wOur doctors are perfect, tireless and always right. But they are only humans. More and more, they are thin, working long hours, growing under immeasurable pressures, and often have limited resources. Of course, better conditions will help, such as more staff and improved systems. But even the best funded clinics with the most dedicated professionals can still lack standards. Doctors, like all of us, work in the Stone Age mind. Despite years of training, the human brain is not optimally equipped to the pace, pressure and complexity of modern healthcare.
Given that patient care is the core objective of medicine, who or what is the best thing to offer it? While AI can still cause doubt, research has increasingly shown how to fix some of the most persistent issues, from misdiagnosis and errors to unequal access to care, and how to correct missed failures.
As patients, we each face at least one diagnostic error in our lifetime. In the UK, conservative estimates suggest that around 5% of primary care visits fail to diagnose properly, putting millions of patients at risk. In the United States, diagnostic errors cause death or permanent injury to approximately 800,000 people a year. Misdiagnosis is a greater risk when you are among one in 1 in 10 people worldwide with rare diseases.
Modern medicine takes pride in being scientific, but doctors don't always practice what the evidence recommends. Research shows evidence-based treatment is only offered about half the time to adults in the US. Your doctor may not agree to the diagnosis either. In a study of over 12,000 radiology imaging, reviewers offering second opinions opposed the original rating in about one in three cases. As work days go up, quality slips slip even more. Inappropriate antibiotic prescriptions increase, resulting in lower cancer screening rates.
This is surprising, but there is an understandable reason for these mistakes. And from a different angle, it's amazing how doctors get it right just as often as they do. The reality of being human – distraction, multitasking, even our body clock – is at the expense of it. However, burnout, depression and cognitive aging do not only take doctors. They increase the risk of clinical mistakes.
Medical knowledge also moves faster than a doctor can catch up. Until graduation, half of medical students are already outdated. Research takes an average of 17 years to reach clinical practice, and in a new biomedical article published every 39 seconds, even a quick read of abstractions takes about 22 hours a day. There are over 7,000 rare diseases, with 250 more identified each year.
In contrast, AI devours medical data at lightning speed 24/7 with no sleep or bathroom breaks. When doctors change in unwanted ways, AI is consistent. These tools also cause errors, but it's hard to dismiss how impressive the latest models are. Some studies show that clinical reasoning involving complex medical conditions is much better than human physicians.
The AI superpower is finding patterns that humans miss, and these tools are surprisingly good at recognizing rare diseases – often better than doctors. For example, a 2023 researcher provided ChatGPT-4 with 50 clinical cases, including 10 rare conditions. They were asked to provide the diagnosis in the form of ranked suggestions. It solved all common cases by the second proposal, earning 90% of the rare conditions by the 8th time. Patients and their families are increasingly aware of these benefits. One child, Alex, saw 17 doctors in three years due to chronic pain. He could not explain his symptoms. Desperately, his mother turned to ChatGpt. This suggests a rare condition called tether cord syndrome. The doctor confirmed the diagnosis and Alex is currently undergoing appropriate treatment.
Next is the access issue. Healthcare is upside down. Most needy people – the sickest, poorest and most marginalized people in society are the most likely to be left behind. A packed schedule and poor public transport means missed the appointment of millions. Parents and part-time workers, including those working in the gig economy, often struggle to attend physical examinations. American Time Usage Survey Data shows patients sacrifice for two hours for a 20-minute doctor visit. The problems for people with disabilities are often worsened for people with disabilities who are about four times more likely to miss care in the UK due to issues with transportation, costs and long waiting lists. Compared to men without disabilities, women with disabilities are more than seven times more likely to have unmet needs due to the cost of care and medication.
Still, it's rare to question the idea of waiting in line for a town doctorate because it's simply the way things have been done all the time. AI could change that. Imagine where your doctor in your pocket is providing information when and when you need it. Under the workers' 10-year plan, Health Secretary Wes Streeting announced that patients can quickly discuss AI and health concerns via the NHS app. It is a bold step, something that can bring practical clinical advice faster than millions.
Of course, this only works for those who can use it. Internet access is improving worldwide, but there are still serious gaps. 2.5 billion people remain offline. In the UK, 8.5 million people lack basic digital skills, with 3.7 million families below the “minimum digital living standard.” This means low connectivity, outdated devices, and limited support. Confidence is also a barrier. 21% of people in the UK feel that technology remains.
At this point, AI healthcare research is almost exclusively stuck to its flaws. Examining the potential of bias and error technologies is a very important task. However, this orientation does not take into account the squeals and sometimes unsafe systems we already rely on. A fair assessment of AI needs to be weighed against the reality of what we currently have.
Charlotte Brees is a health researcher; Dr. Bott: Why Doctors Can Fail Us, and How AI Can Save LifePublished by Yale September 9th.
Read more
Deep Medicine: How Artificial Intelligence Can Make Humans Humans Again Again by Eric Topol (Basic Book, £28)
Co-Intelligence: Ethan Mollick's Living and Cooperation with AI (Wh Allen, £16.99)
Artificial Intelligence: A guide to human reminiscence by Melanie Mitchell (Pelican, £10.99)
