AI Tool Promises ‘Second Sight of Eyes’ to Clinicians, But Does It Improve Patient Outcomes?


Source: media.npr.org

A Revolutionary AI Tool for Healthcare

A 4-month-old boy comes into the clinic with a fever and a stuffy nose. Vyonne Njeri, a registered clinical officer in Nairobi, Kenya, thinks it’s just a cold. But then a yellow box pops up on her computer, telling her to check the infant’s heart because his heart rate is elevated. Njeri credits an AI tool, called AI Consult, with double-checking her work and potentially saving the child’s life.

AI Consult was tested in a randomized trial of nearly 10,000 patient encounters across 16 Kenyan primary clinics operated by Penda Health. The tool uses OpenAI’s GPT-4o, a large language model, to check electronic notes for possible problems or gaps in care. The AI provides three prompts: green for ‘everything is OK’, yellow for ‘click me’ (indicating a small problem), and red for critical concerns that require immediate attention.

The study found that clinicians who used AI Consult produced better diagnoses and treatment plans, according to an independent panel of six Kenyan family physicians. The AI tool also offered welcome feedback to clinicians, such as ‘you need to check on that’ or ‘you’re on the right track.’

However, the study did not find evidence of improved patient outcomes. There was a 23% decrease in treatment failures, but the decrease was not statistically significant due to the low number of cases.

Challenges and Concerns

Dr. Bilal Mateen, a co-author of the study and chief AI officer at PATH, notes that treatment failures are rare in primary care. A larger trial would be needed to detect a meaningful difference in patient outcomes.

Dr. Nicholas Okumu, a healthcare AI researcher and orthopedic surgeon, warns that new AI systems could make mistakes or give bad advice, leading to grave errors. He emphasizes the importance of ongoing oversight and monitoring to ensure the safe use of AI in healthcare.

The Future of AI in Healthcare

Dr. Jonathan Chen, an associate professor of medicine and biomedical data science at Stanford University, praises the study for going beyond simulated tests with AI systems. He notes that the study has the potential to improve healthcare access, particularly in lower-resource countries.

The study’s findings and the potential of AI in healthcare have sparked excitement and caution. As AI Consult and similar tools continue to evolve, it is essential to address concerns about safety, efficacy, and patient outcomes.

Vyonne Njeri remains optimistic about the tool’s potential: ‘Having this tool really adds value. It helps us remember protocols and new guidelines.’

The study’s results and the ongoing development of AI in healthcare highlight the need for continued research and innovation in this field.