Procedure-adjusted incidences of postoperative hospital admissions are not associated with differences in the quality of anesthesiologists’ clinical supervision or nurse anesthetists’ work habits

Abstract

Background

Anesthesia departments may benefit from automated computerized methods to monitor the clinical performance of individual anesthesia practitioners. Hospital admission (>1 night stay) after ambulatory surgery may be a suitable metric, with higher incidences potentially being associated with poor clinical performance. If valid, there should be a small but statistically significant association of postoperative admission with previously validated measures of quality of intraoperative anesthesia care.

Methods

This retrospective cohort study used eight fiscal years of data from one hospital with daily assessments of two different measures of quality of anesthesia clinical care. One variable was anesthesiologists’ clinical supervision evaluated by trainees, principally residents. The second independent variable was nurse anesthetists’ work habits, evaluated by anesthesiologists. These independent variables were binary, the proportions of rater-leniency-adjusted evaluations with maximum performance for all items, calculated annually. The dependent variable was the proportion of ratees’ cases (anesthesiologist or nurse anesthetist) with postoperative length of stay ≤ 1 day, adjusted for surgical suite and procedure category. Thus, for both independent and dependent variables, larger (positive logits) were “good” and smaller (negative logits) were “bad.”

Results

There were no significant associations for either supervision (P =0.14, N=561 anesthesiologist-years) or work habits (P =0.74, N=598 nurse anesthetist-years). Estimated signs of the slopes were for increases in the logits of the quality of clinical supervision to be associated with non-significant decreases in the logits of the probabilities of the patients having lengths of stay ≤1 day. Similarly, increases in the logits of nurse anesthetists’ work habits had negative-signed non-significant associations with the logits of the probabilities of the patients having lengths of stay ≤1 day.

Conclusions

The results show with substantial certainty that higher-performing anesthesia practitioners do not have briefer lengths of stay. Anesthesiologists and nurse anesthetists should not be compared among one another based on whether their patients have a greater than predicted risk of procedure-adjusted hospital admission after ambulatory surgery.