To compare invasive blood pressure measurements recorded using an automated archiving method against clinician-documented values from the same invasive monitor and determine which method of recording blood pressure is more highly associated with the subsequent onset of hypotension.
Retrospective comparative analysis.
Intensive care patients in a university hospital.
Mixed medical/surgical patients.
Using intervals of hemodynamic stability from 2,320 patient records, we retrospectively compared paired sources of invasive blood pressure data: 1) measurements documented by the nursing staff and 2) measurements generated by an automated archiving method that intelligently excludes unreliable (e.g., noisy or excessively damped) blood pressure values. The primary outcome was the occurrence of subsequent “consensus” hypotension, i.e., hypotension documented jointly by the nursing staff and the automated archive. The automated method could be adjusted to alter its operating characteristics (sensitivity and specificity). At a matched level of specificity (96%), blood pressures from the automated archiving method were more sensitive (28%) for subsequent consensus hypotension vs. the nurse-documented values (21%). Likewise, at a matched level of sensitivity (21%), the values from the automated method were more specific (99%) vs. the nurse-documented values (96%). These significant findings (p < .001) were consistent in a set of sensitivity analyses that employed alternative criteria for patient selection and the clinical outcome definition.
During periods of hemodynamic stability in an intensive care unit patient population, clinician-documented blood pressure values were inferior to values from an intelligent automated archiving method as early indicators of hemodynamic instability. Human oversight may not be necessary for creating a valid archive of vital sign data within an electronic medical record. Furthermore, if clinicians do have a tendency to disregard early indications of instability, then an automated archive may be a preferable source of data for so-called early warning systems that identify patients at risk of decompensation.
From the Department of Electrical Engineering and Computer Science (CWH) and Harvard-MIT Health Sciences and Technology (GDC, ATR), Massachusetts Institute of Technology, Cambridge, MA; Institute of Biomedical Engineering (GDC), University of Oxford, Oxford, United Kingdom; and Department of Emergency Medicine (ATR), Massachusetts General Hospital, Boston, MA.
Supported, in part, by the National Library of Medicine (Bethesda, MD) Medical Informatics Traineeship (grant LM 07092), the U.S. National Institute of Biomedical Imaging and Bioengineering (Bethesda, MD) and the National Institutes of Health (Bethesda, MD) under grant R01 EB001659, Philips Healthcare (Andover, MA), and the Information and Communication University (Daejeon, Korea). The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of the National Library of Medicine, the U.S. National Institute of Biomedical Imaging and Bioengineering, the National Institutes of Health, Philips Healthcare, or Information and Communication University.
Dr. Hug and Dr. Clifford received funding from the National Institutes of Health (Bethesda, MD). Dr. Reisner received funding from the National Institutes of Health and consulted for General Electric (Fairfield, CT).
Work performed at the Massachusetts Institute of Technology, Cambridge, MA.
For information regarding this article, E-mail: email@example.com