Dr. Brian Hausfeld has a little Wall Street in his background and he likes to talk about the potential benefits of artificial intelligence for healthcare in economic terms – the law of supply and demand.
Everyone knows how health care works. A patient sees their primary care physician, who then issues a referral. That therapist then issues a subspecialty referral, which may eventually reveal the correct answer. Although a referral to health medicine is probably needed these days.
That’s a lot of thought. A lot of demand. And unfortunately, health care is severely understaffed and underserved.
“From my perspective, it’s not about the tools, it’s really about the access issue,” Hasselfeld said of AI. “How do we care for more patients with the clinical workforce we have today? How do we meaningfully increase productivity? Care for more people on top of existing resources.” ?And not just ask your medical workforce to work more?
“How do we really put meaningful intelligence into what comes first and what comes next in the patient journey?” he continued. “And if we can start to take some of that unnecessary care out of the system, we can unlock some additional supply.”
Hasselfeld is senior medical director of digital health and innovation at Johns Hopkins Medicine and associate director of Johns Hopkins inHealth. He is also a primary care physician focusing on internal medicine and pediatrics at Johns Hopkins Community Physicians.
We interviewed the top voices in health IT as part of our series talking about artificial intelligence. In this, the first part of the interview, he discusses implementing AI in healthcare as a whole. In part two, which will appear tomorrow, he goes deeper into how Johns Hopkins Medicine is using AI today.
Q. As a senior digital health and innovation executive, what forms of artificial intelligence do you have an eye on?
Oh We are at a stage where we are not entirely sure of the magnitude of the problems that will define the new form of artificial intelligence.
Most professionals tracking the general AI industry across all other industry verticals are beginning to recognize that we have what we might call historical or traditional AI, tools that rely on pre-defined computer science. Based on principles, inputs and outputs are built. And now we have our new generative AI, which of course became famous last January with Microsoft and OpenAI’s announcement of ChatGPT, and now all the other competitors on the market.
Technology is going to be truly limitless in how it can be applied to the problems being solved in healthcare. Rather than thinking about this specific type of tool that’s a priority for us, I would describe this as a really important moment in healthcare to solve a big resource problem.
I’m a former economics undergrad who went to Wall Street, so bear with me while we talk economics for a second. Today we have a meaningful supply/demand mismatch in healthcare. Anyone who has tried to get a visit from any institution, large or small, academic or non-academic, surely appreciates the difficulty of running a relatively complex health system and the resulting waiting times.
But from my perspective, technology has not yet done what technology needs to do in health care, what it has done in many other industries, throughout the economy – health care. Increasing productivity and efficiency to help balance all the demand from our patients and the supply we have to offer, which is relatively fixed.
From my perspective, it’s not about the tools, it’s really about the access issue. How do we care for more patients with the medical workforce we have today? How can we increase productivity in a meaningful way? Care for more people on top of the same existing resources? And at the same time, of course, avoid the key balancing component, which is why we can’t ask our medical workforce to do more.
Arguably, many interventions have sought to reduce the amount of work on our clinicians. The tools being implemented really focus on this patient access journey as a key priority – how to get patients to the right type of care, at the right time, quickly.
Certainly, some of the early products being tested on the market help patients identify what type of care they actually need. Now, instead of going through the regular paradigm of referral visits to subspecialty referrals to finally get to that right answer. I also have a role in my health medicine initiative, so it can be called accurate referral or accurate care planning.
How do we really put meaningful intelligence into what comes first and what comes next in the patient journey? And if we can start to take some of that unnecessary maintenance out of the system, we can unlock some additional supply.
On the other hand, we need to live in a paradigm where there isn’t a physician to see a patient for 15 minutes, right? It is not a measure because time and people are fixed. And we need to find a way to care for a larger number of patients with more intelligence between the data entered and the care plans sent back to our patients.
I agree with one of the former leaders. In this series of articles, Dr. John Halamka [at the Mayo Clinic]that patients do not come to doctors to read textbooks.
So, certainly by not advocating we can care for 20 times more patients and remove the clinician from the care journey. But I believe that a visit every three to six to twelve months, for example, is broken in a system that should be based on prevention. And that really means we have a big household data problem to deal with, which I think is a big area of opportunity as tools continue to evolve.
Q. You told me that digital apps, connected devices, wearables and home sensors are all the future of individual health tracking — and yet, broadly speaking, there’s little growth in these approaches, which is doctor/patient. Rarely found in relationships. You believe that the latest iterations of AI will eventually remove significant barriers to access to this new information in clinical care. Please shed light on this topic in detail.
Oh This is actually a great pickup from where we just left off the last question, which is the different historical ways of measuring data at home, from the watch or Fitbit on your wrist to your own personal bedside devices. , such as home blood pressure cuffs, to scales. , glucometer and continuous glucose meter.
We have this wealth of domestic data. Certainly, our own precision medicine group at Johns Hopkins Medicine looking at multiple sclerosis provided an amazing new paradigm for how this data might apply to future diagnosis and maintenance treatment planning.
Acknowledging this motion Movement tracked by a wearable such as a Fitbit or similar advanced movement device can be meaningfully correlated with the progression of the movement disorder, which seems all too well and possibly replaces it. In the long term, MS patients routinely require visits to advanced quaternary neurologic care centers. Expensive MRIs.
But how do we take this measurement example and take it to scale? When we look at our outpatient clinicians today, and I’m a primary care clinician, we can take care of 1,500 to 2,000 patients, if you’re a full-time primary care clinician.
And let’s compare it to a hospital. What is our most intensive area of measurement in the hospital, ICUs and clinical care units? In these units, we have a team of at most 15 or 20 caring physicians, with a nursing ratio of one to one or one to two. So this is the level of staffing it takes to connect patients to devices on a regular basis, certainly on a daily if not hourly basis.
And even on our hospital floor, we have nursing ratios of one to four, one to six, and clinical teams around them, and it’s taking data every four to six hours or every 12 hours.
So how do we go from an environment where we have two patients with nursing support to one clinician, one clinician to thousands of patients with minimal other longitudinal support, and still every day, multiple times a day? Expect to receive data, and not overwhelm our workforce, systems, practice models and payment models that aren’t ready for this level of home enrollment?
This is why we’ve seen things like remote patient monitoring struggle with mass uptake. I think we’ve had Medicare continue to look at how they can improve the conversion, or sometimes even question whether they should remove the RPM coding.
The best possible information known about patients is better known than the transactional nature of a few visits over the course of their month or year. What’s missing in between is systems to take all that data and make it clinically relevant, clinically meaningful and interpretable and put it in the context of that patient.
So, we can create a system where I give you a blood pressure cuff, and I say blood pressure above X and below Y is bad, and we can pick those numbers and they’re most patients. will be valid for But unless I know you, unless it’s specific to your context, it depends on your medical goals and your underlying medical conditions and our mutual treatment goals, it’s bad for you. May or may not be.
Therefore, we need systems that can both handle large amounts of remote data and contextualize it based on everything we know about you, especially which we discuss during our visits and around your treatment plan.
So when we talk. Applications of Generative AI to Solving Problems in Healthcare We often hear about the problem of taking unstructured data in charts, especially written notes, and making some sense of it. C makes it organized and comprehensible to other types. system, to help improve care.
That is the real opportunity here. Part of my job here at Hopkins is to help oversee our virtual care teams. I led these teams through the pandemic. And what we have the opportunity to do is to unlock the value of these remotely connected devices and the visit volume between the data.
If I have a system, know your chart notes and understand what it says about blood pressure, weight, goals, what conditions you have, what medications you are taking, and intelligence. Create an accurate layer. Around that incoming data, so that we don’t reproduce the inpatient alarm fatigue that’s already there on the inpatient side, then I can scale it exponentially on the outpatient side.
Ultimately, we have an opportunity to build a very intelligent layer around home-based data in our clinical workforce, which is not going to grow in size and certainly 1,000 or 2,000 patient home-based data. Cannot measure. Full regular medical day.
I’m excited about the opportunity to finally do what we want for our family members: to have more continuous information about meaningful conditions for our patients, interpreted, ready, available and actionable year after year. It should be.
To watch the video of this interview, Click here.
Editor’s note: This is the seventh in a series of features on the top voices in health IT discussing the use of artificial intelligence in healthcare. To read the first feature, on Dr. John Halamka at the Mayo Clinic, Click here. To read another interview with Dr. Alpin Patel at Geisinger, click here. For a third reading with Meditech’s Helen Waters, click here. To read the fourth, with Rana including Epic, click here. For Mass General Brigham’s fifth reading with Dr. Rebecca G. Mushoris, click here. And for a holiday reading, with Dr. Melek Somai of Froedtert & Medical College of Wisconsin Health Network, click here.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS media publication.