Connect with us

Health

AI Chatbots Make Faster, Less Nuanced Kidney Transplant Decisions Than Doctors

Published

on

Artificial intelligence chatbots may make faster and more confident decisions than human doctors when determining which patient should receive a life-saving kidney transplant, but a new study found that the systems often rely on fewer factors and show less sensitivity to the ethical complexity of such choices.

Researchers from Penn State University in the United States compared decisions made by large language models (LLMs) with choices made by people in earlier studies on kidney allocation. The researchers presented AI models with hypothetical cases based on existing datasets in which human participants had previously chosen between two patients competing for a single available kidney.

The patients were described using characteristics including age, health and drinking habits. Both were considered eligible for the transplant, requiring the decision maker to determine which patient should receive the available organ.

The researchers tested the AI systems in several ways. Some scenarios focused on individual characteristics, while others combined several traits to examine how the models handled competing considerations. In some tests, participants and AI systems were also given the option of flipping a coin, allowing researchers to assess indecision as part of moral decision-making.

Human participants generally placed greater importance on age, often favouring younger patients over older patients. Many of the AI models, however, placed greater weight on lower alcohol consumption.

Hadi Hosseini, the study’s lead researcher at Penn State University, said the findings showed that AI chatbots can diverge from human values when weighing characteristics of patients.

The researchers found that human decisions tended to consider several factors together and were more sensitive to context. LLMs, by contrast, frequently focused on a single characteristic when making their choices.

See also  Health Experts Warn of Winter Dehydration Risks

Another difference involved uncertainty. Human participants were more likely to recognise that organ allocation can involve competing ethical considerations and that there may not be one objectively correct answer. AI models generally showed little hesitation and committed to one choice.

John Dickerson, chief executive officer of Mozilla.ai and a collaborator on the study, said humans recognise ambiguity when allocating scarce resources and use debate to shape how such decisions are made.

The findings come as AI systems are increasingly being used in healthcare for clinical workflows, diagnosis, treatment planning and decisions involving limited medical resources.

The researchers said kidney allocation presents a particularly important test because decisions over deceased-donor and living-donor organs involve medical, ethical and moral considerations.

Hosseini said the ethical stakes are high because decisions involving organ allocation can directly affect patients’ lives. He stressed that the study was not intended to encourage replacing professional medical judgment with AI.

Instead, the researchers said understanding how AI systems behave in high-stakes situations is becoming increasingly important as individuals, organisations and businesses use AI to make decisions or provide recommendations.

Health

Frankfurt Issues Malaria Warning After Six Cases Reported Since July

Published

on

Health authorities in Frankfurt have warned residents and people working near the city’s airport to seek medical advice if they develop an unexplained fever after six malaria cases were reported in the city since July.

Two of the cases have been fatal, while two additional infections have recently been identified in the Schwanheim district, which is located close to Frankfurt Airport. The city’s public health authority is investigating whether the latest cases are connected.

Several of the recent infections have involved airport employees. Authorities suspect that mosquitoes capable of transmitting malaria may occasionally arrive in Germany on aircraft travelling from regions where the disease is widespread. If the insects survive and later bite people, they can potentially transmit the infection.

Malaria is a tropical disease transmitted through the bite of an infected mosquito. It is not normally spread directly from one person to another.

Frankfurt’s health authority said early diagnosis is important because malaria can initially resemble common illnesses such as influenza or a cold. Typical symptoms include fever, chills, headaches, pain in the limbs and gastrointestinal problems.

The Robert Koch Institute has warned that these symptoms can sometimes be mistaken for flu-like illnesses or stomach infections. The most dangerous form of the disease is malaria tropica, caused by the Plasmodium falciparum parasite. Without prompt diagnosis and treatment, it can become life-threatening.

People who live, work or spend significant amounts of time near Frankfurt Airport have been advised to consult a doctor if they develop unexplained fever. They should tell medical staff that they live or have recently spent time in the area around the airport.

See also  Health Experts Warn of Winter Dehydration Risks

The city’s public health authority is also providing information through a telephone hotline, available from 9am to 6pm on 069/212-30030.

Residents can take measures to reduce the risk of mosquito bites by using insect repellents and wearing clothing that covers the skin. Insect nets can also provide protection. Repellents containing DEET or icaridin are among the products recommended by Germany’s Federal Foreign Office.

Cases of so-called airport malaria remain rare in Germany. Research cited by the Robert Koch Institute indicates that more than 99% of malaria cases recorded in Europe are acquired in endemic regions outside the continent.

Between 1969 and early 2024, databases recorded 145 malaria cases in which patients were infected within Europe. More than two-thirds were classified as airport malaria. France recorded 52 cases, Belgium 19 and Germany nine, including eight in Frankfurt and one in Munich.

The two people recently infected in Frankfurt’s Schwanheim district are receiving hospital treatment, according to the city authorities.

The latest cases have prompted renewed attention to the possibility of malaria transmission in areas surrounding international airports, particularly when unexplained fever occurs without a recent trip to a malaria-endemic country.

Continue Reading

Health

Weather Data Could Help Predict Cardiac Arrest Surges Days in Advance, Study Finds

Published

on

Weather forecasts could help predict periods when out-of-hospital cardiac arrests are likely to rise above normal levels by as much as three days in advance, according to a new study by Hungarian researchers.

The study, published in the journal Public Health, found that falling temperatures were among the most significant weather-related factors associated with cardiac arrest. Researchers found that every 1 degree Celsius decline in average daily temperature was linked to an average 1.4% increase in the number of daily out-of-hospital cardiac arrests.

The researchers also found a seasonal difference in cardiac arrest rates. Almost 18% more out-of-hospital cardiac arrests occurred during winter than during summer.

The findings were based on a nationwide analysis of more than 114,000 cardiac arrest cases recorded in Hungary between November 2018 and December 2023. Researchers from Semmelweis University in Budapest, the Budapest University of Technology and Economics and the Hungarian National Ambulance Service compared the cases with daily weather measurements.

The analysis examined several meteorological factors, including temperature, wind speed, air pressure, humidity and air quality.

The researchers found that the risk was not limited to sudden periods of extreme cold. Generally lower temperatures were also associated with higher numbers of cardiac arrests. Rapid changes in temperature could provide another useful warning indicator, with a daily decline of more than 5 degrees Celsius potentially signaling an increase in cases.

One important finding was that the effects of weather changes did not always appear immediately. An increase in cardiac arrests could occur as much as three days after the weather conditions changed.

See also  Saudi Arabia and Mayo Clinic Partner to Develop AI Model for Healthcare

The researchers said meteorological forecasts could eventually be used to develop early warning systems for periods when cardiac arrest cases are expected to increase. Such systems could help people with cardiovascular disease and their families take additional precautions during higher-risk periods.

Forecasting could also help ambulance services and hospitals prepare for a potential increase in emergency cases, allowing them to adjust staffing and resources when needed.

Out-of-hospital cardiac arrest is a major cause of death in developed countries. In Europe, between 67 and 170 cases occur for every 100,000 people each year, while the survival rate remains below 10%.

Researchers said identifying factors that can signal an approaching rise in cardiac arrests could help emergency services prepare for increased demand. Early warnings could also reinforce the importance of immediate assistance and resuscitation when cardiac arrest occurs.

The study suggests that weather information could eventually become one component of wider efforts to anticipate emergency healthcare demand and improve preparedness for cardiovascular emergencies. However, further work would be needed before such forecasting systems could be used routinely in healthcare settings.

Continue Reading

Health

Von der Leyen Says AI Should Support Doctors, Not Replace Them

Published

on

Artificial intelligence could bring major advances to healthcare across Europe, but it should never replace doctors, European Commission President Ursula von der Leyen said during her State of the Union speech at the European Parliament.

Von der Leyen highlighted healthcare as an area where AI could deliver practical benefits for patients while stressing that medical decisions should continue to involve human professionals.

“Do I want my doctor to have instant AI access to all the data needed to make the best diagnostic or treatment choice? Yes, of course. But do I want my doctor to be a robot, AI-assisted? No, of course not,” she said.

The Commission president, who previously worked as a doctor, said she did not want a machine to determine whether a patient has cancer. At the same time, she argued that AI can help doctors process large amounts of medical information and improve the detection and treatment of diseases.

She pointed to AI-supported mammography as an example, saying the technology can identify more cancers at an earlier stage during screening. Earlier detection can improve the chances of successful treatment and survival for patients.

AI is already being developed and used across several areas of healthcare, including medical diagnostics, drug development and administrative tasks. Supporters of the technology say it could help healthcare professionals manage workloads and analyze complex medical information more efficiently.

However, the expansion of AI in medicine has also raised concerns about reliability, accountability, privacy and the handling of sensitive biological information.

Research has found that language models can provide unsafe or inaccurate responses to medical questions and may fail to assess the urgency of certain conditions correctly. Experts have also raised concerns about how AI systems should be allowed to access and process sensitive health and biological data.

See also  Long-Term Air Pollution Exposure Linked to Higher Risk of Common Brain Tumour, Study Finds

The World Health Organization has warned that gaps in legal accountability, uneven investment in workforce development and the risk of excluding some groups require continued attention as healthcare systems adopt AI.

According to a recent WHO report, only 8% of its member states have developed a national strategy specifically focused on AI in health. The organisation said this highlighted the need to match ambitions for AI adoption with concrete policies and safeguards.

Von der Leyen said the European Union should pursue AI development in a “European way”, with technology used to deliver benefits while maintaining human oversight.

She also announced that the European Commission plans to present a series of initiatives in November targeting sectors where industrial AI could have significant potential.

The areas identified include healthcare, transport, agri-food, advanced manufacturing, defence and space. The initiatives are expected to form part of the EU’s wider effort to expand AI use while addressing the risks associated with the technology.

Continue Reading

Trending