Monitoring Medical Interventions for Multidimensional Evaluation of Changes in Patient Test Results with Principal Component Analysis

Monitoring Medical Interventions for Multidimensional Evaluation of Changes in Patient Test Results with Principal Component Analysis

Medical doctors of today are challenged with increasingly large volumes of high-dimensional, heterogeneous, and unstructured data from various sources that pose significant challenges for manual analysis. However, this unstructured data is mainly vital for decision making but there exists a shortage of intelligent tools to extract the hidden knowledge. Given these facts, the application of machine learning methods in healthcare is a growing phenomenon. This paper explores machine learning approaches for interpreting large quantities of continuously acquired, multivariate patient-based medical laboratory data, in intensive care unit (ICU) settings. The research hypothesizes that principal component analysis (PCA) can be able to capture the changes in the outcomes after a medical intervention. We adopted PCA as the main method, to observe and capture the daily changes for intensive care unit patients. The approach will be able to inform the physicians, which laboratory tests are exhibiting variances after an intervention, and their associated epiphenomenon. This can be used as a clue to make decisions on which treatment or diagnosis to apply further. Experimental analysis results indicate that PCA was able to capture patient progression in terms of variances. Permutation tests for the validity and stability of the model exhibit an acceptable significance level with a p-value of 0.001. Results showed that the approach provides promising results for interpreting large quantities of patient data for establishing a cause-effect relationship from medical interventions and be used as an early warning system. The study retrospectively demonstrated the capability of PCA to monitor and provide an alert to the clinicians about the patient's changing conditions, thereby providing opportunities for timely interventions. If coupled with other machine learning models, the approach can also be able to support clinical decision making and enable effective patienttailored care for better health outcomes.

___

  • [1] W. Widanagamaachchi, Y. Livnat, P.-T. Bremer, S. Duvall and V. Pascucci, "Interactive Visualization and Exploration of Patient Progression in a Hospital Setting," AMIA Annual Symposium Proceeding, vol. v2017, pp. 1773-1782, 2017.
  • [2] C.-F. Wang, J. Li, K.-L. Ma, C.-W. Huang and Y.-C. Li, "A Visual Analysis Approach to Cohort Study of Electronic Patient Records," in IEEE International Conference on Bioinformatics and Biomedicine, Belfast, UK, 2014.
  • [3] C. Mathukia, W. Fan, K. Vadyak, C. Biege and M. Krishnamurthy, "Modified Early Warning System improves patient safety and clinical outcomes in an academic community hospital," Journal of Community Hospital Internal Medicine Perspectives, vol. 5, no. 2, 2015.
  • [4] G. B. Smith, D. R. Prytherch, P. Meredith, P. E. Schmidt and P. I. Featherstoned, "The ability of the National Early Warning Score (NEWS) to discriminate patients at risk of early cardiac arrest, unanticipated intensive care unit admission and death," Elsevier Resuscitation, vol. 84, no. 4, pp. 465-470, 2013.
  • [5] D. R. Prythercha, G. B. Smith, P. E. Schmidt and P. I. Featherstone, "ViEWS—Towards a national early warning score for detecting adult inpatient deterioration," Elsevier Resuscitation, vol. 81, no. 8, p. 932– 937, 2010.
  • [6] C. Ye, O. Wang, M. Liu, L. Zheng, M. Xia, S. Hao, B. Jin, H. Jin, C. Zhu, C. J. Huang, P. Gao, G. Ellrodt, D. Brennan, F. Stearns, K. G. Sylvester, E. Widen, D. B. McElhinney and X. Ling, "A Real-Time Early Warning System for Monitoring Inpatient Mortality Risk: Prospective Study Using Electronic Medical Record Data," Journal of Medical Internet Research, vol. 21, no. 7, 2019.
  • [7] X. Cai, O. Perez-Concha, E. Coiera, F. Martin-Sanchez, R. Day, D. Roffe and B. Gallego, "Real-time prediction of mortality, readmission, and length of stay using electronic health record data," Journal of Amarican Medical Information Association, vol. 23, no. 3, pp. 553-561, 2016.
  • [8] F. Babić, M. Vadovský, M. Muchová, J. Paralić and L. Majnarić, "Simple Understandable Analysis of Medical Data to Support the Diagnostic Process," in IEEE 15th International Symposium on Applied Machine Intelligence and Informatics, Slovakia, 2017.
  • [9] H. Mohamadlou, A. Lynn-Palevsky, C. Barton, U. Chettipally, L. Shieh, J. Calvert, N. R. Saber and R. Das, "Prediction of Acute Kidney Injury With a Machine Learning Algorithm Using Electronic Health Record Data," Canadian Journal of Kidney Health, vol. 5, pp. 1-9, 2018.
  • 10] I. Babaoğlu, O. Fındık and M. Bayrak, "Effects of principal component analysis on assessment of coronary artery diseases using support vector machine," Elsevier Expert Systems with Applications, vol. 37, pp. 2182-2185, 2010.
  • [11] S. Kara and F. Dirgenali, "A system to diagnose atherosclerosis via wavelet transforms, principal component analysis and artificial neural networks," Elsevier Expert Systems with Applications, vol. 32, pp. 632-640, 2007.
  • [12] A. E. Johnson, J. T. Pollard, L. Shen, H. L.-W. Lehman, M. Feng, M. Ghassemi, B. Moody, P. Szolovits, L. A. Celi and R. G. Mark, "MIMIC-III, a freely accessible critical care database," Scientific Data, 2016.
  • [13] A. Goldberger, L. Amaral, L. Glass, J. Hausdorff, P. Ivanov, R. Mark, J. Mietus, G. Moody, C. Peng and H. Stanley, "PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals," Circulation 101(23):e215- e220, 2000.
  • [14] J.-L. Vincent, N. D. Nielsen, N. I. Shapiro, M. E. Gerbasi, A. Grossman, R. Doroff, F. Zeng, P. J. Young and J. A. Russell, "Mean arterial pressure and mortality in patients with distributive shock: a retrospective analysis of the MIMIC-III database," Annals of Intensive Care, vol. 8, no. 107, 2018.
  • [15] I. H. Witten and E. Frank, Data Mining: Practical Machine Learning Tools and Techniques, 2005: Elsevier Science.
  • [16] P. D. Tyler, H. Du, M. Feng, R. Bai, Z. Xu, G. L. Horowitz, D. J. Stone and L. A. Celi, "Assessment of Intensive Care Unit Laboratory Values That Differ From Reference Ranges and Association With Patient Mortality and Length of Stay," JAMA Network Open, vol. 1, no. 7, 2018.
  • [17] L. Lebart, "Which Bootstrap for Principal Axes Methods?," in Selected Contributions in Data Analysis and Classification. Studies in Classification, Data Analysis, and Knowledge Organization, Springer, Berlin, Heidelberg, 2007, pp. 581-588.
  • [18] H. Ogasawara, "Asymptotic biases of the unrotated/rotated solutions in principal component analysis," British Journal of Mathematical and Statistical Psychology, vol. 57, p. 353–376, 2004.
  • [19] M. Linting, B. van Os and J. J. Meulman, "Statistical Significance of the Contribution of Variables to the PCA Solution: An Alternative Permutation Strategy," Psychometrika, vol. 76, no. 3, pp. 440-460, 2011.