Development and Validation of a Machine Learning-Based Decision Support Tool for Residency Applicant Screening and Review.

PURPOSE: Residency programs face overwhelming numbers of residency applications, limiting holistic review. Artificial intelligence techniques have been proposed to address this challenge but have not been created. Here, a multidisciplinary team sought to develop and validate a machine learning (ML)-based decision support tool (DST) for residency applicant screening and review. METHOD: Categorical applicant data from the 2018, 2019, and 2020 residency application cycles (n = 8,243 applicants) at one large internal medicine residency program were downloaded from the Electronic Residency Application Service and linked to the outcome measure: interview invitation by human reviewers (n = 1,235 invites). An ML model using gradient boosting was designed using training data (80% of applicants) with over 60 applicant features (e.g., demographics, experiences, academic metrics). Model performance was validated on held-out data (20% of applicants). Sensitivity analysis was conducted without United States Medical Licensing Examination (USMLE) scores. An interactive DST incorporating the ML model was designed and deployed that provided applicant- and cohort-level visualizations. RESULTS: The ML model areas under the receiver operating characteristic and precision recall curves were 0.95 and 0.76, respectively; these changed to 0.94 and 0.72, respectively, with removal of USMLE scores. Applicants' medical school information was an important driver of predictions-which had face validity based on the local selection process-but numerous predictors contributed. Program directors used the DST in the 2021 application cycle to select 20 applicants for interview that had been initially screened out during human review. CONCLUSIONS: The authors developed and validated an ML algorithm for predicting residency interview offers from numerous application elements with high performance-even when USMLE scores were removed. Model deployment in a DST highlighted its potential for screening candidates and helped quantify and mitigate biases existing in the selection process. Further work will incorporate unstructured textual data through natural language processing methods.

Neuropsychological Function Response to Nocturnal Blue Light Blockage in Individuals With Symptoms of Insomnia: A Pilot Randomized Controlled Study.

OBJECTIVES: Insomnia is associated with neuropsychological dysfunction. Evidence points to the role of nocturnal light exposure in disrupted sleep patterns, particularly blue light emitted through smartphones and computers used before bedtime. This study aimed to test whether blocking nocturnal blue light improves neuropsychological function in individuals with insomnia symptoms. METHODS: This study used a randomized, placebo-controlled crossover design. Participants were randomly assigned to a 1-week intervention with amber lenses worn in wrap-around frames (to block blue light) or a 1-week intervention with clear lenses (control) and switched conditions after a 4-week washout period. Neuropsychological function was evaluated with tests from the NIH Toolbox Cognition Battery at three time points: (1) baseline (BL), (2) following the amber lenses intervention, and (3) following the clear lenses intervention. Within-subjects general linear models contrasted neuropsychological test performance following the amber lenses and clear lenses conditions with BL performance. RESULTS: Fourteen participants (mean(standard deviation, SD): age = 46.5(11.4)) with symptoms of insomnia completed the protocol. Compared with BL, individuals performed better on the List Sorting Working Memory task after the amber lenses intervention, but similarly after the clear lenses intervention (F = 5.16; p = .014; η2 = 0.301). A similar pattern emerged on the Pattern Comparison Processing Speed test (F = 7.65; p = 0.002; η2 = 0.370). Consideration of intellectual ability indicated that treatment with amber lenses "normalized" performance on each test from approximately 1 SD below expected performance to expected performance. CONCLUSIONS: Using a randomized, placebo-controlled crossover design, we demonstrated improvement in processing speed and working memory with a nocturnal blue light blocking intervention among individuals with insomnia symptoms. (JINS, 2019, 25, 668-677).