Abstract
IMPORTANCE: Despite substantial advances in perioperative care, patients with cognitive impairment are at increased risk for postoperative delirium. Understanding the association of perioperative best practices with delirium prevention may inform tailored strategies for this high-risk group. OBJECTIVE: To examine associations between anesthesia-related best practices and postoperative delirium in patients with cognitive impairment. DESIGN, SETTING, AND PARTICIPANTS: This observational cohort study involved a secondary, exploratory analysis of a randomized clinical trial of a clinical decision support system designed to identify patients with cognitive impairment and promote adherence to best practices. The study was conducted between 2023 and 2024 at perioperative areas at a large, urban, academic medical center in New York. Participants were adults scheduled for surgery with preoperative cognitive impairment who stayed in the hospital for at least 1 postoperative night and underwent delirium assessment. Patients undergoing organ donation surgery were excluded. EXPOSURES: Twelve perioperative best practices across 5 domains were selected and tracked, reflecting avoidance of potentially inappropriate medications, perioperative glycemic control, hemodynamic management, normothermia maintenance, and anesthetic monitoring. MAIN OUTCOMES AND MEASURES: The primary outcome was postoperative delirium assessed by the 4 A's Test. Associations between best practices and postoperative delirium were assessed using logistic regression models. RESULTS: The study included 1255 patients with cognitive impairment (mean [SD] age, 65.0 [15.1] years; 652 male [52%]), and 426 (33.9%; 95% CI, 31.4%-36.6%) developed delirium. Best practices, such as maintaining glucose less than 200 mg/dL (adjusted odds ratio [aOR], 0.41; 95% CI, 0.24-0.69; P = .001), using a temperature probe (aOR, 0.65; 95% CI, 0.45-0.96; P = .03), and maintaining temperature greater than 36 °C (aOR, 0.64; 95% CI, 0.43-0.94; P = .03), were associated with reduced delirium odds in separate models. No individual practice remained significant when analyzed concurrently, apart from postoperative glucose monitoring (aOR, 1.54; 95% CI, 1.11-2.13; P = .009). CONCLUSIONS AND RELEVANCE: This cohort study of surgical patients with cognitive impairment found that current perioperative strategies derived from established recommendations, while foundational, were insufficient to prevent postoperative delirium. The high rate of postoperative delirium highlights the need to refine perioperative care pathways and identify novel strategies that are most effective for cognitively vulnerable populations. Tailored interventions that integrate electronic health record-derived perioperative data and scalable technological tools in anesthesia workflows can help identify and manage high-risk populations.