Michael Mardini is founder and CEO of National Decision Support Company of Madison, WI.
Tell me about yourself and the company.
National Decision Support Company provides decision support criteria and algorithms that are based on national standards, seamlessly embedded inside of EMRs so that physicians can be aided in making the most appropriate care decisions for their patients at the appropriate time.
What’s the status of Medicare’s advanced imaging requirement?
We’ve gotten some clarity, but there’s still a little bit of fuzziness. It is scheduled to go live on January 1 , 2018. It require physicians to do a consultation with appropriate use criteria for advanced imaging studies for Medicare Part B cases.
What has not been identified yet is the reporting and the claims process. We are going to get some information on that in the next rule-making cycle, which will come out in early July of this year from CMS. There’s still a little bit left to learn, but we think the January 1, 2018 date for the consultation piece is going to hold.
Who doesn’t get paid if the requirement isn’t met?
That is an interesting question. It is the radiologist. It is the radiologist’s responsibility to submit proof that the doctor who gave them the referral did a consultation.
On the back end, ordering clinicians who do not consult appropriately face some penalties by way of prior authorization and further scrutiny around ordering once they get some data over the couple of years, but initially it’s on the radiologist.
Are radiologists willing to accept that change in their workflow in making sure referring physicians went through the mandated steps?
It’s very similar to the commercial prior authorization number. There’s some identifier that is going to be the evidence that there was a consultation done and the clinical decision support mechanisms are required to produce the unique identifier as evidence of a consultation. That number will have to be placed on to the claim that the radiology group submits. They haven’t fully defined what the claims requirements are. There may be some additional data aside from that number, but the workflow is going to be similar.
Your system has to be used by the ordering physician rather than the radiologist, correct?
The ordering physicians are the ones to primarily interact. The radiologists will interact with our system if it’s an unaffiliated referrer to confirm that the decision support number that they have gotten is valid. We think that radiologists will access our solution to confirm that they have a valid number, but with the interaction of AUC and CDS, it’s the ordering doctor, yes.
Is it correct that radiologists are either sent a valid number or they aren’t and they can’t obtain the approval ID themselves?
That’s a very common question that we get. Radiology groups ask us whether they can perform the AUC interaction, even on the phone, so a doctor calls in and they can capture the information. Right now there’s nothing in the regulation or in the statute that would indicate that the radiologists can do that. The onus is on the ordering clinician to do a consultation. It makes sense. This is supposed to be informative and educational to help doctors make the best choices.
What other types of clinical decision support beyond advanced imaging have you added to CareSelect since we last spoke a year ago?
Inside of imaging, we’ve added pretty big sets of criteria for the American College of Cardiology as well as National Comprehensive Cancer Network. Outside of imaging, we’re focusing on some key areas. There’s labs, which is a very similar kind of an issue that’s being faced in imaging. Medications, and when we talk about meds, the entire corpus of meds is impossible to address, but you’ve got some high-cost and specialty meds that need attention.
We are rolling out a solution around opioids, both from a clinical decision support angle as well as a state registry submissions and reviews for opioids. Blood management is also a big topic where there’s some strong criteria out there that needs to be delivered. We’re getting into antibiotic and microbial stewardship, where there’s also some good content out there that absolutely needs to be delivered to help improve decision-making. Admission Level of Care optimization is also a big area of interest.
A year ago, CareSelect was this generic content delivery mechanism focused primarily on the Choosing Wisely initiative. Over the last year, we learned a lot about what the market needs and we’re reacting.
Other companies take the content approach in which the EHR vendor builds their product around a third-party database and handles the user interaction natively within their product. What’s the challenge of offering an integrated service instead?
It’s interesting and it touches the heart of what we do. We start as a hosted content management platform. We use a common web services standards based mechanism to integrate with these EMRs. You can imagine this ability to manage, create criteria, use a single mechanism and a single UI inside of an EMR to deliver thousands of sets of criteria. Whereas all these EMRs have a facility for their customers to build criteria, but these require big build efforts with multiple files created locally that need to be managed.
In our architecture, it’s a common feed. With the CareSelect platform, the technical challenge on the EMR integrations side is simplified. The work on managing the content is taken off the back of the EMR.
Are EHR vendors generally cooperative in adding another company’s product to their systems?
One of the reasons we do well with the EMR vendors is that from a workflow perspective and eyes on the screen, we leverage their platform. There is no CareSelect application. There is no NDSC platform installed locally. We’re leveraging all the native windows that are in the EMRs.
In a sense, we’re adding value to the EMR. The perception to the user is that this is a native EMR alert. There’s nothing foreign about what we are doing, so from the EMR’s perspective, we’re adding value.
Small vendors always complain that the EHR vendors lock them out. Would your approach work with other types of solutions?
There are always challenges around interoperability. I say this all the time — I think these EMR vendors get a bum rap, I honestly do. There is data out there and there are ways to integrate. One of the challenges, or one of the things that I often hear out there with customers, is complaints about vendors that are making offers to solve problems that aren’t reliably solvable, either because the data’s not all there or reliably accessible.
There’s a lot of reasons for that. For us, we stay within ourselves. We understand what we can solve and what we can’t solve and that’s what we deliver. We have good relationships with these EMR vendors. It takes patience. What you ask for today you might not get for another 12 months and that’s fine as long as you can plan for it. These guys have an unbelievable amount of work to do in just delivering everything that these EMRs have to do.
We have our little world, as every vendor does selling their individual solutions. I couldn’t Imagine having to put a ubiquitous system in like a Cerner an Epic or Meditech to satisfy the needs of a couple of thousand doctors and administrators, all with different and sometimes conflicting needs. It’s a challenge and I applaud them for that. Now tack on integrating hundreds of third-party apps all with a different idea of how they want to exist on the desktop. Not fun.
How do you see the future of the company and the ongoing availability of the industry group vetted guidelines that you use?
Sites and hospitals and doctors want to use content for its clinical efficacy. They want to make the right decisions. They prioritize which clinical content sets they use in choosing those that solve a clinical problem, but also address an administrative problem or a business problem. A lot of that has to do with connecting out to payers or their population health platform.
An example would be to ease the prior authorization process, or a notification process, or actively being a part of a population health initiative in an ACO. Using the clinical data and the decision support as a part the workflow to ease the data exchange and communication burdens, for lack of a better term, just to get paid. That’s the cross-section for a decision.
If you’re looking at 40 opportunities to deliver guidance, the 20 that they pick would be the ones that also of have a financial and operational impact. That’s what we’re seeing a lot of. We have hundreds of criteria and the ones that people want to implement are those that are clinically valuable, but that also have an operational and financial impact on their operation.