In this episode of A Moment in Health, Dr. Ashish Jha highlights that 66% of U.S. physicians now report using AI in their clinical practice, reflecting the technology’s rapid adoption across billing, documentation, and patient care. He reviews a new JAMA Network Open study showing that people with type 2 diabetes and opioid use disorder who took semaglutide (Ozempic) had roughly half the risk of opioid overdose compared to those on other diabetes medications, adding to growing evidence that GLP-1 drugs may help reduce addictive behaviors. Dr. Brendan Saloner, Professor of Health Services, Policy and Practice at the Brown University School of Public Health, joins to outline three key policy priorities for curbing opioid overdose deaths.

Dr. Jha discusses:

• Semaglutide and Opioid Overdose Risk in Patients With Type 2 Diabetes and Opioid Use Disorder — JAMA Network Open

About the Guest

Dr. Brendan Saloner is the Donald G. Millar Professor of Alcohol and Addiction Studies in the Department of Health Services, Policy, and Practice at Brown University. One strand of his research focuses on access and quality of care for people with mental health and substance use disorders. A second strand of his research focuses specifically on health care for people in the criminal legal system. At Johns Hopkins University before joining Brown, he co-led the evaluation team for the Bloomberg Overdose Prevention Initiative, a $120 million investment in overdose reduction focused on seven states.

About the Host

Dr. Ashish K. Jha is the dean of the Brown University School of Public Health.

Music by Katherine Beggs, additional music by Lulu West and Maya Polsky

Hey everybody. Ashish Jha here coming at you on a sunny and cool fall day in Providence, Rhode Island, with the next episode of a moment in health, the podcast where we talk about one data point, discuss one study and answer a question, and the data point for today is 66% or about two thirds, two thirds of physicians report using healthcare AI in some form or another in their clinical practice. Can be AI for billing. It can be AI for their visit notes, discharge documents. This number is up very substantially from 2023 and two thirds of doctors in 2024 were using AI. The only last thing I'll say about this is it's not surprising. There's so many AI based tools now that target clinical practice that are trying to make clinical practice more efficient, easier on doctors. I recently saw my primary care physician who recorded the whole visit using an AI based tool, and it automatically generated, generated a very, very high quality note. So these are, these are things that are becoming much, much easier for physicians. You

okay, let's discuss a study. The study of the week that I want to talk about is comes from the September 2520 24 issue of JAMA Network open, and it is entitled semaglutide and opioid overdose risk in patients with type two diabetes and opioid use disorder. You all know semaglutide, more affectionately known as ozempic. This is a weight loss drug. It's great for diabetes, partly through the mechanism of weight loss, but partly, we think through other mechanisms as well. And there has just been a lot of very suggestive evidence, and I'll explain what I mean by that in a second, that these drugs, which are incredible for weight loss, incredible for cardiovascular disease, may actually be helpful for substance use. And these guys did a very high quality cohort study, was not a randomized trial, high quality cohort study, and found that people using semaglutide had substantially lower risk, about, you know, sort of almost 50% the risk of having an opioid overdose compared to people who were on other anti diabetes medications. So that is a very large effect. Let me just finish up by talking about this whole thing of GLP ones and substance use. There is now been study after study after study showing suggesting that GLP ones reduce alcohol overuse, maybe reduce tobacco use, maybe reduce substance use, like opioids. It makes sense to the extent that we think those things are really driven by cravings, and GLP ones help reduce cravings. That's largely how it works with food, it creates a sense of satiety, and people end up eating less, so it may be having those benefits as well. Last final thing on this, look these, these cohort studies, are fantastic. They're very suggestive and helpful. What we need is to go back to the randomized control trials and look much more carefully at whether we see it in the RCTs, or run new, larger randomized trials specifically focus on answering these questions. That's what we're going to need before we start prescribing GLP ones

for substance use disorders. I don't think the evidence is quite there yet, but boy, it's heading in that direction.

All right. And now for the question of the week and for the question of the week, my guest is Brendan Saloner, who is a professor here at the Brown School of Public Health. Brendan, thanks so much for being here. Thank you for having me. All right, so you have been studying opioids and opioid overdoses and what we can do, and obviously we've gone through a horrible period where we've just seen unprecedented number of people dying the last year. I think we've seen that curve turning, thankfully, starting to turn down lots of potential explanations. My question for you is, at this moment in fall of 2025 as we sit here, what are the two three policy things that you would implement if you could implement them tomorrow, to drive those numbers much lower,

it's such a good question. We have been in this incredible public health crisis, and we're not out of it. So I just want to start by saying that numbers have been going down. Very encouraging. We hope to see that continue. So you know, you've asked three things I'm gonna give you unapologetically, very basic answers. Great, not complicated. Number one, every person who uses drugs should have in their possession and know how to use naloxone. Naloxone is the opioid overdose medication. It is a life saving intervention. So I'm really excited, because I have seen in the last few years, Naloxone access get so much better walking around at the Rally for Recovery here in Providence. A few weeks ago, lots of tables with naloxone. The problem is we still see disparate outcomes in who has Naloxone, getting Naloxone into the hands of people who use cocaine where fentanyl is unfortunately creeping into their drug supply, people who are family members, getting it into rural communities, overcoming some of the stigma. I think all of those things would be hugely, hugely beneficial. Okay, I like that. What else? Medications for opioid use disorders save lives. They are medications that when people take them as indicated, their risk of overdose, and particularly dying of overdose, goes down by more than half. So I have two ideas about how we can get more people access to these life saving medications. I think the first thing is, there are reachable moments for people with an opioid addiction. I think oftentimes we succumb to this very unfortunate idea that people don't want help. They do want help, but oftentimes, when they're at their lowest moments, we're not helping them. So I'll give you a couple examples. People go to jail, risk of overdose and dying of overdose after being in jail is astronomically high. A lot of jails still don't treat people with the medications for opioid use disorder. Starting people on medications when they're in jail

can save their lives. I like that another one, and I think this is a real shameful thing about our healthcare system hospitals. Many hospitals do not provide these medications. It would be unthinkable for someone to go into the hospital with a heart attack and not get evidence based care for heart attack, at least the very basic medications that we know help people after cardiac arrest, people who have an opioid overdose when they go to the hospital often leave not having received medications. I've seen numbers that would suggest that as low as about 10% of people in the weeks after hospital discharge on these medications,

and that's because doctors in the hospital think, Oh, this is an outpatient problem, or they think this person's not really ready. What's your best guess of why hospitals aren't

doing okay? So I think there's a few reasons. One is stigma. A lot of people who work in hospitals are not very comfortable treating addiction. They don't have the knowledge base, and the medications are not stocked in hospitals. And I do think that there is a there is a handoff problem between hospitals and outpatient care that exists across our medical care system, but this is a completely different care system that they're getting handed off to, and unfortunately, it's a big dropping

off point. Okay, so that's the second thing. Make these medications much more available, particularly at vulnerable moments. Anything else you would do?

I think that this is a longer term priority, but we need to build capacity in the system, and there's a few ways to do it. One and actually, Brown University, our medical school is has been a leader in this make every person who gets a graduate medical education learn about addiction, learn how to prescribe medications like buprenorphine. It's not rocket science. It's no more complicated than any other chronic disease medication. Many people go through a medical school training without learning about these medicines that should not be allowed. I think that's one way to build capacity. Another is just to make sure that all the places where patients are getting their care, whether it's community health centers, whether it's primary care offices, just across the spectrum, that health systems understand that they need to be doing this and that the payers, Medicaid, Medicare, commercial insurance, are rewarding this kind of care, and oftentimes they should be disincentivizing providers or disadopting them or taking them out of their networks if they're not doing this care. I think that this has got to be set as an expectation in our healthcare system.

Awesome. So make Naloxone much more widely available. Get people treatment, especially when they're most vulnerable and ready for treatment, and third, build up capacity in the system. It is remarkable to me that for a disease that's killing 100,000 people a year, lot of young people that we have lots of doctors who don't know how to treat it, and we have large chunks of the system that ignores it. Those are all really, really helpful answers. Brandon, thank you so much for being here, and thanks for that very clear and helpful set of thoughts. A pleasure to be with you.

Well, there you have it, another episode of a moment in health where we talked about one data point, 66% of physicians, two thirds of all doctors report using healthcare AI in 2024 a number that I suspect has only climbed this year in 2025 and they're using it for a wide, sorry, a wide variety of things, everything from clinical notes to billing et cetera. The study that we discussed was semaglutide and opioid overdose risk in patients with type two diabetes and opioid use disorder. The study suggested that people getting ozempic basically had substantially lower rates of opioid use overdoses, and this. Really comes in the in the wake of a lot of other studies that suggest the same thing. We don't have definitive proof yet, but boy, these data really are pointing in that direction, mostly through cohort studies. We need high quality randomized trials. Our question of the week was answered by Brendan Saloner, who is a professor here at the Brown School of Public Health, just came over to us recently from Hopkins, and he's really one of the leading scholars on opioid overdoses and opioid abuse and what to do about it, and I asked him what policy actions he would take to further reduce opioid overdose sets. Since, in the last year, we finally have seen this horrible, horrible curve of people dying starting to turn. And he laid out what I thought was a really brilliant strategy, and I'm going to give you the big takeaways, widespread availability of naloxone, lots more people ordering medications for opioid use disorders, a point that he made, imagine if 100,000 people were dying of any other disease for which we had treatments and doctors were not prescribing the medicine that would save their lives, we wouldn't stand for it. This is particularly important in high risk situations like jails and hospitals, where people are particularly vulnerable and open to getting treatment, and then kind of a broader building of the healthcare

system capacity, medical education, residency training, I'm really making this a national priority. I would argue it has sort of quasi been one. We have made progress, but if we want to drive these numbers down, we're going to have to do a lot more. So that was all from Brendon. It was fantastic. Thanks so much for listening to another episode of a moment in health. I will be back next week to talk about a data point, a study and answer a question. Thanks so much for listening, and have a great week. Folks.