Integrated Sepsis Surveillance Algorithms

We are currently working on an intelligent sepsis monitoring system to continuously identify and risk stratify patients at increased risk of sepsis. We are working with leading experts in machine learning at the Courant Institute of Mathematical Sciences to develop the machine learning theory to integrate all available demographic, laboratory, radiographic, and continuous hemodynamic data. We will use these algorithms to derive next generation clinical alerts, contextual information retrieval, and targeted decision support. This research is supported by a CIMIT Innovation Grant funded by the Department of Defense. clinical trials protocol 

  • Horng S , Sontag DA, Shapiro NI, Nathanson LA. Machine Learning Algorithms Can Identify Patients Who Will Benefit From Targeted Sepsis Decision Support. National ACEP Research Forum. Denver, CO, Oct 9, 2012.
  • Horng S , Nathanson LA, Sontag DA, Shapiro NI. Prospective Validation of the Angus ICD9-CM Sepsis Abstraction Criteria. National ACEP Research Forum. Denver, CO, Oct 8, 2012.
  • Halpern Y, Horng SNathanson LA, Shapiro NI, Sontag DA. A Comparison of Dimensionality Reduction Techniques for Unstructured Clinical Text. ICML 2012 Workshop on Clinical Data Analysis, July 2012. paper
  • Horng S , Sontag DA, Chiu DT, Joseph JW, Shapiro NI, Nathanson LA. Predicting ICU Admission and Mortality at Triage using an Automated Computer Algorithm. National SAEM Research Forum. Chicago, IL, May 2012.
  • Halpern Y, Horng SNathanson LA, Shapiro NI, Sontag DA. Patient Surveillance Algorithms for the Emergency Department. NIPS Workshop. Sierra Nevada, Spain, Dec 16, 2011. paper
  • Horng S , Sontag DA, Chiu DT, Shapiro NI, Nathanson LA. The Effect of a Triage Nurse's Free Text Assessments on a Machine Learning Algorithm to Identify Infection. National ACEP Research Forum. San Francisco, CA, Oct 2011.
  1. Contextual Information Retrieval
  2. Targeted Decision Support