Can the attending anesthesiologist accurately predict the duration of anesthesia induction?

In a prospective, observational study, the attending anesthesiologists' prediction of anesthesia release time (ART) of the patient to the surgical team was highly correlated with actual ART (r = 0.77; P < or = 0.001). However, this was true only in the aggregate (n = 1265 patients). Indeed, offsetting degrees of under- and over-predicting (24% each) reduced accuracy to only 53% per individual case. For example, under-prediction was associated with ASA physical status IV, a regional anesthetic technique, age >65 yr, and the use of invasive hemodynamic monitoring (P = 0.006). In fact, as the degree of case difficulty increased, the correlation coefficient between predicted and actual ART decreased, indicating a poor predictive value with more difficult inductions (r = 0.82 to r = 0.44; P < or = 0.004). We conclude that knowledge of the presence of specific factors that lead to inaccurate predictions of time required for induction of anesthesia may enhance the accuracy of the operating room schedule.