Managed Care Innovator Mary Barton, MD—HEDIS: What’s Changing and How to Get Ready

Mary Barton, MD, MPP, is vice president of performance measurement at the National Committee for Quality Assurance (NCQA), where she oversees the development, use, and maintenance of NCQA techniques to evaluate health care quality. Dr Barton recently sat down with First Report Managed Care to discuss the latest updates pertaining to the Healthcare Effectiveness Data and Information Set (HEDIS). She provided a look at HEDIS measures that will be rolled out next year, and told us what she thinks organizations should start doing now to get ready. Dr Barton also talked about a new initiative that will allow organizations to implement customized quality care projects, and gave us a preview of what organizations can expect at NCQA’s inaugural Health Care Quality Congress this fall. 

Interviews like this usually end with a glimpse at the future, but let’s cut to the chase. What is the NCQA’s vision for HEDIS beyond 2020?    

Currently, HEDIS measures require retrospective reporting and analysis. They use claims to assess things that were done in the prior year. We want to transform to a system that is more prospective. Clinicians and patients are documenting care as they experience it and using clinical decision support. We think that’s a perfect opportunity for measures to be tied in to the care environment—to remind stakeholders of what needs to be done and when, particularly with regards to preventive care interventions. 

By tying the measures to the care process prospectively, it will take much less effort than the way it’s done today. 

Can you provide an example of how this concept will work? 

Sure, take blood pressure control, which is one of the measures we are changing in 2019. In the past, we required review of the chart in the first 6 months of the year to confirm a diagnosis of hypertension. Then, in the second half of the year, we’d look at measurements to see if the blood pressure was controlled. That is an entirely retrospective view. There is no way to improve performance except by controlling everybody’s blood pressure. I am not suggesting that’s a bad thing, but it’s very broad. 

Under the new process, the plans will receive an algorithm for figuring out who needs to have their blood pressure controlled. Instead, you will receive an algorithm that includes a target goal for blood pressure. This way, a health plan and its analytics team can determine in the first quarter of the year who are going to be the likely people to measure and bring under control. They can send out reminders to those patients to come in and get their blood pressure checked. They can also put them on the roster of the care management team, where nurses can identify ideal candidates who would benefit from the initiative. It’s all prospective.   

Clinicians would likely see this as being more directly related to patient care.

Exactly. We want to build an alliance between the health plan and providers.  Health plans have information that providers can use to make real improvement. 

Can you give another example?

Of course. Consider follow-up after hospital discharge. This is another area where the health plan has data that the clinician should also have. We’ve seen instances where a clinician did not even know their patient was hospitalized. Of course, the health plan knew. That’s not a good thing.  Clinicians would like to have a steady flow of information from the health plan that shows where their patient was treated and when. 

Tell us about brand-new HEDIS measures for 2019? Let’s start with hospitalization following discharge from the skilled nursing facility. NCQA will require that organizations report the number of discharges that result in an unplanned hospitalization within 30 and 60 days. 

That’s right. This is one of the first efforts within the HEDIS data set to look at post-acute care quality.  It’s something that health plans have not really had a direct eye on, in our view. The
Centers for Medicare and Medicaid Services (CMS) has various systems that it uses to measure quality in skilled nursing facilities and other post-acute care settings. But health plans have not been required to let members know which facilities offer the best quality. We think there’s a clear role for health plans to be a part of this ecology of care, especially as the population ages. The rate of hospitalization after discharge from a skilled nursing facility helps us determine which ones offer the best quality.  

Beyond that, the measure holds the plan responsible for what it is doing for the patient during those first 30 and 60 days that a patient is home. To what degree has it done an adequate assessment of the patient’s home situation?  To what degree is it providing the skills, whether it’s occupational therapy, physical therapy, or a visiting nurse? When things like this are done well, rehospitalizations are minimized. 

What should organizations be doing now to get ready to meet this measure?

I suggest running analyses of patients discharged from skilled nursing facilities to determine the adequacy of their care. Who’s keeping an eye on these people? How well have they anticipated and met the needs of patients, particularly high-risk patients? Many health plans already have programs addressing this vulnerable group. This new measure simply shines a light on it and calls attention to the fact that all plans should be doing these kinds of things.  

Another new measure pertains to opioids. Organizations will be required to report the percentage of members whose opioid use lasts 15 or more days within a 30-day period, and the percentage whose use lasts at least 31 days in a 62-day period. The rationale seems fairly obvious. 

Right. Most every health plan is looking at opioid use. We are giving them an opportunity to compare themselves to others. We want to get a handle on extreme situations—people on very high doses, and those getting prescriptions from 3 or more providers or picking them up from 3 or more pharmacies. And we are trying to look at things earlier, which is why we are evaluating use 15 and 30 days out. High use in that time period puts the patient at risk for continued use. In fact, we named the measure “Risk for Continued Opioid Use.” These are patients who should be managed smartly and attentively, lest they get away from you and become addicted. With the opioid epidemic, you don’t just need to drain the tub, you also need to turn off the spigot.  This new measure aims to do that.  

How do you suggest organizations prepare for this new measure?

It’s a simple matter of tracking opioid use by prescription fills and making sure to communicate with treating teams to verify whether or not the patient really needs the prescription. Or are there alternatives?  

The other two new measures for 2019 track prenatal and adult immunization status. Isn’t that something that is likely already occurring in many places? 

I think that remains to be seen. There are many states that have immunization registries that health plans will be able to take advantage of to meet this measure. But other states are not doing much at all, or they make it difficult for health plans to get information.  

In terms of preparing to meet this measure, it seems as if a lot depends on the state or states in which a plan operates. 

It is important to keep in mind that because these measures are in our electronic clinical data source domain, they’re not required. The reporting is optional. Over the next couple of years, we will get a sense of how many plans are able to report. Those who choose to report can draw on the database tied to pharmacies, the clinical network, claims, etc. These are all legitimate sources of information about these immunization measures. 

As for 2019 changes to existing measures, you already spoke about blood pressure control. There is also a change to follow-up after an emergency department visit or hospitalization for mental illness. What will be different? 

Follow-up for mental illness is a very old measure, but follow-up after emergency department visit is just a year or two old. We saw that the codes that we were using to qualify somebody who visited the emergency department who merits follow-up were missing those at risk for suicide attempt. Under the revision, we are adding self-harm codes so that people who come to the emergency department with self-harm who presumably have either depression or some other mental illness diagnosis will be included.  

This fall NCQA is staging its first Health Care Quality Congress (see accompanying sidebar for details). One of the sessions will cover new measures in the pipeline for 2020. Can you give us a sneak preview of what will be covered? 

We will be talking about new measures in the opioid and addiction space beyond what we will introduce in 2019. We will also be introducing changes with regards to all-cause readmissions. 

Can you touch on the high points there? 

I am happy to synopsize the two most important ones related to readmissions. The first is that we plan to no longer distinguish between readmission to the observation department and a regular hospital bed. A return to either will be counted as a readmission. 

Why the change? 

In most institutions, observation beds are indistinguishable from regular hospital beds. The same team takes care of the patient, the risk for healthcare-acquired infection is essentially the same.  We don’t think there is enough distinction to merit excluding those admitted to observation. 

What about the other readmission-related change for 2020? 

The second one concerns the Medicaid population. Last year, we noticed a small number of patients with an outsize number of readmissions that our risk-adjustment model missed. We are not the only ones. AHRQ, Kaiser, and others are having a hard time determining which individuals are likely to have five or six readmissions in a year. For that reason, we are dividing our measure into the all-cause readmission rate for people who have three or fewer readmissions, and separately note the number of people who fall into have four or more readmissions.

One of the other items on the Congress agenda pertains to a new initiative, Flexible HEDIS.  Can you tell us about that? 

This is also known as Allowable Rules for Changes to HEDIS. Health plans and others have been using HEDIS measures in a lot of different ways. We want to make sure that while certain areas of the measures are changeable, some we consider sacrosanct and must stay in. Basically, when modifying a measure, organizations cannot change the numerator of the formula, but they can make adjustments for their needs to the denominator. 

Can you give an example? 

Sure. A health plan can may want to apply a measure to its Hispanic members in three regions. That is an allowable adjustment. The results are not going to be benchmarkable against other health plans unless another plan applied the measure in the exact same way. But it will be benchmarkable against the plan’s overall rate. 

The plan will still have to use its overall rate and the strict denominator criteria when they’re reporting HEDIS measures for health plan accreditation. Any adjustments the plan makes—such as the one I provided in the example above—cannot interfere with the HEDIS measure’s rigor or correctness. 

Besides these allowable adjustments, what else will plans be able to do under Flexible HEDIS? 

Organizations that opt into this will have access to expanded data sets. In the past, if a health plan wanted to map SNOMED [Systematized Nomenclature of Medicine] codes from a clinician’s health records, it had to use our supplemental data rules and map which codes corresponded to the CPT code in our code set. Now we will have an expanded group of code sets. Alongside the CPT and ICD10 codes, we will have SNOMED codes and the RxNorm codes that apply to the measure. This will allow organizations to more easily home in on a particular population.  

What else can attendees expect at the Congress?

We are going to be showing first-year analysis of all of the new measures introduced in the past year. Organizations will also be presenting results of their quality improvement projects. It will be a great opportunity to learn best practices. 

Given the political climate in Washington and the resulting ambiguity about health care delivery, particularly as it pertains to cost, what advice do you have for organizations as they seek to attain quality in the midst of such uncertainty?

We know that the payment system is going to continue to change—Washington, it seems, will always be working on that. The challenge is to deliver high-quality care at whatever price point it turns out that payers and people are willing to pay.  

With that in mind, my advice is to aim to provide all of your members the same care that you want for your mother, aunt, or other family member. 

Quality never goes out of style.