Gregory H. Dorn, MD, MPH is president of First DataBank of South San Francisco, CA.
Tell me about yourself and the company.
I’m a physician. I went to medical school, undergraduate, and medical school at Columbia. I trained in surgery at UCLA and then did a Masters in Health Services Management at UCLA.
During my residency, I became very interested in the process of care and how to improve the clinical process of care. That stemmed from my undergraduate work in operations research, or really industrial engineering. This became a nice marriage of the two.
Throughout my clinical training and subsequent to that, I saw a lot of opportunities where there were clinical practices that weren’t always well substantiated as being best practices. Also, in the hecticness of clinical practice, you would see a lot of errors that would occur, particularly with complex medications in the ICU — I’ve spent a lot of time there.
That passion grew in me. That’s where I helped start a company called Zynx Health. It has grown to become, I think, a standard bearer within the field of evidence-based medicine. Subsequent to that, I moved over to First Databank to take what I’d learned at Zynx, and also prior to that , to bring it to bear within the clinical drug decision support environment. To try to optimize what I think is a really significant opportunity to inform clinical practice at the point of care around drugs. That’s one of the most heavily integrated into the workflow decision support domain today, as opposed to perhaps referential content or medical or nursing or traditional clinical information.
First Databank has been around for about three decades, focused exclusively on integrated drug knowledge. I emphasize the “integrated” piece because there’s a lot of drug knowledge out there, reference and integrated. But from its very inception, FDB has been heavily focused on integrated. That means embedded into the software application used by the clinician, whether she or he be a nurse, a pharmacist, a physician, a nurse practitioner, a physician assistant, and any myriad of care extenders that may come to bear here as the healthcare economy expands tied to the Affordable Care Act.
We’re focused on delivering that clinical content to the point of care. We’re not focused on being a supermarket of information, or being all things to all people and assembling every different type of clinical content you might want, but rather to be true experts at the point-of-care decision-making process such that clinicians get the most value out of that alert, that ordering sentence, or any other type of dosing information or a range of other clinical decision support in the drug domain.
Both First Databank and Zynx have strong brands, to the point that I’m not sure everybody knows that both owned by Hearst. What are the commonalities between the two companies?
Hearst is a very broad, diversified media company. They own the San Francisco Chronicle, Cosmopolitan, parts of ESPN, A&E, and Lifetime. They’re all organized into major operating clusters that are thematic. We’re in Hearst Business Media, which is focused on business-to-business, workflow-embedded content — decision support.
First DataBank, Zynx, and Map of Medicine in the UK are all focused on the medical-clinical side or healthcare side of things. The relationship specifically between Zynx and FDB is that Zynx takes a broad view of clinical decision support and says, “What are all the sources of information I can derive a best practice from? How can I then package that information in a useful clinical format — an order set, a care plan, an intelligent clinical alert?” There’s also a significant amount of forecasters and calculators. Taking a broader approach to distilling best practices.
FDB goes one layer deeper. Zynx can run on the infrastructure of nomenclature data, alerting, drug structured information that FDB provides. We go that layer deeper, where we’re optimizing the exact order sentence. If you have a Zynx order set that’s evidence based that’s going to drive reductions in mortality and you select to execute that order sentence, the next series of steps to make that orderable sentence truly specific to the patient’s context and very intuitive to the clinician but also that it translates into a dispensable that can be handed out by the pharmacy –we have specific data sets that allow that translation to occur seamlessly.
If you think about ordering that medication in the setting of a particular diagnosis or co-morbidity or the setting of another medication or the setting of a particular lab result, our alerts are optimized to make sure that that alert is meaningful to the clinician. That’s where the interplay between Zynx and FDB comes in. Those that use both see significant benefits.
You could argue that most of the value of CPOE and other clinical systems, beyond standardizing what’s available for ordering, is the third-party content such as that offered by Zynx and FDB. Are you actively looking for other areas where critically reviewed literature might come into play to enhance existing clinical systems?
Yes. We think of the clinical decision support environment as a cycle. If you can think of the patient making a transition from healthy to sick and then having to interact … this could be in a chronic sense. I don’t have a chronic disease, I now have a chronic disease. I don’t have an acute condition, I now have an acute condition. At that point, there are three phases where an individual interacts with the healthcare economy with regards to clinical decision support.
There’s something we call a pre-encounter phase, which is before I have an encounter with the healthcare system. There are whole hosts of activities that occur – eligibility, necessity, formularies.
Then there’s the encounter stage, which is when I’m actually in front of the physician. There’s that intimate moment with the nurse, the physician, the pharmacist when the decision is being made. That’s what we’d call the encounter phase of the clinical decision support cycle.
Then there’s the post-encounter phase, all of the activities that relate to what happens after the patient has had an encounter with a health system that are related to clinical decision support. There you’ve got a measurement around data and dashboards and you’ve got clinical billing and just a whole host of activity – claims paying and so on.
We look at the universe with that framework. Today we’re very focused on the encounter phase. As you can see, Zynx and FDB really dominate that encounter phase. When you’re at that moment of receiving care, we can influence the decisions that are being made and reduce mortality and morbidity. We are very interested in looking at types of content that fit the other two domains, whether that be post-encounter and pre-encounter and beyond. Without getting in too many specifics, just know that those are very interesting to us right now.
You recently announced AlertSpace. What are its advantages?
In this encounter phase, there’s this problem of alerts being highly sensitive but not specific. You get lots of alerts, but you don’t know which one is really germane to your patient’s care, so you ignore a lot of them. What we’ve seen in our research is that by clicking through alerts, unfortunately, there’ll be a click-through of the one alert that really mattered. The patient can have an adverse outcome by oversight of that valuable dosing alert, valuable drug –disease interaction, or whatever it may be.
In AlertSpace, we’re allowing institutions to customize their alerts — turn off the alerts that are not as meaningful clinically to them and promote or retain the alerts that are highly clinically meaningful to them. This is done through a web-based tool, a SaaS approach, so it’s pervasive. It’s available to any subscribing institution.
They actually customize their data directly before they get their data load. They’re able to see those alert customizations the next time they publish their FDB data, which can be weekly, monthly, or even daily.
AlertSpace helps reduce the noise factor and highlight the alerts that are truly clinically meaningful, thereby reducing the risk that meaningful alerts are overlooked and patients have adverse outcomes. Right now we have a whole of host of institutions that are using the tool. It’s been our most successful new product launch in the history of FDB.
AlertSpace is a tool, a solution to helping with alert fatigue. But there are also other approaches that we’re taking around the editorial choices we make about which alerts and serve upstream and trying to understand the validity of the content before it has to be adjusted by AlertSpace. There are myriad of approaches we’re taking to optimizing alerts. It’s not just that we’ll keep publishing the same content and give you tool to fix it. It’s more that we’re going to really improve the alert relationships and give you a tool.
That’s an interesting approach. Instead of relying on EMR vendors to repackage your data with the inherent delays, you’re letting customers pre-customize their own. What was the thought process there?
We wanted to close the cycle time gap between new technology reaching the end user. We obviously work with all of our vendor partners because they have to support these customizations, but what I experienced at Zynx, where we have a web-based authoring environment that allows for content to be customized and then published within a myriad of target systems … that paradigm is one we brought over to FDB. We thought FDB had the capabilities to deliver an end user application. We thought that would be very valuable to our brand and to the value we bring the clinicians.
It’s a little bit of what we learned in Zynx. It’s a little bit of trying to close the cycle time between innovation and the end user’s access to that innovation without having to enter into, as you can imagine, a long product cycle or revision cycle. How can we get this alert customization technology into the hand of end users as fast as possible? Through our client base, we know about the mistakes that occur out there around drug CDS. I know personally of hospitals that have had errors that are related to alerts. We’re mission driven about that now.
The rebranding of the company’s image appears to signal that FDB wants end user visibility, not just to the IT folks or people who apply your updates. Are you looking for a brand identity with the end user?
Absolutely. That’s been one of my focuses since I’ve been here. We’ve talked a lot about it, the idea that we are so pervasive throughout all of these different systems — not just with hospitals and medical groups, but PBMs and insurance companies – but yet if you were to go to AMDIS or HIMSS or a whole range of different meetings and ask CMOs or CMIOs, “Have you heard of FDB? Do you know of FDB?” Even the end user clinician, chances are they’re going to say, “No, I haven’t heard of it.”
Based on the impact we’re having and the impact we can have on clinical workflow, we really wanted to have that be more effectively recognized by the marketplace. End user tools that don’t interfere with our relationships with the large system vendors are a very significant strategy going forward for us. I think the reception’s been pretty good. We had a lot of large systems who we’ve met with and they like the approach so far. So I think you are right on there. We’d like to raise that profile.
We’d like to do more around end user tools; help customize the content. The thing I observed when I came in was that pharmacy clinical information was one size fits all. This is across the industry. People just publish a file, the system takes it, puts it in, and you deal with the result. That’s maybe 1.0, or even 0.5 – the first phase of the industry.
Drug CDS 2.0 is going to be about customization and personalization. That’s where we’re headed. Tools and the highly specific content that gets right down to the individual nuances, whether it be their renal function, liver function, physiology, a whole range of things. Eventually and in the not-too-distant future, their genotype and how that’s expressed as a phenotype and how they then metabolize drugs will be a very important area for us.
How do you prepare to start using genomic information?
You have to be vigilant, first and foremost, about the body of evidence — what the body of evidence is telling you about where you can adjust dosages. We’re tracking that. That’s first and foremost. As that grows, we’re compiling it.
The second piece you really need is physicians, nurses, pharmacists, and healthcare institutions to become much broader users of genetic testing. Then using those results to close the loop for a metabolic adjustment with regards to a drug. We can capture the data and develop a dosing tables that say, “If you’re a cytochrome P450 metabolizer, this is your warfarin dose” or whatever it may be in a chemotherapeutic regimen. We can do that. We have people tracking that today.
What we need is the input side, which is doctors becoming reimbursed so that it becomes more common to order a genetic test. That result can be pinged off our data and a more specific dosing parameter can be returned. Our goal is to try to help move that along. Obviously we don’t control all the pieces, but we’re very excited about how that might unfold over the next five years.
You joined the company after average wholesale price lawsuits had come up. What was the impact on the company, and how do you think the industry has changed now that average wholesale price not used to calculate provider drug payments?
That’s a great question. I joined right as we were heading into this cessation of publication, so it was a little bit after my tenure. I spent about six months re-analyzing that challenge. Hearst asked me to do that.
What we realized is that we couldn’t continue to publish AWP, which if you really look into it, is a relatively arbitrary measure. Ceasing publication of AWP had very little impact on the company. We were able to go to our customers and provide them with alternatives, whether it be wholesale acquisition cost or other measures, that they could use to meet their needs. We very successful in being able to provide alternatives that were anchored more directly to data submitted by manufacturers.
What we moved on to is that we are in partnership with the State of New York, doing a survey of average acquisition cost. New York is collecting acquisition costs from pharmacies. We’re averaging those in partnership with Ernst & Young. That’s potentially generating a new benchmark for the State of New York.
You also probably know that Alabama and Oregon both have acquisition price types. California, which we’re very close to, is close to moving forward with an acquisition price type. The federal government has launched an acquisition price type initiative. We’re doing our utmost to push towards this acquisition metric in the hopes of adding transparency around pricing, but still not saying it’s the only measure, but saying there is now a range of price types that can be used. We’ll definitely do our utmost to be first and foremost with the acquisition price type.
I think it’s very exciting. If we can get better transparency on drug reimbursement, it’s better for the patient, it’s better for the healthcare economy, it’s better for employers. There’s a whole host of benefits. I think part of being innovative in that space is what’s been interesting for me.
Any concluding thoughts?
I want to make sure that your audience understands that we’re not just a US-based drug clinical decision support company. We have a division in the United Kingdom — FDB UK — and they have a very, very large position in the UK with drug clinical decision support. We have a significant presence in Asia. We also have a very nicely growing footprint in the Middle East.
We operate as a global drug clinical decision support company. If you look at all the different drug clinical decision support companies, we may be one of the few that do that successfully.That’s an important characteristic of who FDB is. As the healthcare IT market grows globally, we’ll be ready to address the needs that come, wherever they may come from.