Nicole Y. Xu

Purpose: Preoperative anesthesia evaluation is a crucial step in ensuring patient safety and optimizing perioperative care. A heterogenous patient population requiring varying levels of assessment often leads to inefficiencies and additional resource allocation. This study proposes using pre-trained language models to assist in triaging the appropriate degree of preoperative anesthesia evaluation for surgical patients. Methods: Retrospective institutional data were obtained from surgical patients evaluated at a single center preoperative anesthesia care clinic. The performance of four pre-trained language models (RoBERTa, BERT, ClinicalBERT, and PubMedBERT) in the classification of which patients would be appropriate for a nursing preoperative phone call versus in-person clinician evaluation was assessed using F1-score, area under the receiver operating characteristics curve (AUC), specificity, sensitivity, and average precision. For each pre-trained language model, three different data input combinations were assessed: (1) diagnosis codes (D); (2) clinical text data (N); and (3) diagnosis codes and clinical text (D + N). The data were split into training (75%) and test set (25%). Results: There were 1,761 unique patients, with an average of 12 notes per patient and a total of 46,922 clinical documents, included in the analysis. The AUC range between the four language models was highest in the D + N analyses (0.70 – 0.74), lower in the N analyses (0.58 – 0.73) and lowest in the D analyses (0.57 – 0.62). RoBERTa had the highest score compared to the other language models for all data types. Conclusions: Automating integrated analysis using pre-trained language models to aid in preoperative triaging could enhance accuracy and efficiency at scale, reducing manual review and provider burden.

Artificial Intelligence (AI) is rapidly transforming healthcare, and there is a critical need for a nuanced understanding of how AI is reshaping teaching, learning, and educational practice in medical education. This review aimed to map the literature regarding AI applications in medical education, core areas of findings, potential candidates for formal systematic review and gaps for future research. This rapid scoping review, conducted over 16 weeks, employed Arksey and O'Malley's framework and adhered to STORIES and BEME guidelines. A systematic and comprehensive search across PubMed/MEDLINE, EMBASE, and MedEdPublish was conducted without date or language restrictions. Publications included in the review spanned undergraduate, graduate, and continuing medical education, encompassing both original studies and perspective pieces. Data were charted by multiple author pairs and synthesized into various thematic maps and charts, ensuring a broad and detailed representation of the current landscape. The review synthesized 278 publications, with a majority (68%) from North American and European regions. The studies covered diverse AI applications in medical education, such as AI for admissions, teaching, assessment, and clinical reasoning. The review highlighted AI's varied roles, from augmenting traditional educational methods to introducing innovative practices, and underscores the urgent need for ethical guidelines in AI's application in medical education. The current literature has been charted. The findings underscore the need for ongoing research to explore uncharted areas and address potential risks associated with AI use in medical education. This work serves as a foundational resource for educators, policymakers, and researchers in navigating AI's evolving role in medical education. A framework to support future high utility reporting is proposed, the FACETS framework.

Background: The use of continuous subcutaneous insulin infusion (CSII) therapy in pregnancies affected by pregestational diabetes mellitus (DM) has generated mixed outcome data worthy of further investigation. This systematic review and meta-analysis aims to evaluate clinical outcomes associated with CSII versus multiple daily injections (MDIs) in pregnant persons with pregestational DM. Methods: A predefined, systematic, librarian-assisted search of MEDLINE (PubMed), Embase, Cochrane Library, Scopus, ClinicalTrials.gov, and World Health Organization International Clinical Trial Registry Platform (published from 2010 to 2022) yielded 3003 studies describing pregnancy outcomes associated with CSII and/or MDI for pregestational DM. The primary exposure was mode of insulin administration, with cesarean delivery and neonatal hypoglycemia as the primary maternal and neonatal outcomes, respectively. Secondary outcomes included hypertensive disorders of pregnancy, first and third-trimester glycemic control, large-for-gestational age (LGA) neonate, preterm birth, neonatal intensive care unit admission, need for respiratory support, hyperbilirubinemia, 5-minute Apgar <7, shoulder dystocia, and perinatal mortality. We calculated pooled odds ratios (OR) with 95% confidence intervals (CI) using random-effects models. Results: Among 39 eligible studies, 39% of the 5518 pregnancies included were exposed to CSII. Odds of cesarean delivery were higher with CSII (20 studies: 63% vs 56%, odds ratio [OR] 1.3 [95% confidence interval (CI) 1.2-1.5]), but we did not identify a difference in the odds of neonatal hypoglycemia (23 studies: 31% vs 34%, OR 1.1 [95% CI 0.9-1.5]). Among secondary outcomes, only the odds of LGA (20 studies: 47% vs 38%, OR 1.4 [95% CI 1.2-1.6]) were higher in individuals using CSII versus MDI. Conclusions: Use of CSII (vs MDI) for pregestational DM in pregnancy is associated with higher odds of cesarean delivery and delivery of an LGA neonate. Further evaluation of how CSII use may influence neonatal size and delivery route is warranted.

Background: The magnitude of the response of the diabetes professional community to the COVID-19 pandemic is not known. We aimed to examine diabetes technology research trends and resources offered by professional organizations during this period. Methods: We explored patterns of the response from the professional diabetes community to the pandemic by (1) systematically searching for articles related to diabetes, COVID-19, and diabetes technologies; (2) examining publication trends of research protocols (clinicaltrials.gov) and preprints (medRxiv); and (3) reviewing online resources from professional organizations including our website (COVIDinDiabetes.org; an Emory University—Diabetes Technology Society collaboration). Results: We identified 492 articles published between December 2019 and December 2022 meeting our inclusion criteria. Telemedicine and continuous glucose monitoring were the most common reported technologies from most parts of the world. The largest number of preprint articles was published in 2020, with a decline in 2021 and 2022. The number of research protocols related to COVID-19 was the highest in 2020 and declined in 2021 and 2022. Resources from organizations included protocols adapted to treat patients with diabetes and COVID-19, training programs, emergency preparedness, and literature on diabetes and COVID-19. On our website (COVIDinDiabetes.org), there were 12 236 visits and 18 149 pageviews, with 1.6 actions per visits, with most visits coming from North America (N = 7233, 54.2%), South America (N = 2663, 21.8%), and Europe (N = 1219). Conclusions: We conclude that the COVID-19 pandemic promoted unprecedented global research productivity related to diabetes and COVID-19 and that the transition to the use of technology resources has been evident during this period.

Spinal cord stimulation (SCS) technology has been recently approved by the US Food and Drug Administration (FDA) for painful diabetic neuropathy (PDN). The treatment involves surgical implantation of electrodes and a power source that delivers electrical current to the spinal cord. This treatment decreases the perception of pain in many chronic pain conditions, such as PDN. The number of patients with PDN treated with SCS and the amount of data describing their outcomes is expected to increase given four factors: (1) the large number of patients with this diagnosis, (2) the poor results that have been obtained for pain relief with pharmacotherapy and noninvasive non-pharmacotherapy, (3) the results to date with investigational SCS technology, and (4) the recent FDA approval of systems that deliver this treatment. Whereas traditional SCS replaces pain with paresthesias, a new form of SCS, called high-frequency 10-kHz SCS, first used for pain in 2015, can relieve PDN pain without causing paresthesias, although not all patients experience pain relief by SCS. This article describes (1) an overview of SCS technology, (2) the use of SCS for diseases other than diabetes, (3) the use of SCS for PDN, (4) a comparison of high-frequency 10-kHz and traditional SCS for PDN, (5) other SCS technology for PDN, (6) deployment of SCS systems, (7) barriers to the use of SCS for PDN, (8) risks of SCS technology, (9) current recommendations for using SCS for PDN, and (10) future developments in SCS.

Painful diabetic neuropathy is a common vexing problem for people with diabetes and a costly problem for society. The pathophysiology is not well understood, and no safe and effective mechanistically-based treatment has been identified. Poor glycemic control is a risk factor for painful diabetic neuropathy. Excessive intraneuronal glucose in people with diabetes can be shunted away from physiological glycolysis into multiple pathological pathways associated with neuropathy and pain. The first three treatments that are traditionally offered consist of risk factor reduction, lifestyle modifications, and pharmacological therapy, which includes only three drugs that are approved for this indication by the United States Food and Drug Administration. All of these traditional treatments are often inadequate for relieving neuropathic pain, and thus, new approaches are needed. Modern devices based on neuromodulation technology, which act directly on the nervous system, have been recently cleared by the United States Food and Drug Administration for painful diabetic neuropathy and offer promise as next-in-line therapy when traditional therapies fail.

The annual Virtual Hospital Diabetes Meeting was hosted by Diabetes Technology Society on April 1 and April 2, 2022. This meeting brought together experts in diabetes technology to discuss various new developments in the field of managing diabetes in hospitalized patients. Meeting topics included (1) digital health and the hospital, (2) blood glucose targets, (3) software for inpatient diabetes, (4) surgery, (5) transitions, (6) coronavirus disease and diabetes in the hospital, (7) drugs for diabetes, (8) continuous glucose monitoring, (9) quality improvement, (10) diabetes care and educatinon, and (11) uniting people, process, and technology to achieve optimal glycemic management. This meeting covered new technology that will enable better care of people with diabetes if they are hospitalized.

Background: Continuous glucose monitor (CGM) systems were originally intended only for people with diabetes. Recently, there has been interest in monitoring glucose concentrations in a variety of other situations. As data accumulate to support the use of CGM systems in additional states unrelated to diabetes, the use of CGM systems is likely to increase accordingly. Methods: PubMed and Google Scholar were searched for articles about the use of CGM in individuals without diabetes. Relevant articles that included sufficient details were queried to identify what cohorts of individuals were adopting CGM use and to define trends of use. Results: Four clinical user cases were identified: (1) metabolic diseases related to diabetes with a primary dysregulation of the insulin-glucose axis, (2) metabolic diseases without a primary pathophysiologic derangement of the insulin-glucose axis, (3) health and wellness, and (4) elite athletics. Seven trends in the use of CGM systems in people without diabetes were idenfitied which pertained to both FDA-cleared medical grade products as well as anticipated future products, which may be regulated differently based on intended populations and indications for use. Conclusions: Wearing a CGM has been used not only for diabetes, but with a goal of improving glucose patterns to avoid diabetes, improving mental or physical performance, and promoting motivate healthy behavioral changes. We expect that clinicians will become increasingly aware of (1) glycemic patterns from CGM tracings that predict an increased risk of diabetes, (2) specific metabolic glucotypes from CGM tracings that predict an increased risk of diabetes, and (3) new genetic and genomic biomarkers in the future.