Podcast Episode Transcript: Searching for better outcomes and lower costs: The challenges facing state Medicaid directors

With L. Gordon Moore, MD

Dr. Gordon Moore: Welcome to the 3M Inside Angle Podcast. This is your host, Gordon Moore. Today, I am speaking with Billy Millwee, who is the President and CEO of Millwee and Associates. Welcome Billy Millwee.

Billy Millwee: Hey, Gordon, glad to be here.

Gordon: I would like to talk with you about your experience as a Medicaid Director and about issues that are important to them and how Medicaid can vary from state-to-state. Before we get there, tell me a little bit about Millwee and Associates and what you do.

Billy: I was with the Texas Medicaid program for about 20 years, served as the Medicaid Director and Deputy Health and Human Services Commissioner for three years, and was fairly active nationally in Medicaid with the National Association of Medicaid Directors. When I left state-service in August of 2012, decided to form a consulting company, and that led to the creation of Millwee and Associates.

Basically, we are a group that works with states, companies like 3M, to better understand the Medicaid marketplace or to help states understand value-based payments or pursue waivers for changes to the program they might want to make. I do some mentoring work with new Medicaid directors to help them understand Medicaid. I guess I describe us as a general Medicaid consulting firm. It’s myself and I have one full-time employee and about eight or nine 1099 subcontractors that I work extensively with.

Gordon: Your background and what you work on, and your national understanding, is exactly what’s interesting to me. In that, Medicaid is a big amount of health-care spending and it varies a lot across the states, and yet there are probably some common themes there as well. Given your work in Texas as well as your work now around the country, I wonder if you have recognized the kinds of things that are the same and also vary?

Billy: Well, the kind of work that is the same is really the political attention that Medicaid has received. Medicaid is now about 25 percent of every state budget and naturally, it’s a bit contentious, because on one level some states are fairly progressive in looking for coverage options. Other states, like Texas, have not pursued coverage options, so Medicaid is generally the foundation for coverage within any state—many times the largest insurer.

That’s good and that’s bad; it’s good in that the benefit is there and able to provide care for low-income people, but at the same time it really puts a lot of pressure on state-funding. In fact, in my many states, you start seeing where Medicaid-funding is crowding out funding for education and other programs.

Virtually every program you go into, in any state, the spending is top-of-mind for any state Medicaid program and the financing mechanisms are; and Medicaid Directors are always under pressure to reduce costs, but improve quality or improve access and do all those things. So you’re asked to do sometimes what feels to be almost an impossible task—control spending but at the same time, don’t affect quality, don’t affect access, don’t impact the care that we’re delivering now.

Gordon: That’s a really challenging task. I’m asking imaging if you’re going to fix a Medicaid budget we can either cover fewer people, cover fewer services, try to negotiate better prices, or somehow try to fix the way health care is delivered. That last option is typically the one folks are pursing, which also the most challenging.

Billy: Exactly. Going back to the past a bit, the way states used to approach Medicaid in reducing spending, particularly during those years that a state has a real budget problem, was that they would reduce eligibility or they would just do across-the-board provided rate reductions. Neither of those really worked effectively, because you’re going to deny people coverage who are going to be off Medicaid and then when the budget crisis is resolved, the states comes back in and adds those people back in, and they come back in sicker than they were.

Now, there are so many federal requirements, the states are really limited in how much they can reduce eligibility by. I think we’ve all seen that just reducing provider-rates isn’t really effective. I used to try to explain to our appropriations committees that Medicaid is like this big balloon. If you squeeze it here thinking that you’re going to save a few pennies, it’s going to pop up somewhere else. So if you’re trying to reduce one level of service or reimburse rights in one area, it’s just going to create a problem in another area. The best approach is to look at what you’re paying for and how effective is that, and also the incentives that you’re creating through your payment models.

Gordon: Let’s pivot to what’s different across the states and your observations of how states may have to approach this differently because of either local or national issues.

Billy: Well, what’s important, I think, before you go into any state is understanding, believe it or not, the organizational structural because the responsibilities of a Medicaid Director will vary from state-to-state. Depending on where they are in the organization dictates what their concern of the moment may be. For example, in some states, the Medicaid Director is responsible not only for the operations of the Medicaid program but they may also have responsibility for eligibility for all social programs as well as responsibility for the Office of Inspector General.

In those cases, the crisis du jour for them may be the most recent article in the newspaper about a major fraud issue within the Medicaid program, so that changes from state-to-state-to-state what the responsibilities are. In some states, a Medicaid Director may have responsibility for the mental health program, in other states they do not. It’s always helpful to take a look at a specific program and what responsibilities fall within the Medicaid Director’s purview.

Yhe National Association of Medicaid Directors does a survey every two years to look at one, Medicaid Director tenure and what are the top issues within Medicaid Directors—what are they most concerned about? One of the major issues that comes up is around tenure. It’s increasing. I think the average tenure of a Medicaid Director right now is about 20 months, but if you look at a program the size of Medicaid, you really would like to have much more tenured people.

I was very fortunate in being able to spend 20 years in the Medicaid program and headed up every major division within the program, so by the time I became Medicaid Director I really understood how these things worked. If you have a Medicaid Director who is a political appointee and did not have the benefit of growing up in Medicaid, then you have the whole other set of circumstances that you have to deal with.

Universally, Medicaid Directors are generally looking at value-based payments, alternative payment models, and new approaches to delivery stems design. As part of that smarter approach to reducing cost, I think overtime people have realized that just reducing eligibility and cutting provider rates isn’t the answer—that you really have to look at being smarter in the way you pay for serves and in your delivery system design.

Also top-of-mind are systems. You have these legacy Medicaid management information systems that were really designed to process claims and now what Medicaid programs need is more analytic capability, so you’re seeing a lot of growth in most every Medicaid Director. You know, they’ve looked at systems, their confused by systems; they just don’t want more data; they want to understand the data that they have. They are trying to make sense out of data that is flowing out of a fire hose and really trying to get to usable information that they can use to better manage the program.

Gordon: Let’s go back, for a minute, if you don’t mind, to changing the health care delivery system. What levers does a Medicaid Director have to effect that kind of change?

Billy: Well, there are a host of federal waivers that states can get. One is called an 1115 Demonstration Waiver and that will allow a state to do a lot of things—a lot of innovative things. Other waivers allow managed care and probably more universally, Medicaid managed care. Capitated Medicaid managed care has been the lever most frequently pulled. Now, this started in the mid-nineties and the call was budget certainty that the managed-care plans really approaching governors and state legislators with out-of-control Medicaid budgets and saying, “Well, you know if you go with capitated managed care model, you can be certain of a PMPM and you can project your budgets better.”

Now, that has evolved a little bit though because spending is still growing. The utilization is up in many states and so now the focus has shifted a little bit toward, “Well, if we go with managed care, besides budget certainty we also want to get value.” So that’s where the paradigm has shifted over the last few years. I think states are looking for more levers now of saying, “Okay, we bought into this idea that managed care will provide us with budget certainty, but how do we know that we’re getting value?”

An issue that is really getting more and more traction over the past two years are around social determinants of health. As alternative payment models have been really the focus of managed Medicaid Directors over the past five or six years, now you see this social determinants of health come out and increasing focus on how we can manage that with many states saying, “Well, we really can’t get to reducing cost in the Medicaid program without addressing these other factors that impact health but are beyond the ability of the health care system to really resolve, such as housing, transportation, food and security.”

Gordon: The social determinants is such a hot topic. There has been so much discussion around that. It always makes me a little nervous though when I think about health-care delivery system picking up housing or transportation, potentially making it more costly and expensive, as we tend to do when we start, as doctors, are touching things and playing with them.

Billy: Yeah, that’s true and there hasn’t really been a good solution. What I point people to is—there isn’t a good solution right now that I’ve seen on social determinants of health. You can identify them potentially but really, what can you do about them? I would suggest that when you identify someone who is at risk for some of these social determinants, does that not apply a need for greater care management?

If you know you’re dealing with a homeless diabetic schizophrenic, maybe identifying him as being at risk is the important part and then doing additional case management, care manage, and outreach is what is needed, other than creation of a whole new system. There is a host of issues at the federal level that they are going to have to address in the way that flowing flows. There has been some talk about social service block grants—the whole block grant discussion is one that’s a tough one, but blending your Medicaid funding with your housing funding and doing some things like that.

That’s the only way I think that you can really get into that and do something constructive, because you do—is Medicaid going to become a housing agency? I don’t know, but for now I think it’s important to recognize social determinants play a role and in what mobility do we have to manage those. I would suggest that it’s using good tools to care manage those people who are at risk and using the tools to identify and treat that part of the population that really needs the high-utilizers, the small percentage that drive cost. If we focus on those, then I think we can effectively impact the social determinants as best we can.

Gordon: I think about the complexity of putting budgets together. I guess maybe that’s what is sort of an eye-opener for physicians who might be working in federally-qualified health care centers, for instance, and see individuals who are homeless coming back repeatedly to the emergency department and they think, “Boy, if I could just get them housing,” but that’s an entirely different budget that has a whole different set of rules and their various agencies coming together. Is that what the 1115 Waivers are about?

Billy: Yes. What the 1115 Waivers do is try and look at how they can bridge some of those things and bring funding together. There are several states that are working on a combined kind of waiver that will help address housing. There are resources within the community and sometimes I think what is under-utilized is good care management, not necessarily just medical case management, but social case management. For instance, let’s take the provider in the FQHC who has a problem with a schizophrenic diabetic over some housing issues. If that provider had access to a good social work case management person, then they could start linking them up with the resources in the community.

So many of these resources are already available in the community, it’s just there is not one traffic cop directing this. You have the poor provider who is trying to do the best they can in managing the schizophrenic diabetic and it’s not incumbent really on that provider to understand all the options available through HUD. There are many, many social services out there that vary from community to community, so having a good social work case management model that understands the community and their resources available in that community are incredibly helpful.

We use a tool called Aunt Bertha that allows people to find resources within a community. Aunt Bertha is a product that kind of puts all those things together for a community so that they can address some of the social issues around health care—so connecting people with food, connecting people with housing, connecting people with other social supports. So that, I think, is probably where social determinants can be addressed but it has to be at the state level in reassessing its care management models to be more broader in their approach, not just medical case management, but social case management.

Many states are including provisions in their manage care contracts and it’s without deliverables or without measures but requiring the health plan to address SDOH. It’s little concerning, because there really are no performance measures around it, but anxious to see how some of the plans are going to respond to that.

Gordon: I think the hypothesis is that let’s say I’m a care management organization or I’m a group taking some kind of up and down side risk in Medicaid in a state, then if I get a payment to work on care management for high risk individuals and I identify those people through social determinant flags, then I should be able to demonstrate reduced hospitalization, ED utilization, and therefore improve overall costs.

Billy: Exactly.

Gordon: Hopefully, that would be the outcome.

Billy: Right, yeah. The debate becomes, do you do that with a single statewide kind of approach or do you embed that within managed care plans? I don’t know the answer to that. Arizona is looking at a model where they developed a SDOH solution that is managed by the state and, of course, the plans would benefit from it; or do you require each plan to develop its own SDOH model? So that’s going to be something to watch and I’m anxious to see what some of the models are developed look like.

Gordon: I put on my provider hat and I’m thinking here as a doctor, I’m working with people who are covered by Medicaid in the state and I’m trying to be thinking ahead to managing population to better outcomes. I wonder if either the issues for a Medicaid Director is to not make it overly complicated to say, “Look, there is probably lots of inventiveness and things you can do out there. What I want to do is set up parameters and boundaries, and incentives that reward you for doing a good job with that, but let’s not get too nitty-gritty with what I mandate from the state.” Is that typically how folks like you come at it?

Billy: Well, you know you’re going to see different approaches. My approach was, and remains, that if you set the right financial incentives, you’ll create the right outcome; you guide people with incentives versus dictating process. Some states are very much process oriented; they want to dictate and be very prescriptive about what you must do. That doesn’t always work out, because if you dictate the processes then in many way you own the outcome and you limit the innovation and flexibility that a provider can bring to the solution.

There are a set of Medicaid Directors nationally that I kind of look to as leaders and ones that are more forward-thinking thinking those terms, about “Well, let’s talk about setting right incentives versus dictating a specific process.” I think that is the probably the predominant way, although some states do become very prescriptive. I don’t think they generally have a good outcome with that approach.

Gordon: Yeah, I was going to ask, is this a new thing and a trend that’s swinging one way or the other; and then, do you have an insight into which approach tends to demonstrate better outcomes?

Billy: My experience has been if you set the right incentives, then you’ll create the outcomes. I mean, I don’t want to sound callous, but at the end of the day if health care is a business, I want to get the attention of the C-suite involved in solutioning. If I put enough money and incentives at it, then I know I’m going to get the attention of the leadership of the health plan involved and that will influence what happens—the outcomes—if you set the incentives right.

For instance, back in the old days, in Texas anyway, we used an innumerable group of HEDIS measures and we had one percent of the premium at risk based on those HEDIS measures—a very small amount—and also some other measures. You really didn’t get any attention to that at all from the executive team. But when we moved to putting 3 percent of the premium at risk and largely base that on potentially preventable events, well, you got a lot of attention and you saw things happen in months that had either never happened or taken years to address when we used HEDIS measures. So that’s what I mean by setting the financial incentive, you’re speaking the language, I think, of the health plan and you’re getting the attention of the leadership involved in communicating your priorities.

Many times we communicate our priorities in a financial way and I think that helps to tell the vendor community, these are our priorities and we’re willing to put our money in those areas.

Gordon: The one percent to three percent sounds like it made a big difference, but it seems like a surprisingly small change. How is that possible?

Billy: When you look at how a managed care premium is constructed, there are margins built in for profit. In Texas, the profit margin at that point was 2.5 percent. So by going to three percent, we were actually putting at risk more than their entire profit for that plan, so that’s why you got the attention—is that you started out by saying, “The amount that’s built into the premium to recognize profit was placed at risk on performance.” The rationale was, if we’re paying you to do these things and the theory being that if a health plan is performing well, then people are not going to the emergency room to receive their primary care. They aren’t going to the ER but they don’t need to be.

They are not being admitted to the hospital when they don’t need to be, because they’re getting good primary care, and when they are admitted to the hospital they are not being readmitted, and that’s kind of a very, very simple definition of a properly functioning health plan. If they are not performing at that, then to me, they haven’t earned their profit. That was kind of the thinking behind that approach. By doing that, it did get a lot of attention and you started seeing changes very quickly.

The first year we did that, we didn’t put the health plans at risk. We wanted them to live with the data that they had, so that it would be unfair for me to just give you this data and immediately start hitting you with a penalty. It’s important for you to understand your data and where you have opportunities to improve. Then after that, is when the incentives and disincentives start kicking in and a health plan could earn more money. Nothing is better than getting your competitors penalty as a bonus payment. So that’s how that process was constructed.

Gordon: When you cross the profit-margin threshold, I certainly can appreciate how that would capture attention.

Billy: Yeah.

Gordon: Did you have to hire bodyguards at that point? How was the reception?

Billy: It was not well received but at the end of the day, it’s really hard to argue against quality. You can argue unfair measures, but then you dispute that by laying out the measures in a way that people can totally understand. If it’s clinically accurate and people can see how the measure works, and understand that they can effect that measure, then they’re almost forced to live with it or are really embarrassed by it.

You saw some creative things. You we had a health plan in Houston that really took a deep dive on their data and they looked at their hospital admission—potential preventable hospital admissions—and those were coming from emergency departments. They did a gain-share agreement with a group of emergency department physicians that largely staff the ERs that their client were going to and low and behold they saw a reduction in their potential preventable hospital admissions.

Then they took another step and said, “Well, wait, why don’t we have these people going to the emergency department anyway?” They mapped those out on their zip codes and found out that where people lived that there were in their plans, there weren’t a lot of urgent care centers or afterhours care available. So they started working with providers in those zip codes to create afterhours clinics, afterhours care, and you actually saw some actual management of the system in the network happening where you had gain-share agreements, or you also had additional payments being made for afterhours care. The financial incentives worked to create that kind of response that absent those measures, I don’t think that anything would have happened.

Gordon: What was the impact on overall costs?

Billy: We have done some studies—I don’t have those in front of me right now, but we actually saw in Texas where we had a reduction in ED visits, preventable hospital admissions, and readmissions. What was really impressive to me—in Texas we have a couple different models of managed care. One is called STARTSHIP, and that’s basically your pregnant women and children, and we saw the reductions there. More importantly we have a program called START+PLUS, and it’s our integrated managed long-term services and support where you put—that’s your high-utilizers—25 percent of the population that uses 70 percent of the resource. We saw the reductions there as well.

Typically, many people talk about how that people is more difficult to manage, but when you really started looking at preventable events in that population, you had a host of them and you wouldn’t have thought that actually because they should be getting the highest level of service coordination. So what we found was those members probably weren’t get the level of service coordination care management they needed, but when we assigned that three percent at risk amount, that certainly improved.

Gordon: That sounds like a pretty good approach. I like that.

Billy: Yeah, and it gets a little bit better. Right now in Texas managed care, they are developing algorithm—in managed care, members can choose a health plan but many don’t. Maybe 45 to 50 percent of people don’t choose a health plan and they are defaulted into a plan. In some states, they just do this logic that’s fair. If there is four health plans, then each one gets a member and that’s the way the algorithm works.

Well, now they’re working on a model that will assign members based on health plan performance with one of the measures being PPEs. If you think about it, a member that doesn’t choose a health plan is probably very valuable to a health plan because they’re probably non-users. They don’t really use the system that much, they’re not that engaged with health care, they really may not have PCP that they’re really looking for. So that member could be very, very valuable. Now when we not only hold three percent of the premium at risk, now we base the new membership assignment based on your performance on a number of measures, including PPEs, that it also begins to reinforce performance at that level, because that’ a life that’s very valuable to a health plan.

Gordon: I can image that could be a multiplier effect on that. As we move towards the closing, where do you think things are going now—although, that may be too broad—but just in a broad stroke, what’s the next stage for Medicaid Directors?

Billy: We’re at an interesting period. I still participate in the National Association of Medicaid Directors and it’s kind of interesting that you’ve had a bunch of the Medicaid Directors who’ve been around a number of years and have influenced things have left and a new generation is coming into play, particularly at the NAMD Board. We were joking with them at a meeting in the spring about if they’re going to abandon the alternative payment models and delivery system reforms. No, they’re not, but I think you’re going to see more and more emphasis on that. But they’re also really, really embracing the whole concept of social determinants of health.

I think you are going to see more emphasis on SDOH but at the same time, people aren’t going to lose their focus on alternative payment models. I see the new sexy bright shiny object is now going to be SDOH and we need to give that the consideration it needs and factor that in. I think that’s going to pose an interesting discussion on how we do risk adjustment; how we can factor that into things like CRGs, but it’s a discussion that we need to start having, because that’s becoming increasingly important to Medicaid programs. We need to help them understand it better.

Gordon: Well, Billy Millwee, we thank you so much for your time today.

Billy: You bet. Glad to, Gordon.

View Session Spotlight (PDF)