Latest news with #npjDigitalMedicine
Khaleej Times
06-08-2025
- Health
- Khaleej Times
UAE: AI outperforms human radiologists, records zero false negatives
A comprehensive study on an AI model developed by Abu Dhabi-based healthcare group M42 has found that it can safely and accurately detect tuberculosis from chest X-rays — outperforming human radiologists. Published in the prestigious scientific journal npj Digital Medicine – Nature, the study conducted in collaboration with Abu Dhabi's Capital Health Screening Centre (CHSC) is among the largest real-world clinical validations of an AI-driven healthcare solution to date, analyzing over one million chest X-rays to evaluate the efficacy and scalability of AI in TB screening. The study assessed AI Radiology in Screening TB (AIRIS-TB), M42's cutting-edge AI model engineered to streamline routine tuberculosis screenings, allowing radiologists to focus on more complex or urgent cases. The model demonstrated exceptional performance, indicating high diagnostic accuracy and an unprecedented 0 percent of false negatives. This means that the model has the potential to safely automate up to 80 percent of routine chest X-ray assessments, directly reducing the workload of radiologists, minimising the risk of human error, and delivering significant cost efficiencies in high-throughput low-prevalence settings. Labour-intensive Currently, reviews of chest X-rays remains labor-intensive and prone to oversight and error, potentially leading to missed or delayed diagnoses. A prior study has indicated a 26.6 percent increase in missed findings when radiologists double their annotation speed, and a rise in errors after 9 hours into their shift. AIRIS-TB delivered consistently strong performance across a wide range of demographic groups, including variations in gender, age, HIV status, income levels, and a diverse population covering six World Health Organization (WHO) regions, highlighting the model's robustness, fairness and generalizability across diverse global populations. These results indicate the model's potential to significantly improve clinical workflows and drive earlier, more equitable screening of TB in high-volume programs worldwide. "This landmark study marks a pivotal moment in the potential power of AI in the global fight against tuberculosis,' said Dimitris Moulavasilis, Group CEO of M42. 'Our AIRIS-TB model stands as a compelling testament to the unmatched accuracy, safety and scalability that AI can deliver, particularly in resource-limited settings where there is a shortage of radiologists and the need to tackle TB is greatest.' He added that in regions with a high prevalence of TB, the model offers a 'scalable technological solution' that can help save lives. 'These results signal the transformative role AI can play in reshaping global public health and redefining how healthcare is diagnosed, delivered and experienced worldwide,' he said. The study underwent rigorous peer review and ethical oversight by the Department of Health – Abu Dhabi, ensuring transparency, accountability, and the highest standards of clinical integrity. Its publication in a leading scientific journal reinforces M42's position as a global pioneer in AI-led health solutions and solidifies the UAE's growing prominence as a global, data-driven hub for cutting-edge medical innovation and technology.
Associated Press
04-08-2025
- Health
- Associated Press
Waymark Signal Achieves 85% Accuracy in Predicting Medicaid Care Gaps, Study in npj Digital Medicine Shows
Peer-reviewed research demonstrates dramatic improvement in identifying and closing quality gaps for Medicaid patients, outperforming current methods by 35 percentage points SAN FRANCISCO, CALIFORNIA / ACCESS Newswire / August 4, 2025 / Healthcare organizations nationwide face a critical challenge: how to effectively deploy limited staff to close quality measure gaps among patients receiving Medicaid. With health plans facing financial penalties for poor quality performance and providers under value-based contracts unable to reach vulnerable patients through traditional methods, many resort to ineffective and untargeted mass outreach campaigns - helping fewer than 8% of eligible patients. A new peer-reviewed study from Waymark published in Nature's npj Digital Medicine offers a solution. Signal for Quality Improvement, the company's newest predictive analytics technology, identifies with 85% accuracy which Medicaid patients will benefit most from proactive community-based outreach-enabling care teams to prioritize and target patients who need additional support to close gaps in care and improve quality performance. 'Healthcare organizations know that community-based outreach works, but they've been unable to identify and prioritize who will benefit most from these limited resources,' said Dr. Sadiq Y. Patel, lead author of the study and Vice President of Data Science and Artificial Intelligence (AI) for Waymark. 'The challenge isn't just identifying patients with open gaps-it's distinguishing between those who will close gaps on their own versus those who need proactive outreach and support. By incorporating social determinants of health data alongside clinical factors, Signal for Quality Improvement enables care teams to more efficiently prioritize outreach to the patients who need it most.' The study analyzed data from over 14 million Medicaid beneficiaries across 25 states and Washington D.C.. Researchers found that Signal for Quality Improvement achieved 85% accuracy in predicting which patients would miss nine nationally adopted quality measures, as defined by the Healthcare Effectiveness Data and Information Set (HEDIS) developed by the National Committee for Quality Assurance (NCQA). Model simulations showed that using this predictive prioritization to guide outreach would result in a 35 percentage point improvement in helping patients close care gaps compared to traditional, non-predictive methods used by most healthcare organizations, such as alphabetical calling or birthday reminders. Signal for Quality Improvement is the latest addition to the company's proprietary Waymark SignalTM platform, which also includes Signal for Rising Risk. A 2024 peer-reviewed study published in Nature Scientific Reports demonstrated that Signal for Rising Risk predicts patients at risk for future avoidable emergency department (ED) and hospital visits with greater than 90% accuracy. Signal for Quality Improvement builds on this technology by enabling health plans and providers to identify which patients will most benefit from proactive outreach to close quality gaps and improve HEDIS performance. 'Healthcare organizations face significant financial penalties when quality metrics are poor, but current outreach methods are remarkably inefficient,' said Dr. Sanjay Basu, Co-founder and Head of Clinical at Waymark and senior author of the study. 'Signal for Quality Improvement represents a fundamental shift toward evidence-based patient outreach that can improve both health outcomes and resource allocation for the 80 million Americans covered by Medicaid.' The study found that incorporating social determinants of health data not only improved predictive accuracy, but also eliminated pre-existing racial disparities in prediction performance between Black and White patients across four quality measures. 'This paper addresses a persistent challenge in healthcare: efficiently reaching patients who need care most,' said Dr. Michael L. Barnett, study co-author and Professor at Brown School of Public Health. 'The combination of high predictive accuracy with a thoughtful approach that minimized racial bias offers a pathway toward more equitable healthcare delivery.' Waymark works with health plans and primary care providers to deliver community-based early interventions to patients receiving Medicaid. The company's local care teams use the Waymark Signal platform to identify and support patients who are at risk of avoidable ED and hospital visits or have open gaps in care. A 2024 peer-reviewed study published in NEJM Catalyst found that Waymark reduced all-cause hospital and ED visits by 22.5% for rising risk patients, and improved care quality by an average of 11.8 absolute percentage points across seven quality measures in its first year of service. The full article titled " Predicting Quality Measure Completion Among 14 Million Low-Income Patients Enrolled in Medicaid " was published in npj Digital Medicine, a peer-reviewed journal published by Nature. The authors were Sadiq Y. Patel MSW PhD of Waymark, Michael L. Barnett MD of Brown School of Public Health, and Sanjay Basu MD PhD of Waymark. About Waymark Waymark is a public benefit company dedicated to improving access and quality of care for people receiving Medicaid. We partner with health plans and primary care providers-including health systems, federally qualified health centers (FQHCs), and independent practices-to provide community-based care to people enrolled in Medicaid. Our local teams of community health workers, pharmacists, therapists and care coordinators use proprietary data science technologies to deliver early interventions to hard-to-reach patient populations. Waymark's peer-reviewed research has been published in leading journals including the New England Journal of Medicine (NEJM) Catalyst, Nature Scientific Reports, and Journal of the American Medical Association (JAMA)-demonstrating measurable improvements in health outcomes and cost savings for Medicaid populations. For more information, visit Contact Information Iman Rahim Communications [email protected] SOURCE: Waymark press release
Yahoo
25-06-2025
- Health
- Yahoo
People Are Asking ChatGPT for Advice on Injecting Their Own Facial Filler, a Cosmetic Procedure That Should Only Be Carried Out by Licensed Medical Professionals
Since OpenAI first introduced ChatGPT to the public back in 2022, people have done all sorts of ill-advised things with the AI tool — from attorneys filing court documents that cite hallucinated caselaw to everyday users spiraling into severe mental health crises as the chatbot affirms delusional thoughts. Now add to that list: asking ChatGPT for advice on how to inject facial filler — a trendy cosmetic procedure intended to puff up features like lips and cheeks — at home, without the assistance of a medical professional. "I'll be injecting myself tonight," one Redditor wrote in a recent post. "I have all things needed on hand and I'm trying to research the best way of keeping things as sterile/clean as possible. I asked ChatGPT and it said I should absolutely not use normal gloves, I googled and can't find any specific info on it." Needless to say, this is a resoundingly terrible idea. Please don't do this procedure at home, and instead go to a qualified medical facility so you don't hurt yourself. (While pros can screw up this process too, at least they can be held liable.) Unfortunately, nobody chastised the Redditor for asking ChatGPT for advice. In fact, a quick perusal of the same subreddit, where thrifty beauty aficionados swap tips on administering cosmetic procedures on their own, finds a huge number of similarly alarming situations. "I used ChatGPT to help me map my tox and PN placements, how to dilute my tox facial and depth of injections, etc," one commenter enthused. "If you send it annotated photos it can view your mapping and correct it." Another user turned to AI after problems with a DIY cosmetic procedure. "Asked [ChatGPT], and it said that since a small amount likely migrated to cheek area through tear trough [sic]," they wrote. "But since it migrated, likely was dissolved into bloodstream. Fibrosis possible but may resolve. If fat was dissolved it should be very negligible." AI models may be set to revolutionize medicine in certain ways, such as at the Icahn School of Medicine at Mount Sinai, which is incorporating AI into training doctors. Researchers are excited about AI being used to diagnose diseases such as prostate cancer and heart disease earlier than before. But the jury is still out on how effective AI chatbots will be in dispensing useful medical advice. For example, a recent npj Digital Medicine paper in March revealed that while large language models such as ChatGPT are more accurate than search engines, they are still going to spew out more than 30 percent of incorrect advice under certain circumstances. In addition, the quality of output is reliant on the quality of the prompt. "We found that some input prompts, which guide the models towards reputed sources, are much more effective than basic prompts (or prompts that give no context at all)," the researchers wrote. "However, lay users would hardly resort to sophisticated prompts or complex interactions with the LLMs." In a nutshell, sure you can ask ChatGPT questions — but please confer with a real doctor before undertaking any treatment, especially if you're doing it at home. More on ChatGPT: Man Annoyed When ChatGPT Tells Users He Murdered His Children in Cold Blood
The Sun
29-04-2025
- Health
- The Sun
How your eyes could reveal if you have ADHD – plus 9 signs of the condition to look out for
WHILE your eyes are often called the window to your soul, they could also offer clues to conditions like ADHD. While your vision might appear fine, a study published last month found certain characteristics at the back of the eye might point to the behavioural condition. Scientists in South Korea trained computers to spot signs of ADHD, short for attention deficit hyperactivity disorder, by looking at changes in the eye. An AI computer model was able to predict the condition with 96 per cent accuracy just by analysing images of the light-sensitive layer at the back of the eye, also called the retina. The team found key signs in the retina, such as more blood vessels, thicker vessels, and smaller optic discs (how the eye connected to the brain), which could show someone has ADHD. These eye changes may reflect how the brain develops in people with the condition since the retina is closely linked to the brain. "Our analysis of retinal fundus photographs demonstrated potential as a noninvasive biomarker for ADHD screening," the researchers, led by a team from Yonsei University College of Medicine, wrote in their paper. It's estimated that up to 2.6million children and adults in the UK have been diagnosed with ADHD. While globally, around five per cent of people are thought to be affected. It can make people seem restless and cause them to get easily distracted. Most cases are diagnosed in children under 12, but the number of adults being diagnosed is on the rise. Experts think this is due to a combination of increased awareness, decreased stigma, and improved access to healthcare. Do you or your child have ADHD- Here's the NHS test as Brits waiting two years for diagnosis For the new study, published in npj Digital Medicine, experts tested the AI model on eye images from 323 children and teens with ADHD and 323 without. The AI system performed well in predicting ADHD and identifying key signs of the disorder, including differences in the retina and how well someone can focus their attention. Next, the researchers want to try these tests across larger groups of people and wider age ranges. The average age of participants in this study was 9.5 years, but ADHD can present quite differently in adults. A faster and more accurate diagnosis could help many people get the support they need sooner. This is important as undiagnosed ADHD has been linked to struggles at work, relationship breakdowns, and poor mental health. "Early screening and timely intervention can improve social, familial, and academic functioning in individuals with ADHD," the researchers said. The 9 'hidden' signs of ADHD in adults ADHD has long been associated with naughty schoolkids who cannot sit still in class. And that is part of it. Fidgeting, daydreaming and getting easily distracted are all symptoms of the behavioural condition, which is why it is often spotted in children. However, attention deficit hyperactivity disorder is far more complex than simply having trouble focusing. Henry Shelford, CEO and co-founder of ADHD UK, says: 'If it isn't debilitating, it isn't ADHD.' In recent years, social media has given rise to trends which conflate specific personality traits or single behaviours with ADHD. You might be thinking, 'I'm always losing my keys, forgetting birthdays and I can never concentrate at work — I must have ADHD'. But it's not as simple as that. Though these may all point to the condition, Dr Elena Touroni, a consultant psychologist and co-founder of The Chelsea Psychology Clinic, says: 'The key distinction lies in how much a behaviour impacts a person's daily life. 'Genuine ADHD symptoms affect multiple areas of life - work, relationships and emotional wellbeing - whereas personality traits are typically context-dependent and less disruptive.' ADHD UK's Henry, who has the condition himself, adds: 'Having ADHD is hard. One in ten men with ADHD and one in four women with ADHD will at some point try to take their own lives.' So how can ADHD manifest in someone's life? While hyperactivity is a common indicator, here are nine other subtle signs: Time blindness - losing track of time, underestimating how long tasks will take, regularly being late or excessively early Lack of organisation - a messy home, frequently misplacing items, forgetting deadlines Hyperfocus - becoming deeply engrossed in activities for hours Procrastination - feeling overwhelmed by to-do lists and struggling to determine what needs your attention first so focusing on less important tasks Heightened emotions - emotional struggles can manifest in angry outbursts, feeling flooded with joy or shutting down because you feel too much at once Being a 'yes man' - agreeing to new projects at work or dinner dates with friends when you're already busy (a desire to please) Impatience - interrupting people mid-conversation, finding it painful to stand in a queue, being overly-chatty Restlessness - tapping, pacing, fidgeting or feeling restless on the inside Easily distracted - by external things, like noises, or internal things like thoughts
