Search Articles

View query in Help articles search

Search Results (1 to 10 of 594 Results)

Download search results: CSV END BibTex RIS


Identifying Psychosocial, Self-Management, and Health Profiles Among Women With Chronic Pain Who Have Experienced Intimate Partner Violence and Those Who Have Not: Protocol for a 2-Phase Qualitative and Cross-Sectional Study Using AI Techniques

Identifying Psychosocial, Self-Management, and Health Profiles Among Women With Chronic Pain Who Have Experienced Intimate Partner Violence and Those Who Have Not: Protocol for a 2-Phase Qualitative and Cross-Sectional Study Using AI Techniques

Machine learning, as a subset of artificial intelligence (AI) techniques, analyzes large datasets to identify patterns and make predictions [154]. These methods are transformative in gender studies, uncovering patterns in gender-related data and deepening the understanding of social inequalities and biases [155]. These methods have been successfully applied to areas such as forecasting gender-based violence, further demonstrating their relevance in social sciences [156].

Ainara Nardi-Rodríguez, Sónia Bernardes, María Ángeles Pastor-Mira, Sofía López-Roig, Lidia Pamies-Aubalat, Andrés Sánchez-Prada, Victoria A Ferrer-Pérez, Ignacio Rodríguez-Rodríguez, Purificación Heras-González

JMIR Res Protoc 2025;14:e66396

Improving AI-Based Clinical Decision Support Systems and Their Integration Into Care From the Perspective of Experts: Interview Study Among Different Stakeholders

Improving AI-Based Clinical Decision Support Systems and Their Integration Into Care From the Perspective of Experts: Interview Study Among Different Stakeholders

The literature describes a further wide range of problems and barriers in the context of AI-based CDSS [10-12]. These relate to AI or CDSS and a combination of both, AI-based CDSS. While some of the problems relate to technical integration and operational use [10,11], others relate to the legal and ethical framework [12].

Godwin Denk Giebel, Pascal Raszke, Hartmuth Nowak, Lars Palmowski, Michael Adamzik, Philipp Heinz, Marianne Tokic, Nina Timmesfeld, Frank Martin Brunkhorst, Jürgen Wasem, Nikola Blase

JMIR Med Inform 2025;13:e69688

Privacy-Preserving Glycemic Management in Type 1 Diabetes: Development and Validation of a Multiobjective Federated Reinforcement Learning Framework

Privacy-Preserving Glycemic Management in Type 1 Diabetes: Development and Validation of a Multiobjective Federated Reinforcement Learning Framework

RL offers a more dynamic and adaptive solution, as it allows artificial intelligence (AI) models to learn optimal insulin dosing strategies through trial and error. Unlike traditional ML, RL does not require explicit supervision and can adjust insulin delivery based on real-time feedback from CGM data. Several RL-based frameworks have demonstrated improvements in personalized insulin dosing [4-7,11].

Fatemeh Sarani Rad, Juan Li

JMIR Diabetes 2025;10:e72874

Perspectives of Health Care Professionals on the Use of AI to Support Clinical Decision-Making in the Management of Multiple Long-Term Conditions: Interview Study

Perspectives of Health Care Professionals on the Use of AI to Support Clinical Decision-Making in the Management of Multiple Long-Term Conditions: Interview Study

Participants typically felt unsure of exactly what AI is and how it differs from existing non-AI computer tools (15/20, 75%): To me AI is almost something that is talked about and I don’t understand. Participants typically had limited experience of the use of AI within the health care setting, and even those GPs who had experience working with AI technology companies did not feel that they fully understood AI.

Jennifer Cooper, Shamil Haroon, Francesca Crowe, Krishnarajah Nirantharakumar, Thomas Jackson, Leah Fitzsimmons, Eleanor Hathaway, Sarah Flanagan, Tom Marshall, Louise J Jackson, Niluka Gunathilaka, Alexander D'Elia, Simon George Morris, Sheila Greenfield

J Med Internet Res 2025;27:e71980

Artificial Intelligence to Improve Clinical Coding Practice in Scandinavia: Crossover Randomized Controlled Trial

Artificial Intelligence to Improve Clinical Coding Practice in Scandinavia: Crossover Randomized Controlled Trial

Modern CAC approaches often include AI since LLMs like Chat GPT have demonstrated impressive capabilities in natural language processing tasks. However, these generative AI models tend to perform poorly when applied to clinical coding [11,12]. This poor performance is perhaps largely explained by the vast and intricate label space of ICD codes (with thousands of specific options), and lack of localized, domain-specific, clinical data for training purposes.

Taridzo Chomutare, Therese Olsen Svenning, Miguel Ángel Tejedor Hernández, Phuong Dinh Ngo, Andrius Budrionis, Kaisa Markljung, Lill Irene Hind, Torbjørn Torsvik, Karl Øyvind Mikalsen, Aleksandar Babic, Hercules Dalianis

J Med Internet Res 2025;27:e71904

Design of a Mobile App and a Clinical Trial Management System for Cognitive Health and Dementia Risk Reduction: User-Centered Design Approach

Design of a Mobile App and a Clinical Trial Management System for Cognitive Health and Dementia Risk Reduction: User-Centered Design Approach

These mock-ups encompassed (1) an overview page featuring details on all study participants, (2) a dedicated view for an individual study participant, (3) the conceptualization of an AI simulation, and (4) a data entry page aligned with the study protocol. The feedback on the mock-ups included the exclusion of real names given that all study participants would be collectively visible on one page in the mock-up.

Hannes Hilberger, Bianca Buchgraber-Schnalzer, Simone Huber, Theresa Weitlaner, Markus Bödenler, Alara Abaci, Jeroen Bruinsma, Ana Diaz, Anna Giulia Guazzarini, Jenni Lehtisalo, Seungjune Lee, Vasileios Loukas, Francesca Mangialasche, Patrizia Mecocci, Tiia Ngandu, Anna Rosenberg, Elisabeth Stögmann, Konsta Valkonen, Elena Uhlik, Helena Untersteiner, Laura Kneß, Helmut Ahammer, Sten Hanke

JMIR Aging 2025;8:e66660

Evaluating the Quality of Psychotherapy Conversational Agents: Framework Development and Cross-Sectional Study

Evaluating the Quality of Psychotherapy Conversational Agents: Framework Development and Cross-Sectional Study

In order to effectively interact with AI-based chatbots and apply the CAPE framework, we used a persona-based approach. In this approach, the researcher interacts with the chatbot via a dynamic script representing a fictional client called a persona. The personas are written as client evaluations in the biopsychosocial framework, providing holistic information on the fictional client.

Kunmi Sobowale, Daniel Kevin Humphrey

JMIR Form Res 2025;9:e65605

Enhancing Magnetic Resonance Imaging (MRI) Report Comprehension in Spinal Trauma: Readability Analysis of AI-Generated Explanations for Thoracolumbar Fractures

Enhancing Magnetic Resonance Imaging (MRI) Report Comprehension in Spinal Trauma: Readability Analysis of AI-Generated Explanations for Thoracolumbar Fractures

Patients with lower health literacy may benefit from structured AI-generated summaries, potentially reducing anxiety and misunderstandings regarding their diagnosis. However, given the residual complexity in some explanations provided by GPT-4o, integrating human oversight in AI-assisted patient education remains crucial until further improvement in LLMs is seen in the future.

David C Sing, Kishan S Shah, Michael Pompliano, Paul H Yi, Calogero Velluto, Ali Bagheri, Robert K Eastlack, Stephen R Stephan, Gregory M Mundis Jr

JMIR AI 2025;4:e69654

Proposal for Using AI to Assess Clinical Data Integrity and Generate Metadata: Algorithm Development and Validation

Proposal for Using AI to Assess Clinical Data Integrity and Generate Metadata: Algorithm Development and Validation

Machine learning is a subfield of artificial intelligence (AI) that enables information technology systems to recognize patterns in existing data and develop solutions autonomously. This research involves classifying data quality variables for clinical data. The classification requires a training dataset with examples for each data source entity’s input and output variables. All machine learning models and classifiers were developed in Python (Python Software Foundation) using scikit-learn libraries.

Caroline Bönisch, Christian Schmidt, Dorothea Kesztyüs, Hans A Kestler, Tibor Kesztyüs

JMIR Med Inform 2025;13:e60204

Enhancing Diagnostic Accuracy of Ophthalmological Conditions With Complex Prompts in GPT-4: Comparative Analysis of Global and Low- and Middle-Income Country (LMIC)–Specific Pathologies

Enhancing Diagnostic Accuracy of Ophthalmological Conditions With Complex Prompts in GPT-4: Comparative Analysis of Global and Low- and Middle-Income Country (LMIC)–Specific Pathologies

Sight Bot, a research chatbot, uses both Open AI and Pub Med’s application programming interface (APIs) to restrict the information available to GPT-3.5 [8]. This limits the data that the AI can access in the hopes that this will reduce “AI hallucination”–the fabrication of data [8]. Bio Med LM is built upon the Hugging Face GPT model with 2.7 billion parameters and is also trained upon biomedical data from Pub Med [9]. However, there is limited research on AI used as ophthalmological diagnostic tools.

Shona Alex Tapiwa M'gadzah, Andrew O'Malley

JMIR Form Res 2025;9:e64986