Outcomes of Shared Decision-Making for Low-Dose Screening for Lung Cancer in an Academic Medical Center.

Shared decision-making (SDM) helps patients weigh risks and benefits of screening approaches. Little is known about SDM visits between patients and healthcare providers in the context of lung cancer screening. This study explored the extent that patients were informed by their provider of the benefits and harms of lung cancer screening and expressed certainty about their screening choice. We conducted a survey with 75 patients from an academic medical center in the Southeastern U.S. Survey items included knowledge of benefits and harms of screening, patients' value elicitation during SDM visits, and decisional certainty. Patient and provider characteristics were collected through electronic medical records or self-report. Descriptive statistics, Kruskal-Wallis tests, and Pearson correlations between screening knowledge, value elicitation, and decisional conflict were calculated. The sample was predominately non-Hispanic White (73.3%) with no more than high school education (53.4%) and referred by their primary care provider for screening (78.7%). Patients reported that providers almost always discussed benefits of screening (81.3%), but infrequently discussed potential harms (44.0%). On average, patients had low knowledge about screening (score = 3.71 out of 8) and benefits/harms. Decisional conflict was low (score = - 3.12) and weakly related to knowledge (R= - 0.25) or value elicitation (R= - 0.27). Black patients experienced higher decisional conflict than White patients (score = - 2.21 vs - 3.44). Despite knowledge scores being generally low, study patients experienced low decisional conflict regarding their decision to undergo lung cancer screening. Additional work is needed to optimize the quality and consistency of information presented to patients considering screening.

Predicting reoperation and readmission for head and neck free flap patients using machine learning.

BACKGROUND: To develop machine learning (ML) models predicting unplanned readmission and reoperation among patients undergoing free flap reconstruction for head and neck (HN) surgery. METHODS: Data were extracted from the 2012-2019 NSQIP database. eXtreme Gradient Boosting (XGBoost) was used to develop ML models predicting 30-day readmission and reoperation based on demographic and perioperative factors. Models were validated using 2019 data and evaluated. RESULTS: Four-hundred and sixty-six (10.7%) of 4333 included patients were readmitted within 30 days of initial surgery. The ML model demonstrated 82% accuracy, 63% sensitivity, 85% specificity, and AUC of 0.78. Nine-hundred and four (18.3%) of 4931 patients underwent reoperation within 30 days of index surgery. The ML model demonstrated 62% accuracy, 51% sensitivity, 64% specificity, and AUC of 0.58. CONCLUSION: XGBoost was used to predict 30-day readmission and reoperation for HN free flap patients. Findings may be used to assist clinicians and patients in shared decision-making and improve data collection in future database iterations.