A Few Words With ForeSee Medical CEO, Dr. Seth Flam
ForeSee Medical CEO and Co-Founder Dr. Seth Flam was recently invited to speak at AAPC’s “Listen Up” Virtual Conference Series. Dr. Flam spoke on best practices for software technology, challenges for risk adjustment workflows and the future of risk adjustment (RA).
What Makes Great RA Software?
Before we developed our risk adjustment software, we attended RISE Conferences to listen to the actual coders and physicians that would use our software. Because if you don’t listen, you can’t create great software, and I like to believe that we created great software. We listened, and the outcome was a five tenet plan.
The first part of that plan was to reject retrospective review as the primary method of optimizing RAF scoring. In other words, our software is built to move the industry from the messy retrospective review process to a more intuitive prospective or concurrent review method. Retrospective review is not optimal for health plans since many have been subject to federal audits that are sometimes linked to retrospective review and likewise the process is a huge hassle for providers who must fax records or in some way participate in the chart acquisition process. If things are done right and more organically at the point of care, with the assistance of software, we can clean that up. So that was our first tenet.
The second was to build a product that encouraged and facilitated collaboration between the coding community and the provider, because we don’t think that software replaces coders. It extends the use of coders across more providers by making it faster for coders to review clinical documentation. Since our industry can’t afford one coder per doctor, we need to improve that ratio. So one of the tenets of our software is to be able to facilitate collaboration and communication between coder and provider.
The third is to be able to link the clinical documentation that’s already in the chart with the recommendations that our software makes. Doctors are evidence based beings. We need to show them the evidence, and be able to point it out to them, on the interface and within seconds, so the original documentation that exists in the patient chart gets promoted into the interface in such a way that it’s easy for the physician to see why the coder and software is recommending that code.
The next tenet, and I heard this over and over again at RISE Conferences was to avoid false positives. I attended sessions with CRC’s where they all raised their hand saying “we’re rejecting natural language processing software because it’s more work than it’s worth, the number of false positives is too high”. You could imagine, if your email inbox was 90% junk mail and only 10% good mail you would probably give up on email right? So we worked very hard to find the right balance between the least amount of false negatives, and the least amount of false positives. So when we make a recommendation in our software, we want it, (for the most part, and we are very good, we’re in the very high nineties) to be accurate. So coders and the physicians don’t lose confidence in the software.
Finally, physicians don’t want to jump around between too many applications. They don’t want to go from their EHR to a third party product. So we’ve been able to listen to them and work with them so that we can insert our clinical decision support inside the most modern EHRs.
And so those are the five tenets of what we think makes great software. If you can do that well, then you can get physicians to work quickly, either concurrently (at the point of care), or prospectively just a day or two before, in collaboration with their coding partners.
How Do Your Clients Measure ROI?
Our risk adjustment software has a per member per month pricing structure so it does not require a huge upfront financial commitment. The return happens quickly. Workflows improve on day one and the number of charts a coder can review improves by an order of magnitude of approximately ten. Of course the way HCC coding works is that if you improve today, you may not feel the complete benefit for several months. But with regard to perfecting clinical coding, you’ll start to see results right out of the box. I would say it takes us approximately two months or so to implement a medical group. Our winning formula is an affordable per member per month fee, ten to twenty times return on investment and all with a fairly simple implementation.
Really when you think about it, a lot of hard work is being performed at the payer end. What those payers really should be doing is sponsoring products like ForeSee in the marketplace because the only way you get prospective or concurrent review is if you're engaging with the medical group before a claim is submitted. By the time you’re doing retrospective review, you’re into a different kind of natural language processing (NLP) product. First pass in India, second pass here in the states, chart acquisition, now you’re into the billions of dollars that the payers must pay. But what the payers really need to do is work with the provider community to push tools out right at the point of care, and work with the coders that the medical groups are employing, so that when they get data, the data is almost perfect, and retrospective review slowly dwindles away.
Your View of The RA Future?
Well I want to say this, it’s not technology, well I guess it is in a way. When I started to look at technology and NLP and computers analyzing medical charts I started to look at physician notation. I’m a doctor, so I could come at this from my own experience, and I see a flaw in medical documentation. I started to see that flaw as I looked at the opportunity for technology to read through charts and maybe one day, by understanding our clinical records, contribute to creating better treatment choices. That data could then be fed to pharmaceutical companies to create opportunities for cures that we would have never had with hand written notes. I look at the documentation today and in a way it makes me sad because we’re not training our medical students and our residents to document in a way that could help technology achieve the goal that we aspire to. So when a doctor says, past medical history of diabetes, or they note in the past medical history section; Diabetes Type 1, and you’re a coder, and you’re going to get audited, you think what is this doctor really telling us? I scratch my head. What I’d like to see is that risk adjustment programs help us as an ecosystem and as a community, work with physicians to improve documentation standards for the betterment of our patients.
So I’m a patient right? I don’t want garbage on my problem list. So when I take my smartphone, and login to my patient portal and see something on my problem list that I don’t have, it scares the crap out of me! Because I don’t want another doctor looking at that and saying, he has this disease, when I simply don't. Or, I may have had it, but it was transient. So I’m looking at things in a bigger way and saying the use of NLP technology in risk adjustment allows us to analyze charts for their integrity also.
I would like medical schools, and I know this is a pipe dream, but I would like medical schools and residency programs to help doctors understand that the more organized and structured their documentation is, the better chance that a computer could read that chart, and one day contribute to the development of a treatment that may help that very same patient.
So if you say, what is the future of risk adjustment and closing care gaps? The answer is that much of the work being done at the payer needs to move to the provider, and it needs to be done at the point of care concurrently, the right way, the first time. I also want to say this. It’s not a complete failure by payers why we’re in the situation that we’re in today. It’s that technology like NLP wasn't available years ago. But the technology is here today, and now it’s time for decision makers, like payers, to be able to accept that technology and motivate providers to be doing things the right way, the first time, at the point of care. So that’s my view of things.
ForeSee Medical is a specialized software platform designed to increase the profitability of Medicare Advantage risk contracts. Using AI including natural language processing and machine learning, our disease detection algorithms rationalize patient data across the healthcare system. It’s simple, using AI we discover diseases from text notes and EHR data you already have. Then, we empower you with insightful HCC risk adjustment support, at the point of care, and integrated seamlessly with your EHR.
Blog by: The ForeSee Medical Team