32 min
ENCORE EP 263 - The Start-up Who Won Medicare’s AI Contest, Beating Out IBM, Deloitte, and Mayo—A Conversation With Andrew Eye Relentless Health Value™
-
- Medicine
If I had a nickel for every guest on this show who went on to achieve wild success … TJ Parker from PillPack three years before they were bought by Amazon. Anyway, let me introduce this show with a clip from the recent podcast (EP325) with Dr. Mai Pham. We were talking about the rampant and very open secret of excessive upcoding in Medicare Advantage (MA) that is costing American taxpayers a fortune and is very not correlated with actual spend. Here we go with Dr. Mai Pham:
Stacey: Do you have any thoughts relative to how you ensure that these MA plans that are becoming vast are still accountable to not game the system? How do you plug loopholes in a way that doesn’t invite additional and more nefarious gaming?
Dr. Pham: My fantasy has always been that CMS can develop, or somebody can develop, a black box machine learning–driven, risk-adjustment algorithm that no one can see into—not even the payer. It would very much level the playing field, assuming that it was developed correctly, appropriately, and you used unbiased data; but that’s the kind of system and extreme solution that I think starts to sound almost necessary given the state of things and the rate of acceleration in upcoding.
So, people may not have noticed that CMS had put out a request for — I think it was a challenge grant, maybe? And they recently announced a couple of winners. They were asking for artificial intelligence–driven approaches to predicting health outcomes, which I believe is just the first shadow approach, the first step that you take in thinking about artificial intelligence–driven risk adjustment.
I also want the audience to understand, it’s not like we’re talking about replacing a really superlative gold standard, right? The majority of the most commonly used risk-adjustment approaches today produce a correlation with actual spend of only like 0.2. This is the best we can do? This is how we’re deciding how we’re going to spend a trillion dollars each year? Surely, we can do better.
And, by the way, the winner of that CMS AI contest was ClosedLoop.ai; and Andrew Eye from ClosedLoop.ai was on the show. Cue Encore Episode here!
In the original version of this show, there was a whole prelude about whether AI is or is not anything beyond an overused marketing pitch; but I think, in the time-space continuum, we’re beyond that conversation now. Don’t get me wrong, everybody still has AI in their cloud analytics platforms. And some of them are still, as they say, programmed in PowerPoint (that was a joke); but real deals are emerging from the fray.
As mentioned, in this health care podcast I talk with Andrew Eye about AI. (He was born for this job.) Andrew is CEO over at ClosedLoop.ai. ClosedLoop.ai beat out over 300 rivals with their system that forecasts adverse health events and then plops warnings even in the EHR with action steps for clinicians to avoid the calamity in the making.
You can imagine many things that CMS might be contemplating using this tool for, including as a control for false upcoding and all of the financial toxicity that goes along with that. By the way, keep in mind all the top-performing Medicare Advantage plans are using today, right now, some form of advanced analytics and artificial intelligence to risk stratify their populations and predict which members will, without intervention, become high cost in the near term. Others are using AI right now to do the kind of predictive analytics that you need to excel at population health.
I get to ask Andrew some of the hard questions that have been bothering me about all the AI hype, and he set me straight a couple of times. Love it when that happens.
You can learn more at closedloop.ai or by following Andrew (@andreweye) on Twitter. Andrew Eye’s executive and entrepreneurial experience spans over 20 years in business to consumer and business to business for start-ups and Fortune 500 companies. Andrew founded and sold three technol
If I had a nickel for every guest on this show who went on to achieve wild success … TJ Parker from PillPack three years before they were bought by Amazon. Anyway, let me introduce this show with a clip from the recent podcast (EP325) with Dr. Mai Pham. We were talking about the rampant and very open secret of excessive upcoding in Medicare Advantage (MA) that is costing American taxpayers a fortune and is very not correlated with actual spend. Here we go with Dr. Mai Pham:
Stacey: Do you have any thoughts relative to how you ensure that these MA plans that are becoming vast are still accountable to not game the system? How do you plug loopholes in a way that doesn’t invite additional and more nefarious gaming?
Dr. Pham: My fantasy has always been that CMS can develop, or somebody can develop, a black box machine learning–driven, risk-adjustment algorithm that no one can see into—not even the payer. It would very much level the playing field, assuming that it was developed correctly, appropriately, and you used unbiased data; but that’s the kind of system and extreme solution that I think starts to sound almost necessary given the state of things and the rate of acceleration in upcoding.
So, people may not have noticed that CMS had put out a request for — I think it was a challenge grant, maybe? And they recently announced a couple of winners. They were asking for artificial intelligence–driven approaches to predicting health outcomes, which I believe is just the first shadow approach, the first step that you take in thinking about artificial intelligence–driven risk adjustment.
I also want the audience to understand, it’s not like we’re talking about replacing a really superlative gold standard, right? The majority of the most commonly used risk-adjustment approaches today produce a correlation with actual spend of only like 0.2. This is the best we can do? This is how we’re deciding how we’re going to spend a trillion dollars each year? Surely, we can do better.
And, by the way, the winner of that CMS AI contest was ClosedLoop.ai; and Andrew Eye from ClosedLoop.ai was on the show. Cue Encore Episode here!
In the original version of this show, there was a whole prelude about whether AI is or is not anything beyond an overused marketing pitch; but I think, in the time-space continuum, we’re beyond that conversation now. Don’t get me wrong, everybody still has AI in their cloud analytics platforms. And some of them are still, as they say, programmed in PowerPoint (that was a joke); but real deals are emerging from the fray.
As mentioned, in this health care podcast I talk with Andrew Eye about AI. (He was born for this job.) Andrew is CEO over at ClosedLoop.ai. ClosedLoop.ai beat out over 300 rivals with their system that forecasts adverse health events and then plops warnings even in the EHR with action steps for clinicians to avoid the calamity in the making.
You can imagine many things that CMS might be contemplating using this tool for, including as a control for false upcoding and all of the financial toxicity that goes along with that. By the way, keep in mind all the top-performing Medicare Advantage plans are using today, right now, some form of advanced analytics and artificial intelligence to risk stratify their populations and predict which members will, without intervention, become high cost in the near term. Others are using AI right now to do the kind of predictive analytics that you need to excel at population health.
I get to ask Andrew some of the hard questions that have been bothering me about all the AI hype, and he set me straight a couple of times. Love it when that happens.
You can learn more at closedloop.ai or by following Andrew (@andreweye) on Twitter. Andrew Eye’s executive and entrepreneurial experience spans over 20 years in business to consumer and business to business for start-ups and Fortune 500 companies. Andrew founded and sold three technol
32 min