Using Machine Learning to Evaluate Attending Feedback on Resident Performance.

BACKGROUND: High-quality and high-utility feedback allows for the development of improvement plans for trainees. The current manual assessment of the quality of this feedback is time consuming and subjective. We propose the use of machine learning to rapidly distinguish the quality of attending feedback on resident performance. METHODS: Using a preexisting databank of 1925 manually reviewed feedback comments from 4 anesthesiology residency programs, we trained machine learning models to predict whether comments contained 6 predefined feedback traits (actionable, behavior focused, detailed, negative feedback, professionalism/communication, and specific) and predict the utility score of the comment on a scale of 1-5. Comments with ≥4 feedback traits were classified as high-quality and comments with ≥4 utility scores were classified as high-utility; otherwise comments were considered low-quality or low-utility, respectively. We used RapidMiner Studio (RapidMiner, Inc, Boston, MA), a data science platform, to train, validate, and score performance of models. RESULTS: Models for predicting the presence of feedback traits had accuracies of 74.4%-82.2%. Predictions on utility category were 82.1% accurate, with 89.2% sensitivity, and 89.8% class precision for low-utility predictions. Predictions on quality category were 78.5% accurate, with 86.1% sensitivity, and 85.0% class precision for low-quality predictions. Fifteen to 20 hours were spent by a research assistant with no prior experience in machine learning to become familiar with software, create models, and review performance on predictions made. The program read data, applied models, and generated predictions within minutes. In contrast, a recent manual feedback scoring effort by an author took 15 hours to manually collate and score 200 comments during the course of 2 weeks. CONCLUSIONS: Harnessing the potential of machine learning allows for rapid assessment of attending feedback on resident performance. Using predictive models to rapidly screen for low-quality and low-utility feedback can aid programs in improving feedback provision, both globally and by individual faculty.

The "Difficult Learner" in anesthesiology: Challenges, pitfalls, and recommendations.

Struggling learners often require interventions that are time-consuming and emotionally exhausting for both the trainee and faculty. Numerous barriers, including lack of resources, faculty development, and fear of legal retribution, can impede medical educators from developing and implementing robust remediation plans. Despite the large volume of literature citing professionalism education and the "hidden curriculum" as problem areas in medical education, frontline educators lack practical tools and empowerment to address unprofessionalism in trainees. The paucity of resources in this arena has led to decreased job satisfaction, increased burnout, and an exodus from academic medicine. Department leadership acknowledgment and investment in training faculty on remediation strategies for learners in difficulty and providing tools to meet these challenging job demands will improve faculty's job satisfaction and overall well-being. The authors review salient literature and methodology for diagnosing learners in difficulty, with focus on a high-yield, pragmatic approach that can be taken by medical training programs, including those that lack a robust medical education infrastructure.

Less pain and earlier discharge after implementation of a multidisciplinary enhanced recovery after surgery (ERAS) protocol for laparoscopic sleeve gastrectomy.

BACKGROUND: Laparoscopic sleeve gastrectomy (LSG) may be complicated by postoperative pain, nausea, and vomiting, with consequent increases in length of stay (LOS), decreased patient satisfaction, and higher costs. While enhanced recovery after surgery (ERAS) protocols have been in circulation for many years, there is no standard ERAS protocol for bariatric surgery. METHODS: Data were collected prospectively and compared to a historical control. All patients undergoing LSG, ages 18 to 75, were included in the pathway; those with preoperative chronic opioid use were excluded from our results. Statistical analysis was performed using t-statistics and chi-squared test. Ninety patients undergoing LSG, performed by a single surgeon, were included in our ERAS group from November 26, 2018, to April 30, 2019, and were compared to a historical control of 570 patients who underwent LSG over the previous 5 years (pre-ERAS). Measured outcomes included discharge opioid prescriptions issued, hospital length of stay, 30-day readmissions, reoperations, morbidity, and mortality. RESULTS: Ten (11%) ERAS patients vs 100% of pre-ERAS patients received opioid prescriptions upon, or after, discharge (p < 0.001). The ERAS group LOS decreased to 1.36 days vs 2.40 days in the pre-ERAS group (p < 0.001). 30-day readmission rates were 0% for ERAS patients vs 3.09% for pre-ERAS patients (p = 0.149). 30-day reoperation rates were 0% for ERAS patients vs 0.54% for pre-ERAS patients (p = 1). Thirty-day morbidity rates were 3.33% (3) for ERAS patients vs 3.27% for pre-ERAS patients (p = 1); there was no 30-day mortality in either group. CONCLUSION: ERAS for LSG results in a clinical and statistically significant reduction in postoperative opioid use and LOS, without increasing 30-day readmissions, reoperations, morbidity, or mortality.

Be Active and Be Well? A Cross-sectional Survey of US Anesthesia Residents.

BACKGROUND: Wellness among resident physicians is important to their well-being and ability to provide clinical care. The relationship between physical activity and wellness among anesthesia residents has not yet been evaluated. We surveyed anesthesia residents to evaluate their levels of physical activity and self-perceived wellness scores. We hypothesized that residents with high self-reported physical activity levels would be more likely to have higher wellness scores. METHODS: Three hundred and twenty-three anesthesia residents were invited to participate in this cross-sectional survey study. The survey included questions regarding demographics (age, gender, clinical anesthesia year, work hours), physical activity (based off the US Department of Health and Human Services [USDHHS] guidelines), and wellness (using the Satisfaction With Life Scale). The relationship between wellness and physical activity levels was evaluated. RESULTS: One hundred forty-one residents responded (43.6% response rate). Thirty-eight (27.1%) residents met our activity threshold for physically active. Eighty-six respondents (61.4%) were classified as having high wellness based on their survey answers. No significant associations were found between demographic data and wellness, including age or clinical anesthesia training year. Among those residents who described physical activity consistent with USDHHS guidelines, 29 (76.3%) had high wellness scores. After logistic regression analysis, residents who achieved the physical activity guidelines were more likely to have high wellness scores (odds ratio 2.54, 95% confidence interval 1.13-6.20, P value .03). CONCLUSIONS: Anesthesia resident physicians with high physical activity levels had higher self-perceived wellness scores.

Learning Preferences of First-Year Anesthesiology Residents During Their Orientation Month: A Single-Institution Study.

As a step toward understanding whether learning preferences may benefit trainees and instructors, we studied the learning preferences, based on the VARK questionnaire, of anesthesiology residents during their orientation month to determine whether knowing their preferences influenced the educational resources they used. While resource utilization was similar for residents who knew their preferences and those who did not, residents overall used aural and read/write resources frequently. The learning preferences of 13 residents (72.2%) changed during the month. While further research is needed, this study provides insight into anesthesiology residents' learning styles.