Blog

AI-supported reading in breast cancer screening reduces workload and improves detection rates

AI-supported reading in breast cancer screening reduces workload and improves detection rates

Out of all the medical artificial intelligence promises made over time, a single instance has crossed the entire finish line: a randomized trial, a prespecified primary endpoint, and a published outcome. Focus on breast cancer screening, specifically through the MASAI trial. Spanning 105,934 women randomly assigned across the national screening initiative in Sweden, artificial intelligence-assisted analysis uncovered 29% more malignancies compared to traditional double interpretation—all while keeping false positives steady and slashing reading burdens by 44%. Furthermore, 2026 brought forward the previously missing metric: interval cancers, which serves as the definitive benchmark for evaluating whether a screening protocol remains safe.…
Read More
Why solitary arts activities do not reduce loneliness or improve health

Why solitary arts activities do not reduce loneliness or improve health

A GP prescribing a painting class instead of an anxiolytic is a powerful image, and arts-on-prescription programmes have spread across half of Europe on the back of it. The most complete systematic review finds a significant improvement in wellbeing. And, at the same time, that the evidence on anxiety and depression comes down to practically a single study.There are three further findings that ought to change how these programmes are designed and rarely do. Among older adults, arts-based activities outperform both physical and social ones for reducing loneliness. In a British household panel of 23,706 people, the benefit appears only…
Read More
Do doctors lose skill when they use AI? Detection fell from 28.4% to 22.4%, and the next trial found nothing

Do doctors lose skill when they use AI? Detection fell from 28.4% to 22.4%, and the next trial found nothing

In August 2025, The Lancet Gastroenterology & Hepatology published a finding that travelled the world: at four Polish centres, endoscopists who had spent months working with AI-assisted detection found fewer adenomas when performing a colonoscopy without it. The rate fell from 28.4% to 22.4%. Four months later, a Norwegian prospective trial with 5,013 colonoscopies and three phases found exactly nothing: neither loss nor gain of skill when the tool was withdrawn. And a third study, over three years, found sustained improvement. This article sorts out a scientific controversy that is live right now, and explains what separates the three designs.The…
Read More
mental health chatbot effectiveness is real but limited

mental health chatbot effectiveness is real but limited

Meta-analyses consistently indicate that mental health chatbots alleviate symptoms of both anxiety and depression. Although an effect size hovering around 0.3 remains modest within this academic field, it is nonetheless tangible. The core issue surfaces upon reviewing individual clinical trials one by one. Specifically, when a specialized therapy application was pitted against standard ChatGPT, the dedicated app failed to surpass the generic bot across any evaluated metric. Meanwhile, a six-week study tracking 540 participants revealed that the artificial intelligence variant boosted interaction frequency by 2.4 times and engagement duration by 3.8 times, yielding symptom reduction comparable to traditional static workbooks.…
Read More