The classification of diabetes and prediabetes by static glucose thresholds obscures the pathophysiological dysglycemia heterogeneity, primarily driven by insulin resistance (IR), beta-cell dysfunction, and incretin deficiency. This review demonstrates that continuous glucose monitoring and wearable technologies enable a paradigm shift towards non-invasive, dynamic metabolic phenotyping. We show evidence that machine learning models can leverage high-resolution glucose data from at-home, CGM-enabled oral glucose tolerance tests to accurately predict gold-standard measures of muscle IR and beta-cell function. This personalized characterization extends to real-world nutrition, where an individual's unique postprandial glycemic response (PPGR) to standardized meals, such as the relative glucose spike to potatoes versus grapes, could serve as a biomarker for their metabolic subtype. Moreover, integrating wearable data reveals that habitual diet, sleep, and physical activity patterns, particularly their timing, are uniquely associated with specific metabolic dysfunctions, informing precision lifestyle interventions. The efficacy of dietary mitigators in attenuating PPGR is also shown to be phenotype-dependent. Collectively, this evidence demonstrates that CGM can deconstruct the complexity of early dysglycemia into distinct, actionable subphenotypes. This approach moves beyond simple glycemic control, paving the way for targeted nutritional, behavioral, and pharmacological strategies tailored to an individual's core metabolic defects, thereby paving the way for a new era of precision diabetes prevention.

In the pursuit of Artificial General Intelligence (AGI), automating the generation and evaluation of novel research ideas is a key challenge in AI-driven scientific discovery. This paper presents Relative Neighbor Density (RND), a domain-agnostic algorithm for novelty assessment in research ideas that overcomes the limitations of existing approaches by comparing an idea's local density with its adjacent neighbors' densities. We first developed a scalable methodology to create test set without expert labeling, addressing a fundamental challenge in novelty assessment. Using these test sets, we demonstrate that our RND algorithm achieves state-of-the-art (SOTA) performance in computer science (AUROC=0.820) and biomedical research (AUROC=0.765) domains. Most significantly, while SOTA models like Sonnet-3.7 and existing metrics show domain-specific performance degradation, RND maintains consistent accuracies across domains by its domain-invariant property, outperforming all benchmarks by a substantial margin (0.795 v.s. 0.597) on cross-domain evaluation. These results validate RND as a generalizable solution for automated novelty assessment in scientific research.

Clario, a technology company that delivers the leading endpoint solutions for decentralized, hybrid and site-based clinical trials, today published a manuscript in Nature Medicine that outlines how AI can improve patients' experience in DCTs. This comes as remote trials continue to rise in popularity, placing increased responsibility on participants. The study explores how AI automation can support improvements in digital health user interfaces. "AI is being used to enhance user experience in customer-facing applications across many industries," said Kevin Thomas, Ph.D., Director of Artificial Intelligence at Clario, one of the researchers. "Adopting this approach in clinical trials means we can help more patients to enroll, empower them to complete trials without undue burden, and ensure they are able to submit high-quality health assessments throughout their participation."

Computer vision is a domain of AI that enables the automatic assessment of images and videos. Many mobile banking apps use computer vision to coach customers to take photos of their checks for electronic deposits. If you hold the camera too far from the check or the lighting is too dark, the app will provide real-time advice about the required adjustments. The same could be done for user-submitted photos and videos in clinical trials; this approach has already been tested in telemedicine8. If a patient's entire body needs to be visible in the video, this could be confirmed automatically while the video is being captured in the trial's mobile application, allowing for immediate adjustments.

The general public is requested to help eradicate biases based on race and other underprivileged communities in artificial intelligence (AI) algorithms for healthcare. Health scientists are seeking support to resolve how'minoritized' communities, who are actively deprived because of social constructs, would not see future advantages from using AI in healthcare. The scientists, guided by the University of Birmingham and University Hospitals Birmingham, recently reported in Nature Medicine about the introduction of a consultation on a set of principles that they anticipate will cut biases that are said to be present in AI algorithms. There is increasing proof that certain AI algorithms do not work as well for specific groups of people - mainly those in minoritized racial/ethnic communities. A few of these come from biases in the datasets used to create AI algorithms.

At the beginning of the COVID-19 pandemic, before the widespread adoption of reliable point-of-care assays to detect SARS-CoV-2, one highly active area of research involved the development of ML algorithms to estimate the probability of infection. These algorithms based their predictions on various data elements captured in electronic health records, such as chest radiographs. Despite their promising initial validation results, the success of numerous artificial neural networks trained on chest X-rays were largely not replicated when applied to different hospital settings, in part because the models failed to learn or understand the true underlying pathology of COVID-19. Instead, they exploited shortcuts or spurious associations that reflected biologically meaningless variations in image acquisition, such as laterality markers, patient positioning or differences in radiographic projection6. These ML algorithms were not explainable and, while appearing to be at the cutting edge, were inferior to traditional diagnostic techniques such as RT-PCR, obviating their usefulness.

Bayesian Health and Johns Hopkins have announced ground-breaking results showing that many lives have been saved with a new clinically deployed AI platform called Targeted Real-Time Early Warning System (TREWS). The AI platform activates state-of-the-art AI within the electronic medical record and tracks patients from the moment they are admitted to hospital until they are discharged. The early warning system is designed to send alerts to healthcare providers when there is cause for concern. A real world study - conducted in 5 hospitals - demonstrated that the TREWS AI system led to the detection of sepsis on average almost 6 hours earlier than traditional methods, with a sensitivity rate of 82%. This is significant because sepsis is responsible for 20% of all deaths globally and early detection could save over 11 million lives every year.

Patients are 20% less likely to die of sepsis because a new AI system developed at Johns Hopkins University catches symptoms hours earlier than traditional methods, an extensive hospital study demonstrates. The system, created by a Johns Hopkins researcher whose young nephew died from sepsis, scours medical records and clinical notes to identify patients at risk of life-threatening complications. The work, which could significantly cut patient mortality from one of the top causes of hospital deaths worldwide, is published today in Nature Medicine and npj Digital Medicine. "It is the first instance where AI is implemented at the bedside, used by thousands of providers, and where we're seeing lives saved," said Suchi Saria, founding research director of the Malone Center for Engineering in Healthcare at Johns Hopkins and lead author of the studies, which evaluated more than a half million patients over two years. "This is an extraordinary leap that will save thousands of sepsis patients annually. And the approach is now being applied to improve outcomes in other important problem areas beyond sepsis."

First, the IoT provides a platform that allows public-health agencies access to data for monitoring the COVID-19 pandemic. For example, the'Worldometer' provides a real-time update on the actual number of people known to have COVID-19 worldwide, including daily new cases of the disease, disease distribution by countries and severity of disease (recovered, critical condition or death) (https://www.worldometers.info/coronavirus/). Second, big data also provides opportunities for performing modeling studies of viral activity and for guiding individual country healthcare policymakers to enhance preparation for the outbreak. Using three global databases―the Official Aviation Guide, the location-based services of the Tencent (Shenzhen, China), and the Wuhan Municipal Transportation Management Bureau―Wu et al. performed a modeled study of'nowcasting' and forecasting COVID-19 disease activity within and outside China that could be used by the health authorities for public-health planning and control worldwide8. Similarly, using the WHO International Health Regulations, the State Parties Self-Assessment Annual Reporting Tool, Joint External Evaluation reports and the Infectious Disease Vulnerability Index, Gilbert et al. assessed the preparedness and vulnerability of African countries in battling against COVID-19; this would help raise awareness of the respective health authorities in Africa to better prepare for the viral outbreak9.

Artificial intelligence (AI) is poised to broadly reshape medicine, potentially improving the experiences of both clinicians and patients. We discuss key findings from a 2-year weekly effort to track and share key developments in medical AI. We cover prospective studies and advances in medical image analysis, which have reduced the gap between research and deployment. We also address several promising avenues for novel medical AI research, including non-image data sources, unconventional problem formulations and human–AI collaboration. Finally, we consider serious technical and ethical challenges in issues spanning from data scarcity to racial bias. As these challenges are addressed, AI’s potential may be realized, making healthcare more accurate, efficient and accessible for patients worldwide. AI has the potential to reshape medicine and make healthcare more accurate, efficient and accessible; this Review discusses recent progress, opportunities and challenges toward achieving this goal.