External Validation of an Algorithm to Guide Opioid Administration at the End of Surgery-Protocol for an Observational Cohort Study of the OPIAID Algorithm.

backgroundDespite advances in pain management, inadequate pain relief and opioid-related adverse events remain common challenges in perioperative care, often contributing to prolonged recovery and reduced quality of life. The perioperative opioid algorithms for individualized dosing (OPIAID) project aims to develop machine-learning algorithms tailored to provide patient-specific opioid dosing across the different phases of perioperative care. For each phase, eight models are trained on granular data from 1.1 million surgical procedures, including demographic and surgical details, vital signs, administered analgesics, pain, and opioid-related adverse events. The two most accurate models will proceed to external validation. The best-performing model will subsequently be tested as a decision support against current standard of care.

objectivesThis protocol describes the design and external validation of the intraoperative OPIAID algorithm, which suggests the end-of-surgery opioid dose intended for postoperative analgesia by approximating clinical performance and evaluating reliability, agreement, and calibration.

methodsIn this multicenter, TRIPOD+AI-adherent, prospective observational cohort study, we will collect data from a diverse surgical population of 656 adult patients undergoing elective or acute surgery under general anesthesia. All patients will require intraoperative opioid administration at the end of surgery for postoperative pain management and a subsequent stay in the post-anesthesia care unit. The cohort will be used to externally validate two machine-learning models through standardized measures of reliability, agreement, and calibration, and thereby designate the intraoperative OPIAID algorithm. Subsequently, the cohort will be used to approximate the clinical efficacy, safety and overall performance of the intraoperative OPIAID algorithm's recommended doses versus the clinician-administered doses. These comparisons will be based on each approach's proximity to a golden standard "optimal dose," which is calculated based on a predefined generic ruleset incorporating intraoperative opioid dosing, postoperative pain, opioid-related adverse events, and need for rescue opioid administrations.

conclusionThe intraoperative OPIAID algorithm is intended as a clinical decision aid for anesthesiologists and nurse anesthetists in providing adequate postoperative pain management.