Hasil untuk "Medical philosophy. Medical ethics"

Haoan Jin, Han Ying, Jiacheng Ji et al.

Recent advances in large language models have enabled their application to a range of healthcare tasks. However, aligning LLMs with the nuanced demands of medical ethics, especially under complex real world scenarios, remains underexplored. In this work, we present MedES, a dynamic, scenario-centric benchmark specifically constructed from 260 authoritative Chinese medical, ethical, and legal sources to reflect the challenges in clinical decision-making. To facilitate model alignment, we introduce a guardian-in-the-loop framework that leverages a dedicated automated evaluator (trained on expert-labeled data and achieving over 97% accuracy within our domain) to generate targeted prompts and provide structured ethical feedback. Using this pipeline, we align a 7B-parameter LLM through supervised fine-tuning and domain-specific preference optimization. Experimental results, conducted entirely within the Chinese medical ethics context, demonstrate that our aligned model outperforms notably larger baselines on core ethical tasks, with observed improvements in both quality and composite evaluation metrics. Our work offers a practical and adaptable framework for aligning LLMs with medical ethics in the Chinese healthcare domain, and suggests that similar alignment pipelines may be instantiated in other legal and cultural environments through modular replacement of the underlying normative corpus.

Chutian Huang, Dake Cao, Jiacheng Ji et al.

Background: While Large Language Models (LLMs) have achieved widespread adoption, malicious prompt engineering specifically "jailbreak attacks" poses severe security risks by inducing models to bypass internal safety mechanisms. Current benchmarks predominantly focus on public safety and Western cultural norms, leaving a critical gap in evaluating the niche, high-risk domain of medical ethics within the Chinese context. Objective: To establish a specialized jailbreak evaluation framework for Chinese medical ethics and to systematically assess the defensive resilience and ethical alignment of seven prominent LLMs when subjected to sophisticated adversarial simulations. Methodology: We evaluated seven prominent models (e.g., GPT-5, Claude-Sonnet-4-Reasoning, DeepSeek-R1) using a "role-playing + scenario simulation + multi-turn dialogue" vector within the DeepInception framework. The testing focused on eight high-risk themes, including commercial surrogacy and organ trading, utilizing a hierarchical scoring matrix to quantify the Attack Success Rate (ASR) and ASR Gain. Results: A systemic collapse of defenses was observed, whereas models demonstrated high baseline compliance, the jailbreak ASR reached 82.1%, representing an ASR Gain of over 80 percentage points. Claude-Sonnet-4-Reasoning emerged as the most robust model, while five models including Gemini-2.5-Pro and GPT-4.1 exhibited near-total failure with ASRs between 96% and 100%. Conclusions: Current LLMs are highly vulnerable to contextual manipulation in medical ethics, often prioritizing "helpfulness" over safety constraints. To enhance security, we recommend a transition from outcome to process supervision, the implementation of multi-factor identity verification, and the establishment of cross-model "joint defense" mechanisms.

Dawan Jamal Hawezy

Abstract Introduction Patient satisfaction is a significant measure of healthcare service quality as the patient is the center of any surgical procedure. Patient satisfaction refers to the extent to which a patient’s expectations of optimal care align with their perception of the care received. Patient satisfaction during informed consent is enhanced when written informed consent is accompanied by verbal consent in the preoperative period. Satisfied patients are more inclined to adhere to therapy, engage actively in their care, utilize healthcare services, willingly partake in decision-making, and remain with a healthcare provider. This research examines the practical and ethical considerations of obtaining informed consent during surgical procedures. To better understand and make informed decisions, this study aims to assess the efficacy of present consent methods and pinpoint obstacles patients encounter. Methodology A cross-sectional study was conducted from April to December 2024. Data were gathered by second-year students from Koya University’s Faculty of General Medicine by interviewing postoperative patients who had undergone general surgical procedures. The results were entered into a Google form and analyzed using SPSS27. Results In interviews with participants, 430 out of 572 patients (75%) indicated trust in their surgeons performing the surgery, while 525 patients (91%) expressed respect for their surgeons’ opinions. However, 41% (239 patients) reported not reading the informed consent form, and a similar percentage denied that the details of the form were explained by the medical staff responsible for the surgery as there are some medical terms or situations in the form that are challenging to assume if not explained. Conclusion Compared to others, participants with a higher educational level sought extensive time from the responsible surgeons to discuss every detail of the surgery before signing the informed consent, with a statistically significant difference observed. A similar difference was noted between private and public hospitals.

Jingjing WANG, Li KE

A review and discussion of related research on music therapy in China reveals four primary areas of focus: operational methods, therapeutic effects, mechanisms of action, cultural and philosophical dimensions. These studies exhibit diversity and localization in terms of research perspectives, methods, theoretical construction, and operational techniques, gradually improving and expanding the discipline system of music therapy in China. Although some achievements have been made in current research, some studies still face challenges in theoretical basis, research methods, and research perspectives. Future research should explore more rigorous experimental paradigms and diverse theoretical frameworks to further promote the development of music therapy as a localized discipline in China.

Yafang QIAN, Zhenliang WANG

The rapid development of intelligent healthcare is profoundly influencing pre-diagnosis, intra-diagnosis, and post-diagnosis stages. However, as medical advancements grow, so do associated risks. Proactive measures are preferable to reactive remedies. Focusing on the most controversial issues in intelligent healthcare: the subject of intelligent healthcare responsibility and the principle of liability, the authors view the ambiguous subject of intelligent healthcare responsibility and the principle of liability as the main obstacles to the effective legal regulation of intelligent healthcare. Based on an analysis of the criteria for selecting the subject of intelligent healthcare liability and the principle of liability, the authors hence propose improvements such as reasonable allocation of the burden of proof, the adjustments and alignment of the existing laws, as well as the optimisation of evaluation standards and institutions.

Maroun Badr, Fabiano Nigris

En las últimas décadas, el comportamiento sexual de las mujeres ha cambiado al invocar el principio de autodeterminación/autonomía y la gestión de sus propios cuerpos. La autodeterminación se ejerce a través de la libertad de elección. La capacidad de decidir requiere madurez, que se adquiere en la edad adulta, generalmente a los 18 años. El inicio temprano de la actividad sexual a una edad temprana puede resultar en embarazos no planificados y no deseados. Numerosos factores llevan a las menores a optar por la interrupción del embarazo a través de los servicios de salud estatales, lo que cuestiona la validez de dicha decisión. El objetivo de este artículo es realizar un análisis comparativo de esta práctica en Francia e Italia, destacando los desafíos y las cuestiones bioéticas que conlleva.

Dingshan FENG, Jian HAN

China's welfare and ethical review system for laboratory animals faces significant gaps in legislative hierarchy, clarity of responsibilities for review bodies, procedural standardization, and process supervision. Current regulations lack dedicated provisions for animal welfare, leading to ambiguous roles among review entities, insufficiently standardized supervision procedures, and an absence of independent third-party evaluation mechanisms and unified oversight standards. Consequently, ethical reviews often become mere formalities, with supervision of the implementation phase of animal experiments being particularly weak. This paper proposes a regulatory framework focused on the implementation phase：at the legislative level, revising the Regulations on the Management of Laboratory Animals to include explicit animal welfare sections, elevating the legal status of the review system; at the standards level, refining supervision procedures, establishing an animal welfare evaluation indicator system and cross-supervision mechanism, and strengthening the implementation of national standards; at the technological level, enhancing ethics and operational training, promoting information-based monitoring technologies to enable quantifiable dynamic supervision of the entire animal experimentation process.

Joshua Hatherley, Anders Søgaard, Angela Ballantyne et al.

Federated learning (FL) is a machine learning approach that allows multiple devices or institutions to collaboratively train a model without sharing their local data with a third-party. FL is considered a promising way to address patient privacy concerns in medical artificial intelligence. The ethical risks of medical FL systems themselves, however, have thus far been underexamined. This paper aims to address this gap. We argue that medical FL presents a new variety of opacity -- federation opacity -- that, in turn, generates a distinctive double black box problem in healthcare AI. We highlight several instances in which the anticipated benefits of medical FL may be exaggerated, and conclude by highlighting key challenges that must be overcome to make FL ethically feasible in medicine.

Blake VanBerlo, Alexander Wong, Jesse Hoey et al.

Data augmentation is a central component of joint embedding self-supervised learning (SSL). Approaches that work for natural images may not always be effective in medical imaging tasks. This study systematically investigated the impact of data augmentation and preprocessing strategies in SSL for lung ultrasound. Three data augmentation pipelines were assessed: (1) a baseline pipeline commonly used across imaging domains, (2) a novel semantic-preserving pipeline designed for ultrasound, and (3) a distilled set of the most effective transformations from both pipelines. Pretrained models were evaluated on multiple classification tasks: B-line detection, pleural effusion detection, and COVID-19 classification. Experiments revealed that semantics-preserving data augmentation resulted in the greatest performance for COVID-19 classification - a diagnostic task requiring global image context. Cropping-based methods yielded the greatest performance on the B-line and pleural effusion object classification tasks, which require strong local pattern recognition. Lastly, semantics-preserving ultrasound image preprocessing resulted in increased downstream performance for multiple tasks. Guidance regarding data augmentation and preprocessing strategies was synthesized for practitioners working with SSL in ultrasound.

Xin Wang, Yin Guo, Kaiyu Zhang et al.

This work presents a novel Bayesian framework for unsupervised domain adaptation (UDA) in medical image segmentation. While prior works have explored this clinically significant task using various strategies of domain alignment, they often lack an explicit and explainable mechanism to ensure that target image features capture meaningful structural information. Besides, these methods are prone to the curse of dimensionality, inevitably leading to challenges in interpretability and computational efficiency. To address these limitations, we propose RemInD, a framework inspired by human adaptation. RemInD learns a domain-agnostic latent manifold, characterized by several anchors, to memorize anatomical variations. By mapping images onto this manifold as weighted anchor averages, our approach ensures realistic and reliable predictions. This design mirrors how humans develop representative components to understand images and then retrieve component combinations from memory to guide segmentation. Notably, model prediction is determined by two explainable factors: a low-dimensional anchor weight vector, and a spatial deformation. This design facilitates computationally efficient and geometry-adherent adaptation by aligning weight vectors between domains on a probability simplex. Experiments on two public datasets, encompassing cardiac and abdominal imaging, demonstrate the superiority of RemInD, which achieves state-of-the-art performance using a single alignment approach, outperforming existing methods that often rely on multiple complex alignment strategies.

Ahmad Chaddad, Jihao Peng, Yihang Wu

The use of deep learning (DL) in medical image analysis has significantly improved the ability to predict lung cancer. In this study, we introduce a novel deep convolutional neural network (CNN) model, named ResNet+, which is based on the established ResNet framework. This model is specifically designed to improve the prediction of lung cancer and diseases using the images. To address the challenge of missing feature information that occurs during the downsampling process in CNNs, we integrate the ResNet-D module, a variant designed to enhance feature extraction capabilities by modifying the downsampling layers, into the traditional ResNet model. Furthermore, a convolutional attention module was incorporated into the bottleneck layers to enhance model generalization by allowing the network to focus on relevant regions of the input images. We evaluated the proposed model using five public datasets, comprising lung cancer (LC2500 $n$=3183, IQ-OTH/NCCD $n$=1336, and LCC $n$=25000 images) and lung disease (ChestXray $n$=5856, and COVIDx-CT $n$=425024 images). To address class imbalance, we used data augmentation techniques to artificially increase the representation of underrepresented classes in the training dataset. The experimental results show that ResNet+ model demonstrated remarkable accuracy/F1, reaching 98.14/98.14\% on the LC25000 dataset and 99.25/99.13\% on the IQ-OTH/NCCD dataset. Furthermore, the ResNet+ model saved computational cost compared to the original ResNet series in predicting lung cancer images. The proposed model outperformed the baseline models on publicly available datasets, achieving better performance metrics. Our codes are publicly available at https://github.com/AIPMLab/Graduation-2024/tree/main/Peng.

Shaikh Kaniz Sayeda

Plagiarism defines as the act of using another person work or ideas without giving proper crediton, is a widespread problem in both academic and professional settings. This abstract investigates the many facets of plagiarism, including its numerous ramifications and the potential preventative steps. Plagiarism can manifest itself in a variety of ways, including the outright copying of content or the paraphrasing the text without correct or lack of referencing. Plagiarism brings about a wide range of consequences from academic penalties such as failing grades to legal repercussions in professional contexts. In addition to this, it is detrimental to the principles of intellectual honesty and stifles the development of skills in critical thinking. Lack of awareness about citation procedures is one of the cause. Both employers and educational institutions have a responsibility to adopt preventative actions in order to solve this problem. implementation of instructional programs on appropriate ways of citation, the cultivation of a culture of academic integrity, and the utilization of systems that detect instances of plagiarism are the solution.

Laura Arbelaez Ossa, Giorgia Lorenzini, Stephen R. Milford et al.

Abstract Background While the theoretical benefits and harms of Artificial Intelligence (AI) have been widely discussed in academic literature, empirical evidence remains elusive regarding the practical ethical challenges of developing AI for healthcare. Bridging the gap between theory and practice is an essential step in understanding how to ethically align AI for healthcare. Therefore, this research examines the concerns and challenges perceived by experts in developing ethical AI that addresses the healthcare context and needs. Methods We conducted semi-structured interviews with 41 AI experts and analyzed the data using reflective thematic analysis. Results We developed three themes that expressed the considerations perceived by experts as essential for ensuring AI aligns with ethical practices within healthcare. The first theme explores the ethical significance of introducing AI with a clear and purposeful objective. The second theme focuses on how experts are concerned about the tension that exists between economic incentives and the importance of prioritizing the interests of doctors and patients. The third theme illustrates the need to develop context-sensitive AI for healthcare that is informed by its underlying theoretical foundations. Conclusions The three themes collectively emphasized that beyond being innovative, AI must genuinely benefit healthcare and its stakeholders, meaning AI also aligns with intricate and context-specific healthcare practices. Our findings signal that instead of narrow product-specific AI guidance, ethical AI development may need a systemic, proactive perspective that includes the ethical considerations (objectives, actors, and context) and focuses on healthcare applications. Ethically developing AI involves a complex interplay between AI, ethics, healthcare, and multiple stakeholders.

Jamil Zaghir, Marco Naguib, Mina Bjelogrlic et al.

Prompt engineering is crucial for harnessing the potential of large language models (LLMs), especially in the medical domain where specialized terminology and phrasing is used. However, the efficacy of prompt engineering in the medical domain remains to be explored. In this work, 114 recent studies (2022-2024) applying prompt engineering in medicine, covering prompt learning (PL), prompt tuning (PT), and prompt design (PD) are reviewed. PD is the most prevalent (78 articles). In 12 papers, PD, PL, and PT terms were used interchangeably. ChatGPT is the most commonly used LLM, with seven papers using it for processing sensitive clinical data. Chain-of-Thought emerges as the most common prompt engineering technique. While PL and PT articles typically provide a baseline for evaluating prompt-based approaches, 64% of PD studies lack non-prompt-related baselines. We provide tables and figures summarizing existing work, and reporting recommendations to guide future research contributions.

J. Aultman, Patricia L Raimer, D. Grossoehme et al.

There is a critical need to establish a space to engage in careful deliberation amid exciting, important, necessary, and groundbreaking technological and clinical advances in pediatric medicine. Extracorporeal membrane oxygenation (ECMO) is one such technology that began in pediatric settings nearly 50 years ago. And while not void of medical and ethical examination, both the symbolic progression of medicine that ECMO embodies and its multidimensional challenges to patient care require more than an intellectual exercise. What we illustrate, then, is a person-centered framework that incorporates the philosophy and practice of palliative care and care-based ethical approaches. This person-centered framework is valuable for identifying and understanding challenges central to ECMO, guides collaborative decision-making, and recognizes the value of relationships within and between patients, families, healthcare teams, and others who impact and are impacted by ECMO. Specifically, this person-centered approach enables caregivers to provide compassionate and effective support in critical, and often urgent, situations where conflicts may emerge among healthcare team members, families, and other decision makers. By reflecting on three cases based on actual situations, we apply our person-centered framework and identify those aspects that were utilized in and informed this project. We aim to fill a current gap in the pediatric ECMO literature by presenting a person-centered framework that promotes caregiving relationships among hospitalized critically ill children, families, and the healthcare team and is supported through the philosophy and practice of palliative care and clinical ethics.

N. Boychenko

Advantages and disadvantages of one of the most influential methodological approaches in bioethics – principlism – are considered. The practical origins of principlism in bioethics and its theoretical origins in the philosophy of principles are revealed. The main philosophical sources of bioethics are indicated, which include utilitarianism, Kantian ethics, rule-based theory of general morality, virtue ethics, and feminist ethics. The irreducibility of principlism to any of these theories in particular and at the same time its appeal to each of them was revealed. In medical ethics, principlism appears primarily as an important practical approach that allows instrumentalization of decision-making in ethically complex situations of treating patients, conducting biomedical and clinical research. The sources of the original four principles, which are recognized by the majority of theorists of principlism, have been clarified. Among such sources, the so-called Belmont report, which names three principles, is of secondary but big importance: the principle of respect for the individual; the principle of beneficence and the principle of justice. The main arguments, classical formulation and typical contexts of application of the four principles of bioethics presented by Tom Beaucamp and James Childress as the generally accepted basis of principlism are presented: respect for autonomy, nonmaleficence, beneficence and justice. European bioethicists subsequently complement and develop these four principles with other principles: dignity, integrity and vulnerability. The main directions of criticism of principlism are revealed: lack of specificity, uncertainty of practical application of ethical theory, possible contradiction between principles. The need for constant and systematic practice of applying ethical principles in making medical decisions is defined as an antidote to these shortcomings of principlism. It is noted that Beaucamp and Childress see such permanent practice through specification and balancing procedures.

Souichi Ishigame

Although the relationship between cancer and electric potential has been studied for a long time, there are few published studies regarding this relationship. There is an electric potential difference between cancer and normal tissue, which disappears during chemotherapy. Furthermore, this electric potential difference increases in fast-growing cancers. Conversely, cancer may develop around abnormal electric potential differences in the body. These abnormal charges and electric potentials in the body can be removed by simply hugging a live tree growing in the ground. Alternatively, a tree growing in the ground and a conductive mat may be connected by conductors at several points, and a person may sleep on the conductive mat to dispel these charges. Humans and apes are genetically similar, and as apes live in trees, humans may be less likely to develop cancer if they are potentially balanced with trees growing on the ground.

Ilke Coskun Benlidayi, Kubra Tuncer, Tunay Sarpel

Hypertrophic osteoarthropathy (HOA) is a condition characterized by aberrant skin and osseous tissue proliferation in the distal extremities. Mutations in the 15-hydroxyprostaglandin dehydrogenase gene (HPGD) and the soluble carrier organic anion carrier family member 2A1 gene (SLCO2A1) were associated with primary HOA. Secondary HOA, which is also called as 'hypertrophic pulmonary osteoarthropathy' is responsible for 95-97% of cases. Herein, we present a 19-year-old female patient with primary HOA and heterozygous mutation in the SLCO2A1 gene.

Lizethe Vanessa Bernal Santana, Ivo Queiroz, Luis Alberto Sánchez Alfaro

Resumen: Objetivo: comprender la percepción del principio de beneficencia construida por pacientes de una clínica integral del adulto, de la Facultad de Odontología en la Universidad Nacional de Colombia. Metodología: investigación cualitativa; se utilizaron tres herramientas de recolección de datos: la observación participante, el diario de campo y diez entrevistas semiestructuradas, grabadas y transcritas. Se realizó un análisis de contenido a la información obtenida. Resultados: se establecieron cinco categorías sobre las cuales gira la percepción y construcción del sentido común respecto al principio bioético de beneficencia en los participantes: vivencia durante la atención, valoración de la calidad del tratamiento, necesidades en salud oral, comunicación y balanza riesgos/ beneficios, y beneficios obtenidos. Conclusiones: la percepción y construcción del sentido común respecto al principio bioético de beneficencia, según los participantes, más allá de beneficios técnicos y terapéuticos, está mediada por la experiencia objetiva del proceso de atención y se valora mediante la resolución de sus necesidades y de los beneficios obtenidos en términos terapéuticos, de estética y autocuidado, así como económicos, de accesibilidad, profesionalismo e intersubjetividad, en la triada docente estudiante-paciente, por la consolidación de la confianza.

Ange Lou, Shuyue Guan, Murray Loew

Segmenting medical images accurately and reliably is important for disease diagnosis and treatment. It is a challenging task because of the wide variety of objects' sizes, shapes, and scanning modalities. Recently, many convolutional neural networks (CNN) have been designed for segmentation tasks and achieved great success. Few studies, however, have fully considered the sizes of objects, and thus most demonstrate poor performance for small objects segmentation. This can have a significant impact on the early detection of diseases. This paper proposes a Context Axial Reverse Attention Network (CaraNet) to improve the segmentation performance on small objects compared with several recent state-of-the-art models. CaraNet applies axial reserve attention (ARA) and channel-wise feature pyramid (CFP) module to dig feature information of small medical object. And we evaluate our model by six different measurement metrics. We test our CaraNet on brain tumor (BraTS 2018) and polyp (Kvasir-SEG, CVC-ColonDB, CVC-ClinicDB, CVC-300, and ETIS-LaribPolypDB) segmentation datasets. Our CaraNet achieves the top-rank mean Dice segmentation accuracy, and results show a distinct advantage of CaraNet in the segmentation of small medical objects.