Why AI Will Have Limited Impact on Population Health

A Sensible Medicine Guest Post

As is to be expected in the 2020s, Sensible Medicine published a long list of articles on AI. Today, we are pleased to publish a commentary by Dr. Howard Bauchner, who is betting against AI in one area of healthcare, population health. Although it seems unwise to bet against our future overlords at this moment, I think Dr. Bauchner is on to something. I agree with him in part because many of the interventions he discusses, such as managing hypertension and hypercholesterolemia, require genuine patient-doctor relationships and effective communication skills, which might take hundreds of years for AI and humans to adapt to.

Adam Cifu

AI is all around medicine. It is helping clinicians with administrative tasks, and there are dozens of articles suggesting it is as good as, if not better than, physicians at diagnosis. And it can certainly help in reading radiographs, ECGs, MRIs, and videos from colonoscopy. It is quite likely that it will help refine therapy. But the gains from these advances, if studies demonstrate an impact on improving health outcomes, will be at the margins. More patients may be diagnosed with breast, prostate, or colon cancer, but this is likely to have a limited impact on the entire population and thus have little or no impact on population health. By population health, I mean, the health, well-being, and healthy life-expectancy of the entire population.

Why? Because the keys to population health are known, but as a healthcare and political system, we continue to fail. Diagnosing and treating hypertension, hypercholesterolemia, and obesity, better nutrition, and more exercise are the medical keys to improving population health. Yet, the US and many countries do poorly with diagnosing and treating both hypertension and hypercholesterolemia, neither of which requires complicated diagnostic algorithms or therapies. Obviously, the drug revolution is helping with obesity. Nutrition has become a focus of the Trump Administration, but whether they can have any more success than previous administrations is unclear. For decades, the medical community has urged the population to eat healthy foods and exercise more. It is not clear, and I would suggest, unlikely that AI will help improve the diagnosis and treatment of hypertension or improve the nutritional status of the population.

The non-medical keys to improving population health – eliminating poverty, and better education and housing – are not the focus of the Trump Administration, and it is not likely that AI will help with any of these challenges. Indeed, it is quite likely, given recent reforms, that more people will be categorized as living below the poverty line, and millions will lose their health insurance, making diagnosing and treating hypertension, hypercholesterolemia, and obesity even more difficult.

So we should enjoy this AI revolution. But we should not expect much from it with respect to population health, which requires a more focused commitment from the healthcare system and state and federal governments. For the healthcare system, a focus on treating hypertension and hypercholesterolemia is critical. We failed at the primary prevention of obesity, but researchers and drug companies have come to our rescue with the introduction of a growing list of anti-obesity drugs, which appear to have limited side effects but huge health benefits. Over time, these drugs will become more affordable for people with health insurance. From a governmental standpoint, it is possible to move people out of poverty, as demonstrated by the Child Tax Credit, And that should be applauded. Housing remains a national crisis, which will require a concerted effort by local, state, and federal governments. It has been said many times that in a country as rich as the US, we must ensure a reasonable standard of living for everyone. But that requires political will – something that seems in short supply at the moment.

Howard Bauchner, MD, is a Professor of Pediatrics and Public Health at the Boston University Chobanian & Avedisian School of Medicine. He is also a Visiting Scholar at the National University of Singapore. Dr. Bauchner is also the former Editor in Chief of JAMA and the JAMA Network and former Editor in Chief of Archives of Disease in Childhood.