Upgrade to Pro — share decks privately, control downloads, hide ads and more …

ATS 2016: Crowdsourcing Research Communities to Solve Problems in Critical Care

ATS 2016: Crowdsourcing Research Communities to Solve Problems in Critical Care

Presentation given at the American Thoracic Society 2016 Conference.



May 18, 2016


  1. Crowdsourcing Research Communities to Solve Problems in Critical Care Tom

    J. Pollard MIT Laboratory for Computational Physiology, Institute for Medical Engineering and Science @tompollard
  2. Evolution of MIMIC, a freely available critical care database

  3. None
  4. Research opportunity ▪ Huge volumes of data are captured daily

    ▪ …data that could be used to discover new knowledge for the benefit of patients 4
  5. ▪ but, most of this data ends up in a

    black hole, either lost or locked away 5
  6. • In 2003, awarded NIH funding to establish a research

    program… • …with the objective of developing and evaluating ICU monitoring and decision systems. 6
  7. • Critical for the project was the development of a

    comprehensive clinical database. • In retrospect, this database was the most important output of the project. 7
  8. 8

  9. 9 • Close collaboration between clinicians and engineers • Support

    by the manufacturers of the monitoring systems • Supportive institutional review board • Trust between MIT and the hospital
  10. Publicly released in 2010 (MIMIC-II), and last updated in 2015

    (MIMIC-III). Notable because: • Freely available • Over 55,000 hospital admissions • Highly detailed • Decade of data (2002-2012)
  11. Some details…

  12. Relational database (a collection of linked tables) 12

  13. Accessing MIMIC Two key steps to gaining access to MIMIC:

    • complete a recognized course in protecting human research participants that covers Health Insurance Portability and Accountability Act (HIPAA) requirements • sign a data use agreement, which outlines appropriate data usage and security standards, and forbids efforts to identify individual patients. 13
  14. Heart rate O2 saturation NIBP, mean Respiratory rate Intake volume,

    dL Output volume, dL 0 0 10 30 50 40 20 60 120 100 80 60 40 20 Time after admission to the intensive care unit, hours Measurement, absolute value Code status Full code Comfort measures Incomprehensible sounds Flex-withdraws None Oriented Obeys commands Spontaneously Oriented Obeys commands Spontaneously Oriented Obeys commands Spontaneously Confused Obeys commands To speech Confused Obeys commands To speech GCS: Verbal GCS: Moto GCS: Eye Platelet, K/uL Creatinine, mg/dL White blood cell, K/uL Neutrophil, % Morphine Sulfate Vancomycin (1 dose) Piperacillin (1 dose) NaCl 0.9% Amiodarone Dextrose 5% 48 0.7 9.1 37 53 12.4 46 0.7 16.8 45 0.8 23.2 10.0mL/hour 10.0mL/hour 10.0mL/hour 1mg/min 0.5mg/min 0.5mg/min 50mL/hour 25mL/hour 25mL/hour
  15. 15

  16. MIMIC is widely used in research, teaching, and industry

  17. None
  18. None
  19. None
  20. None
  21. In summary…

  22. If you are… • a new investigator: use MIMIC for

    your research or to replicate your own studies. • a department head: use this in teaching. • an information officer: work with us to develop a multi-center dataset. • a grant reviewer: consider the value of data sharing with the research community. 22
  23. http://mimic.physionet.org DATATHONS London: Dec 2016 Melbourne: Feb 2017 MIT-Philips eICU

    Database Release: Aug 2016
  24. None