addresses the problem of decision-making in relation to the administration of noninvasive mechanical ventila tion (NIMV) in intensive care units.

Methods: Data mining methods were employed to find out the factors influencing the success/failure of NIMV and to predict its results in future patients. These artificial intelligence-based methods have not been applied in this field in spite of the good results obtained in other medical areas.

Results: Feature selection methods provided the most influential variables in the success/ failure of NIMV, such as NIMV hours, PaCO₂ at the start, PaO₂ / FiO₂ ratio at the start, hematocrit at the start or PaO₂ / FiO₂ ratio after two hours. These methods were also used in the preprocessing step with the aim of improving the results of the classifiers. The algorithms provided the best results when the dataset used as input was the one containing the attributes selected with the CFS method. Conclusions: Data mining methods can be successfully applied to determine the most influential factors in the success/failure of NIMV and also to predict NIMV results in future patients. The results provided by classifiers can be improved by preprocessing the data with feature selection techniques.

Keywords

Noninvasive ventilation - respiration disorders - respiratory insufficiency - data mining - feature selection methods - classifiers - multi-classifiers

• References

• 1 Peek N, Swift S. Intelligent Data Analysis for Knowledge Discovery, Patient Monitoring and Quality Assessment. Methods Inf Med 2012; 51 (Suppl. 04) 318-322.
• 2 Wróbel L, Sikora M. Censoring Weighted Separate-and-Conquer Rule Induction from Survival Data. Methods Inf Med 2014; 53 (Suppl. 02) 137-148.
• 3 Paoin W. Lesson Learned from Data Mining of WHO Mortality Database. Methods Inf Med 2011; 50 (Suppl. 04) 380-385.
• 4 Bellazzi R, Diomidous M, Sarkar IN, Takabayashi K, Ziegler A, McCray AT. Data Analysis and Data Mining: Current Issues in Biomedical Informatics. Methods Inf Med, 2011; 50 (Suppl. 04) 536-544.
• 5 Spreckelsen C, Spitzer K, Honekamp W. Present Situation and Prospect of Medical Knowledge Based Systems in German-speaking Countries. Methods Inf Med 2012; 51 (Suppl. 04) 536-544.
• 6 Beuscart R. Know Thyself. Methods Inf Med 2013; 52 (Suppl. 06) 536-537.
• 7 Tobin MJ. Mechanical ventilation. N Engl J Med 1994; 330 (Suppl. 15) 1056-1061.
• 8 Tobin MJ. editor. Principles & Practice of Mechanical Ventilation. 2nd ed. McGraw-Hill; 2006
• 9 Carrillo A, Gonzalez-Diaz G, Ferrer M, Martinez-Quintana ME, Lopez-Martinez A, Llamas N. et al. Non-Invasive ventilation in community-acquired pneumonia and severe acute respiratory failure. Intensive Care Med 2012; 38 (Suppl. 03) 458-466.
• 10 Nava S, Hill NS. Non-Invasive ventilation in acute respiratory failure. The Lancet 2009; 374 9685 250-259.
• 11 Nava S, Ceriana P. Causes of Failure of Noninvasive Mechanical Ventilation. Respiratory Care 2004; 49 (Suppl. 03) 2095-2303.
• 12 Lightowler JV, Wedzicha JA, Elliott MW, Ram FS. Non-Invasive positive pressure ventilation to treat respiratory failure resulting from exacerbations of chronic obstructive pulmonary disease: Cochrane systematic review and meta-analysis. BMJ 2003; 326 7382 185.
• 13 Shirakabe A, Hata N, Yokoyama S, Shinada T, Kobayashi N, Tomita K, Kitamura M, Nozaki A, Tokuyama H, Asai K, Mizuno K. Predicting the success of noninvasive positive pressure ventilation in emergency room for patients with acute heart failure. Journal of Cardiology 2011; 57: 107-114.
• 14 Martí S, Pallero M, Ferrer J, Ríos J, Rodríguez E, Morell F, Muñoz X. Predictors of mortality in chest wall disease treated with noninvasive home mechanical ventilation. Respiratory Medicine 2010; 104: 1843-1849.
• 15 Babuskaa R, Alica L, Lourensb MS, Verbraak AFM, Bogaard J. Estimation of respiratory parameters via fuzzy clustering. Artificial Intelligence in Medicine 2001; 21: 91-105.
• 16 Li Q, Lin N. The Bayesian Elastic Net. Bayesian Analysis 2010; 5 (Suppl. 01) 151-170.
• 17 Montero MA, Ruíz R, García-Torres M, Sarro LM. Feature Selection Applied to Data from the Sloan Digital Sky Survey. Trends in Applied Intelligent Systems. Lecture Notes in Computer Science Volume 6096. 2010 pp 611-620.
• 18 Hall MA. Correlation-based Feature Selection for Machine Learning. PhD Thesis, University of Waikato Hamilton, New Zealand: 1999
• 19 Hall M, Frank E, Holmes G, Pfahringer B, Reutemann P, Witten IH. The WEKA Data Mining Software: An Update. SIGKDD Explorations 2009; 11 (Suppl. 01) 10-18.
• 20 Quinlan JR. C4.5: Programs for Machine Learning. San Mateo, CA, USA: Morgan Kaufmann; 1993
• 21 Cooper GF, Herskovits E. A Bayesian Method for the induction of probabilistic networks from data. Machine Learning 1992; 9: 309-347.
• 22 Aha D, Kibler D. Instance-based learning algorithms. Machine Learning 1991; 6: 37-66.
• 23 Breiman L. Bagging predictors. Machine Learning 1996; 24 (Suppl. 02) 123-140.
• 24 Freund Y, Schapire RE. Experiments with a new boosting algorithm. Proceedings 13th International Conference on Machine Learning. 1996 pp 148-156.
• 25 Breiman L. Random Forests. Machine Learning 2001; 45 (Suppl. 01) 5-32.
• 26 Wolpert DH. Stacked Generalization. Neural Networks 1992; 5: 241-259.
• 27 Gama J, Brazdil P. Cascade Generalization. Machine Learning 2000; 41 (Suppl. 03) 315-343.
• 28 Honrubia T, García López FJ, Franco N, Mas M, Guevara M, Daguerre M. et al. Noninvasive vs conventional mechanical ventilation in acute respiratory failure: A multicenter, randomized controlled trial. Chest 2005; 128 (Suppl. 06) 3916-3924.
• 29 Schettino G, Altobelli N, Kacmarek RM. Noninvasive positive-pressure ventilation in acute respiratory failure outside clinical trials: Experience at the Massachusetts General Hospital. Crit Care Med 2008; 36 (Suppl. 02) 441-447.
• 30 Bower KM. The Paired t-Test. Scientific Computing and Instrumentation 2001; 18 (Suppl. 03) 62-63.
• 31 Guyon I, Elisseeff A. An Introduction to Variable and Feature Selection. Journal of Machine Learning Research 2003; 3: 1157-1182.