Key Takeaways
- Prioritize AI as a complementary tool for data processing and pattern recognition, explicitly maintaining clinician accountability for final medical decisions.
In an era where technology is infiltrating nearly every aspect of our lives, the healthcare industry is no exception. Artificial intelligence (AI) promises to deliver faster diagnoses, more efficient care coordination, and improved patient outcomes. As the field of AI in healthcare continues to evolve, numerous artificial intelligences in healthcare research papers and scholarly articles are being published to explore its potential. Yet, as AI tools and large language models (LLMs) continue to gain attention, many medical professionals are left pondering how to incorporate these innovations into an already complex healthcare environment without losing the crucial human touch.
Recently, we had the opportunity to sit down with Jay Anders, MD, a seasoned internist and Chief Medical Officer who has spent decades in roles spanning private practice, information technology, and medical leadership. He has been at the frontlines of healthcare's digital transformation since the early 2000s, providing both clinical and administrative perspectives on emerging technologies. In our conversation, he offered a candid view of AI's promise, its pitfalls, and how healthcare leaders might strike the best balance between technology and compassionate patient care in this AI-influenced clinical world.
Dr. Anders's journey in healthcare began in a multi-specialty group practice, where he spent 20 years as a general internist and took on administrative duties as president and chief medical director. His first foray into healthcare IT was spurred by a simple logistical challenge: How do we stop transporting paper charts between offices?
His passion for healthcare IT led him to become the Chief Medical Officer (CMO) of multiple organizations, both on the payer side and within the health IT field, including stints at Christie Clinic, InteGreat, Med3000, McKesson, and eventually his current position, where he has spent over a decade exploring advanced technologies for health information management at Medicomp Systems, Inc. This experience has given him unique insights into the relationship between AI and doctors, as explored in various AI in healthcare articles, particularly in the realm of electronic health records and telemedicine.
When discussing AI, one of Dr. Anders's first points is the need to demystify what large language models (LLMs) and other AI tools actually do. This understanding is crucial for both clinicians and patients as we navigate the landscape of AI in healthcare, as highlighted in numerous artificial intelligence scholarly articles.
Although AI does not think in the human sense, it excels at certain repetitive, data-heavy tasks and can generate plausible-sounding responses. AI algorithms have shown promising results in AI diagnostics and clinical decision support systems. However, as Dr. Anders warns, there is a danger in assuming that AI's convincing language equates to correctness.
This oversight, he emphasizes, lies at the heart of the clinician's role in an AI-driven world. Just as no physician would accept a new medication without understanding its clinical studies, so too must they rigorously evaluate AI output to ensure it meets standards of quality and accuracy. This critical evaluation is essential for building trust in AI within the healthcare community and addressing the ethical considerations of AI in healthcare, particularly in the context of AI-enabled clinical decision support systems (CDSS).
The practice of medicine, according to Dr. Anders, is far more than diagnosing conditions and prescribing treatments. Much of a clinician's expertise lies in subtle observations—body language, emotional state, lifestyle context, and more. This human element is crucial when considering the AI impact on medical decision-making and the development of personalized medicine.
He worries that if AI tools are deployed haphazardly, especially in settings like chatbots making independent decisions, the all-important human element could be diminished. The art of medicine requires empathy, judgment, and the responsibility that comes with directly interacting with a living, breathing human being—none of which can be replaced by an AI's text predictions.
Instead, Dr. Anders envisions AI as a complementary tool, providing clinicians with immediate access to vast amounts of data they can then contextualize with their professional expertise. This vision aligns with the concept of human-AI collaboration, where AI's ability to identify patterns in large datasets, sort through clinical histories, or flag unusual findings can free up clinicians to do what they do best: treat patients holistically. This approach is particularly promising in the field of AI-powered patient monitoring, where continuous data collection can enhance personalized care strategies.
One of the greatest barriers to widespread AI adoption is the question of trust. Can clinicians trust an AI's outputs if its training data is incomplete, inaccurate, or biased? According to Dr. Anders, trust in AI is built through:
Dr. Anders recommends leveraging AI's strengths in non-clinical areas first, such as scheduling, materials management, and operational logistics. By proving its value in these low-risk areas, AI can build a track record of reliability, which might then ease its acceptance in more critical applications like diagnosis and treatment planning. This approach also allows for the assessment of AI reliability and AI robustness in healthcare settings, particularly in AI-driven healthcare management systems.
Healthcare organizations are constantly striving to meet quality metrics—blood pressure checks, diabetes management, preventive screenings, and more. These are typically tracked through numerator/denominator calculations requiring myriad billing codes and clinical documentation details.
AI can potentially sort and analyze these huge datasets, identifying trends or overlooked details that clinicians might miss amid their workloads. This capability aligns with current AI in healthcare trends, particularly in the realm of clinical decision support and predictive analytics in healthcare.
Still, there is significant promise in using AI to identify patients at risk, highlight missing screenings, and detect patterns in population health. The key is combining AI's data analysis with the clinician's informed perspective on what to do next, exemplifying the potential of human-AI collaboration in improving patient care. This approach has been documented in various AI in healthcare articles, with notable progress seen in AI in healthcare 2019 and AI in healthcare 2020 reports, particularly in the development of AI-enabled precision medicine techniques.
A long-standing challenge in healthcare is interoperability—the ability of different systems to seamlessly exchange and make use of patient information. Dr. Anders points to recent strides by the Office of the National Coordinator (ONC) and their rules against "information blocking" as major steps forward.
Yet, simply enabling data exchange is only half the battle. Clinicians often find themselves deluged with PDFs or non-standardized data that are too cumbersome to parse in day-to-day workflows. This challenge highlights the need for AI in hospital operations to streamline data management and improve efficiency.
This is another space where AI can shine: sifting through the noise and distilling clinical insights. However, leaders must address cultural barriers that hinder true data sharing—such as competition among hospital systems, concerns about patient retention, and the lack of proper resources in rural settings. Additionally, ensuring patient data privacy in AI healthcare applications remains a critical concern that must be addressed to build trust in these systems. The development of explainable AI and addressing AI limitations are crucial steps in this process, particularly in the context of machine learning in medicine and AI for medical research.
One of Dr. Anders's passions is improving healthcare in rural America, where patients often lack close access to specialist services and advanced medical resources. AI-enabled tools and data-sharing platforms can make a huge difference—if these communities are granted the funding and support they need.
In many rural areas, the biggest hurdle is not a lack of interest, but a lack of funding for broadband, updated electronic health record (EHR) systems, and ongoing technical support. These issues represent potential barriers to efficiency in implementing AI solutions. By championing these issues, healthcare leaders can ensure that AI's benefits reach some of the most underserved populations. This includes exploring the potential of AI-assisted surgery and medical imaging AI to bring advanced care to remote areas.
At the close of our conversation, Dr. Anders circled back to the essence of leadership in this quickly changing environment. Technology moves fast, but the principles of quality care and patient well-being cannot be compromised. Preparing medical professionals for an AI-influenced clinical world is a crucial task for healthcare leaders.
He encourages clinician leaders to remain vocal advocates for patient-centered AI deployments:
Artificial intelligence has the potential to reshape modern healthcare—streamlining administrative tasks, boosting diagnostic accuracy, and enhancing the quality of care that patients receive. Yet, success hinges on prudent leadership and a deep understanding that AI is an aid, not a substitute, for the human clinician.
Dr. Anders's journey from paper charts and trucking records between clinics to exploring advanced AI applications underscores just how far healthcare has come. His insights remind us that while technology continues to evolve, the fundamental mission of healthcare—to care for and heal people—remains unchanged. By embracing AI thoughtfully, ensuring it is well-trained, fostering interoperability, and prioritizing the patient-physician relationship, clinicians and leaders can harness the power of AI to forge a brighter, more equitable future in healthcare.
As we move forward, it's crucial to continue monitoring and analyzing AI's impact through comprehensive artificial intelligence in healthcare reports. These reports, along with ongoing research in areas like AI in medical imaging and AI in mental health, will help guide the ethical and practical implementation of AI in healthcare settings. The future of healthcare lies in the synergy between human expertise and technological innovation, with AI-enabled CDSS and human-AI collaboration paving the way for more efficient, accurate, and personalized patient care.
<p>it's not a thinker it doesn't do it's not really intelligent it just picks out the next word now it's very good at it because it's billions and billions and billions of [Music] words hello this is Zach with the American Journal of healthc care strategy and you are listening to the clinicians and Leadership podcast series where we seek to empower clinicians from bedside to boardroom I'm joined today by a by a very esteemed and and accomplished guest who's who's served in our Health Care system in in the United States of America in in a wide variety of of uh areas in in in uh uh Industries and so I'm I'm very excited to welcome today Dr J Anders Dr Anders why don't you take a second to introduce yourself and tell us a little bit about your experience and how you got to where you are now thanks Zach and thanks for having me um I have been in this business quite a while and done several things as you as you mentioned I started out being a general internist uh in a multip special degree practice uh did that for 20 years including administrative roles in that practice as president and chief medical director and things like that um I got interested in healthcare it at that point in time we were at the very Forefront of of electronic health records and we were looking for some way believe it or not just to stop transporting charts across five or six different offices so we had trucks and charts and things going everywhere and tracking systems so I thought maybe it's time we look at some kind of electronic readon system and that's where it all started um I got involved with a company called integrate which provided that system to our Clinic um subsequently came their medical director in the process of doing that I also did a St at Coventry Health as one of their medical directors that's on the payer side of of the fence um was the CMO of integrate CMO of something called mid 3000 which was a billing company primarily but they also did electronic health records they were purchased by messen um and I became their chief medical officer of their business Performance Services Division which was again a more billing and the electronic healthc care stuff um then I came to metop i' I've known Dave Loro at metop for now almost 20 years uh We've collaborated back and forth his product was actually in our old integrate product years and years years ago uh and I've I've always liked what metacomp has done very fascinated by their technology and I've been working with them now for the last 11 years so that's kind of my uh short story well and I I I'm just thrilled to have you on today I mean you you've just walked through briefly just your your experience and and the the the wealth of experience and knowledge that you have um and I'm I'm I'm something that was really interesting to me just looking over your resume prior to this meeting was into this interview was your your involvement in the information technology sector of healthc care from since the you know 2007 since the since the earlier 2000s so you've been involved in the the information technology side of healthcare for for quite some time now now and fast forward to 2025 and the the prevalence and presence of AI and artificial intelligence becoming more and more integrated into Healthcare I'm I'm interested to hear your perspective on how you see the role of clinicians evolving as artificial intelligence becomes more integrated into Healthcare you know I F that's a very good question um one of the things that's come out recently if you read all the I'm going to call it hype most of it HP about what AI is actually addressing in healthcare right now it is trying to actually outdo clinicians and what I mean by that is they can take tests AI can do a test better than a medical student can better than a resident can better than a attending physician can um but I will say this the practice of medicine is not taking a multiple choice exam it's a holy different thing so what I see the clinician role in all of this and I think this is not going to end if it does it will be a very sad state for American Healthcare I believe is that clinicians need to oversight these systems that are being created and what I mean by that is the human is the one that actually is going to take the responsibility for the care of that patient I'll give you an example so a woman comes in and she's got a breast problem she's got a bleeding nipple okay that's something that's very rare actually um I read all these wonderful uh stories about how AI can augment breast exam and imaging analysis and all of that but what happens if they miss it what happens if AI is wrong this is something that's not seen very often so you're going to have humans involved to try to oversight that AI input or augmentation um that comes basically at almost every level so I I see the clinician role is to using it as tools where the tools are appropriate and then oversi it because the rubber finally meets the road with the clinician and the patient I think that's a I think that's a great point and just and speaks to just the importance of having someone with that clinical experience in exercising oversight over new technology and and tools and artificial intelligence to recognize and to to promote proper usage in there and it it should make jobs and and lives easier um and and not complicate it further and that's kind of a delicate balance that's that's hard to find especially in in new as technology is is innovating and there's there's new things that are being implemented and and without oh without experience someone may not know how to implement those effectively and so I'm I'm curious as well kind of going off what you said you answered this a little bit but I I want to I want to hit on home a little bit more on it is in in a world where where data algorithms and and Technology especially in healthcare can often overshadow the human element and the human component to healthare how how can the the clinician leaders ensure that artificial intelligence AI enhances rather than diminishes the the patient clinician relationship how do we prevent artificial intelligence from from how do we how do we ensure it's a bridge between the patient and the clinician not a not a barrier here that's also a very good question one of the things that that has been talked about recently is these medical chatbots that are actually they think could actually start to make decisions on their own regarding patient care well again it goes back to the responsibility and algorithms guidelines all of those things have been around for a very long time very long time long before chat GPT ever poked its head up in the world so people patients are not algorithms they are a person they have their own specific makeup when you take that interaction between the physician clinician and the patient away that goes away and I will tell you this in all my years of practicing one of the things I learned a whole lot about folks just walk watching them walk from the waiting room into an exam just that that particular observation there is very little way that particular type of art of medicine will get into an AI system I'm not a lite but AI can what it can do is it can augment when Physicians have questions or they think they need to know something else so it's like having a medical library at your fingertips if you ask it the right questions and it gives you the right output there we'll get into that I think a little bit later but as long as you can trust it it will actually augment your ability to care for patients when somebody comes in with a low-grade fever and a runny nose and a sore throat and I take a look at them and I do a throat swab I don't need AI to help me with that I can treat that patient without ever going to a computer so we've got to kind of worry about or or think about where and how much and how it gets poked into the healthcare Continuum in the treatment of patients and I think that's a big concern well I think that I think you describing it as a medical library at your fingertips is is something that is is attractive to to any physician or clinician to that is in is taking care of patients is having the opportunity to be like hey I I've got access to this resource right in front of me if you describe it as a medical Library yet when you describe it as or when when you mention artificial intelligence sometimes that's met with this this skepticism um and and you may have some individuals in over you know across the medical field that are that are quick to introduce new technology and AR artificial intelligence as a tool into their practices and you may have others that are that are dragging their feet that that know how they have functioned over the years and it's worked really well for them and and they don't want to introduce artificial intelligence or or even even more so are skeptical of of artificial intelligence and you mentioned a key word in your last answer just if you can trust it if you can trust artificial intelligence that's that's a key thing so I'm I'm curious to hear your perspective on on how can a leader how can a clinician leader how can a physician leader Foster a a culture of confidence and and competence around artificial intelligence while addressing those those valid concerns whether that's concerns about accuracy concerns about bias how how can you Foster that culture that uses AI as the tool that it is well a lot of that will come about when AI is more consistent and more able to be trained well with data that is trusted data so all of these systems are trained on data and you've got to trust data you're giving it to actually give you the output you want the problem right now is that Ai and that's I'll back up and give you a definition of what an llm actually is it is an a program algorithm that predicts the next word that's what it does that's all it does it's not a thinker it doesn't do it's not really intelligent it just picks out the next word now it's very good at it because it's billions and billions and billions of words but really doesn't replace valid clinical data being fed as training for these things because we have that and there was an article in the last three months where 50% of the medical devices that are incorporating AI are trained on artificial data artificial that means some other computer thought it up that that right there that statement right there if you put that in front of a clinician they're going to say what because we've been reading studies clinicians have been reading studies for ever since they started training you get a drug study you get what you know you get studies you have to read and discern the same thing goes with AI you can read it but you have to discern what it's telling you and if it gives you something completely off base you've got to be able to recognize that the other issue is those systems are very very good at convincing you that they're right so as long as we keep perspective that this is a big ball of information out there that can be accessed in certain ways trained on real data with real patients and real situations and intermediated by a clinician that's where I think it's going to end up and where it should end up well and you you also just you mentioned the the skill set that particularly your your Physicians have to develop on reading a study reading an evaluating data and letting that drive their decision um not just taking not just taking a study at its at its abstract and and agreeing with the conclusions but but reading through the actual data and the process and then evaluating that and um and it I love how you just take that and you apply it directly to implementing AI is is you're not we're not just blindly trusting what it says but we're we're verifying and we're evaluating its process and and and and seeking to to have that verification um I I think that's a great way that Physicians and particularly Physicians but other clinicians can use this skill set that they're already developing over the course of their careers and and apply it in in in a different way so um something that something we've heard a lot about uh is is the the capability of artif AI and and Technology to to assist you know hospitals Healthcare Systems in in meeting quality measure compliance levels um I'm I'm interested how how can it is a tool like like we've said how can AI technology assist not just not just in meeting those those measures those levels but but redefining what high quality Care looks like do you see a do you see capabilities of AI in implementing in those ways I do um to break it down just a little bit quality measures you know are kind of a surrogate for quality care that's there if you follow the measure you should be doing quality care so let's put that on the parking lot for a moment one thing that AI can do is go through a lot of data and organize that data for you because quality measures really should lead to better care which means you have to do something as a clinician to make sure you've met the measure whatever and then make sure the patient's okay with it um one of the downfalls of llms right now is they're not very good coding machines they they don't know the codes and if you look at what a quality measure actually is there's a numerator and a denominator and a whole Gob of codes so you've got to be able to really discern not only what the clinician's doing but the coding behind it so if you take just a regular old text note and you're trying to move through that system you got to make sure all that is correct um so we're not quite there yet but what I do see is that it can go through data and analyze okay this patient may have this that you didn't know about you didn't see it it's buried in this Myriad of data in a health record oh maybe I should pursue that AI is very good at doing some of those tasks the other thing I think we've got to keep in mind here is these systems when they when they first came out every everybody aimed it right at diagnosis diagnosing care doing treatment plans all that stuff there are a whole host of things in healthc care that AI can do that are not direct patient care making sure facilities are scheduled correctly making sure materials are where they need to be anticipating how much of this or that you need to order as a hospital or health system so when you take it poking in Ai and healthare there's a lot of lwh hanging fruit that doesn't directly impact a patient so why not vet these systems on their ability to do that kind of work before we turn them loose on something that's much more complex I think I think that's a fantastic fantastic answer um particularly just I mean you you mentioned just the all the benefits that Ai and this technology can bring Beyond just the diagnosing of a patient and and how that the diagnosing of the patient he's mentioned is is something that catches makes a lot of people pause and and and is can be a focus of the skepticism that AI faces yet the ability of it to evaluate analyze and synthesize data in a very quick Manner and doing this with large amounts of data is is a tool that that can be leveraged really well and really effectively and I I think that you you hit the nail on the head just in your answer there so shifting gears a little bit Dr Anders I'm I'm I'm interested something something that is has long been an issue in um the healthare industry and and field is this this interoperability and how that has just all always been a challenge and and for for those of you that don't maybe aren't familiar with that term it's it's it's the ability of your your software your your technology systems to to exchange and make use of information but it's this ex Exchange of information of of different systems and and um and using that information and sharing that information so I'm I'm I'm interested to hear your perspective on how clinicians in leadership and can influence and and what role they should play and driving standardization across systems that make these AI tools more impactful more helpful we've talked about how they can be impactful and how they can be used rightly so how how do your clinician leaders your physician leaders and other clinician leaders uh what what what role do they play in driving this standardization across these systems to make AI tools more impactful well first off I'd like to congratulate the office of the national coordinator and Mickey tripathi's crew who have come up with some I think fairly solid guidelines on interoperability and how that whole system is going to work one of their focuses right now or has been uh information blocking meaning I'm not going to share my data because it might be something that's business proprietary to me because it's the big Clinic down the street and I'm not going to send my patients to the big Clinic down the street because they might take them and do take them away and now I lose my patience um that's been a problem forever the thing about interoperability is when it comes down to it if it is done well if it is done well you can save a tremendous amount of money in healthcare and you can get better pay PA care because you're not having to repeat something that's already there in a format that can be used big issue here these things right now it's it's like a fire hose the hoses are there they're connected water is Flowing but when you get all use of my old patients I had patients that would have in a paper record you know eight inches of chart I mean we're talking chart Big chart how do I find what I need in that I mean it's impossible there you go that's what AI can do for you it can sift through all that nonsense and say okay give me everything in there that is related to diabetes provided it's trained on what diabetes is so when it comes interoperability to enhance that is the usability of that data that goes from one place to another right now the like I said the pipes are open and people are connected up and people are sharing things but using that data after you get it is another whole story and my fear is that Physicians when they get it are going to say I can't find anything here I'm not going to read through 47 pages of PDFs I mean that's just not going to happen so we got to really work on that part of it so connecting it up getting it exchanged and finding what you need for the patients that's in front of you is the key factor and You' you've touched on this next next question um and this is this is the last question before I let you go Dr Andre so I appreciate your time today and you you in your last answer you touch you touched on that a little bit but beyond technology what what are and what are some cultural or organizational barriers that that must be addressed to achieve true interoperability in healthc care and and how can your clinician leaders tackle and and address those effectively you you've touched on this a little bit but when I'm curious to hear some more of your thoughts on that well one of the problems we have if if you want to roll back to the AI question for a minute and actually sorting this data that raises a Spectre of the social determinance of health and and the social biases that some of these systems do have um I mean one of the things I would encourage leadership to do um we did it at my old Clinic back in the paper days is we never stopped anything from going anywhere if you requested information you got it we didn't care about the business aspect but we should have I don't my board did not poke against me on that but today we we've got to be able to share information across the board and and that means everything you have one of the problems with the covid-19 pandemic is the CDC could not get accurate information as to the spread of this disease why because people weren't sending there wasn't the capability of sending that so there's a lot of Social and cultural aspects to that the other thing or one of my soap boxes has been rural health I kind of grew up in that part of Illinois uh I spent a lot of times in emergency rooms on a 25 bed hospital so I understand that delivery of care these folks don't have a lot of resources to make some of these decisions I mean we have to start leveraging what we have but there needs to be a funding mechanism for Rural health so that information can actually get there so leaders can work with other leaders to kind of get coalitions together and that's happening in my old home state of Illinois um it happens a little bit here in Pennsylvania but not as much as I'd like but where those those people that live in rural areas can benefit most from this type of interoperability and type of connected care because they see their Primary in a little town of 15,000 and they go to the big city and they see the specialist there's all kinds of things done that's got to flow back into that primary Care's office to really deliver the health care that patient deserves so we've got to start really looking at the funding we've got to start looking at the cultural cross connectivity in parts of our country nobody has a problem with the healthc care done here in Pittsburgh we got two major centers things flow all over the place no big deal you go out in central Pennsylvania another story so leaders need to need to step up and make their wishes known because if patient care is the top of our list and it is you still have to be a business but if patient care is the top of the list then we've got to start coming together and say okay how can we really make this work for everybody regardless of your resources because the resources will grow if we can get the cost down and the care elevated well Dr Anders I'm grateful just for the opportunity to interview you today thank you thank you for joining us on the clinicians and Leadership podcast um I I I've greatly benefited from this conversation and and I know that that anyone that has the opportunity to listen to it and and take take from what you have said and your experiences your insights will benefit as well and you just you at the end there you spoke to the importance of just continuing having these conversations and and working together in the collaboration Within within our Health Care system is is always critical but especially when we're seeking to implement changes that are going to benefit the patient which is the ultimate goal of any Healthcare organization um especially in those rural areas there there needs to be advocacy and help done Beyond just those rural areas and and so I'm I'm I'm grateful for you and your passion um and the way that you are you are uh seeking to to encourage those kind of conversations so thank you Dr Anders for joining us today um and that is all I've got for you thanks Zach I appreciate being here</p>
Want to reach healthcare executives and decision-makers? Join industry leaders like HealthMap Solutions on our podcast.
Become a Guest
