Episode 8: Stay in Sync with Innovations in Medication Decision Support

Drug Topics®: Hello! I’m Gabrielle Ientile, and you’re listening to Over the Counter, the podcast from Drug Topics®.

In this episode, we’re going to hear from Anna Dover - a seasoned pharmacist and IT expert from First Databank - about the new and exciting modernizations in the medication decision support space.

Anna Dover: I actually have been practicing pharmacy since 2000. And I spent most of my time in direct patient care in a hospital setting. I had some management roles worked primarily in critical care.

After about 11 years, I had the opportunity to move over to the IT department at the hospital I was at - we were going through a large EHR implementation, and I moved into more of an analyst role, became the CDS lead, was the instructional designer, and worked a lot on medication warning optimization as well as general CDS projects.

After 6 years, I decided it was time for a new challenge and joined FDB actually to work on a new product improving medication decision support. It's really been great. I've enjoyed FDB as a customer and I really enjoy working for the company.

Drug Topics®: According to Anna and First Databank, traditional medication decision support is pretty limited, since it prioritizes the medication over the context in which the patient is utilizing the drug, and fails to implement considerations that are specific to that patient.

This creates an overflow of alerts, since pharmacists receive alerts for all of their patients. This makes it really difficult for health care professionals to identify which of their patients are at a higher risk for drug-related problems.

Here’s Anna to delve deeper into the limitations of traditional mechanisms for medication decision support.

Dover: The traditional way to provide medication decision support is looking primarily at the drug regimen. The challenge there is: it's not black and white in the case of each patient. You have general guidelines and guidance. And in a critically ill patient, for example, the experience that I had as a clinician, things that may not be appropriate in the general population may be appropriate in that critically ill patient.

So the idea that taking in additional patient information, whether it's the clinical setting they're in, what their lab values are, information about their age, becomes very important to really get to the most important information, preventing fatigue and really getting that meaningful information for the provider to act on.

Drug Topics®: And when pharmacists have to deal with a constant barrage of alerts, many of which don’t need to be acted upon, or aren’t necessary, the consequence is alert fatigue.

Anna offered examples of alert fatigue from her experience as a pharmacist, where information that’s really meaningful for a pharmacist gets drowned out by all of the noise in the system.

Dover: I've heard several analogies for this. One could be thinking about driving in a blizzard. It’s the idea that you have so many different things popping up in an EHR or a software system.

And it's not just medication decision support, there could be other decision support in there around regulatory requirements or other safety concerns. There can be things just to make sure an order is entered properly and not missing any information. That, in essence, can create several pop ups that a provider that it's just trying to take care of the patient and finish up the ordering for that patient, stops them as an interruptive path, and they need to address it to move on. This creates this page where it's hard to see what is important.

So in my past experience entering medication orders, I knew that I would not be able to see any of the alerts, in particular, when I had a patient come out of, say cardiac surgery. Because everything that they're on, while classically could interact, in the setting of that particular patient, it's the appropriate regimen, it's in compliance with the guidelines, it's safe to do.

So I knew, for example, in a patient coming out of heart surgery, that they would have many orders that would traditionally alert, but it would be appropriate in the context of that patient that was following guidelines and safe for the patients to do – the benefit outweighed the risks. So because I was unable to even see any important alerts, knowing I'd have a flurry of them, I had my own process to go through: review the patient, review the existing medication regimen, and make sure I didn't miss any important safety checks.

I know as a pharmacist, I was fatigued. There was maybe 1 order a week, and I took care of I probably entered thousands of orders in a week, there would be maybe 1 order a week that I wouldn't get an alert. And that was the 1 that I noticed, because I wasn't prompted with alerting just to share that. And I think the big shift now with EHR, pharmacy had been used to this, had been using the traditional medication decision support for many, many years. The shift to putting these in front of a physician now, moving from paper and to off the shelf EHR that may not have the same customization as a homegrown system, has now pushed that problem out into a space that's not as comfortable.

As pharmacists, we’re known and trained to be medication experts. And physicians, while they have great expertise in medications, they're doing many other things in the care of the patient. And it can be more complex to think about the potential cytochrome p450 interaction. That's a more comfortable space for pharmacists than for a physician.

I think the shift in the type of software we're using to take care of patients, and how widespread the usages has really surfaced the issue and made it much more challenging for the seeing these alerts than it was in the past.

Drug Topics®: The most effective way that FDB has discovered to address alert fatigue and drug-centered medication decision support is to flip the script to patient-centered alerts, which take into account things like lab values and clinical risk scores, which work to prioritize high impact clinical scenarios.

Dover: That's the exciting thing that we're working on now. We have a platform that we're calling patient first internally, and the idea is that we're putting the patient first.

It's almost thinking more like a clinician, when you're thinking about the patient, you're thinking about the factors, and then you're going back and looking at the medications to see if any of them could be causing or augmenting the problem. What we're doing with the new platform, a web service-based platform, we're providing a product called targeted medication warnings.

And what it allows folks to do is actually present say, for example, an alert for hyperkalemia. When you think about the way drug-drug information is curated, it's looking at the interaction of the 2 drugs together, but the outcome could be the same. In the example of hyperkalemia, we have more high potassium level, we have a lot of content around the 2 drugs together that could cause it. But very likely when a physician, for example, prescribes a drug that could interact, they know this, and they've likely already reviewed the lab results.

So in that case, we actually can provide this more targeted alert when it matters, when the patient has a high potassium lab as opposed to just at the point of prescribing when it's not a concern yet.

We've done this based on not just scholarly articles and research, but also based on feedback from our customers and others in the healthcare IT space. And even more exciting is that we can look at some of the alerting statistics from our customers. We have another product, alert space analytics, that customers can provide us with their alerting patterns.

We can actually look and find the things that are likely fatiguing, and our clinicians can look at it to decide: how can we make this alert work better? So the hyperkalemia high potassium levels is 1. Another area of concern is QT prolongation, which can lead to a lethal arrhythmia potentially in a patient. It’s a very important alert but it's hard to get the required specificity for it.

Our clinicians have actually evaluated not just what would be required clinically to identify those at higher risk based on research, but also evaluated what data is likely available in the software so that they can have a safe path to allow customers and clinicians to use relevant information about the patient and highlight only those patients with the greatest risk.

Drug Topics®: But do the targeted medication alerts require the use of an electronic health record (EHR)?

Dover: Not necessarily. That's basically started in because I think - it's funny, as a pharmacist, I think we were so used to alert fatigue that we were less likely to complain than physicians who were newer to that type of phenomenon. And we can provide this outside of EHR, any software system that would be interested in using it.

And even more exciting, we're looking at a data version of this that we can provide as well. It's another newer product that we're working on. Because it's been so exciting in the provider space, we're looking at other ways that we can deploy this content not just the web service model, but other more traditional methods that MTV has provided content.

Drug Topics®: Here’s what Anna had to say about the future of medication decision support within the pharmacy space:

Dover: This is where I could wax philosophical for a while, because it's pretty exciting to me.

I think where we're heading is moving away from interruptive alerts for things that could be potential problems, and now deploying a way to surveil so that pharmacists who have always managed these things could more easily identify potential risks, and then only use interruptive alerting at the point that it's a concern.

So only stopping someone when it really matters, but allowing pharmacists to go in and find those patients at greatest risk proactively. More of that medication review process, looking for drug related problems that is part of our clinical practice in pharmacy.

Drug Topics®: Here’s the biggest takeaways from this episode:

1. Limitations to medication decision support stems from it being drug-centered rather than patient centered, where alerts are creating fatigue in pharmacists and clinicians.
2. Innovations that Anna and First Databank are currently working on flip  the script to patient-centered medication decision support that functions to prioritize high impact clinical scenarios where interruptive alerts are only used if there is a concern.
3. FDB’s Targeted Medication Warnings don’t need to go through an EHR, and the company is also working to introduce a data version that can be deployed into more traditional methods as well.

If you enjoyed this episode, make sure to subscribe to Over the Counter wherever you listen to your podcasts.

And don’t forget to check out DrugTopics.com for more expert interviews, innovations in pharmacy, and the latest news on the most pressing pharmacy issues.

Thanks for listening, and we hope to see you next time at the counter.