Ethical Considerations in the Use of Artificial Intelligence and Machine Learning in Health Care: A Comprehensive Review

Cureus, Год журнала: 2024, Номер unknown

Опубликована: Июнь 15, 2024

Artificial intelligence (AI) and machine learning (ML) technologies are revolutionizing health care by offering unprecedented opportunities to enhance patient care, optimize clinical workflows, advance medical research. However, the integration of AI ML into healthcare systems raises significant ethical considerations that must be carefully addressed ensure responsible equitable deployment. This comprehensive review explored multifaceted surrounding use in including privacy data security, algorithmic bias, transparency, validation, professional responsibility. By critically examining these dimensions, stakeholders can navigate complexities while safeguarding welfare upholding principles. embracing best practices fostering collaboration across interdisciplinary teams, community harness full potential usher a new era personalized data-driven prioritizes well-being equity.

Язык: Английский

---

AI-based differential diagnosis of dementia etiologies on multimodal data

Nature Medicine, Год журнала: 2024, Номер 30(10), С. 2977 - 2989

Опубликована: Июль 4, 2024

Abstract Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an artificial intelligence (AI) model that harnesses broad array data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations multimodal neuroimaging, identify the etiologies contributing individuals. The study, drawing on 51,269 participants 9 independent, geographically diverse datasets, facilitated identification 10 distinct etiologies. It aligns diagnoses with similar strategies, ensuring robust predictions even incomplete data. Our achieved microaveraged area under receiver operating characteristic curve (AUROC) 0.94 classifying individuals normal cognition, mild cognitive impairment dementia. Also, AUROC was 0.96 differentiating demonstrated proficiency addressing mixed cases, mean 0.78 two co-occurring pathologies. In randomly selected subset 100 neurologist assessments augmented by our AI exceeded neurologist-only 26.25%. Furthermore, aligned biomarker evidence its associations different proteinopathies were substantiated through postmortem findings. framework has potential be integrated as screening tool clinical settings drug trials. Further prospective studies are needed confirm ability improve patient care.

Язык: Английский

---

Artificial Intelligence in Drug Discovery and Development

Springer eBooks, Год журнала: 2024, Номер unknown, С. 1461 - 1498

Опубликована: Янв. 1, 2024

Язык: Английский

---

Artificial Intelligence in Drug Discovery and Development

Springer eBooks, Год журнала: 2023, Номер unknown, С. 1 - 38

Опубликована: Янв. 1, 2023

Язык: Английский

---

AI-enhanced patient-centric clinical trial design

AIP conference proceedings, Год журнала: 2025, Номер 3262, С. 020020 - 020020

Опубликована: Янв. 1, 2025

---

The role of artificial intelligence in electrodiagnostic and neuromuscular medicine:Currentstate and future directions

Muscle & Nerve, Год журнала: 2023, Номер 69(3), С. 260 - 272

Опубликована: Дек. 27, 2023

Abstract The rapid advancements in artificial intelligence (AI), including machine learning (ML), and deep (DL) have ushered a new era of technological breakthroughs healthcare. These technologies are revolutionizing the way we utilize medical data, enabling improved disease classification, more precise diagnoses, better treatment selection, therapeutic monitoring, highly accurate prognostication. Different ML DL models been used to distinguish between electromyography signals normal individuals those with amyotrophic lateral sclerosis myopathy, accuracy ranging from 67% 99.5%. also successfully applied neuromuscular ultrasound, use segmentation techniques achieving diagnostic at least 90% for nerve entrapment disorders, 87% inflammatory myopathies. Other successful AI applications include prediction response, prognostication intensive care unit admissions patients myasthenia gravis. Despite these remarkable strides, significant knowledge, attitude, practice gaps persist, within field electrodiagnostic medicine. In this narrative review, highlight fundamental principles draw parallels intricacies human brain networks. Specifically, explore immense potential that holds studies, other aspects While there exciting possibilities future, it is essential acknowledge understand limitations take proactive steps mitigate challenges. This collective endeavor advancement healthcare through strategic responsible integration technologies.

Язык: Английский

---

AI-based differential diagnosis of dementia etiologies on multimodal data

medRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Фев. 11, 2024

Abstract Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an AI model that harnesses broad array data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations, multimodal neuroimaging, identify the etiologies contributing individuals. The study, drawing on 51, 269 participants 9 independent, geographically diverse datasets, facilitated identification 10 distinct etiologies. It aligns diagnoses with similar strategies, ensuring robust predictions even incomplete data. Our achieved micro-averaged area under receiver operating characteristic curve (AUROC) 0.94 classifying individuals normal cognition, mild cognitive impairment dementia. Also, AUROC was 0.96 differentiating demonstrated proficiency addressing mixed cases, mean 0.78 two cooccurring pathologies. In randomly selected subset 100 neurologist assessments augmented by our exceeded neurologist-only evaluations 26.25%. Furthermore, aligned biomarker evidence its associations different proteinopathies were substantiated through postmortem findings. framework has potential be integrated as screening tool various clinical settings drug trials, promising implications person-level management.

Язык: Английский

---

Laboratory Data as a Potential Source of Bias in Healthcare Artificial Intelligence and Machine Learning Models

Annals of Laboratory Medicine, Год журнала: 2024, Номер 45(1), С. 12 - 21

Опубликована: Окт. 24, 2024

Artificial intelligence (AI) and machine learning (ML) are anticipated to transform the practice of medicine. As one largest sources digital data in healthcare, laboratory results can strongly influence AI ML algorithms that require large sets healthcare for training. Embedded bias introduced into models not only has disastrous consequences quality care but also may perpetuate exacerbate health disparities. The lack test harmonization, which is defined as ability produce comparable same interpretation irrespective method or instrument platform used result, introduce aggregation with potential adverse outcomes patients. Limited interoperability at technical, syntactic, semantic, organizational levels a source embedded limits accuracy generalizability algorithmic models. Population-specific issues, such inadequate representation clinical trials inaccurate race attribution, affect erroneous conclusions based on literature.

Язык: Английский

---

Future prospective of AI in drug discovery

Advances in pharmacology, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Язык: Английский

---

Machine learning and clinical EEG data for multiple sclerosis: A systematic review

Artificial Intelligence in Medicine, Год журнала: 2025, Номер 166, С. 103116 - 103116

Опубликована: Апрель 29, 2025

Multiple Sclerosis (MS) is a chronic neuroinflammatory disease of the Central Nervous System (CNS) in which body's immune system attacks and destroys myelin sheath that protects nerve fibers, leading to wide range debilitating symptoms causing disruption axonal signal transmission. Accurate prediction, diagnosis, monitoring treatment (PDMT) MS are essential improve patient outcomes. Recent advances neuroimaging technologies, particularly electroencephalography (EEG), combined with machine learning (ML) techniques - including Deep Learning (DL) models offer promising avenues for enhancing management. This systematic review synthesizes existing research on application ML DL EEG data MS. It explores methodologies used, focus architectures such as Convolutional Neural Networks (CNNs) hybrid models, highlights recent advancements technologies have significantly improved diagnosis monitoring. The addresses challenges potential biases using ML-based analysis Strategies mitigate these challenges, advanced preprocessing techniques, diverse training datasets, cross-validation methods, explainable Artificial Intelligence (AI), discussed. Finally, paper outlines future applications trends underscores transformative ML-enhanced improving management, providing insights into directions overcome limitations further clinical practice.

Язык: Английский

---