Q&A article

Q&A with Dr. James Tcheng, Duke University

How structured reporting has transformed cardiology workflows and outcomes for all

When he’s not involved in cardiac catheterization and coronary angioplasty procedures as an interventional cardiologist, Dr. James Tcheng might be spotted in front of a computer. It’s there that he tools around with code and programming to engineer something functional, operationally sound, and downright fascinating.

It makes sense, then, that the Duke University School of Medicine professor’s clinical work pairs well with his personal affinity for technology. As a result, he’s discovered a special passion for clinical informatics that helps cardiologists work better together, improving workflows, communication, and outcomes as well.

With much of that work grounded in structured reporting, a topic that Dr. Tcheng teaches to Duke medical students and will speak to during CentricityLIVE in Las Vegas (May 16-18), we asked the John Hopkins University School of Medicine graduate how he believes data and technology will shape the future of the field:

So, what is structured reporting, and how does it relate to cardiology?

I'll start by dispelling the concept that it's actually about a report. Instead, it’s about the team-based capture of information that’s integrated cleanly and effectively into clinical workflows — distributing responsibilities for data management across all members of the healthcare team. So if you're caring for a patient and you happen to be the closest individual to a piece of data about that patient, it’s your responsibility to capture and manage that piece of data at that point in time.

To do that, you’re presented with a forms-based approach that helps you capture the information as data in a way that can be reused for other purposes. Plus, it gives an opportunity to prompt physicians for additional information, if needed.

So in the case of a procedure, the computer takes the pieces of information, extracts them from the procedure log, and recompiles them to summarize what happened during the procedure. The physician, the person who is actually documenting the procedure, is then only responsible for the assessment of the procedure findings. There is no longer a need to recapitulate all the information that was in the chart or to describe the procedure itself.

Is there a particular area of medicine for which structured reporting is most beneficial?

In the procedure arena, structured reporting is actually very effective, because you can prompt physicians by this forms-based motif for information relevance to that particular context. It helps make sure you don't forget to ask about a certain piece of information. If everybody is doing their component pieces, the approach helps define what it is you should be doing as a best practice.

When all of the members of the team are doing components of the documentation, you actually end up saving time, especially for the physicians. You can cut the amount of time to document a procedure by more than 50%, and you can make the final reports available almost immediately.

Is that how you run procedures at Duke? What are the clinical and financial benefits you’ve seen?

Yes. At Duke, we have structured reporting in our procedure areas, and we always complete our final procedure notes before the patient has left that environment. You can then communicate to the next provider who opens up that document. You can also bill a lot faster, so it actually saves time and money, improves the data accuracy, and increases the responsibilities of all team members.

That way, everybody practices at the top of their license. When different individuals are responsible for different areas of documentation, they know they become responsible for that piece of data — and that set of information is going to be used by others. These dependencies are then self-reinforcing so that everybody feels better about their contributions to the processes of care. It's actually a job satisfier.

So how did you, as an interventional cardiologist, get involved in clinical informatics?

I've been a computer geek as long as I can remember. Through high school and especially into college, I did everything I could to figure out how computers worked. I took computer programming classes, and had declared my major as an electrical engineer, but ended up with a biology major.

Then with my career at Duke, I spent the first 17-18 years in the clinical trials arena. I then had the epiphany: I can develop all this science, follow the scientific method, develop evidence all I want, but in order to disseminate it, it needed to be embedded into clinical processes, and especially clinical systems — computers, if you will. That's how I became focused much more on the informatics side.

It seems that some practices have been a little slow to adopt structured reporting as their model. What's holding them back?

There are several aspects of it. Probably the most important one is cultural: Old habits die hard. We’re taught in medical school that being verbose (as opposed to being structured), and writing an extensive dissertation about patients is the best way to achieve a high grade. So when you're taught this way in medical school, it can be a leap to go the other direction, because what you're taught is what you typically continue to manifest later in life.

The second barrier is complexity. These systems are fairly complex because there are so many dependencies involved in them. But I think that as electronic health records are brought online and people realize that we can't keep doing things the way we used to, there’s now a growing chorus for structured reporting.

We’ll get there, but it’s just going to take time, effort, a resolve, commitment, and resources.

Do you see a future where medical schools are teaching structured reporting, like you do at Duke?

I think so, but it’s going to take time as well. This is not just something that happens overnight, but medical school curricula evolve all the time. The fact that I actually have an opportunity to interact with medical students is a manifestation of that. It'll happen as the drivers align, and the drivers are already there.

In terms of breakthroughs, where do you see structured reporting headed in the future?

Two things: real-time and predictive analytics. If you're capturing information as data and the system is monitoring that data as you capture it, you can actually return information in real-time that supports decision-making. And even if it isn't complete, you can be building a profile of each patient that helps physicians know how to move forward.

The other piece is predictive analytics. You can imagine aggregating lots of patients’ data, understanding the time course of disease. So let's say you're in the echocardiography laboratory and you’re following a cohort of patients who have some degree of aortic stenosis. The tech performs the echocardiogram, and the system could then understand what the values were from six months ago, and a year ago, and anticipate the trajectory of that particular patient's aortic stenosis. From there, it could determine, for example, that in January 2020 you should be thinking about replacing this patient's aortic valve, because in February 2020 he or she is going to get into trouble.

What role do regulatory bodies play in structured reporting?

From the standpoint of CMS, there are so many things that give them an interest in structured reporting — customizing the approach to the patient, the aggressiveness of therapeutics, decision-making for diagnostics, length of stay, etc. All of that is enabled by a paradigm where structured reporting has a key role by capturing information and data.

Another regulatory body that's extremely interested in this is the FDA. They’re committed to using real-world data to generate real-world evidence for regulatory decision-making purposes. Are devices safe? Are they effective? Should this device be approved? Should this device be removed from the market because of evidence? When you capture the information as data, then you can use that information to improve reporting, informing the total product life cycle for medical devices in particular.

I've been talking with the FDA about this very concept. What is it going to take to enable the capture of information at the clinical care level that can then inform regulatory processes? It’s a whole series of things, but one of them is that we should be capturing data in a structured or semi-structured way at the point of care that can then be harvested for regulatory purposes for assessment of devices. The stronger the data you have, the better off either an individual patient will be, or the health of all patients will become based upon the implementation of structured reporting.

How do you think that could potentially impact research?

The mantra is capture once, use it many times. If you capture information as data, that one piece of information can serve multiple purposes — including performance improvement and quality. This is an environment that is an enabler for clinical research as much as it is for directly benefiting patients, supporting regulatory bodies, and reducing the overall cost of healthcare.