RESUMEN
Objectives: This study seeks to determine the potential use and reliability of a large language learning model for answering questions in a sub-specialized area of medicine, specifically practice exam questions in otolaryngology-head and neck surgery and assess its current efficacy for surgical trainees and learners. Study Design and Setting: All available questions from a public, paid-access question bank were manually input through ChatGPT. Methods: Outputs from ChatGPT were compared against the benchmark of the answers and explanations from the question bank. Questions were assessed in 2 domains: accuracy and comprehensiveness of explanations. Results: Overall, our study demonstrates a ChatGPT correct answer rate of 53% and a correct explanation rate of 54%. We find that with increasing difficulty of questions there is a decreasing rate of answer and explanation accuracy. Conclusion: Currently, artificial intelligence-driven learning platforms are not robust enough to be reliable medical education resources to assist learners in sub-specialty specific patient decision making scenarios.
RESUMEN
Objective: To examine the differences in opioid prescribing by cognitive status following common elective surgical procedures among Medicare beneficiaries. Background: Older individuals commonly experience changes in cognition with age. Although opioid prescribing is common after surgery, differences in opioid prescribing after surgery by cognitive status are poorly understood. Methods: We conducted a retrospective analysis of patients ≥65 years participating in the Health and Retirement Study (HRS) linked with Medicare claims data who underwent surgeries between January 2007 and November 2016 and had cognitive assessments before the index operation. Cognitive status was defined as normal cognition, mild cognitive impairment (MCI), or dementia. Outcomes assessed were initial perioperative opioid fill rates, refill rates, and high-risk prescriptions fill rates. The total amount of opioids filled during the 30-day postdischarge period was also assessed. Adjusted rates were estimated for patient factors using the Cochran-Armitage test for trend. Results: Among the 1874 patients included in the analysis, 68% had normal cognition, 21.3% had MCI, and 10.7% had dementia. Patients with normal cognition (58.1%) and MCI (54.5%) had higher initial preoperative fill rates than patients with dementia (33.5%) (P < 0.001). Overall, patients with dementia had similar opioid refill rates (21%) to patients with normal cognition (24.1%) and MCI (26.5%) (P = 0.322). Although prior opioid exposure did not differ by cognitive status (P = 0.171), among patients with high chronic preoperative use, those with dementia had lower adjusted prescription sizes filled within 30 days following discharge (281 OME) than patients with normal cognition (2147 OME) and MCI (774 OME) (P < 0.001; P = 0.009 respectively). Among opioid-naive patients, patients with dementia also filled smaller prescription sizes (97 OME) compared to patients with normal cognition (205 OME) and patients with MCI (173 OME) (P < 0.001 and P = 0.019, respectively). Conclusions: Patients with dementia are less likely to receive postoperative prescriptions, less likely to refill prescriptions, and receive prescriptions of smaller sizes compared to patients with normal cognition or MCI. A cognitive assessment is an additional tool surgeons can use to determine a patient's individualized postoperative pain control plan.
RESUMEN
Purpose: Despite established opioid-free protocols for postoperative analgesia after common orthopaedic sports procedures, many patients continue to request opioids postoperatively. The purpose of this study was to elucidate patient factors influencing preferences for opioid versus nonopioid postoperative analgesia. Methods: Patients (age >/ = 15) without a history of a documented chronic pain disorder who were scheduled for one of ten sports procedure types from August 2020 to May 2021 were eligible for inclusion. Patients were excluded if undergoing revision surgery, had concomitant injuries, had opioids use >3 months preoperatively, or unable to read English. Recruitment ended after 100 patients enrolled. At the patients' preoperative visit, patients were administered a written survey assessing pain medication preferences. Participants completed the Opioid Risk Tool survey, as well as Visual Analog Scale and Patient-Reported Outcome Measurement Information System surveys. Results: One hundred patients participated in the study. Forty-two patients preferred opioids versus 58 patients preferring nonopioid postoperative analgesia. Patients preferring opiates were more likely to have had previous surgery (90.2% vs. 69.6%, p = 0.023) with post-operative pain managed with opiates (87.5% vs 55.4%, p = 0.003), higher preoperative Visual Analog Scale score (6±3.5 vs. 3±2, p < 0.001), reported post-operative pain as a reason for opioids preference (88.1% vs 20.0%, p < 0.001), and were less concerned about addiction (4.8% vs. 45.5%, p < 0.001) and side effects (11.9% vs. 52.7%, p < 0.001). For every unit increase in Visual Analog Scale score, the odds of preferring opioid pain control increased 1.41 times. Conclusions: Patients with a history of prior surgery utilizing opioid pain control, higher Visual Analog Scale scores preoperatively, and concern for inadequately managed postoperative pain were more likely to prefer opioid pain control following common orthopaedic sports procedures. Patients may benefit from increased preoperative education about opioid risks and the role of multimodal pain management regimens.
