The George Washington University School of Medicine and Health Sciences (abbreviated as GW Medical School or SMHS) is the professional medical school of the George Washington University, in Washington, D.C. SMHS is one of the most selective medical schools in the United States based on the number of applicants.
Patients with Parkinson’s disease, often accompanied by osteoporosis, have a higher risk of osteoporotic fractures. However, a large cohort study assessing the preventive effect of bisphosphonates for osteoporotic fracture on patients with Parkinson’s disease is scarce. This retrospective cohort study included nearly whole population of South Korea during 2002 to 2018 from the national patient database of Korea National Health Insurance Service. Of 608,202 patients with Parkinson’s disease, a total of 7335 patients were selected for index regression model and 5806 patients were further selected for landmark model of 1 year after diagnosis of osteoporosis. Occurrence of osteoporotic fracture was set defined as outcome variable for both regression models. Regression analysis with a landmark model one year after the index diagnosis of osteoporosis showed that the use of bisphosphonates had a significant preventive effect on osteoporotic fractures (hazard ratio [HR] 0.83, 95
BACKGROUND:Abrocitinib treatment previously showed clinically meaningful improvements in skin signs and symptoms in patients with moderate-to-severe atopic dermatitis (AD) up to 48 weeks, with a manageable safety profile. OBJECTIVE:To further evaluate long-term efficacy and patient-reported outcomes (PROs) with abrocitinib in patients with moderate-to-severe AD up to 112 weeks of treatment. METHODS:JADE EXTEND (NCT03422822), an ongoing, phase 3, long-term extension study of previous abrocitinib trials, includes patients from qualifying trials MOA (NCT03915496), MONO-1 (NCT03349060), MONO-2 (NCT03575871), COMPARE (NCT03720470), TEEN (NCT03796676) and DARE (NCT04345367). Efficacy assessments include the proportion of patients achieving IGA score of 0 or 1 (IGA 0/1); absolute Eczema Area and Severity Index ≤7 (EASI ≤ 7), ≥75%/≥90%/100% improvement in EASI (EASI-75/-90/-100); ≥4-point improvement in Peak Pruritus Numerical Rating Scale (PP-NRS4); PP-NRS score of 0 or 1 (PP-NRS 0/1); EASI-90 + PP-NRS 0/1. PROs include Dermatology Life Quality Index score of 0 or 1 (DLQI 0/1), Children's Dermatology Life Quality Index score of 0 or 1 (CDLQI 0/1), Patient Global Assessment score of 0 or 1 (PtGA 0/1), and ≥4-point improvement in Patient-Oriented Eczema Measure (POEM ≥ 4). Data cut-off: 5 September 2022. RESULTS:At data cut-off, 403 (42%) and 425 (56%) patients received abrocitinib 200 mg and abrocitinib 100 mg, respectively, for ≥112 weeks. Week 112 efficacy responses were IGA 0/1 57% and 52%; EASI ≤7 81% and 77%; EASI-75 84% and 78%; EASI-90 61% and 54%; EASI-100 26% and 21%; PP-NRS4 70% and 58%; PP-NRS 0/1 43% and 33%; EASI-90 + PP-NRS 0/1 37% and 30%. Week 112 PROs were DLQI 0/1 43% and 41%; CDLQI 0/1 59% and 39%; PtGA 0/1 75% and 72%; POEM ≥4 88% and 81%. CONCLUSIONS:Abrocitinib treatment resulted in clinically meaningful outcomes and improvements in PROs, including high-threshold endpoints, in substantial proportions of patients with moderate-to-severe AD up to 2 years.
While large language models (LLMs) have shown promise as learning tools for medical education, their reported accuracy on multiple-choice questions (MCQs) varies widely across studies, necessitating synthesis. This meta-analysis synthesizes LLM accuracy on text-based MCQs from biomedical disciplines and USMLE Step 1-level content and explores study characteristics that moderate LLM performance. Studies published between January 1, 2022, and August 5, 2025, were identified using seven databases. Titles and abstracts were screened against eligibility criteria, which included testing LLM performance on text-based MCQs in English. Extracted data included accuracy rates (correct responses/total questions) and study characteristics. A random-effects proportional meta-analysis was used to calculate the pooled accuracy of LLM versions across studies, and a meta-regression tested the effects of study characteristics on accuracy. The final analysis included 41 articles comprising 207 proportions. Newer OpenAI models (GPT-4o through GPT o1 = 90%; GPT-4 = 82%) demonstrated significantly higher accuracy than earlier versions (GPT-3/3.5 = 59%; p < 0.001). Model version explained 56% of the total variance (p < 0.001). The pooled accuracy of newer Anthropic (Claude 3 through 3.7 = 86%) and DeepSeek (R1 = 86%) versions fell within the range of performance for the newer OpenAI versions, while newer Google (Gemini series = 76%) and Microsoft (Copilot = 69%) versions demonstrated lower performance. A discipline-level analysis demonstrated that OpenAI's GPT-4 through GPT-o1 models achieved pooled accuracy rates between 83 and 87% for most disciplines, including the anatomical sciences. These findings demonstrate that newer LLMs are achieving higher accuracy than older models on biomedical and Step 1-level MCQs, supporting their potential integration as supplementary study tools in foundational biomedical education.
OBJECTIVE:Artificial intelligence (AI) and robotics are transforming neurosurgical care; however, the application of these technologies in low- and middle-income countries (LMICs), where the burden of neurosurgical diseases is highest, is not well-characterized. In this scoping review, we sought to map the current landscape of AI and robotics in global neurosurgery to identify themes and gaps that can catalyze the expansion of safe neurosurgical care worldwide. METHODS:A systematic search was conducted across MEDLINE/PubMed, Embase, Web of Science, and IEEE Xplore. This review followed the Population, Concept, and Context (PCC) framework, including studies that utilized AI or robotics for neurosurgical diagnosis, decision-making, or treatment in global developing or low-resource settings. RESULTS:A total of 22 studies were identified, with a rapid acceleration in research observed since 2023. Geographically, the majority of applications were in Africa (n = 13), followed by Asia (n = 6) and South America (n = 4). Trauma was the predominant field (n = 11), where machine learning was frequently used for triage and prognostication. Other common applications included AI for brain tumor classification and segmentation, automated stroke detection, and image-guided surgical robotics, with models trained on local data outperforming those imported from other settings. Nineteen of the 22 studies included coauthors from the region of application, and 16 included coauthors from a HIC. CONCLUSION:AI and robotics are demonstrating increasing utility in supporting global neurosurgery by augmenting workforce capacity and improving efficiency of care. However, success depends heavily on data representation. Future work should focus on supporting local research capacity and leadership and expanding applications to postoperative care.
Background The Predicting Risk of cardiovascular disease (CVD) EVENTs (PREVENT) equations accurately estimate CVD risk, but how different combinations of ages and risk factors translate into risk estimates, including combinations required to meet clinically meaningful risk thresholds, is not intuitive. Methods We calculated sex-specific estimatesion of 10-year risk of total CVD (composite of atherosclerotic CVD and heart failure) for a hypothetical person between the ages of 30–79 years with the PREVENT-CVD base equations and varied individual or multiple risk factors simultaneously, with all other risk factor levels set at sex-specific, population-based average values. Next, we examined how specific age and risk factor combinations would exceed clinically-meaningful thresholds. Secondary analyses included the PREVENT-CVD equations with add-on predictors and the 30-year PREVENT-CVD equations. Results Ten-year risk estimates for a hypothetical person with average risk factor levels ranged from 0.3 %-17.4 % for a female and 0.7 %-22.8 % for a male with the PREVENT-CVD base equations, exceeding the clinically-relevant risk threshold of ≥7.5 % at age 68 years if female and at age 63 years if male. Additionally, a hypothetical person with Stage 3 CKD and diabetes would exceed the 10-year risk threshold of ≥7.5 % with the PREVENT-CVD equations at age 43 years if female and at age 36 years if male. Similar patterns were observed with add-on predictor equations. With average risk factor levels, 30-year PREVENT-CVD risk estimates ranged from 2.5 %-20.5 % for a female and 4.8 %-26.5 % for a male. Conclusions These results can support clinicians and patients in the interpretability of the PREVENT equations and can inform clinician-patient discussions on preventive efforts.