Editor’s Note: Dr. Westenkow was a recipient of an APSF Research Grant for 1987. He reports here on a portion of his work.
An anesthesia workstation is more than just a gas machine with monitors on it. The workstation concept should help improve the safety of anesthesia by gathering data to one central location, by detecting critical events and by helping to control the delivery of anesthesia.
A prototype workstation, assembled under a grant for the Anesthesia Patient Safety Foundation, has nine pressure sensors, a pneumotach, an infrared anesthetic agent analyzer, a mainstream C02 analyzer, a polarographic oxygen monitor, a ventilator bellows position sensor, and three mass flow controllers. PASCAL programs acquire, filter and plot the data. When a critical event is found, one of the 43 heuristic rules used by the computer is employed and it actually draws a diagram of the anesthesia machine and breathing circuit to highlight the faulty component and prints an alarm message.
The prototype alarm system was tested, measuring the accuracy with which it identified 2 6 critical events. The alarm system correctly identified 94% simulated critical events. The failures occurred when the alarm system did not see fresh gas hose disconnects and an open inspiratory valve. The system did not specifically identify an “open inspiratory valve” but instead saw the “high endtidal C02”. An open expiratory valve was seen as “Ventilator Failure”.
The alarm system in the workstation produced alerts that are more meaningful, and more diagnostic than are the traditional oxygen, C02, agent or pressure alarms. The graphic presentation of the alarm condition and text should reduce reaction time and help in rapidly resolving critical events. The alarm thresholds are set automatically, thus averting mistakes sometimes made when setting alarm limits manually.
In the prototype workstation, the computer electronically controls the fresh gas flows to achieve the user specified fresh gas flow rate, F02 and endtidal volatile anesthetic concentration. This “autopilot” feature reached the desired end-tidal volatile agent concentration after 4.4 + 1.4 min (mean + SD). It kept the inspired oxygen concentration, within .002 vol% of the desired value. The autopilot appears to make anesthesia delivery easier to manage.
- Loeb R, Westemkow D, Brunner 1, Feldman B, Pace N: The Utah Anesthesia Workstation. I Clin Mon 4:148-149, 1988.
Dr. Westenskow is from the Department of Anesthesiology, University of Utah, Salt Lake City