If one could accurately predict who, when and why patients develop shock then effective preemptive treatments could be given to improve outcome and more effectively use healthcare resources. But signs of shock often occur late once organ injury is already present. We have previously shown that an integrated monitoring system (IMS) when coupled to a bedside care plan improves outcome in step down unit (SDU) patients and reduces the duration and intensive of cardiorespiratory insufficiency. We also showed that by advanced HR variability analysis (sample entropy of RR interval variance), we could identify at-risk patients within 2 minutes and if monitored for 5 minutes could predict with a high degree of precision those SDU patients who would remain stable or deteriorate within the next 48 hr. The problem is in defining the minimal monitoring data set needed to initially identify those patients across all possible processes and then specifically monitor their response to targeted therapies known to improve outcome. To address these issues we developed multivariable models through machine learning (ML) data-driven classification techniques [multivariate regression analysis, Fourier analysis, principal component analysis, artificial neural networks, random forest classification, etc.] to predict parsimoniously cardiorespiratory insufficiency. We initially used our porcine models of controlled hemorrhagic shock. Preliminary analysis of our clinically relevant porcine model of hemorrhagic shock ML modeling allowed us to characterize the patterns seen in hemorrhage, hypovolemia to decompensation and resuscitation and predict which animals would collapse during severe hypovolemia and which would not. We found that during hemorrhage that deviation from normal variance could be identified with a high degree of certainty minutes before any measurable change in hemodynamic variables was observed. What is not known is the minimal data set (number and type of independent measures, the sample frequency and the lead time which are necessary to create a robust algorithm) that will: 1) predict impending shock, 2) select the most effective treatments, 3) monitor the response to those treatments, and 4) determine when resuscitation has effectively restored tissue perfusion and can be stopped. We refer to this concept as Monitoring Parsimony. Also, it is not known if additional monitoring parameters or longer lead time before which insufficiency is recognized will improve the ability to achieve the above four targets with the best sensitivity and specificity once an alert is raised. We envision a basic monitoring surveillance that identifies those patients most likely to become unstable for targeted focused monitoring and treatments to deliver highly personalized medical care.

The seminar presents the results of the paper: Pinsky MR, Dubrawski A."Predicting clinically relevant events using real-time data streams and data libraries", Department of Critical Care Medicine, University of Pittsburgh and the Auton Lab, Carnegie-Mellon University, Pittsburgh, PA.

Dr. Michael R. Pinsky is a Visiting Professor of Anesthesiology at UC San Diego on sabbatical leave from the University of Pittsburgh School of Medicine where he has been since 1981 and is a Professor of Critical Care Medicine, Bioengineering, Anesthesiology and Cardiovascular Diseases. His primary research interests are cardiovascular and pulmonary physiology, shock resuscitating, sepsis, energy metabolism, healthcare informatics, and outcomes research. He received his MD, CM from McGill University in 1974, Internal Medicine and Pulmonary fellowship at Stanford University in 1979 and Cardiopulmonary Physiology at Johns Hopkins Medical Institutions in 1981. He has published over 300 manuscripts, 26 textbooks and 250 chapters. He received the 1st International Vicenza Award for Critical Care Nephrology (1997), Docteur honoris causa (Dr hc) from the Université de Paris V (2002), Sodalem honoris causa from the Bohemian Society of Anesthesiology and Acute Care Medicine, Czech Republic (2004), Professor Honoaria from the Universidad de Iberoamérica, Costa Rica (2010), Master of Critical Care Medicine (MCCM) from the College of Critical Care Medicine (2012), Honorary Member of the European Society of Intensive Care Medicine (2012), and the SMART award for outstanding achievement in intensive care (2013). He is PI on two NIH R-01 awards and program director for his departments T-32 training grant which he initiated and has run for the past 18 years.