ABSTRACT
Prediction of severe exacerbation triggered by uncontrolled asthma is highly important for patients suffering of asthma, as avoiding maleficent symptoms that could need special treatment or even hospitalization, can protect patients from the aftereffects of bronchostilation. In this study 4 machine learning algorithms are assessed with respect to their accuracy in predicting severe asthma exacerbations using various parameters. The approach proposed for personalized decision support and patient guidance considers two classification tasks: a 2-class prediction task, which allows for estimation of risk of exacerbation 7 days ahead, and a 4-class prediction task, which allows for monitoring the on-going exacerbation events with respect to their duration. The preliminary results concerning the 2-class prediction task reveals that daily spirometry and tracking of medication usage are of the most significant metrics on the analysis of the study group of patients who tackled severe exacerbation events during the on-going evaluation campaign. All tested algorithms demonstrate comparable high accuracy, with the SVM and the Bayesian Naive ones demonstrating better sensitivity, while the Random Forests demonstrates better negative prediction power (very high specificity).
- Eric D. Bateman, Roland Buhl, Paul M. O'Byrne, Marc Humbert, Helen K. Reddel, Malcolm R. Sears, Christine Jenkins, Tim W. Harrison, Santiago Quirce, Stefan Peterson, and Göran Eriksson. 2015. Development and validation of a novel risk score for asthma exacerbations: The risk score for exacerbations. Journal of Allergy and Clinical Immunology 135, 6 (2015), 1457--1464e4.Google ScholarCross Ref
- Joseph Finkelstein and In cheol Jeong. 2017. Machine learning approaches to personalize early prediction of asthma exacerbations. Annals of the New York Academy of Sciences 1387, 1 (2017), 153--165.Google ScholarCross Ref
- Anne M. Fitzpatrick, Benjamin M. Gaston, Serpil C. Erzurum, and W. Gerald Teague. 2006. Features of severe asthma in school-age children: Atopy and increased exhaled nitric oxide. Journal of Allergy and Clinical Immunology 118, 6 (2006), 1218--1225. arXiv:NIHMS150003Google ScholarCross Ref
- Global Initiative for Asthma. 2014. Global strategy for asthma management and prevention Âl' 2014. (2014), 148 pages. http://www.ginasthma.org/local/uploads/files/GINAGoogle Scholar
- J. B. Hagan, S. A. Samant, G. W. Volcheck, J. T. Li, C. R. Hagan, P. J. Erwin, and M. A. Rank. 2014. The risk of asthma exacerbation after reducing inhaled corticosteroids: a systematic review and meta-analysis of randomized controlled trials. Allergy 69, 4 (2014), 510--516.Google ScholarCross Ref
- Persijn J. Honkoop, Andrew Simpson, Matteo Bonini, Jiska B. Snoeck-Stroband, Sally Meah, Kian Fan Chung, Omar S. Usmani, Stephen Fowler, and Jacob K. Sont. 2017. MyAirCoach: The use of home-monitoring and mHealth systems to predict deterioration in asthma control and the occurrence of asthma exacerbations; Study protocol of an observational study. BMJ Open 7, 1 (2017).Google Scholar
- Zuleyha Kaya, Feyza Erkan, Mine Ozkan, Sedat Ozkan, Nazmiye Kocaman, Banu Aslantas Ertekin, and Nese Direk. 2009. Self-management plans for asthma control and predictors of patient compliance. Journal of Asthma 46, 3 (2009), 270--275.Google ScholarCross Ref
- Kirsten M. Kloepfer, Wai Ming Lee, Tressa E. Pappas, Theresa J. Kang, Rose F. Vrtis, Michael D. Evans, Ronald E. Gangnon, Yury A. Bochkov, Daniel J. Jackson, Robert F. Lemanske, and James E. Gern. 2014. Detection of pathogenic bacteria during rhinovirus infection is associated with increased respiratory symptoms and asthma exacerbations. Journal of Allergy and Clinical Immunology 133, 5 (2014), 1301--1307.e3. arXiv:NIHMS150003Google ScholarCross Ref
- Otilia Kocsis, Gerasimos Arvanitis, Aris Lalos, Konstantinos Moustakas, Jacob K. Sont, Persijn J. Honkoop, Kian Fan Chung, Matteo Bonini, Omar S. Usmani, Stephen Fowler, and Andrew Simpson. 2017. Assessing machine learning algorithms for self-management of asthma. 2017 E-Health and Bioengineering Conference, EHB 2017 (2017), 571--574.Google ScholarCross Ref
- Rik Jb Loymans, Thomas Pa Debray, Persijn J Honkoop, Evelien H Termeer, Jiska B Snoeck-Stroband, Tjard Rj Schermer, Willem Jj Assendelft, Merel Timp, Kian Fan Chung, Ana R Sousa, Jaap K Sont, Peter J Sterk, Helen K Reddel, and Gerben Ter Riet. 2018. Exacerbations in adults with asthma: A systematic review and external validation of prediction models. The journal of allergy and clinical immunology. In practice (2018).Google Scholar
- R. J. B. Loymans, A. Gemperli, J. Cohen, S. M. Rubinstein, P. J. Sterk, H. K. Reddel, P. Juni, and G. ter Riet. 2014. Comparative effectiveness of long term drug treatment strategies to prevent asthma exacerbations: network meta-analysis. Bmj 348, may13 3 (2014), g3009--g3009.Google ScholarCross Ref
- Gang Luo, Bryan L. Stone, Bernhard Fassl, Christopher G. Maloney, Per H. Gesteland, Sashidhar R. Yerram, and Flory L. Nkoy. 2015. Predicting asthma control deterioration in children. BMC Medical Informatics and Decision Making 15, 1 (2015), 1--8.Google ScholarCross Ref
- Matlab. {n. d.}. MATLAB: The Language of Technical Computing. ({n. d.}). http://www.mathworks.com/products/matlab/Google Scholar
- Sandeep Puranik, Erick Forno, Andrew Bush, and Juan C. Celedón. 2017. Predicting severe asthma exacerbations in children. American Journal of Respiratory and Critical Care Medicine 195, 7 (2017), 854--859.Google ScholarCross Ref
- R. {n. d.}. The R Project for Statistical Computing. ({n. d.}). https://www.r-project.org/Google Scholar
- Helen Reddel, Sandra Ware, Guy Marks, Cheryl Salome, Christine Jenkins, and AnnWoolcock. 1999. Differences between asthma exacerbations and poor asthma control. Lancet 353, 9150 (1999), 364--369.Google Scholar
- Daniel Sanchez-Morillo, Miguel A. Fernandez-Granero, and Antonio Leon-Jimenez. 2016. Use of predictive algorithms in-home monitoring of chronic obstructive pulmonary disease and asthma: A systematic review. Chronic Respiratory Disease 13, 3 (2016), 264--283.Google ScholarCross Ref
- Ryuji Sato, Katsuyuki Tomita, Hiroyuki Sano, Hideo Ichihashi, Shigeyoshi Yamagata, Akiko Sano, Toshiyuki Yamagata, Takayuki Miyara, Takashi Iwanaga, Masato Muraki, and Yuji Tohda. 2009. The Strategy for Predicting Future Exacerbation of Asthma Using a Combination of the Asthma Control Test and Lung Function Test. Journal of Asthma 46, 7 (2009), 677--682.Google ScholarCross Ref
- Andrew Skeggs, Tricia McKeever, Lelia Duley, Eleanor Mitchell, Lucy Bradshaw, Kevin Mortimer, Samantha Walker, Steve Parrott, Andrew Wilson, Ian Pavord, Chris Brightling, Mike Thomas, David Price, Graham Devereux, Bernard Higgins, Tim Harrison, and Rebecca Haydock. 2016. FourFold Asthma Study (FAST): A study protocol for a randomised controlled trial evaluating the clinical cost-effectiveness of temporarily quadrupling the dose of inhaled steroid to prevent asthma exacerbations. Trials 17, 1 (2016), 1--9.Google ScholarCross Ref
- Stephen J. Teach, Peter J. Gergen, Stanley J. Szefler, Herman E. Mitchell, Agustin Calatroni, Jeremy Wildfire, Gordon R. Bloomberg, Carolyn M. Kercsmar, Andrew H. Liu, Melanie M. Makhija, Elizabeth Matsui, Wayne Morgan, George O'Connor, and William W. Busse. 2015. Seasonal risk factors for asthma exacerbations among inner-city children. Journal of Allergy and Clinical Immunology 135, 6 (2015), 1465--1473e5.Google ScholarCross Ref
Index Terms
Exploring machine learning for monitoring and predicting severe asthma exacerbations
Recommendations
Investigating the Impacts of Weather and Personalisation on Asthma Exacerbations using Machine Learning
ICSCA '22: Proceedings of the 2022 11th International Conference on Software and Computer ApplicationsThis paper reports the results of investigating the impacts of weather and personalisation on asthma exacerbations. Additionally, several machine learning classification algorithms were applied and evaluated to develop an intelligent mHealth solution ...
Technological innovations in asthma patient monitoring and care
Recent studies have shown that long-term monitoring of asthma severity can reduce asthma exacerbations, optimise drug therapy and decrease the cost of asthma management [National Heart, Lung and Blood Institute, 1997]. The management of a chronic ...
Comments