Hacking Your Way Into Medicine as a Clinical Informatics

Apple Podcast | Google Podcast

SS 184: Hacking Your Way Into Medicine as a Clinical Informatics

Session 184

Dr. David Liebovitz specializes in general internal medicine and clinical informatics. Listen to learn about this exciting field full of data, coding, and much more!

It’s an up and coming field that’s becoming more necessary because we are gathering so much data in the medical world with all the research going. Everything that goes into an electronic medical record could potentially help your patients in the future. If this is something you’re interested in, check out the American Medical Informatics Association.

For more podcast resources to help you with your medical school journey and beyond, check out Meded Media.

Listen to this podcast episode with the player above, or keep reading for the highlights and takeaway points.

[01:47] Interest in Clinical Informatics

During his chief resident year way back in 93-94, David says he had contemplated going to a variety of specialties, but had an opportunity to become a chief resident for a year. During that year, he worked administratively with a lot of teaching and seeing patients too.

Then he realized there was just so much more we could do with information effectively to manage patients than we had then. And that’s what sparked his interest.

David also studied electrical and computer engineering as an undergraduate. Although he had always planned to go to medical school, he enjoyed a variety of the science aspects of engineering. And that lent itself also to him realizing all of the potential in this area.

[02:52] What is Clinical Informatics?

Clinical informatics is a clinical specialty that focuses on how we acquire information from patients. It involves how we store that information for later use and how we apply information to help patients obtain optimal outcomes – both at the single patient level as well as at a population level.

'Clinical informatics is how we're managing information effectively, gathering, storing and how we're using this information.'Click To Tweet

There is enough information available now about how to do this effectively. Hence, it was created as its own ACGME (Accreditation Council for Graduate Medical Education) to be its own subspecialty. It’s also recognized by the American Board of Medical Specialties. Just like cardiology, geriatrics, or gastroenterology, clinical informatics is now a board-approved subspecialty as well.

It has a two -year clinical informatics fellowship. After medical school, there’s residency of variable length depending on the specialty. And then following residency, one can go into a fellowship for additional subspecialty training, and that’s where clinical informatics falls.

For clinical informatics, one can complete any of the other accredited residency programs and then begin this fellowship. So they could have surgeons together with non-surgeons in these training programs. All paths potentially could lead to clinical informatics if somebody wanted to.

[06:22] The Scope of Practice

There are lots of opportunities now including electronic medical records and everything attached to them in any different way. This would include patient portals and how patients submit information. They can look at what sort of terminology should be mapped to so that alerts can percolate through the system. So that when a patient responds in a certain way that requires additional attention.

Now, an individual physician can understand who they should reach out to and who’s most in need or who hasn’t been seen in a long time. They can leverage all this information more effectively and within the scope of practice of a clinical informatics physician.

Physicians will have a better understanding of the techniques, approaches, and the standards of how information is stored. They will have tools available and analytics to actually solve those sorts of problems.

For instance, during COVID, they were able to implement risk scores for patients that have been validated across many patients at different hospitals using an electronic health record. They could then see which patients should they round on extra or be extra vigilant about in terms of deterioration. A lot of that is being implemented in different stages now as well.

[08:13] Is Technology Going to Replace Physicians?

It’s a common fear among physicians that they will be replaced with technology. But being a builder of the future technology, David explains we are even nowhere close when that will be the case.

In recent years, medical schools have focused so much attention to appropriate bedside training, interactions, and understanding that communication with patients is critical.

The ability to extend hope to patients and provide appropriate understanding and context for what they’re going through is not going to be replaced by artificial intelligence. 

“Being able to be effectively empathic consistently, open-minded, listening and demonstrate cultural humility – all of these aspects are not replaceable outside of a human in any foreseeable future.”Click To Tweet

On the other hand, physicians can’t keep track of all the various drug interactions that exist. A physician could be adding medication for a sick patient who is already on eight or nine medications. And if you look at the potential implications of an additional chemical added to that mixture and what might interact with each other, physicians can’t keep that straight. And so, we need tools to help us. This is where medical records can look at that.

Moreover, there’s a lot of effort now in what’s called “explaining the black boxes” for even machine learning and artificial intelligence.

So if it says do X, why? We need an understanding of that. And in medicine, there’s a significant effort to ensure that even when we have algorithms in place that guide decision-making, the patient and the physician will have to come to an appropriate understanding of what to do next. Shared decision-making needs to be involved.

[11:03] Should Doctors Rely Solely on the Algorithms?

David stresses that the final decisions between the patient and the physician to proceed include a personalized contextual understanding and risk benefit assessment. For one patient, taking even a medication a day may be too much. For another, adding in another may potentially interact. These are value decisions that aren’t cut and dry decisions in almost all cases.

'Where the art comes in is the effective communication and understanding.'Click To Tweet

The patients are not trained in medicine, the physicians are. And a lot of the training is going to also focus on understanding how to craft these sorts of algorithms. How does one have an understanding that the guidance one receives isn’t just some expert coming up with it? These are weighted based on how thousands of other patients that have responded in similar circumstances. So it’s not a cookbook medicine anymore.

That’s why this guidance is being there and how much should it be weighted still has to be a personal decision.

[13:28] Typical Day and Types of Patients

David is a practicing internal medicine physician. He is a general internal medicine physician so he sees adults. His patients are intermixed with supervising them, taking care of patients, and providing guidance, and finishing notes.

He was one of the members of their clinical decision support committee. With his informatics background, he helped with the guidance for the rules and alerts they put into their electronic health record. So they were reviewing a variety of the rules and alerts and things to change – the sensitivity and specificity, how often they’ll trigger and the value of these alerts at that meeting.

Separately, he was involved with some work on a research project related to informatics. That includes trying to avoid some errors and care for patients and preventing wrong patient and wrong medication errors.

He’s at a medical center now with 12 hospitals. Imagine the number of orders being put in every day across thousands of patients across the system and a busy emergency room.

At any time by mistake, someone might order the wrong medication out of all those orders that were entered in. Or they might enter the wrong chart or enter a medication and how might they address it. So they’re doing research on that.

David is currently involved in a variety of educational efforts related to informatics. He has been reaching out to their school’s students and interest group for digital health and data science.

[16:50] What Does the Future Look Like?

“In the specialty of clinical informatics, the need is not going to go away, it is only going to increase.”Click To Tweet

David talks about information being more interoperable, where in effect, there’s something called the 21st Century Cures Act. It was originally passed in 2016 and it had provisions that are now active as of April 5, for example.

Some of the active provisions there was that patients should have immediate access to their electronic records. There are also other provisions that are going to come into effect in the coming next couple years. In a staggered fashion, is interoperability of data more effectively.

In this context, individual records will be more portable. Patients can take records in their entirety. But even more to the point and a potential threat to large vendors in a sense, is the mandate that organizations should be able to export all the records. they will be able to import them into another system as well.

The interoperability also allows a whole new ecosystem – think like an app store for Apple, or for Google.

There’s a variety of vendors that have app stores coming up right now, however, there’s going to be convergence and the ability to generate apps. There’s a likelihood that they should work with different systems. The apps need to be tailored according to the standards to ensure physicians are going to have better tools to work with.

“The whole environment is evolving in such a way that data becomes more portable.”Click To Tweet

There is a huge need for continued ongoing, ever-growing informatics work from a clinical informatics fellowship graduate to consider getting into.

There’s other data complimentary to the EHR, which is also going to help with personalized care. There’s genomic data, how to analyze, interpret, and continue to update as new information about specific nucleotide sequences and how that’s wrapped into personalized care.

David adds they have people in their medical center who now focus full-time in informatics and don’t see patients anymore. But many will have a blended approach of seeing patients and having informatics-related responsibilities.

[21:42] The Role of Informatics in Merging Patient Information with the Latest Research

David explains he continues to tweak his own approaches to staying up to date on different topics. Through supervised machine learning, one can craft or have curated reviews that come out in different topics. Those curations have some automated searches curated by an individual and then fed up.

This would then allow physicians and students to keep up to date on specific topics. It provides an overview of the likely impactful clinical articles that have come out. 

In the research space, there could be individual search queries that one can do semantic search, for instance. Microsoft has created a nice tool that helps support automatic retrieval of research articles and specific subjects. Google Scholar is also Google’s approach and they have a variety of these tools available.

“These tools are important and clinical informaticians are directly involved in this sort of work to recognize specific structured content.”Click To Tweet

[24:46] What Makes a Good Informatician

Once you’ve completed residency training, you could go into it. Now, one can also work on informatics as a field without going to medical school and develop this. There are opportunities if one chooses not to or if one chooses to go to medical school.

Physicians can get into fellowship if they want to gain an in-depth understanding of what’s possible from individual data elements to managing patients populations more effectively with information.

It’s not expected to complete a two year fellowship to be an expert programmer, for example. One would learn some coding – enough to be able to supervise some projects and understand how it works and the tools available to improve individual patients and populations. But not necessarily coding for a livelihood.

The kinds of job you could get for this would be quality safety informatics roles, and chief medical information officer. You could be working with vendors or at startups to leverage this knowledge or just in your day to day practice.

From an introductory perspective, be able to understand the concepts and how things work. That’s going to be expected during the fellowship program itself. You also have to have a mindset of unintended consequences, unexpected findings, and being sensitive to all possible permutations of the sorts of system changes.

David explains that when they’re making changes from an informatics perspective, they’re impacting thousands and more patients at once.

For custom coding to address very specific issues, one needs a skilled programmer. The automatic generated solutions are not going to work. Similarly in machine learning, there are now automated tools that let you pick the best algorithm to apply to a data set. Then you do the data cleansing and transformation prior to doing a predictive algorithm.

[31:09] Final Words of Wisdom

David recommends students who might be interested in this field to go take a look at the American Medical Informatics Association. He encourages students to investigate this as a career option.

In terms of technology, medicine is a little bit behind in some aspects and way ahead in others. But this is going to continue to grow exponentially in terms of information as well as in the possibility of really helping people.

'Informatics will let one within medicine become one of the experts helping drive medicine forward.'Click To Tweet


Meded Media

American Medical Informatics Association