Health

Doctors Warn Against Reliability of At-Home Gut Microbiome Tests

Published

5 months ago

November 4, 2025

Medical experts are cautioning consumers about the growing popularity of at-home gut microbiome tests, saying the products are largely unregulated, often unreliable, and offer little practical medical insight beyond personal curiosity.

The tests, marketed by private companies for €90 to €460 or more, claim to analyse the trillions of bacteria living in a person’s intestines and provide personalised health insights. But gastroenterologists say the science behind these promises is still too limited to make meaningful recommendations.

“There’s a world in your gut that we still don’t know a lot about,” said Dr. Mark Benson, a gastroenterologist at the University of Wisconsin-Madison. “While there’s growing evidence that changes in the gut microbiome are associated with diseases like diabetes, liver disease, and obesity, we don’t yet know if those changes are a cause or a consequence.”

The gut microbiome — a complex ecosystem of bacteria that aids digestion, immune function, and even mental health — varies widely between individuals, and even within the same person over time. “Most of the variability between people, we don’t understand,” said Dr. Eamonn Quigley, head of gastroenterology and hepatology at Houston Methodist Hospital.

Most consumer microbiome tests require a stool sample, which is analysed by a private lab. Customers receive a report comparing their bacterial mix to what the company defines as an “ideal” microbiome. But experts note that no official medical standards exist for such comparisons. Unlike clinical stool tests ordered by doctors to diagnose infections or inflammation, these direct-to-consumer products are not federally regulated and vary widely in quality.

“When patients bring in these reports, there’s often little we can do with them,” said Dr. Sean Spencer, a physician-scientist at Stanford University. “Beyond antibiotics and diet, we don’t have reliable tools to change the microbiome.”

Some companies also blur the line between testing and sales. Nearly half of microbiome testing firms also sell supplements they recommend based on test results, according to Science Policy Forum. For example, wellness brand Thorne includes supplement suggestions with its reports. While its chief scientific officer Nathan Price denies any conflict of interest, critics warn that such practices can mislead consumers into unnecessary spending.

Doctors stress that people don’t need expensive tests to improve gut health. “Eat more plant-based fibre and protein, get good sleep, and exercise,” Benson advised. “Those habits have proven benefits — the tests don’t.”

Researchers remain hopeful that microbiome science will one day lead to targeted treatments for digestive and metabolic disorders. But for now, Quigley said, “We’re still in the early stages. Most people are simply wasting their money.”

Health

Novo Nordisk Teams Up With OpenAI to Accelerate Drug Discovery Using AI

Published

10 hours ago

April 14, 2026

Web Reporter

Danish pharmaceutical giant Novo Nordisk has announced a new partnership with OpenAI aimed at integrating artificial intelligence across its drug development and business operations.

The collaboration, revealed on Tuesday, is expected to help the company identify new treatments more quickly and improve how medicines are developed, produced and delivered to patients. Novo Nordisk said the use of advanced AI tools will allow it to analyse vast and complex datasets, uncover patterns that were previously difficult to detect, and shorten the timeline from research to patient access.

Chief executive Mike Doustdar said the agreement marks an important step in positioning the company for the future of healthcare. He noted that millions of people living with chronic conditions such as obesity and diabetes still require better treatment options, adding that new therapies remain to be discovered.

Novo Nordisk is widely known for its leading treatments in these areas, including Ozempic and Wegovy, which have seen strong global demand in recent years. The company said integrating AI into daily workflows will allow its teams to test ideas more rapidly and bring innovations to market at a faster pace.

The partnership will not be limited to research and development. Both companies plan to apply AI tools to manufacturing processes, supply chains and commercial operations, with pilot programmes already set to begin. Full integration is expected by the end of the year.

Sam Altman said artificial intelligence is transforming industries and has the potential to significantly improve outcomes in life sciences. He added that the collaboration would support faster scientific discovery and more efficient global operations, helping to shape the future of patient care.

The move comes as pharmaceutical companies increasingly turn to AI to gain an edge in drug discovery. Novo Nordisk has already invested in innovation through initiatives such as the Danish Centre for AI Innovation, developed in partnership with Nvidia and Denmark’s export and investment fund.

Competition in the sector is intensifying. US-based Eli Lilly, a key rival in the weight-loss drug market, recently announced its own AI-focused collaboration with Insilico Medicine to develop new treatments. The agreement, valued at up to $2.75 billion, highlights the growing role of AI in reshaping pharmaceutical research.

Industry analysts say such partnerships reflect a broader shift toward data-driven innovation in healthcare, where the ability to process and interpret large volumes of information is becoming increasingly important.

For Novo Nordisk, the partnership with OpenAI signals a commitment to staying at the forefront of this transformation, as companies race to harness technology in the search for new and more effective treatments.

Health

Study Finds AI Models Fall Short in Early Medical Diagnosis

Published

11 hours ago

April 14, 2026

Web Reporter

A new study has found that artificial intelligence language models still struggle with one of the most critical aspects of medical care, raising concerns about their use without human oversight.

Researchers from Mass General Brigham reported that AI systems failed to produce an appropriate early diagnosis more than 80 per cent of the time. The findings, published in JAMA Network Open, highlight ongoing limitations in how these systems reason through complex clinical scenarios.

The study examined 21 large language models, including systems developed by OpenAI, Google and xAI. Among those tested were versions of GPT, Gemini, Claude, Grok and DeepSeek.

Researchers used a structured evaluation tool known as PrIME-LLM to assess how well the models handled different stages of clinical reasoning. These stages included forming an initial diagnosis, ordering tests, reaching a final diagnosis and planning treatment. The models were tested using 29 standardised clinical scenarios, with information introduced gradually to mirror real-life patient cases.

While the systems showed relatively strong performance when identifying a final diagnosis, their ability to generate a differential diagnosis — a key step in distinguishing between conditions with similar symptoms — remained limited. This early-stage reasoning is widely regarded as essential in medical decision-making.

Marc Succi, a co-author of the study, said current models are not ready for independent clinical use. He noted that differential diagnosis represents a core part of medical practice that AI has yet to replicate effectively.

Another researcher, Arya Rao, said the findings show that AI performs best when given complete information but struggles when cases are still developing. She explained that the models are less reliable in situations where doctors must make judgments based on limited or uncertain data.

Despite these shortcomings, the study identified a group of higher-performing systems, including advanced versions of GPT, Gemini, Claude and Grok. These models achieved final diagnosis success rates ranging from around 60 per cent to over 90 per cent when provided with detailed clinical data such as lab results and imaging.

Experts not involved in the research also stressed the importance of caution. Susana Manso García said the findings reinforce that AI should not replace professional medical judgement. She advised that patients continue to seek guidance from qualified healthcare providers when dealing with health concerns.

The study concludes that while AI has made progress, it still requires close human supervision in clinical settings. Researchers say the technology shows promise as a support tool, but its current limitations mean it cannot yet be trusted to make independent medical decisions.

Health

Genetic Differences May Shape Effectiveness of Popular Weight-Loss Drugs, Study Finds

Published

4 days ago

April 10, 2026

Web Reporter

Genetic variations may help explain why some patients respond better than others to widely used weight-loss medications, according to new research that points to the potential for more personalised treatment approaches.

Drugs such as Ozempic, Mounjaro and Zepbound have transformed the treatment of obesity in recent years. These medications belong to a class known as GLP-1 receptor agonists, which mimic a natural hormone that regulates appetite and blood sugar, helping people feel full for longer. Despite their growing use, patient outcomes vary widely, with some individuals losing less than 5 percent of their body weight while others achieve reductions exceeding 20 percent.

The study, conducted by researchers at the 23andMe Research Institute and published in Nature, examined genetic data alongside patient-reported experiences to better understand these differences.

Researchers analysed information from nearly 28,000 participants who had taken GLP-1 medications for a median period of just over eight months. Their findings identified specific genetic variants that appear to influence how individuals respond to these treatments.

One such variation in the GLP1R gene was linked to improved effectiveness. Individuals carrying a particular version of this gene lost an average of 0.76 kilograms more than those without it during the study period. Another variant in the GIPR gene was associated with an increased likelihood of side effects such as nausea and vomiting among patients taking tirzepatide-based drugs, though it did not affect weight loss outcomes.

Noura Abul-Husn, chief medical officer at the research institute, said current approaches to weight management often rely on trial and error. She noted that patients frequently begin treatment without clear expectations about how effective a drug will be or what side effects they might experience.

Experts not involved in the study said the findings offer useful insight but should be interpreted with caution. Marie Spreckley of the University of Cambridge said the genetic effects identified are relatively small in clinical terms, especially compared with the typical weight loss of 10 to 15 percent seen in trials of these medications. She added that factors such as dosage, treatment duration, sex and drug type likely play a larger role in determining outcomes.

Still, researchers believe the results could mark a step toward more tailored therapies. Cristóbal Morales, a specialist in metabolic health in Spain, said the ability to predict how patients will respond to treatment through pharmacogenomics could improve both drug selection and safety.

The findings highlight the growing interest in personalised medicine, where treatments are adapted to an individual’s genetic profile, though further studies are needed to confirm how these insights can be applied in clinical practice.