Abstract
The clinical assessment of a specific patient’s condition can be a very difficult process as multiple variables/risk factors may be involved. Thus, clinical guidelines frequently recommend the use of models that were developed with the objective of aiding the clinical decision. However, these models still present some significant flaws that must be overcome. On the other hand, recent clinical datasets that result from patients’ data gathered directly in the hospital or through telemonitoring systems are available.
The conjugation of this evidence, lead to different perspectives in the improvement of clinical decision: enhancement of the representation of clinical knowledge (current clinical models); extraction of new useful knowledge from recent clinical datasets; flexible combination of these two elements. This paper presents some achievements in relation to the improvement of current clinical models as well as to the extraction of knowledge from recent clinical datasets. In relation to the first issue an approach based on decision trees complemented with an optimization procedure to adjust the respective decision thresholds was applied while the latter is based on clustering theory in order to derive simple and interpretable rules.
This work is validated in the context of cardiovascular disease namely with coronary artery disease patients, assessing the risk of death 30 days after the admission. The largest Portuguese coronary artery disease patients dataset (13902 patients with acute coronary syndrome), provided by the Portuguese Society of Cardiology is used for validation purposes.
Some preliminary results were achieved, showing the potential of the proposed strategies to aid the clinical decision. This is an ongoing research with several possible research paths that are being pursued.
On behalf of the Investigators of the National Registry on Acute Coronary Syndromes, Portuguese Society of Cardiology
Preview
Unable to display preview. Download preview PDF.
References
1. Berner E. et al. (2009) Clinical Decision Support Systems: State of the, Agency for Healthcare Research and Environment, nº 09-0069-EF, 2009.
2. Perk J. et al., European guidelines on cardiovascular disease prevention in clinical practice (2012). European Heart Journal 33 (13); 1635–701, 2012.
3. Bellazi R. et. al., Data Analysis and Data Mining: Current Issues in Biomedical Informatics, Methods of Information in Medicine, doi.org/10.3414/ME11-06-0002, 2011.
4. Mendes D. et al., “Improving Clinical Models based on Knowledge Extracted from Current Datasets: a new Approach”, EMBC, 2016.
5. Hammouda K. et al. A Comparative Study of Data Clustering-Techniques. doi:10.1.1.126.3224, University of Waterloo, 2000.
6. Kohavi R. et al., Wrappers for feature subset selection, Artificial intelligence, vol. 97, no. 1, pp. 273–324, 1997.
7. Goncalves P. et al., TIMI, PURSUIT, and GRACE risk scores: Sustained prognostic value and interaction with revascularization in NSTE-ACS., European Heart Journal, vol. 26, pp. 865–872, May 2005.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Mendes, D., Paredes, S., Rocha, T., Henriques, J., Carvalho, P., Morais, J. (2018). Knowledge and Data Driven Approaches Applied to Clinical Assessment. In: Eskola, H., Väisänen, O., Viik, J., Hyttinen, J. (eds) EMBEC & NBC 2017. EMBEC NBC 2017 2017. IFMBE Proceedings, vol 65. Springer, Singapore. https://doi.org/10.1007/978-981-10-5122-7_34
Download citation
DOI: https://doi.org/10.1007/978-981-10-5122-7_34
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-10-5121-0
Online ISBN: 978-981-10-5122-7
eBook Packages: EngineeringEngineering (R0)