Subscribe to The Podcast by KevinMD. Watch on YouTube. Catch up on old episodes!

Your ER doctor has about 25 minutes to figure out your medical history and decide what to do next. Hamed Husaini, an emergency physician and physician executive, explains why so much of that data never reaches the bedside and what AI can do about it. This episode is based on his article “AI in health care data management: Curing the EHR overload,” published on KevinMD. You will hear why records from skilled nursing facilities, primary care, and home health rarely get read in time, why duplicate medications and missed end-of-life directives slip through, and how a one-page AI synopsis pushed into the native EHR before you walk into the room changes what the next 25 minutes look like. Hamed argues the bottleneck is not data volume; it is the pull model that asks busy clinicians to fetch records they never have time to read. If the system already feels like it should know your records and still doesn’t, this episode names why and what changes when the data starts flowing the other direction.

True team-based care starts with you. At ChenMed, we believe the best way to care for patients is to change the way we practice medicine.

When you join our team, you are empowered to lead. We’ve moved beyond the traditional volume-heavy model to focus on true value-based care. Our model gives you the time and resources to manage complex cases and make a lasting impact on your community.

Whether you are applying for a primary care physician, nurse practitioner, or medical director position, you will feel supported by a physician-led culture that understands your challenges. Your dedication doesn’t go unnoticed here. You’ll be rewarded with a career that offers both professional fulfillment and a better quality of life. Visit ChenMed.com/physicians-KevinMD to learn more.

VISIT SPONSOR → https://ChenMed.com/physicians-KevinMD

Partner with me on the KevinMD platform. With over three million monthly readers and half a million social media followers, I give you direct access to the doctors and patients who matter most. Whether you need a sponsored article, email campaign, video interview, or a spot right here on the podcast, I offer the trusted space your brand deserves to be heard. Let’s work together to tell your story.

PARTNER WITH KEVINMD → https://kevinmd.com/influencer

SUBSCRIBE TO THE PODCAST → https://www.kevinmd.com/podcast

RECOMMENDED BY KEVINMD → https://www.kevinmd.com/recommended

Transcript

Kevin Pho: Hi, and welcome to the show. Subscribe at KevinMD.com/podcast. Today, welcome Hamed Husaini. He’s an emergency physician and physician executive. Today’s KevinMD article is “AI in Health Care Data Management: Curing the EHR Overload.” Hamed, welcome to the show.

Hamed Husaini: Great to be here. Thanks for having me, Kevin.

Kevin Pho: All right, so just tell us a little bit about your story, then we’ll jump right into your KevinMD article.

Hamed Husaini: Yeah, absolutely. I come from a family of physicians, a lot of head down, just seeing patients. Early on in medical school I realized that I looked at medicine more holistically.

And so I became an emergency physician, practiced, led some teams, and during that time, started looking at the things that we were doing inside of health care that really could be eliminated or potentially drove very little, if any, value. And so that led me into a career of doing some things on the payer side to really understand that side, and then now into technology to really see how it can drive quality and improve the lives of those who provide the care every day.

Kevin Pho: All right. Your KevinMD article talks about how AI specifically can help with health care data management. As everyone listening knows, AI is really disrupting a lot of fields in medicine, but I’m interested in hearing your lens and your perspective. So just tell us a little bit about your article for those who didn’t get a chance to read it.

Hamed Husaini: Sure. So when we think about AI, I think there’s a lot of tactics of being scared. Is it going to replace physicians? Is it going to replace the health care providers? But in reality, the step one and where we’re at today, it’s how can we take the things that really make our practice of medicine less exciting, less joyful, that are just sort of onerous and tasking of sorts, and how can we use AI to really get rid of some of those tasks?

And this isn’t like you go from zero to 100. There’s opportunities to really sort of, over time, just do a little bit at a time to improve your AI literacy and to be able to drive value in your practice. Last thing I’ll say is that, you know, as a clinician, hour one of your shift is different than hour 10 of your shift.

And so how do we better leverage technology and AI to reduce the absolute number of words that we have to read to be able to effectively take care of patients in a quality way?

Kevin Pho: So specifically in your field of emergency medicine, tell us the problems that AI could solve, some of the obstacles and some of the resistances that you face every day in emergency departments.

Hamed Husaini: So by the time patients arrive to the emergency department, most patients have had some kind of visit or transfer or recent hospitalization or stay at a skilled nursing facility. And so we have all of these different inputs of significant data. A lot of times they’re multiple pages, they’ve got multiple different perspectives from a different specialty, or a nurse versus a home health provider or a physical therapist.

And so you’re trying to collate all of this to create a story and, oh, by the way, you have 15 minutes to do so. So to date, what we’re doing is really taking that data, having AI create a story for us, bringing all the different inputs, and so effectively creating a synopsis, so before I go see the patient, I can now have sort of a fighting chance to be able to understand where they’re coming from, what the problem is, and what I can do about it now to really assist them.

Kevin Pho: In the emergency department, I can only imagine that there are so many different EHR systems from patients who come from so many different systems. And it’s not just hospital systems, but you said skilled nursing facilities, nursing homes, rehabilitation facilities, various outpatient clinics, and they all run on different systems, right? So sometimes if you can’t read where they’re coming from because your EHR is not connected to theirs, you’re essentially flying blind, right?

Hamed Husaini: 100 percent, and I think it’s only going to get worse, because with the advent of AI and Claude and all these organizations that allow for the building of technology that’s customized, and things that can be done quicker and sooner, I think what the future looks like is not fewer groups that are providing these EHRs. It’s going to be more.

Kevin Pho: So it doesn’t sound like the issue is too little data, but the data is just too fragmented, right? And whenever a patient comes to you and you have to just consolidate and find where they’re coming from, just give us a sense of how long that takes you typically.

Hamed Husaini: Yeah, I mean, I was actually talking with my mother just this week who went to the ER last week, and she was very upset because the doctor didn’t know about her rheumatology appointment and her cardiology appointment and the care she’s getting through her primary care doctor. And I asked her, I said, “Was it busy in the emergency room?” She goes, “There was people everywhere.”

And so the reality is, when you look at national data, most ER doctors are seeing between 2.2 and 2.5 patients per hour. As the primary clinician, you think about that, that means that you at best have about 20 to 25 minutes to spend with each patient. And be able to synthesize all of this data, have a really good understanding. And then remember, you can’t walk out of that room without a plan. And so it’s going from the beginning to the end, and this timeframe is shortened.

And so when you think about other aspects of health care, other specialties and other occupations, the time to decision-making is a lot longer. And so being able to synthesize this data in a way that you feel good, you have a really good understanding of that patient’s care or that patient’s history, and then you can apply that to their care and then articulate that back to the patient, I think that’s sort of the home run we’re trying to hit.

Kevin Pho: So let’s get a little more granular. So when you say that AI can surface a lot of this data from different electronic medical records, exactly what are we talking about? Just tell me about the process of an ideal workflow.

Hamed Husaini: Absolutely. So what we’re talking about today is really working with electronic health records and creating connections between them to be able to pull information into the EHR when a patient is registered. And so this isn’t a retrospective sort of review. This is a prospective opportunity to be able to change care.

And so what we’re talking about here is, how do we build those bridges? How do we improve interoperability in a way that we actually are bringing that data to the clinician’s hands prior to them making all of these really expensive and often unnecessary decisions in caring for this patient? And so taking one level deeper, what we’re talking about is having that patient register, getting that patient’s data pulled from their electronic health record, whether it’s the skilled nursing facility, ambulatory, and then creating that synopsis in your native EHR that you’re caring for that patient. Because the last thing we need is multiple different sign-ons and multiple different systems. That isn’t solving for anything.

Kevin Pho: So when things work the way they’re supposed to, from your perspective as an emergency physician, and a patient comes from somewhere outside your system, what kind of information are you giving that AI is synthesizing?

Hamed Husaini: So, for example, reason for transfer. You take a patient from a skilled nursing facility that comes to the ER, they’ve got two, three, four weeks of data. They’ve got ups and downs. They’ve had improvements and regressions. And now they come to the ED, and maybe they have some baseline chronic problems that you may think are acute or not. You don’t really know. And so being able to synthesize all the data, say, “This is why we sent the patient to the hospital,” the reason for transfer.

Being able to take a medication list, and oftentimes those medication lists have multiple duplications of the same medicine, and it takes time to try to figure out which one’s active, which one’s inactive. So being able to mitigate any kind of mistakes across the medications.

And then the last thing is accuracy of data. So, for example, being able to convey things like end-of-life desires, POLST forms, and DNR orders if they’re there. So really it’s just being able to bring a lot of data and make it concise, make sure it’s accurate, and help synthesize the thought process for the clinician when they’re looking at it.

Kevin Pho: And how can you make sure that the information that the AI is synthesizing is in fact accurate and it’s not making stuff up, not hallucinating? Because in a busy ER, I’m sure you don’t have time to check up on the data that the AI is giving you.

Hamed Husaini: Sure. And sometimes it does sound odd, right? And when you’re reading it, you go, “Hey, man, that doesn’t really sound right.” And so having the ability to reference where that came from is important. And I think that’s the first step in our AI literacy, is being able to take a synopsis, and when you have questions, when you have concerns about something maybe not adding up, being able to go back and reference the location.

Kevin Pho: And I want to emphasize this, that this process that we’re describing, I think a lot of patients assume this is already happening, that when they go to the emergency department, you as the emergency physician have all that data. But in fact, you don’t, right? So what sounds like a basic step, it sounds like AI can really help us bridge that gap for something that sounds so rudimentary.

Hamed Husaini: You hit the nail on the head, Kevin. The reality is that there’s this concept between a push and pull. It sounds so simple. It’s two words that start with P and are four letters. But the reality is that if I have to go into a health record and go pull data for a patient, it’s just not going to happen to the degree necessary.

So when we take AI, when we summarize things, now we’re talking about taking that synopsis, pushing it to you, and going, “Here it is. Take 25 to 30 seconds to review it, highlight it with the patient, discuss it with the patient, and then move on.” That concept between push/pull is so significant, but I think it’s often misunderstood and underappreciated.

Kevin Pho: So I know that a lot of EHR systems, they’re heavily siloed with proprietary information, and they’re hesitant to share information, especially with other EHRs. So how does this AI solution that we’re describing today, how does it overcome those silos?

Hamed Husaini: Yeah, I mean, I think most EHRs are in relationships of how to sort of give and get, right? Like, here at PointClickCare where I work, we are sharing data with lots of other organizations, even some that are competitive to us. But the point is, how do we give information but also get information to help the health system improve?

If you take a quick example, if I go to West Virginia, for example, and I want to build a network and be able to provide value for the clinicians and to use AI and integrate all of the different layers of the health care system, the more people and the more organizations that I bring on, the better that data. So again, it comes back to you can’t always ask to be receiving and not giving. And so I think that’s the impetus and the charge for EHRs today, and I think that is improving significantly with all the names that you’re aware of and know.

Kevin Pho: And in terms of clinical decision-making in the emergency department, if you have that information, it really cuts down on a number of duplicate tests and unnecessary tests that you may order, right?

Hamed Husaini: 100 percent. I mean, we’ve seen case studies of where a lot of it’s about awareness, a lot of it’s about initiatives. I sort of think back to my training, and I think we have a problem with academic medical centers in the United States. And large part is because, as clinicians, they want us to learn medicine, and they don’t want us to learn the practice of medicine.

And I don’t think that’s nefarious per se, but what I think it is, is that there’s so many opportunities to understand the nuance of how payers work and how organizations work. But at the end of the day, the opportunities are abound. And I think that sharing technology is prudent.

Kevin Pho: So what about after the emergency visit in terms of information and tests that you order in the emergency department, how does that get transmitted to a patient’s native EHR?

Hamed Husaini: Yep. So what happens is basically, with national care networks, when there has to be what we call a TPO relationship. So there has to be a treatment or a payer or an operational relationship created. And once that’s created, then we can share that data across all of those people, all of those organizations that are included.

So for example, a patient comes from a skilled nursing facility to an emergency department, has an ER visit, maybe doesn’t get admitted. They subsequently go back to the skilled nursing facility. That data is put into a national care network, and the skilled nursing facility then goes and gets that information. It’s sent to them. So now the whole ecosystem is connected.

Kevin Pho: So what do we have to look forward to when it comes to AI synthesis of patient records in the emergency department? What’s on the horizon?

Hamed Husaini: So I think there’s opportunities at looking at trends. I think that’s probably the next thing, looking at the horizon, being able to look at the longitudinal care and longitudinal outcomes of patients. So for example, being able to recognize that a 65-year-old female who presents with hip pain, 94 percent would get like a hip X-ray. So being able to predict models in how care is delivered and recognize trends in health care. I think that’s being able to synthesize many, many visits across emergency medicine and other specialties. And I think that’s a future opportunity that a lot of us are investigating.

Kevin Pho: We’re talking to Hamed Husaini. He’s an emergency physician and a physician executive. Today’s KevinMD article is “AI in Health Care Data Management: Curing the EHR Overload.” Hamed, let’s end with some take-home messages that you want to leave with the KevinMD audience.

Hamed Husaini: Yeah. I think AI is here, and we all know it’s here. I think it’s a bit intimidating, and I would encourage any and all physicians to take their first step in, because you really can’t be completely an expert or literate until you take your first step, and you’re going to see the growth and the opportunity in your practice for years ahead.

Kevin Pho: Hamed, thank you so much for sharing your perspective and insight, and thanks again for coming on the show.

Hamed Husaini: Appreciate it, Kevin. Thank you.