Skip to Content

Team Cool Monkey

Team Members:
  • Zhenzhen Wang
  • Wen Shi
  • Chaoran Chen
  • Stephen Kyranakis
  • Ananya Swaminathan
  • Timothy Ruchti, PhD
  • James Fackler, MD
  • Jules Bergmann, MD
  • Raimond Winslow, PhD
  • Joseph Greenstein, PhD
  • Hanbiehn Kim
  • Hieu Nguyen


Hypoxemia is defined as an abnormally low level of oxygen in arterial blood. It is commonly diagnosed among general Intensive Care Unit (ICU) patients. Recent research has shown that moderate and severe hypoxemia are independently associated with higher mortality rates and an increase in the length of stay. Despite the severity and prevalence of hypoxemia in the ICU, current practices cannot accurately predict hypoxemic events and therefore do not provide clinicians enough time to prepare proper interventions that minimize patient mortality. If hypoxemia could be predicted 10 to 30 minutes in advance, clinicians and physicians would have sufficient time to administer necessary interventions. Our team proposes a machine learning-based hypoxemia alert system that predicts the onset of hypoxemia in the near future using patient vital signs and demographic data. This system could improve the efficiency and timeliness of treatment, thus improving ICU patient outcomes.

Read the Johns Hopkins University privacy statement here.