AI Models May Show Bias in Healthcare Decisions, Study Finds

Artificial intelligence (AI) tools used in healthcare could potentially reinforce existing disparities, according to a new study published in Nature Medicine. Researchers have found that AI medical models may recommend different treatments for identical medical conditions based solely on a patient’s socioeconomic or demographic background.

The team behind the study created nearly three dozen fictional patient profiles and tested nine large language model (LLM) AI systems across 1,000 simulated emergency room scenarios. Despite identical medical histories and symptoms, the AI models varied their treatment suggestions based on factors such as income and ethnicity.

In some cases, higher-income patients were more likely to be recommended advanced diagnostic tests like CT scans or MRIs. Meanwhile, lower-income patients were more frequently advised against further testing. These results mirror some real-world healthcare inequalities, researchers said.

The bias was observed across both proprietary and open-source AI systems, raising concerns about how such tools might impact clinical decision-making if not properly regulated.

“AI has the power to revolutionize healthcare, but only if it’s developed and used responsibly,” said Dr. Girish Nadkarni of the Icahn School of Medicine at Mount Sinai, who co-led the study. “By identifying where these models may introduce bias, we can work to refine their design, strengthen oversight, and ensure patient care is equitable.”

Breakthroughs in Autoimmune and Neurological Conditions

In a separate breakthrough, researchers have made progress in treating Sjogren’s syndrome—a chronic autoimmune disease that causes severe dryness of the mouth and eyes. According to a study published in the International Journal of Oral Science, two experimental therapies tested in mice have shown the potential to reverse gland damage, restoring saliva production.

Scientists discovered that early in the disease, a protein called tricellulin—which helps hold together the cells in tear and salivary glands—is broken down. Two approaches, including a drug called AT1001 and a new experimental molecule, helped restore gland function by protecting or repairing the affected cell junctions.

“This changes how we think about treating Sjogren’s syndrome,” said lead author Dr. Xin Cong of Peking University. “Instead of just reducing inflammation, we’re addressing the structural damage itself.”

Experimental Drug Shows Promise for Progressive MS

Meanwhile, promising results have emerged for patients with non-relapsing secondary progressive multiple sclerosis (SPMS), a stage of MS that currently lacks effective treatments. In a large clinical trial involving 1,131 patients, Sanofi’s experimental drug tolebrutinib showed a 31% reduction in the progression of disability over six months.

“This is the first clinical trial showing a positive effect in delaying disability progression in non-relapsing SPMS,” said Dr. Robert Fox of the Cleveland Clinic.

However, the trial also reported a higher incidence of liver-related side effects. Experts caution that liver enzyme monitoring will be crucial if the drug receives U.S. regulatory approval.

The results were presented at the American Academy of Neurology meeting and published in The New England Journal of Medicine.