Analytics Practicum Project at Rutgers University, New Jersey

Our project used the MIMIC Dataset to predict 30 day- critical care readmissions. The machine learning model that predicted whether a patient upon admission was at risk of readmission. This is a problem which costs Medicare $15 million dollars a year. Our results were comparable with leading reasearch papers on this topic as identified in the literature survey using Area under the ROC curve and recall as evaluation metrics.

Features deemed to be important towards the predictive power of the algorithms were: Hemoglobin levels Creatinine levels White blood cell count Hematocrit Heart Rate Platelet Count Red blood cell count Systolic Blood Pressure Blood Oxygen Saturation PTT (partial thromboplastin time)

Plus two we added in: Is the admission a readmission? Number of disease categories involved in patient diagnosis (based on ICD- 9 Codes)

Rutgers Universtiy: Advanced Analytics and Practicum

Michael Albuquerque Miranda So Sonal Oberoi

[1] MIMIC-III, a freely accessible critical care database. Johnson AEW, Pollard TJ, Shen L, Lehman L, Feng M, Ghassemi M, Moody B, Szolovits P, Celi LA, and Mark RG. Scientific Data (2016). DOI: 10.1038/sdata.2016.35. Available at: http://www.nature.com/articles/sdata201635

[2] Pollard, T. J. & Johnson, A. E. W. The MIMIC-III Clinical Database http://dx.doi.org/10.13026/C2XW26 (2016).

[3] Physiobank, physiotoolkit, and physionet components of a new research resource for complex physiologic signals. Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov P, Mark RG, Mietus JE, Moody GB, Peng C, and Stanley HE. Circulation. 101(23), pe215–e220. 2000.

[4] Johnson, Alistair EW, David J. Stone, Leo A. Celi, and Tom J. Pollard. “The MIMIC Code Repository: enabling reproducibility in critical care research.” Journal of the American Medical Informatics Association (2017): ocx084.