Large language models (LLMs) are emerging as powerful tools in healthcare, with a growing role in global health, particularly in low- and middle-income countries. This Perspective examines the current...

In a prospective study involving 579,583 women in multiple sites across the USA, a multistage AI-driven workflow for digital breast tomosynthesis mammograms assisted radiologists in identifying 21.6% ...

In a prospective study involving 579,583 women in multiple sites across the USA, a multistage AI-driven workflow for digital breast tomosynthesis mammograms assisted radiologists in identifying 21.6% ...

A model developed with data from 19,403 patients with colorectal cancer for prediction of 1-year mortality is used as a decision support tool in a prospective cohort, showing promising results in redu...

By matching antigenicity prediction with a forecast of circulating strains in the next season, a model is shown to outperform standard recommendations for vaccine design in terms of vaccine effectiven...

Trained and validated on multimodal data from 14.5 million images from multicountry datasets, a foundation model is shown to increase diagnostic and referral accuracy of clinicians when used as an ass...

Going from model development to a pilot implementation study, a deep learning model shows that acute aortic syndrome can be diagnosed directly from noncontrast CT, increasing accuracy and decreasing t...

By analysing the smartphone data of 2,112,288 participants, in particular observing and comparing the activity of the same individual in two different environments, we find that increases in the walkability of environments result in increases in daily physical activity, providing evidence of the importance of the built environment on physical health.

Analysis of a deeply phenotyped, longitudinal cohort of 249 participants elucidates the network of concurrent molecular and macroscopic patterns related to insurgence and course of ME/CFS.

Analysis of a deeply phenotyped, longitudinal cohort of 249 participants elucidates the network of concurrent molecular and macroscopic patterns related to insurgence and course of ME/CFS.

Using data from six population-scale cohorts involving more than 10 million individuals, a large language model-based aging signature is shown to outperform traditional aging biomarker indices and sig...

In a silent trial involving 197 patients with lung cancer over a span of 4 months, a fine-tuned pathology foundation model was able to identify the presence of EGFR mutations with high accuracy, poten...

Using mathematical models calibrated to clinical trial results on vaccine efficacy, this study projects the potential impact of the vaccine at population level, showing the conditions for it to be eff...

A reinforcement learning-based virtual reality system, implementing a series of sport sessions with coaching, was as effective as standard physical activity in reducing fat mass in adolescents, with p...

Nature Genetics - This study integrates polygenic risk scores with four emulated clinical trials using FinnGen data and shows the feasibility of this approach while highlighting potential pitfalls.

Trained on large and multicenter datasets with different imaging modalities, a foundation model is shown to have strong performance on the full spectrum of clinically relevant tasks and to increase us...

Job Title: Associate or a Senior, Nature Medicine (Digital Medicine and Artificial Intelligence) Location: New York or Milan (Hybrid Working) Application Deadline: May 30, 2025 About Springer Nature G...

Salvi et al. find that GPT-4 outperforms humans in debates when given basic sociodemographic data. With personalization, GPT-4 had 81.2% higher odds of post-debate agreement than humans.

Job Title: Associate or a Senior, Nature Medicine (Digital Medicine and Artificial Intelligence) Location: New York or Milan (Hybrid Working) Application Deadline: May 30, 2025 About Springer Nature G...

A panel of nine LLMs was exposed to simulated clinical cases with switched sociodemographic features exploring ethnic, social, sexual orientation and gender dimensions and showed differences in recomm...

In an evaluation involving 125 standardized patient cases, open-source DeepSeek large language models are shown to perform at least on par with state-of-the-art proprietary large language models in di...

In a quasi-experimental, pre–post-implementation study involving 51,760 hospitalizations, an electronic health record-based risk prediction model informed clinicians of patients needing consultation o...

In a large pragmatic cluster-randomized trial involving 74 clinical units in 2 health systems in the United States, a real-time alert system assisting clinical care teams in identifying patients at ri...

A randomized study highlights the potential of large language models to assist clinicians in patient care; future research should harness our understanding of the cognitive processes underlying clinic...

An algorithm that learned from tens of thousands of nutrition prescriptions for premature babies could reduce medical errors and better identify what nutrients the smallest patients need.

An artificial intelligence model defines a data-driven set of total parenteral nutrition compositions to assist clinicians in personalized treatment of neonates in intensive care and is able to adapt ...

The alternate splicing small molecule risdiplam, administered to a fetus with spinal muscular atrophy, had an apparent ameliorating effect on the disorder through 30 months.

In a randomized controlled trial in individuals with persistent atrial fibrillation, an individualized ablation procedure, in which areas with abnormal electrophysiological characteristics—as detected...

By simulating the implementation of airport-based wastewater surveillance sites at the global level, a modeling study shows how this early warning system would perform in identifying sources of pandem...

By simulating the implementation of airport-based wastewater surveillance sites at the global level, a modeling study shows how this early warning system would perform in identifying sources of pandem...

In a large-scale analysis of ambulatory electrocardiographic recordings, a deep learning algorithm had a substantially higher sensitivity for the detection of critical arrhythmias as compared to techn...

In a prospective study involving 92 physicians from multiple institutions, access to large language model assistance on top of conventional resources increased a score expressing the quality of their ...

In a prospective study involving 92 physicians from multiple institutions, access to large language model assistance on top of conventional resources increased a score expressing the quality of their ...

An individual-based mathematical model evaluating the public health benefits of RSVpreF vaccination showed that vaccinating older adults and pregnant women can substantially reduce respiratory syncyti...

Trained on a large corpus of medical text and patient records and tested across diseases, with specific focus on rare presentations, an open-source medical language model demonstrates higher accuracy ...

With an improved framework for model development and evaluation, a large language model is shown to provide answers to medical questions that are comparable or preferred with respect to those provided...

Large language models can be manipulated to generate misinformation by poisoning of a very small percentage of the data on which they are trained, but a harm mitigation strategy using biomedical knowl...

Trained on a large corpus of medical text and patient records and tested across diseases, with specific focus on rare presentations, an open-source medical language model demonstrates higher accuracy ...

Using multiple datasets from real-world evidence and completed trials, a machine learning model using routine blood and clinical data is shown to be predictive of patient response to immune checkpoint...

By simulating realistic doctor–patient conversations, a framework can be applied to large language models to investigate shortcomings and bias in patient interactions, providing insight before actual ...