P030: EQUITY CONSIDERATIONS IN ARTIFICIAL INTELLIGENCE–ASSISTED PERIOPERATIVE CARE: IMPLICATIONS FOR HISPANIC PATIENTS IN FLORIDA
Osvaldo Conde, MD; Tyler Chonis, MD; Ali Yashback, MD; Fernando Mendez, MD; Brandon To; Jerry Kourkomelis, MD; Sebastian Camargo, DO; Aurelio Varona, MD
HCA FL Kendall Hospital
Introduction: Artificial intelligence (AI)–based clinical decision support systems are increasingly incorporated into perioperative medicine. Current applications include predictive risk stratification, intraoperative hemodynamic monitoring, closed-loop anesthetic delivery, and natural language processing tools. While these technologies may improve efficiency and standardize care, their effect on health equity has not been fully evaluated. Hispanic patients in the United States experience documented perioperative disparities, including lower rates of guideline-concordant prophylaxis, higher postoperative complication rates in some studies, and barriers related to limited English proficiency. In Florida, where a substantial proportion of surgical patients identify as Hispanic, it is important to consider whether AI-supported tools reduce or unintentionally reinforce these differences.
Methods: A narrative review of published literature on AI applications in anesthesiology was performed, with emphasis on studies assessing algorithm performance across demographic subgroups. Evidence describing perioperative disparities affecting Hispanic patients was examined to identify areas in which AI-based tools may influence care delivery. Clinical domains reviewed included intraoperative hypotension prediction models, perioperative risk stratification systems, postoperative nausea and vomiting (PONV) prediction tools, closed-loop anesthetic platforms, and language-support technologies. Reports addressing algorithmic bias and subgroup performance monitoring were also analyzed.
Results: Published studies describe disparities affecting Hispanic patients across the perioperative continuum. Compared with non-Hispanic White patients, some analyses report longer length of stay, increased postdischarge complications, and lower rates of guideline-concordant PONV prophylaxis. Limited English proficiency remains common and may contribute to communication barriers in preoperative counseling and postoperative instructions.
AI-assisted tools have demonstrated improved risk prediction accuracy and reduced intraoperative hypotension in selected populations. However, algorithmic bias remains a concern. Underrepresentation of Hispanic patients in training datasets may affect model calibration and predictive performance. In addition, machine learning systems trained on historical practice patterns may reproduce existing disparities in treatment recommendations. Language-dependent interfaces may further limit access to AI-enhanced decision support.
Conversely, models developed using diverse datasets and routine subgroup performance analysis have shown improved predictive accuracy without widening disparities when equity safeguards are incorporated.
Discussion: For anesthesiologists practicing in Florida, integrating AI into perioperative workflows requires careful oversight. Predictive analytics may reduce variability in care and improve safety, but implementation should include routine review of model performance across racial and ethnic groups. Practical steps include requesting subgroup performance data from vendors, evaluating dataset representation before adoption, incorporating multilingual patient education tools, and monitoring postoperative outcomes by ethnicity after implementation.
AI should support clinical judgment rather than replace it. When used thoughtfully and monitored consistently, these systems can help standardize care while maintaining accountability for equitable outcomes.
Conclusion: AI-based decision support has the potential to improve perioperative care. At the same time, without structured evaluation, these tools may perpetuate existing disparities affecting Hispanic patients in Florida. Anesthesiologists play an important role in ensuring that AI is implemented responsibly, with transparent performance reporting and ongoing equity monitoring. Careful integration can align technological advancement with the specialty’s commitment to safe and equitable patient care.