Latest News

Health care uses many kinds of help. Some help works well. Some help does not work well. The way help is used can shape the final result. Artificial intelligence can act as a smart helper in this system. It can look at many kinds of health facts at once. Those facts can come from patient records, signs, images, and costs. This may assist doctors in making care choices that are more careful and better timed. The idea is simple. Better data can lead to better care choices. Better care choices can lead to better use of time and money. The system may then avoid low value care. It may also give strong reasons to build care for smaller groups of patients. That could change how health care invests in new options. Predictive AI can guide two main kinds of choices. The first choice can spot patie

Artificial intelligence now drafts replies to patient messages inside electronic health records. The goal was to cut down doctor workload. But a recent study found that physicians still spend extra time reviewing and changing these drafts. The research looked at over fourteen thousand responses from more than a thousand clinicians at one major medical center between April twenty twenty-four and August twenty twenty-five. Each reply was generated by AI, then touched by a human doctor before sending. Researchers built a fifteen-part system to categorize every edit doctors made. The most common changes involved scheduling appointments and giving lifestyle tips. Many edits were moderate, shifting the message's meaning without altering the actual treatment plan. Editors frequently rephrased in

International researchers ran a retrospective study to see if continual learning could help an AI model that checks endotracheal tube placement on chest X‑rays. They tested three ways to use the model: keep it as‑is, fine‑tune it at each new hospital, or let it learn continuously across sites. The dataset included 2,313 intensive‑care chest radiographs from 23 hospitals in 12 countries across five continents. One hospital was set aside for external validation. Patients were adults (≥18) who had tubes placed in 2021 or earlier when there were too few recent cases. The team gathered cases through keyword searches in hospital systems and cross‑checked them with radiology reports. Each hospital contributed roughly 100 anteroposterior images, of which 50 were used for training. The images were

Researchers looked at health data for American Indian and Alaska Native people. They found a scary pattern. Deaths from alcohol and drugs are way too high in many areas. The numbers cover 2018 through 2023. Some counties see rates near 1,850 per 100,000 people. That is a lot higher than the national average. The team used mapping tools to spot trouble zones. They checked 78 counties across 13 states. These places got flagged for deeper review. The study pulled info from the CDC and the Indian Health Service. It also looked at highway rest stops to guess where risks might cluster. All this happened between February and August 2025. Access to care is a huge piece of the puzzle. During normal business hours, over 22,000 Native people live far from help. After hours, that number jumps to mor

The researchers took 200 real‑world clinical notes from patients seen in outpatient and emergency settings. They fed each note to 24 different AI assistants before the treating physicians wrote their own plans. Both the AI suggestions and the doctors’ recommendations were later parsed by three separate AI tools and priced using 2026 Medicare rates. The main goal was to see how much the AI‑suggested care cost compared with what the physicians ordered. The numbers were striking. Across the 20 general AI models, the average bill for AI‑suggested care was about $314 per visit, while doctors’ plans averaged $97. That meant AI added roughly $216 per encounter, or about 3. 2 times the physician cost. The extra expense broke down into about $105 for diagnostic tests and $111 for specialist referra

Many cities in the Netherlands try out game-based apps to get people active. Most of these experiments stop after the trial phase. A recent study looked at how local governments choose, launch, and expand these digital activity programs. The researchers used one app called Missie Master as their main example. In the first part, the team interviewed city officials, health workers, and the business that made the app. In the second part, they held group workshops and talked to cities new to this idea. The results show that cities often adopt an app because a staff member loves it or because another city recommended it. They also need to believe the app is reliable. Generic sales messages from companies are usually ignored. Getting the program off the ground requires early preparation. A cit

Researchers looked at how much time American teens spend outside. Being outdoors offers chances for exercise, sunlight, and time with friends. They used a big national survey called the American Time Use Survey. The data covers twenty years, from 2003 to 2023. More than fourteen thousand teenagers, ages fifteen to nineteen, kept time diaries. This helps us see if outdoor habits are changing across the country. For a long time, boys spent more time outside than girls. That gap has disappeared. Boys' outdoor time dropped a lot. Now boys and girls spend about the same small amount of time outdoors. Both groups are at low levels. The old pattern where boys led the way is gone. Spending more time on computers for fun links to less time outside. This connection is stronger for boys. On the oth

Hospitals are testing a new AI system that could spot dangerous kidney problems in patients before they get worse. The system uses a type of smart computer program called a transformer, which is good at understanding patterns in complex data over time. Doctors know that catching kidney injury early can save lives. But patient records have so much information that it can be hard to find the right clues. This new tool tries to make sense of all that data by looking at lab results, medical notes, and other health details all at once. The researchers built something called BAT-AKI, which stands for Biomarker-Aware Transformer for AKI. It pays special attention to changes in body chemicals that doctors check through blood and urine tests. These biomarkers often change before a patient gets re

Researchers tested how well the Indonesian version of a children's intelligence test holds up. They looked at 16 subtests from the WISC-V-ID with over 1,500 Indonesian kids aged 6 to 16. The goal was to see if the test's internal structure makes sense when used in Indonesia. The team ran two types of statistical analyses. First, they used exploratory factor analysis to find natural groupings in the data. This approach clearly supported a four-factor model. A five-factor version did not work as well. One subtest stood out as problematic. The Picture Concepts section performed poorly in the Indonesian adaptation. When researchers removed it, the results improved significantly. Confirmatory factor analysis showed both four- and five-factor higher order models fit the data adequately. But t

Researchers in Hartford, Connecticut gave us a key idea about sickness. They looked at a mix called SAVA. That stands for HIV, substance use, and violence. These three issues fed into each other. People facing one often faced the others too. This showed that illness does not happen in isolation. The city became the birthplace of syndemic thinking. Older health models treat diseases like separate boxes. One box for HIV. One box for addiction. One box for violence. That view misses how life works. Social conditions shape who gets sick and who stays well. Syndemic theory pushes us to see the whole picture. It asks us to connect health with the world around us. Later, the focus shifted to HIV. Harm reduction and better treatment lowered HIV's role in the mix. Yet the deeper conditions did no

Typhoid fever is a serious illness. It comes from a germ named Salmonella Typhi. Many countries have seen fewer cases over the last thirty years. But it is still a big problem around the world. In 2021, about 7. 1 million people got sick with typhoid. The true number could be between 5. 6 million and 9. 2 million. The World Health Organization recommends a vaccine called TCV. Ten countries have already added it to their programs. Indonesia is not one of them yet. Leaders in Indonesia want to know if they should add it too. To help them decide, researchers ran a cost study. They visited 23 hospitals and local health centers across six provinces. They looked at patients who stayed in the hospital with typhoid. The patients were children from 6 months to 17 years, and adults 18 years and older.

On July 23, 2025, the administration that assumed office in January 2025 released a bundle of executive orders and policy papers about artificial intelligence. The documents cover many areas of tech and set the stage for a new direction in U. S. strategy. They were released with little fanfare but quickly drew attention from analysts and industry watchers. One of the orders tells federal agencies to stop buying or using AI systems that push diversity, equity, and inclusion goals. The language avoids the term "woke AI" but makes the stance clear. A separate 28‑page action plan follows, declaring that the United States is locked in a global contest for AI supremacy. It calls for more research, faster innovation, and less regulation, especially in health care, where the plan says rules are ho