Automatic control in anesthesia: a comparison in performance between the anesthetist and the machine

This report is divided into two parts. First, we developed two new servo control systems by modifying an existing one. The original system was designed to control inspired halothane concentration using mean arterial pressure; the two new systems were designed to control inspired halothane concentration using end-tidal concentration or to control mean arterial pressure using the automated infusion of nitroprusside. Second, we compared the performance of experienced physician and nurse anesthetists (nine, six, and six experiments, respectively). The experiments incorporated a standardized testing sequence of two changes in desired blood pressure (set point) and two pharmacologically induced disturbances in blood pressure (perturbations). The scoring was designed to examine how fast blood pressure changed (90% response time), how far past the set point it went (overshoot), how long it took to eliminate most of the fluctuations in blood pressure (settling time), and the degree of fluctuation of blood pressure after settling (stability). Given three systems to be tested, there were (3 X 14) 42 possible mean scores for the machine and 42 for the anesthetists. The machine scored better than the anesthetists in 38 out of 42 of the mean scores; the differences were statistically significant in 19 out of 42 scores. The wide scatter in performances of the anesthetists prevented the achievement of significance in nine cases with large differences between means. Thus when the scores from the three systems were combined to achieve a larger n value, the machine outperformed the anesthetist in 12 out of 14 scores.(ABSTRACT TRUNCATED AT 250 WORDS)