July 23, 2009: Morning Session (continued)

Jeffrey Schiff: Thank you. We regret everyone has to talk fast. I think we will have time for refinement and questions.

Sarah Hudson Scholle: I'm going to talk even faster because I have even more specific things to tell you about. I'm Sarah Hudson Scholle from the National Committee for Quality Assurance (NCQA), and what I want to do is tell you about the measures we have now, not that we are really that excited and comfortable with them, as well as the works that we are doing to try to expand the measure set for children. Unfortunately, those great ideas about new measures are not ready for you for now, but they will be soon. I want to give you a perspective of the implementation issues that we have to deal with this. We are developing measures, and why we do not have the best set now, but where we are headed to try to improve that.

NCQA is a not-for-profit. We work for measurement transparency and accountability. Currently, accredited Medicaid plans cover about half of Medicaid beneficiaries who are in full-risk managed care organizations (MCOs), and any accredited health plan is scored based on how they are organized but also their performance on HEDIS [Health Plan Employer Data and Information Set] measures so, essentially, it is required. We work with States. We have 25 State Medicaid programs that recognize or require Medicaid-NCQA accreditation.

Reporting HEDIS measures means that you follow NCQA specifications, that you have a certified HEDIS auditor audit your measures, and that you report it to NCQA through our data system. So our data cover about half of the MCO Medicaid population, but we only have this system for reporting for managed care. We have been approached and actually increasingly, I have had four or five States call in and say, "What about fee-for-service in PCCM? Are you not doing that yet?" So we know there is a lot of interest in use and we know that States are using the HEDIS specifications, but we do not call them HEDIS data unless they have been audited, and that is because one of our foundational principles is that we want to make fair comparisons across health plans, across organizations.

I did want to point out we are really involved in a lot of work on Medicaid. I just wanted to mention we are re-evaluating our accreditation model for Medicaid plans. There is going to be something out for public comment next week that describes our plans, and we would love for you to take a look at that on our Web site and I'll make sure you get the note about it.

We are also working with the Centers for Medicare & Medicaid Services (CMS) on a benchmarking project to try to see how we could incorporate data that follow HEDIS specifications but do not go through our auditing and submission process, and so we are working on that. In general, what we have found is the specs are pretty much followed pretty closely. There are actually pretty small ones, so we have a number of States, again, only managed care at this point in looking at that.

I did want to mention we have been working for about 2 years now on trying to think about how we are going to move the agenda forward on child health quality measurement. We have been supported by the Commonwealth Fund to do this work. And our strategic panel said, "You have to be thinking about a tie to outcomes." And we see that as being really important. I do not think the measures that we have on tap get to all of these outcomes yet, but they are getting closer.

What are the challenges for quality measurement in children's health? Anne Beal talks about the unique aspects of childhood that make it challenging to measure. A quality we are looking at a developmental trajectory. We are looking at care in multiple sectors, which is less common in sort of the typical adult population. The evidence base, typically, NCQA works from guidelines that are based on randomized controlled trials. And that is where we get our measures.

Well, as you know, there are fewer RCTs for children's health care. It is harder to study those long-term outcomes. Often, we are using an evidence base about the potential for problems for children later on. It is a little bit harder to get past—I know that the U.S. Preventive Services Task Force is thinking about how to use this kind of thinking in their reviews—but if we do not have Preventive Services Task Force guidelines, it is hard for us to create measures. And we have spent an awful lot of time thinking about how to rectify the situation. In the past, we have not had a lot of public- and private-sector demand for standardized measures. Nikki just talked about the issues on the public sector side. I have to say children are not a big concern for most of our private-sector clients and supporters.

I want to get into the issues around data sources and what data we have and how that affects the measures that are in our current data set. Administrative data, the claims data that health plans have from paying for care, typically used for quality measurement, they are available but they are limited. We know the limitations in diagnoses and what is documented there.

We have a number of measures that are called hybrid measures where you use the administrative data, the claims data to identify a sample to find out what you can about performance on the measure from claims, and then you do a chart review on a sample of patients. It is viewed as the gold standard. You will find we have very few current measures for children that use this approach because they are expensive. But sometimes, surveys would be a better approach, and I think we have about 160 plans that submit HEDIS data, Medicaid plans that submitted HEDIS data last year.

About 80 submitted CAHPS® [Consumer Assessment of Healthcare Providers & Systems] data to NCQA last year, either using the children with chronic conditions measures or the CAHPS® for children. Again, if they are going to submit HEDIS CAHPS® data to NCQA, they have to use a certified survey vendor and the State cannot—the plan cannot do it themselves. They have to do that, so we find that that is important but again, expensive, and the plans that do this are typically doing it because it is a State requirement.

We are all about using electronic health records. We have been working to specify measures. As we are moving forward on new child health measures, we are going to be looking at the potential for electronic health records to support that measurement. We see that as being sort of the new administrative data or electronic data. But just because you have an electronic health record, you do not know that you can do it.

Accountability—the measures that we have now in HEDIS are specified at the health plan level because that is who reports the data. As we think about the accountability, we are thinking about a population that is responsible to the health plan. That is why we have continuous enrollment criteria because if you are a health plan, you do not want to be dinged for the patient that showed up on the last day of November, and you only have 2 months to get them caught up in their immunizations.

As you think about accountability at different levels, you might have different expectations. As you go into smaller levels of accountability thinking about providers or even health plans, the size of the eligible population gets smaller and it is harder to—you might not have a sample size to be able to do fair comparisons, and you might need to think about risk adjustment for some measures, particularly things I heard yesterday, readmissions and ambulatory-care-sensitive hospitalizations. We are trying to look at those measures for Medicare right now. It is hard, very hard.

I'm skipping over financing.

Let me just point out one measure that we have—followup care for children prescribed with ADHD medication. This is an admin-only measure. It is for children who are on medications. We are looking to see did they have followup visits. Our committee did not want to have that measure. They wanted a measure that looked at whether children really needed the ADHD medication, whether they have an adequate diagnosis. The problem was the information was not in the chart, so it is a bigger issue about changing practice to make sure things get documented. That the information comes from the schools or from families, and it is in the chart where you can see it.

So good measurement also depends on good practice, payment for that good practice and so we measure what we can measure. Performance rates for this measure right now are only 35 to 40 percent. While it is not a very good measure, it is still showing places where we can improve.

I did want to just tell you quickly about our approach for measurement in the future. It is a patient-centered approach. We are looking at children at specific ages. We want to make an efficient chart review approach where we say if you are turning 2, well, we are looking to see if your immunizations are up to date. Did you also have a lead screening? Did you have iron supplementation? Did you have developmental screening using a standardized tool, et cetera? If you have a chronic disease, is there an individualized care plan? Much of this is going to represent a very big change in practice and documentation, and we are getting a lot of pushback on this measure, on this, because of the burden that it is going to represent. In today's charts, often it could be a 1- to 3-hour chart review to find all of these components. We are working with the American Academy of Pediatrics (AAP). I think it is going to require some other things.

And the last thing I want to say is that we are working on care coordination and thinking about how measurements should be combined and sort of stacked so that you are thinking about the accountability at the primary care practice, especially practices, community and State; that you are thinking about structure process and outcomes. And we are working on updating current measures as well.

Rita Mangione-Smith: Sorry, you all had to go so fast.

Denise Dougherty: You want to come up here and sit?

Female Voice: Thank you. My first question is for Helen. I'm wondering if you can tell me where NQF efforts stand on sort of ensuring that there is one measure for a condition. Remember you and I had a discussion that you have like 30 measures for influenza vaccine over 65 years of age, if we are talking about pan-Medicaid measures, and the reason I'm really asking is, you had two measures that sounded like catheter-associated bloodstream infections for NICUs. I'm wondering are they really the Centers for Disease Control and Prevention (CDC) specs, or did you make different specs which is probably not helpful for these kinds of efforts? Then I have a question for Nikki after that.

Helen Burstin: Okay. I'll start. We are trying very hard to try to get the best in class and the one distinction is at times we will endorse two measures when they are based on different data sources. For example, the one example I gave you, I believe is one that is registry-based and one that is not, so I think in that instance we think it is appropriate, again, trying to get people where they are to try to bring those measures in.

We are trying to do harmonization projects. That effort we did for example, we truly did have 30 measures submitted or endorsed around pneumococcal and influenza vaccine. Most of those are gone or have been indicated that at measure maintenance they either have to conform to the standard specifications or demonstrate their superiority, or they are gone. But it has been a struggle. I think we are finally at the point now where we have sort of the heft to sort of push some of those changes through, but it has been a bit of a struggle.

Female Voice: Okay. Nikki, when you did your presentation, one of the things that I have thought about is that you say you are working with 47 States helping grow the new Medicaid directors. What I did not hear in your presentation was—how are you focusing them to help them understand how Medicaid is a possibility even for hospitalized children and for children in emergency rooms in terms of efforts? A lot of your efforts are HEDIS-based. They do not have those kinds of bang-for-the-buck cost savings that can come with some hospitalization measures. Can you give me a sense of where you think Medicaid directors, at least in your organization's counselling and growing of leaders, think about their role as policy leaders? And that graphic you had, the one quality assurance to quality improvement has to go [indiscernible] and this in that last category of having Medicaid be a policy leader for that State in terms of efforts and quality.

Nikki Highsmith: We try not to segment provider groups or institutions or populations, but we tend to do that sometimes based on funding. I actually think CMS and NICHQ are doing some work with States across the country in the area of neonatal care, and they can speak about that. We train Medicaid directors and leaders to try to cut across populations and institutions. We do not have any specific work right now in the area that you mentioned, but I think CMS and NICHQ can speak to that as well.

Jeffrey Schiff: Xavier?

Xavier Sevilla: This is a question and a request for Helen. You mentioned that there are 55 endorsed measures by NQF specifically for children. I think it would be very useful if you could circulate those if you already have them as a set; you could circulate that to the subcommittee for us to deliberate from now until the next meeting. Helen Burstin: Sure. Yes, that is right.

Xavier Sevilla: That will be great. The other question is a question for the whole subcommittee and for the speakers and everybody else is really looking at how going into the future can we really reduce duplication in terms of having lots of different organizations involved with measurement. As an outsider, I just get the impression that sometimes we are doing things and duplicating the same work, so this is just something for everybody just to kind of ponder how we can really all work for the same thing.

Helen Burstin: If I could just briefly respond, I think that is very much our reason for being at NQF. We are trying to be the standard organization and trying to make sense of it, which is why I think some of my concerns it would be really nice to at least start as a starting point and saying these are the ones that people have agreed have met these criteria move from there as opposed to starting at an earlier stage; where measures exist at least go through that process. Most of these NCQA measures are NQF-endorsed. There are no other competing ones, for example. But I think it is a real opportunity.

Sarah Hudson Scholle: If I could just say, NCQA is working with AMA's Physician Consortium for Performance Improvement (PCPI)—which is a group that represents all the physician organizations. So we are working with them to try to make sure that our measures are in parallel and complementary. As we have measures, we specify them at the physician level working with that group.

Nikki Highsmith: Just maybe from a State perspective, I mean you often hear, "If you have seen one Medicaid program, you have seen Medicaid program." I think that paradigm is shifting among Medicaid directors. They are trying to say, "We do want some national standardization." We are looking to the types of entities that are sitting here today to help create that standardization, and we as purchasers want to use those measures to drive standardization across States. We are hearing more States talk about how they can create standardization at a State by State or national level.

Rita Mangione-Smith: I think we have Paul next and then the other Paul.

Paul Miles: This afternoon, we are going to be considering importance, and you mentioned that small is better, so what can you share with us from the measures that have been developed that you think are the most important in terms of data you could give to the States for improving outcomes and affecting cost of care? Is that something you track with your measures?

Nikki Highsmith: Sarah, do you want to talk first?

Sarah Hudson Scholle: If I had to choose, I would choose immunizations because we know it has a clear link to outcomes, a good evidence base. I think having some sort of patient experiences report, and we have standardized measures for that. Among the other measures, the measures that people have used most frequently are the well-child visit measures. And the reason is they are feasible, but we are concerned that you cannot get to the content which is why we are creating new measures to look at the content of well care. So I think the real challenge you have is whether you want—knowing how hard it is to get chart review measures and the challenges. I think BMI percentile, this is the first year it is going to be reported for NCQA because we know that is a huge problem, and that would probably be one of the chart review measures I would go for. So those would be tops on my list to think about.

Paul Miles: Let me rephrase my question. I was not asking about the evidence base for—because I think we are assuming that these are valid. Is there evidence that the measures are then used to actually improve outcomes, and is there a path for where you could then give them to the States to say if you measure this, here is how you could use that measure to actually impact outcomes and decrease cost?

Nikki Highsmith: I mean I think most States use the HEDIS measures for their own—holding health plans accountable. They also use it to drive improvement, improvement projects within their States. Most of the time, it is sort of how many projects can you talk about at one time, and it is a priority-setting process within the States, but all the measures that were just mentioned had oral health in there. Jim and Mary would not let me leave the table, but I think there is evidence particularly in managed care where States are using HEDIS measures to drive quality improvement, projects and programs within the States. The question has been more around priority setting, and how many of them can we tackle at once. But of any of the measures that were just mentioned are activities going on within the States to try to improve care for the population.

Helen Burstin: Just one more point—I flashed up the national priorities and goals at the very end there—but I think it is a very useful framework to think about the way you would pick somebody's measures knowing those are some of the really critical issues at a national level we should consider. That list is wonderful, but it does not particularly get at safety. It does not particularly get a coordination of care, so I would really encourage you to think as well about we know these are the big ticket issues and think about ways. Some of those other measures can be done with administrative data, like some of the safety measures from AHRQ, so I would just encourage you to think broadly about what we would know as important at the national level and ground some of that in those six priorities.

Rita Mangione-Smith: So we have Paul and then Carroll and then Denise, then we will get you guys after that.

Paul Melinkovich: This is a question for Sarah. As an advocate for children's health, I'm really excited about the content of well child care measures that you are working on. As a medical director for a large network of primary care clinics, I'm terrified by the thought of it.

The question I have of you is, have you gotten any acceptance of this? And in the realm of feasibility of collecting the data because it seems to be a significant challenge absent really electronic health records with little check-up boxes where you can prove that these things got done, which I know do not exist in most places.

Sarah Hudson Scholle: That is why we are going to field testing in the next month with the measures. I have to say we have talked about this approach with lots and lots of people, Medicaid, health plans, pediatricians, other clinician groups. They love it. They say this really makes sense; it is really getting into the content of care.

But the second step is, you know, this is going to be really hard to find and the chart is really a change in practice. I think this really represents a new approach for NCQA to go in and say we are going look in the chart for multiple things, and we are going to be thinking—we are really trying to push on a number of activities at the same time.

We are working with the AAP as we develop these measures because one of the things we think is going to have to happen is that there would be templates and tools to help practices document these items. All of the items that are screening items or assessment items have a followup piece that is connected to it trying to get at the coordination of care issues. We are pretty convinced that performance in our field test will be low, and that this is a stretch for the future.

On the other hand, if you look at the NPP (National Priorities Partnership) priorities, they are all about population health and looking at a total population. It is all about patient engagement and care coordination, so these are the measures that will drive it. We are also actively working with electronic health records, organizations that are using electronic health records to identify health. These measures could be specified and whether the specifications would work in an EHR environment, and how you can report it. I guess what I have to say is that this is the future, and if we do not specify the measures for the future today, we are not going to have them in 10 years.

That being said, it is going to be a very big stretch to try to document this information from charts today in most practices, and we recognize that but we are kind of—you know, our choice is do we just use administrative claims data, or do we say what we want to see in the chart going forward?

Female Voice: Okay, Carroll?

Carroll Carlson: The one thing that you did not talk about was the cost aspect, and if you could expand on that some more, I think that that is a discussion we will probably entertain this afternoon as well. I hate to put a blanket on everything, but dollars are important here; it is a finite resource. So anyone of you three probably could talk about it.

Sarah Hudson Scholle: So the cost of data collection and chart review and reporting. Clearly, that relates to why we have the current measures that we have and why—I have not gotten the data for reporting of the BMI percentile measure—which is a chart review measure—yet, so our quality compass is going to be out in a couple of weeks, so I do not know how many plans reported it, but the word on the street is that plans are going to report either the measure for kids or the measure for adults but not both because of the cost of chart reviews. So we are very cognizant that that is an issue, and we are getting a lot of pushback of this and actually were asked to explain what our transition approach would be and how we are going to manage these burden issues.

We are considering several different approaches for trying to manage the transition. One is to drop some of the measures that we have and replace these measures, so get rid of the well-child visit measures and focus on this measure. One would be stratifying our sampling so that instead of five measures, five different age groups, we really just have one measure with a stratified sample from all of the different age groups. That means you cannot report any of the—we will use the field test to try to identify whether we can report any of the indicators. So it is really going to be a challenge to think about this, and we understand that.

Nikki Highsmith: Maybe one comment. Do you think you will need to think about the measures based on administrative data? If we are looking at kind of the current state of the field and the current state of the measure sets, but again you have an opportunity to have a couple of stretch measures where you are pointing towards the future in the balance between how many are administrative versus how many are stretches, something I know you will talk about this afternoon.

Getting to the burden issue and the comment that was made just previously, I think AHRQ and CMS have an opportunity here to think about what sort of technical systems they are going to provide to States, both in terms of the measure collection, validation, and reporting, which is enormous, but also in terms of the improvement piece. A lot of the improvement work that is being done by States is driven by some State priorities, and there are not a lot of national cross-State initiatives; it is an opportunity as Barbara knows.

Rita Mangione-Smith: Denise and then we will go to Cathy, Linda, and Marlene.

Denise Dougherty: Just real quickly, Helen. I'm really looking forward to getting the evidence base and the specs for the measures, 55 measures. One question is, though, do you have any information on who is using those measures now and particularly for the State, kind of level of comfort whether any State programs are using those measures?

Helen Burstin: That is not something that we would have available to us unless we sort of know it by—somebody has indicated that to us, but I think many of those measures for children I think since so many of them are NCQA-based, NCQA could provide. I think the key here is to go to the individual measure developers for that information. I mean AHRQ should certainly be able to provide information on who is using the pediatric quality indicators as an example so that might be one strategy. Anything we can provide that we have, we would be happy to, but it is not something we have informally [sounds like].

Rita Mangione-Smith: Okay, so next is Cathy Caldwell.

Cathy Caldwell: I know yesterday I commented several times and Anne did as well that a lot of what we are talking about fits in managed care, and it does not fit in fee for service, and that issue is much bigger than just limited State resources, and I think you all did a great job explaining that. Is there any current—and I know you all best—high priority trying to move in that direction, but are there any meaningful current measures, even if we had a few, because States actually want to be able to measure these things? I do not think Anne and I are saying, you know let's not talk about it because it does not fit our system. I would love to be able to meaningfully measure the number of well-child visits delivered through our program, so are there any good measures right now that actually measure what we want to measure in a fee-for-service environment where we are probably going to have to use claims data because we really do not have the resources to go out and do 1- to 3-hour chart reviews and that type of thing?

Sarah Hudson Scholle: So we use our measures in managed care because that is the reporting construct, but we know that the same measures are being used in PCCM States. In Pennsylvania, they regularly run reports of these measures off of their claims data, so I think, you know, we talked last night about some of the issues with reporting; maybe the well-child visits are a little harder, that is actually a hybrid measure in HEDIS for Medicaid because of the kind of billing situation that you mentioned yesterday. We use these measures also. I think most of these could work at a physician level, so I mean if they make sense for the population, I think the specs could be adopted, but there are some adaptations in the specs that you probably want to do that have to do with enrollment periods and things like that.

Rita Mangione-Smith: Linda?

Linda Lindeke: Thanks for the opportunity to speak for 7,000 pediatric nurse practitioners, and the tyranny of the visit checklist is just looming in my mind, and when you try to link these vital measures to outcomes, and we know how very difficult it is to build relationships with high-risk populations or any family these days, I'm glad continuity is something we are talking about but data is only as good as what really happens. And as someone who teaches newbies how to do well-child visits, I used to have an hour for ESPDT visits, and we could make it in an hour. We are talking about 15 minutes of a monologue from the provider just to get through all the topics.

And then if you want to see that you have actually made an impact on the health of that child by measuring the checklist, I just have to bring the reality of what this means, and I have no solution, except to CMS. Please pay us more for this work. If you pay us more, they will give us more time because that payment directs how much time we are given in the course of that visit, and then somehow the continuity of visits, if you know the family, you do not have to spend so much time on each item. So CMS, you have a big role for us.

Helen Burstin: Just one brief response, and I do not know how much you guys have looked at the list that has just come out of the Health IT Policy Committee around meaningful use, but there are several measures on there, and the plan is to have all those, several measures for children actually, but there is a plan to have all those retooled and available for electronic health record to electronic data source system in short term.

So if nothing else, I think it is fine to think about where you are now, but it is really important to think through how we get those electronic specifications out there. And it is not just EHR-based. A lot of this work, I think, really if you think at the community level, it involves health information exchanges, so it is about thinking about the bits and bytes of data, for example, on the States. It may have immunization registries—how does that link to the EHRs to make some of this more real and more real-time? And that is the future clearly.

Female Voice: Just so I could say this is—I agree. Some of this is going to be in a checklist, but this is not supposed to be a checklist for a specific visit. This is by a key age how these things happen, so we are really trying to allow practices the opportunity to try to capture these things over the course of time, and we give them ample time. It is like the immunizations, it is up to date by age. All those immunizations are supposed to happen by age 18 months if you look at the AAP or the recommendations. So we try to allow for that, but I know that is a concern.

One of the reasons we are focusing on this followup issue is that we do not want it to be just a checklist that somebody checks off and never does the followup, so it is not—did it happen? What was the result, and if it is abnormal, is there a followup?

Female Voice: Because the emergency room is a place where many children get care off the Medicaid population, and we started talking about the immunizations in the emergency room, the State rates went up at least in Minnesota, so I appreciate that.

Rita Mangione-Smith: We can only take one more question. Marlene, do you mind if we let Marina—yeah, thanks.

Marina Weiss: Marlene, I'll try to divine what you are thinking.

Obviously, most children are healthy and are going to be seen in an ambulatory setting. We are interested in promoting good health outcomes for children and developmental progress and so on, but we also need to be focused on the fact that there is great cost to the Medicaid program and the CHIP program associated with inpatient care. And so back to what you guys were talking about earlier, and hopefully Marlene has kept her [indiscernible] head is as well, could you talk a little bit about the measures that you think are most robust, let us say, and important in the inpatient setting from each of your perspectives?

Return to Contents
Proceed to Next Section