The Accountable Care Organization regulations were first promulgated under authority of the ACA’s Medicare Shared Savings Program in 2011. Three years later, the regs are in the shop for a tune-up. Farzad Mostashari MD was one of the authors of the Brookings Institution ACO issue brief released in the spring, suggesting some changes to the program that would keep current ACOs engaged past the end of their three-year contract term, and improving the program overall. Dr. Mostashari, former National Coordinator for Health IT, is now the founder and CEO of Aledade, a startup focused on helping physician organizations develop ACOs. With a level of excitement shared only by a small coterie of health wonks — and usually reserved for video recordings of unboxing the latest hi-tech toy — Farzad livetweeted his reading of the 429-page typewritten version of the proposed ACO rule when it was released late last Monday. (See the CMS Fact Sheet on Proposed Changes to the MSSP and the Aledade post on the proposed reg.
The rule was published officially on December 8, with a 60-day comment period. I had the opportunity to interview Dr. Mostashari about the new rule. As he noted in our conversation, CMS is calling for input on a variety of issues, so don’t be shy, especially if you have some data to back up your suggestions on the choices that remain to be made in this rulemaking process.
Key pieces of the proposed ACO / MSSP regulation include changes to patient attribution rules and addressing the issue of what happens after year three (both with new benchmarks and new risk sharing options). There are ways in which the regs could go further. CMS has demonstrated its flexibility given some of the restrictive legislative language (patient attribution based on nonphysician encounters); one wishes it had taken similar policy leaps in other areas (e.g. prospective member attribution based on member opt-in, until member opts out — now that would honor beneficiaries’ right to choose even more than the current and proposed mostly retrospective attribution rules).
Have a listen to our conversation here:
Farzad Mostashari HealthBlawg Interview 12 2014
(it runs about 20 minutes).
If you prefer, read the transcript below or download the PDF: Farzad_Mostashari_Interview on MSSP NPRM.
One fundamental question we really didn’t delve into (beyond the discussion of Year 4 gainsharing) is: How does one maintain long-term viability of an approach to health care payment and delivery that is predicated on spending less each year than the year before? At some point, you start cutting bone, or burning muscle, or whatever metaphor you like.
There is certainly a lot of fat, but we have a serious Wanamaker problem in health care. Big data analytics and genomics may solve this problem in health care, but we’re not all there yet. (Y’know: The future is already here; it’s just not evenly distributed.) As at least one med school professor has expressed the Wanamaker problem: Half of what we teach in medical school is wrong; we just don’t know which half.
Better data analytics hold the promise of overcoming this problem. If we can grasp that holy grail, then using the MSSP to achieve real improvements in the realms of health care access, cost and quality becomes more likely. Aledade thinks it has the inside track on the analytics. We’ll see how their ACO clients do over time.
What do you think? Are the proposed changes to the regs sufficient? How would you respond to the CMS call for input on improving the Medicare Shared Savings Program? What are the key data points that we need to be tracking and influencing?
The Harlow Group LLC
Health Care Law and Consulting
HealthBlawg :: David Harlow’s Health Care Law Blog
Interview of Farzad Mostashari, Founder and CEO of Aledade
December 9, 2014
David Harlow: This is David Harlow at HealthBlawg, and I have with me today Dr. Farzad Mostashari, who is the founder and CEO of Aledade and former National Coordinator for Health IT at the U.S. Department of Health and Human Services. Hello, Farzad and welcome.
Farzad Mostashari: Hi, David. Thanks for having me.
David Harlow: Today we’re going to talk about the new proposed ACO regs; and as folks know, the Medicare Shared Savings Program regulations that govern ACOs — the Accountable Care Organizations and other related programs — first came out in 2011, and were finalized in November of that year. So, here we are three years later, and there are proposed revisions to those regs are out for consideration, and I guess the questions that I have at this point in time are: What have we learned over the past three years? Where’s this program headed in the future? ACOs are supposed to create incentives for healthcare providers to improve quality, to reduce costs, thereby breaking the cycle that’s tied to fee-for-service reimbursement, and other sorts of “old thinking.” So, the question to kick this off, Farzad, is: The Medicare Shared Savings Program is at once a very ambitious program, and also one that’s tied up in knots, thanks in parts to its enabling legislation. So, hundreds of organizations have signed up as ACOs; over time some of them have very publicly withdrawn from the program. There are good points and bad points to the program as it has been rolled out to date. So, what do you say, what’s your assessment of the difficulties with the existing rules, and what do the proposed amendments do to solve these problems?
Farzad Mostashari: I guess I’m going to be speaking both as someone who’s studied ACOs as a part of Brookings, the group that came up with the idea with particular focus on physician-led ACOs and we published a policy, an issue brief, that CMS cited in their NPRM [Notice of Proposed Rule Making]. So, both as kind of a researcher and a visiting scholar at Brookings, but, also now, as someone whose new company, Aledade, is setting up ACOs, particularly physician-led, independent physician-led ACOs, across the country, and what I’ll tell you is, that it is wonderful to have the ability to look at the United States of America, and say “Anywhere we want to go we can set up an ACO. We know what the rules are. We don’t have to think about different states, different regulations, different insurers and everyone having their own approach.” So, the fact that there is, as part of the Medicare Program, a permanent program that is a Medicare Shared Savings Program that has the same rules all over the country is an incredible asset, and it has been a major push forward, to encourage this movement — as you pointed out — from volume to value, and to let healthcare providers, doctors and clinics and hospitals and others, do this transition into this new type of thinking, where your income doesn’t depend on how many people you can get in and out, and how many procedures you can do with them, but actually on measures of outcome. Income depends on outcome, right? Are the patients actually going to the hospital? Are they being admitted to the hospital? Are they racking up costs? How you doing on the quality measures? How are you on patient experience and satisfaction? So, that’s a major thing and I think that, in that respect, the fact that we now over five million Medicare patients in these arrangements just quickly over the past two, two-and-a-half years, is a great accomplishment, but what the CMS was — the policy challenge that it had was, a lot of those groups were saying, “We don’t know what’s going to happen next. This is a three year contract that we signed with you. What happens in year four, and are you going to make the incentives such that you’re encouraging people to take higher levels of risk?” If I were to identify one overwhelming issue that’s a policy challenge that CMS had in this was, almost everybody who entered into the ACO program entered in under one-sided risk, meaning if the cost came down they shared in any savings, but if costs went up they did not have to share in any losses.
David Harlow: Right, and if I remember correctly that meant that the exposure to savings was somewhat lower; the shared savings percentages would have been higher if people accepted risk.
Farzad Mostashari: That’s right. It was a little bit higher like from 50 % sharing if you’re in the one-sided program to a 60% share, in the two-sided program, and basically almost no one took them up on the two-sided option, and there is a lot of belief and I actually sign on to this, that you have more commitment to change if you have downside, as well as upside opportunity, risk and opportunity –
David Harlow: Sure –
Farzad Mostashari: — and so a major challenge for CMS was a lot of people were saying “Look, if you force us to go to-two sided risk, we’re just dropping out because we’re not ready.” A lot of people were saying that. So, obviously that would not be a good thing for the program. So, how do you maintain momentum while going, keeping movement up the escalator, right? It sounds familiar to me. So, CMS, I think again in this NPRM, and we have to note, this is the notice of proposed rule making, and everyone who has been kind of a policy wonk, following this stuff, knows that a lot can change between the proposal and the final rule, and the regulators really, really do listen to comments. So, if you have, particularly, if you have evidence or data or some important perspective, you really should, between now and 60 days from now put in your comments to CMS, they really do read and respond to them. But in the overall feeling, I think, that I had when I read those 429 pages, I felt that they had not quite done enough to encourage people to step up to the two-sided risk, where they want people to go. And the CMS actuaries themselves, in the proposed regulation, estimated that under the system that they put in place, only 10% of ACOs would be in the two-sided program. So, that to me is the big issue that they still need to work on; they can fix it, but they still need to work on that.
David Harlow: Sure, now the fact of the matter is that many of these Medicare ACO operations are also providing services to other patients as well, and the numbers haven’t changed, the Medicare requirement is that there be at least 5000 Medicare beneficiaries who are attributed to an ACO, and my thought is that the actions of a system that’s trying to make an ACO work under Medicare, are going to be felt by other patients in the system as well. They’re not going to provide one method of caring for patient X because she is member of a Medicare ACO, and a completely different method of caring for patient Y who is not.
Farzad Mostashari: A lot of the primary care doctors I work with have that very concern. They say, “This is great stuff that we’re doing for the Medicare patients. It’s almost not fair for me not to be doing it for all my patients.” The challenge is that, two things, one, the other health plans can be a little bit of free riders on this, where they don’t have to — maybe feel that they don’t have to, also ante up some shared savings, and that the practice will make some changes anyway, but more than even being able to do it because you have shared risk, you got to have shared data, and one of the big things that Medicare is doing to make this actually work, whereas you could say, “Why doesn’t a primary care doctor do this today?” For one thing they don’t have the incentives, the other thing is, they don’t have the data to understand what the costs are, what the patient’s utilization is of the emergency room of the hospital or specialist, what scans they’ve had, what medications they are filling, all that stuff, and CMS is saying “Okay if you’re an ACO you can get that for the Medicare patients.” They can’t give that to me for my Aetna patients or my United patients or Blue Cross patients. So, just practically speaking it is definitely a barrier to these practices being able to do this sort of population health work for all of their patients. So, our strategy has been that we identify the primary care providers who have a significant, substantial enough, portion of their practice being Medicare fee-for-service patients, so that A, it’s worth their while, and B, they can make these changes and affect a significant, substantial, portion of their population, not just a small part, and then three, once we get them up and running on the Medicare patients, we would then go to other health plans and say “Hey look, we have this whole infrastructure in place. We’re doing good things, we’re getting good results. Why don’t you give us shared risk and shared data for your panel of patients as well?”
David Harlow: And here you’re talking about what you’re doing as Aledade?
Farzad Mostashari: What we’re doing as Aledade, and what I expect most ACOs are thinking — that theyned to expand their population beyond just Medicare to the commercials.
David Harlow: Right, so there are couple of areas in which you have shared already some comments on the proposed regulations and I wonder if we could highlight a couple of those. I’m thinking specifically, starting off with the patient attribution, there’s some changes in the regulation that allow for attribution to an ACO in a couple of new different ways. Could you speak to that a little bit?
Farzad Mostashari: One of the challenges in population health is: Who are my patients? Who is the denominator? I wrote a piece for HIMSS talking about how the denominator is like one of the greatest inventions ever, right? Who is the total number of patients? Who is the cohort? Who is the population that I need to focus on? — Not just the people who have a visit scheduled with me today – Who’s not in the office today that I should care about? And to do that you need to know, If I’m a shepherd who’s my flock, right? So, that’s the challenge here, and the way that they’re doing it right now, is at the end of the year they say “Okay, people vote with their feet. If they don’t like the care you’re giving, they can go somewhere else and then they’re not attributed to you anymore, and on the other hand, if you attract them, and you’re sticky to them, and they come in and see you, then they’re yours.” This is a little challenging for some ACOs. I actually kind of like this because it says to you, “You have to treat every patient as if they are in your ACO,” because they could be. Other organizations, particularly those who have experience with managed care and Medicare Advantage say “No, no, give me my roster,” like “I want a list of names that’s not going to change.” So, CMS did this, proposed that for Track 3, they include that capability.
David Harlow: But the roster is based on an expected panel of attributed patients that could be adjusted.
Farzad Mostashari: That’s right. Whoever was your patient last year, that’s your panel. If they move, if they don’t like you, if they change doctors, they go to Florida, too bad, and if you bring in new patients, too bad, your roster is set. What they should have done — and you’ll appreciate this, David — is they should have let the patient vote, not just with their feet, but with their voice. They should have let patients affirmatively say “This is my primary care doctor. I want to have my care coordinated by this person,” and to have the patient be the one to make that enduring connection until they choose to revoke it, in which case they should be able to choose not to be in an ACO. So, what they, what CMS missed in this, was the opportunity to affirm patient choice as overriding everything else, and it’s a concept that the patient advocate groups has been pushing for as well as ACOs themselves have being saying and even MedPAC has said “Why not let the patient choose if they want to be in an ACO?”
David Harlow: It’s very interesting problem because the legislation put us into the situation by being very respectful of the notion of patient choice.
Farzad Mostashari: Yeah. So, let’s double down on patient choice.
David Harlow: But you raised a very interesting approach to changing the way in which we measure patient choice, and allow for patient choice, which would be more productive all around.
Farzad Mostashari: Yes, and right now the only choice patients have is to have their data not shared, and I think we should go a step further and say if you want to choose you can choose to be in this ACO, it doesn’t matter if I go to a Walgreens or if I go to a nursing home and I get care there, my primary care doctor is this person, and contrarily, if I’m uncomfortable with this whole thing, I shouldn’t just be able to out of my claims data being shared, I should be able to opt out of having my healthcare have financial implications in this way for my doctor.
David Harlow: Right. Now one of the other things we touched on is the idea of making things more user-friendly for physicians, and I think one of the aspects of the changes to patient attribution may be doing that in term of effectively opening up the possibility for some specialists to belong to more than one ACO. Do you see that as a positive?
Farzad Mostashari: I think it’s fine, I mean honestly I believe that the driving heart of accountable care is primary care. So, it makes sense to me that if you’re a specialist, it’s fine, you can belong to multiple, or you can have patients from multiple ACOs, and it won’t affect the attribution of those ACOs, but really, the attention should be on the people who provide the primary care, and the other thing the rule did, which was really terrific, was they allowed for nurse practitioners and PAs who deliver primary care, where in many parts of the country, and particularly rural parts of the country, deliver the majority of primary care to also be included in terms of that attribution. So, that was a good thing.
David Harlow: Yeah, so that was sort of an acrobatic feat going beyond the language of the statute.
Farzad Mostashari: Well the, I’m sure the lawyers made sure . . .the language…
David Harlow: I’m sure it fits in somewhere…
Farzad Mostashari: Yeah, but they did work hard, I guess we would say in reg-speak they worked hard to make it possible.
David Harlow: So, what are some of the other highlights that you see as creating opportunities as result of the proposed changes here, assuming that they do get finalized as written?
Farzad Mostashari: The main things that we touched on, patient attribution, and the two-sided risk that we talked about. Another issue is, now what happens in year four in terms of the benchmark, if you’ve reduced cost, are you constantly chasing your own tail? At some point, you can’t keep going down lower. So, they talked about a bunch of sensible approaches to that, but they didn’t propose any one of them. So, the commenters are going to have to do a lot of the heavy lifting here in term of creating consensus, through groups like Brookings or others, in terms of what one path CMS should offer on that. Kudos due to CMS for making more data available on more ACO beneficiaries, so they streamlined the data opt-out process and they included more information, that’s a good thing, and the nurse practitioner attribution was good thing, and then the final thing that’s notable is, that they said and I certainly support this, there’s a lot of payment rules in place to protect against self-referrals and basically getting more — too much — services to patients. So, telehealth restrictions, nursing home three-day stay, those kinds of things, and those kinds of policies make sense, if you’re in a fee-for-service world, where the more you to do the more you make money, but in a world where the issue is taking on shared risk, it doesn’t make sense any more. You’re only restricting me, for example, from directly admitting the patients to a nursing home instead of having to have a three-day hospital stay. Right, I’m not going to do that…
David Harlow: It makes no sense in this context.
Farzad Mostashari: That’s right. In this context it doesn’t make any sense, applying telehealth outside rural areas, well why not, if that’s the most efficient way to do it? I’m not going to abuse it, because I’m not, I’m on the hook here just like you are, Medicare. So, allowing for those waivers for ACOs that are taking on two-sided risk also makes sense as something that CMS has proposed.
David Harlow: Yeah, so it’s encouraging and where I guess my overall sense of all this is that it would be nice if it could have been done in a cleaner way, but we’re left with all of these vestiges of fee-for-service reimbursement including just the way that ACOs are paid in the first place, and there ar
e all these adjustments that have to be taken in order to make this work, certainly a lot of work here for the lawyers and consultants, and hopefully it is all going to benefit patients and provider organizations as well —
Farzad Mostashari: A triple shot, right?
David Harlow: Yeah, that’s right.
Farzad Mostashari: A triple shot is what we’re looking for: good for society, good for patients and good for the providers.
David Harlow: Yes, absolutely. Well, thank you very much. I’ve been speaking with Farzad Mostashari about the proposed ACO rule, and we will see how this rule gets finalized in the months ahead. Thank you again, Farzad.
Farzad Mostashari: Thanks again.
David Harlow says
Comment from the mailbag:
This is exactly why this trend is likely to go the way the MCOs did 20-30 years ago. It’s a problem of diminishing returns – eventually, as you point out, ya can’t get more without affecting quality of care.
Do I think there will be no changes in the business of medicine? Of course not. There will be quality analytics – some of which will be relevant to actual patient care, much of which won’t.
So the big question is, will these steps really bend the cost curve? Temporarily, maybe. But over the long haul, patients/employees/consumers of health insurance will make demands, and doctors, who are smarter than average but may not be aware of nuances in payment initially, will learn to exploit the system.
It’s already obvious in the ACO environment – those ACOs that maximized their coding were successful, and those who did not, failed or were marginal and quit.
We (all of us: medical, insurers, patients, policymakers) have to get our heads out of the clouds and remember the dirty, basic facts about human nature: we are self-centered, self-interested beings. If you want different behavior, get real and incentivize the behavior you want. It won’t be cheap, but it’s more likely to be effective.
Sue Ann Jantz, administrator