Patient and clinician acceptability of automated extraction of social drivers of health from clinical notes in primary care

Journal of the American Medical Informatics Association, Год журнала: 2025, Номер unknown

Опубликована: Март 14, 2025

Artificial Intelligence (AI)-based approaches for extracting Social Drivers of Health (SDoH) from clinical notes offer healthcare systems an efficient way to identify patients' social needs, yet we know little about the acceptability this approach patients and clinicians. We investigated patient clinician through interviews. interviewed primary care experiencing needs (n = 19) clinicians 14) their "SDoH autosuggest," AI-based SDoH notes. presented storyboards depicting asked participants rate discuss rationale. Participants rated autosuggest moderately acceptable (mean 3.9/5 patients; mean 3.6/5 clinicians). Patients' ratings varied across domains, with substance use most employment least acceptable. Both groups raised concern information integrity, actionability, impact on interactions relationships, privacy. In addition, transparency, autonomy, potential harm, whereas usability. Despite reporting moderate envisioned approach, expressed multiple concerns AI that extract SDoH. emphasized need high-quality data, non-intrusive presentation methods, clear communication strategies regarding sensitive needs. Findings underscore importance engaging mitigate unintended consequences when integrating into care. Although like hold promise efficiently identifying notes, they must also account ensure these are do not undermine trust.

Язык: Английский

---

Utilizing Technology to Enhance Cancer Education and Support Services

IGI Global eBooks, Год журнала: 2025, Номер unknown, С. 203 - 220

Опубликована: Март 7, 2025

This chapter explores how technology is transforming cancer education and support for survivors healthcare providers. It covers innovations like mobile apps, telemedicine, AI, virtual reality, social media, highlighting their impact on sharing information, offering support, delivering care. Through literature reviews, case studies, research, the assesses effectiveness of these tools in improving patient outcomes, enhancing provider education, creating a more connected care community. also addresses challenges digital literacy, accessibility, data privacy, insights professionals, policymakers, tech developers to optimize

Язык: Английский

---

Integrating Multiplex Immunohistochemistry and Machine Learning for Glioma Subtyping and Prognosis Prediction

MedComm, Год журнала: 2025, Номер 6(5)

Опубликована: Апрель 22, 2025

ABSTRACT Glioma subtyping is crucial for treatment decisions, but traditional approaches often fail to capture tumor heterogeneity. This study proposes a novel framework integrating multiplex immunohistochemistry (mIHC) and machine learning glioma prognosis prediction. 185 patient samples from the Huashan hospital cohort were stained using multi‐label mIHC panel analyzed with an AI‐based auto‐scanning system calculate cell ratios determine proportion of positive cells various markers. Patients divided into two cohorts (training: N = 111, testing: 74), model was then developed validated subtype classification The identified distinct subtypes significant differences in prognosis, clinical characteristics, molecular profiles. high‐risk subtype, associated older age, poorer outcomes, astrocytoma/glioblastoma, higher grades, elevated mesenchymal scores, inhibitory immune microenvironment, exhibited IDH wild‐type, 1p19q non‐codeletion, MGMT promoter unmethylation, suggesting chemotherapy resistance. Conversely, low‐risk characterized by younger better astrocytoma/oligodendroglioma, lower favorable profiles (IDH mutation, codeletion, methylation), indicated sensitivity. mIHC‐based enables rapid classification, facilitating tailored strategies accurate prediction, potentially improving management outcomes.

Язык: Английский

---

Comparative Analysis of Generative Artificial Intelligence Systems in Solving Clinical Pharmacy Problems：A Commentary on AI's Performance on the Clinical Pharmacy (Preprint)

Опубликована: Апрель 16, 2025

In recent years, the implementation of artificial intelligence (AI) in health care is progressively transforming medical fields.However, there remains a gap between technological potential and practical application, necessitating establishment scientific evaluation system.Despite some existing research beginning to conduct clinical application assessments generative AI dialogue systems, these efforts are largely limited testing individual models on single tasks, lacking horizontal comparative analysis across multiple validation continuous decision chains real scenarios.As systems play an increasingly extensive role field Medicine Pharmacy, we need more explore this area. To systematically evaluate compare performance eight mainstream both domestic international, four core pharmacy practice scenarios: medication consultation, education, prescription review, case with pharmaceutical care. This study aims quantitatively assess their capabilities addressing common problems. Assessment questions were extracted from consultation clinic records, cases, pharmacist standardized training examination databases. Three researchers tested different same day using "inquiry prompts." A double-blind scoring design was employed, six experienced pharmacists backgrounds evaluating responses 0-10 scale dimensions: accuracy, rigor, applicability, logical coherence, conciseness, universality. Statistical used one-way variance (ANOVA) score differences comparison tests for significant results, intraclass correlation coefficient (ICC) calculations inter-rater consistency.Systematic descriptive evaluations AI-generated also conducted. DeepSeek-R1 demonstrated best overall all task categories. Qwen, GPT-4o, Claude-3.5-Sonnet, Gemini-1.5-Pro performed slightly inferior DeepSeek-R1. Doubao Kimi showed inconsistent performance, while ERNIE Bot poorest. Comprehensive indicated that still have certain limitations should be as reference tools rather than independent decision-making bases. Inter-rater consistency good agreement (ICC>0.75) review. However, lowest level (ICC=0.70) observed assessing conciseness care, reflecting cognitive among raters regarding standards complex issues. The model demonstrates supportive tool practice. overall, current require systematic improvement refinement ability handle multidimensional none

Язык: Английский

---

Artificial Intelligence in Vascular Neurology: Applications, Challenges, and a Review of AI Tools for Stroke Imaging, Clinical Decision Making, and Outcome Prediction Models

Current Neurology and Neuroscience Reports, Год журнала: 2025, Номер 25(1)

Опубликована: Май 9, 2025

Язык: Английский

---