Dr. Zach Kohane is the founding editor-in-chief of NEJM AI, a new journal from the publishers of the New England Journal of Medicine.
He is also the dean of the Department of Biomedical Informatics at Harvard Medical School, and has the rare dual expertise of being a physician and a doctorate in computer science.
He talked with Daniel about how to cut through the noise and decide which of the AI health hype is worth your attention.
This interview has been edited for length and clarity.
Will AI really revolutionize care?
I think there's a tremendous amount of hype. And by any standard, I think the number of companies that will make a meaningful contribution in five years will be small.
Still, I believe that minorities, important and decisive minorities, can be very influential.
Where can we see its impact?
Currently, primary and preventive healthcare is collapsing. That's what we're facing.
In that context, the question is not “Will this make a difference?” That's exactly what I'm saying: “In any case, things are changing, but there are huge disparities.”
How do we make the most of the workforce we have, rather than the workforce we want?
We have fully trained physician assistants and nurse practitioners. How can AI be used in conjunction with these populations to provide primary and preventive care?
What are you looking forward to in the coming months and years?
The pace of change is accelerating, so many of the problems we know – prejudice, illusion, lack of up-to-date information, lack of common sense – will not look the same next year.
I think it's going to be different than what we've seen in medicine over the last 30, 40 years.
Is the government ready?
no.
Frankly, most of us wouldn't even know what to do with the government in this situation.
So I think it's appropriate that the government is saying, as part of the Biden executive order, that we need to get people who know about this together and start looking at it.
Here, we explore the ideas and innovators shaping healthcare.
The power of “exaltation.” For simple activities, Improve mood after traumatic loss, according to a new study. So-called highs such as completing tasks, getting enough sleep, eating out, spending time with family, and connecting with friends served as a positive balance for people dealing with the death of a loved one.
Bonus: This finding held true regardless of socioeconomic status or age.
Share your thoughts, news, tips and feedback with Carmen Paun. [email protected]Daniel Payne [email protected]Ruth Reeder [email protected] or Erin Shoemaker [email protected].
Send your tip securely Through SecureDrop, Signal, Telegram, or WhatsApp.
AI could create doctors You'll be better at your job, but don't expect that to eliminate the stigma that has long plagued our health care system.
How is it? New research in natural medicine Researchers from Northwestern University, Massachusetts Institute of Technology, Stanford University and others have found that AI systems could help doctors make the right diagnosis more often.
However, the effects were not always evenly distributed, with patients with lighter skin showing greater improvement.
This study investigated how AI systems impact the diagnosis of dermatologists and primary care physicians.
According to the study authors, the AI system was excellent at diagnosing a variety of skin tones, and 459 primary care physicians and 389 dermatologists used the system to more accurately diagnose diseases. It is said that it became
However, primary care physicians improved their diagnosis of light-skinned patients compared to darker-skinned patients.
remove: AI has the potential to be a powerful tool to improve care and allow healthcare providers to check their blind spots.
But that doesn't mean AI will reduce bias. In fact, this study suggests that AI has the potential to cause significant disparities in care.
Scientists use A study by researchers at the University of Virginia shows that AI can help identify promising drugs and, importantly, better understand how and why those drugs work. .
research, Published in the Proceedings of the National Academy of Sciencesdescribes the development of a computer model to identify cardiac drugs that may reduce scarring after a heart attack and produce better patient outcomes.
Why? The researchers based their machine learning model on decades of human knowledge about the effects of drugs on fibroblasts, a type of connective tissue cell that helps the heart repair after injury but can leave scars. I trained.
Previous attempts to identify agents that reduce scarring have generally been unsuccessful due to gaps in knowledge. This is because it is often unclear. how The medicine worked.
“that way, Machine learning helps identify cellular features produced by drugs,” study co-author Jeffrey Sauserman, a professor of biomedical engineering at UVA, said in a statement.
“Bridging machine learning and human learning will not only predict drugs for fibrosis; [scarring] as well as how they work. This knowledge is necessary to plan clinical trials and identify potential side effects. ”
In the paper, the research team studied the effects of 13 promising drugs on human fibroblast cells. They used that data to train a model to predict how the drug would affect cell behavior. The model predicted how a drug approved by the Food and Drug Administration would suppress the fibroblast fibers that make the heart stiff.
Reality check: Sausserman is optimistic about the possibility of understanding how the new drug works, but more research is needed to see if the drug is effective in animal models and human patients.
