Association between VA providers’ reported use of care strategies and emergence delirium: a national retrospective cross-sectional survey

BACKGROUND: Emergence delirium (ED) is a short-term behavioral condition associated with risk of harm that may occur when a patient awakens from anesthesia. A national survey was conducted to evaluate our hypothesis that Veterans Health Administration (VHA) anesthesia providers who reported routinely screening patients for ED risk and selecting medication strategies associated with lower ED risk would be less likely to report estimated ED occurrence. METHODS: A retrospective cross-sectional self-report survey was designed by staff at Veterans Affairs' Pittsburgh Healthcare System (PHS) and launched nationally to VHA anesthesia providers. The survey was conducted after an intervention at PHS to increase routine patient screening for ED risk and use of the "exact medication" strategy promoted by the intervention regardless of location: administration of propofol, dexmedetomidine, and ketamine and avoidance of midazolam and volatile anesthetics. Comparisons of percentage of PHS and non-PHS anesthesia providers who reported use of the "exact medication" strategy and estimated ED occurrence were analyzed using tests of independence and multivariable logistic regression models. RESULTS: Based on survey data from 690 VHA anesthesia providers (PHS n = 34, non-PHS n = 656), both routine screening (PHS: 88.2%, non-PHS: 58.4%, p < 0.001) and the "exact medication" strategy (PHS: 44.1%, non-PHS: 1.4%, p < 0.001) were reported by more PHS than non-PHS providers. Survey results indicated 37.3% more providers at PHS (38.2%, n = 13) reported using the "routine screening plus exact medication" strategy than non-PHS (0.9%, n = 6, p < 0.001) and providers who used that care strategy were less likely to report estimated ED occurrence (PHS: 46% vs non-PHS: 50%, p < 0.05). Using multivariable modeling analysis, the "routine screening plus exact medication" care strategy was associated with lower odds of estimated ED occurrence regardless of location (OR = 0.36, p < 0.05). CONCLUSIONS: Bivariate comparisons and multivariable adjusted modeling of survey findings supported our hypothesis that providers who reported routinely screening patients for ED risk and employing the "exact medication" strategy were less likely to have estimated ED occurrence. Use of these descriptive analyses to generate hypotheses for future assessments is recommended.