Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Although large language models (LLMs) show promise in controlled settings, a study now exposes their limitations in real-world clinical applications and points the way towards robust evaluation and benchmarking before clinical use.
As growing evidence shows that engagement with politics is associated with worsening mental and physical health, a bumper election year and increasing polarization could affect millions.
SARS-CoV-2 variant-specific neutralizing antibodies (nAbs) mediate protection against infection by the cognate variant to distinct extents, while the majority of protection elicited by natural infection is not mediated by nAbs.
A natural experiment of over 200,000 people who received a shingles vaccine revealed that the recombinant vaccine is associated with lower risk of dementia than the live vaccine, within 6 years of vaccination
For patients with shortness of breath, a new, simplified screening tool — relying only on basic clinical information — can estimate the probability of heart failure with preserved ejection fraction, providing the opportunity for improved detection in non-specialist settings.
A translational study and phase 1 trial show that the JAK inhibitor ruxolitinib stimulates responses to nivolumab in patients with refractory Hodgkin lymphoma.
In two preregistered studies involving 2,280 participants using simulated clinical scenarios, advice labeled as involving AI was evaluated as significantly less reliable and less empathetic compared with ‘human’-labeled advice.
In this large interventional, randomized controlled trial, structured and individualized exercise in patients with metastatic breast cancer led to lower levels of fatigue and improved health-related quality of life.
Measuring between 5 and 20 proteins in the blood predicts the future risk of developing a diverse range of 52 diseases across clinical specialties. These protein signatures perform better than standard risk factors and blood tests used in clinical practice for risk prediction. Examples of well-predicted diseases include multiple myeloma and idiopathic pulmonary fibrosis.
Trained on 1.5 million whole-slide images from 100,000 patients, a pathology foundation model is shown to improve performance of specialized models in detection of rare cancers.
Proteomic prediction models developed using a large-scale dataset from the UK Biobank Pharma Proteomics Project were superior to clinical models for assessing the 10-year risk of 67 diseases across different types of pathology, including multiple myeloma, motor neuron disease, pulmonary fibrosis, celiac disease and dilated cardiomyopathy.
As healthcare becomes more data-driven and precise, promising greater accuracy, targeted use of resources and healthier populations, we ask what it will take to realize these benefits in local and global contexts.
Ventral striatal activity in individuals with OCD has strong circadian periodicity and high predictability in the severely symptomatic state. After deep brain stimulation, these features decrease substantially in clinical responders but remain raised in non-responders, thus providing a measurable predictor of clinical status.
In a cohort of 6,785 participants from the All of Us Research Program whose sleep was monitored by a Fitbit over a median of 4.5 years, sleep duration, stages and irregularity were associated with the incidence of obesity and a number of cardiovascular and psychological disorders.
Tailored to provide diabetes management recommendations from large training and validation datasets, an artificial intelligence system integrating language and computer vision capabilities is shown to improve self-management of patients in a prospective implementation study.
The BePRECISE (Better Precision-data Reporting of Evidence from Clinical Intervention Studies & Epidemiology) consortium, comprising 23 experts in precision medicine, cardiometabolic diseases, statistics, editorial and lived experience, provides a checklist tool and recommendations for reporting precision medicine research.
Avian influenza vaccines in the US strategic stockpile elicit immune responses that recognize the clade of avian influenza circulating in cows and could be a resource before newer vaccines are approved and deployed.