top of page

Health Equity Conversations

Dana Gelb Safran

Through Health Equity Conversations, Joshua Liao hosts a series of discussions featuring people and groups around the country working to improve equity and health through systems change. 


In this episode, Josh spoke with Dana Gelb Safran, ScD, President & CEO of the National Quality Forum.

This interview is a condensed version of the conversation between Dr. Liao and Dr. Gelb Safran, and has been edited for clarity and length. For full-length discussion, please access the audio recording (available episodes accessible via Apple Podcasts and Spotify).

Joshua Liao: Thank you so much for joining us for this episode of Health Equity Conversations, we're excited to have you.

Dana Gelb Safran: I’m very excited to be here Josh. Thanks for inviting me.

JLWhat led you to your current role as president and CEO of NQF?


DGS: I’ve been in the field of quality measurement and improvement for 30 years. Roughly the first half of that 30 years, I was a measure developer. I was focused in particular on patient-reported measures, both of health

Gelb Safran.jpg

Dana Gelb Safran, ScD

President & CEO

National Quality Forum

and of care. Then the second half of that 30 years, I’ve been in a series of different roles in the private sector as more of a consumer of measures: using measures to try to drive better quality, outcomes, affordability, and equity in health care.


When this role at NQF came open, really one of the things that attracted me was that NQF had recently published a new strategic plan that envisioned the organization beginning to play in some new spaces, in addition to its original core remaining functioning around being the nation’s steward of its portfolio of quality measures; beginning to play what I call upstream and downstream from that steward of quality measures space. Upstream, to where we decide what are the measures that we need, that we don’t have, what’s the data infrastructure we need for those measures to be successful, what should be our priorities for improvement based on performance on the measures that we have today, and health equity strongly among those areas of improvement. And downstream, to how are you going to put [endorsed measures] to work, to make care better, more affordable, more equitable. Those upstream and downstream spaces are really what have animated my career and got me very excited. And that strategic plan put a large emphasis on health equity, both in our work outward facing and our work inward facing as an organization and with respect to our own employees and stakeholders.


JL: You mentioned the strategic plan at NQF, are there specific ways that the strategic plan was already being implemented when you came aboard or is being developed now to really address health equity?


DGS: There are two of the 12 goals in the strategic plan that directly address health equity, and one of those is outward facing and one of those is inward facing. Both were already very significantly in motion. We can talk over this time together about some of the specifics of the outward facing work that NQF is doing around health equity.


On the inward facing work, we already had a really ambitious and robust diversity, equity, and inclusion road map that had begun to be implemented. Now, after about a year of my tenure, at NQF, were even further down the road and very proud of that work that the organization has been doing, and frankly has been doing even before that strategic plan was implemented.


JL: How do you think about stratifying measures by important equity dimensions, and how can we implement this type of strategy in a way that accounts for multiple dimensions?


DGS: I think that it’s correct that right now, we’re in a moment where health equity finally has moved center stage in terms of a goal that every stakeholder in the health care ecosystem – whether it’s payers, providers, purchasers, patient advocacy organizations, policymakers – everybody has their eyes focused on health equity.


In 2001, when the Institute of Medicine, now the National Academy of Medicine published its seminal report on crossing the quality chasm and named health equity as one of the six pillars of a high performing health system, that was one of six pillars, all six of which were embraced. But I think most of us would say that health equity really just languished over the years from then until now –that’s 21 years as something that was an aspiration but nobody was really taking action.


And now there’s clarity on the part of all these organizations that we need to take action. But as you say, the issue of measurement then comes front and center because we have to think about how do we actually measure health equity. There are a number of ways that we can get at the measurement of health equity, but all of us agree that a place to start is with our existing measures of health care performance. Whether those are measures of patient care experience, whether those are measures of the adherence to guidelines around screening for preventive care or chronic care management, stratifying those measures to see how are we doing with respect to equity, what are the disparities that are there? And you asked, by what variables should we stratify? We should stratify by all the variables for which we think there could be inequities and we’re concerned about that. We need to stratify by race and ethnicity. We need to stratify by age. We need to stratify by gender. We probably need to stratify by sexual orientation and gender identity. We need to stratify by disability status, any of the areas: region, rurality is a very important issue with respect to inequities and disparities in care.


So one of the challenges that raises is, do we have the data that we need to stratify and actually understand how performance looks today for different population subgroups? The answer in many cases, and quite specifically the answer with respect to race and ethnicity, is no. We do not systematically have those data today. Much of our focus around health equity, and many of the stakeholders that we’re interacting with every day are hard at work thinking about to what extent can you use imputed data, or proxies, while you’re building out the information on patient self-reported race and ethnicity data.


JL: How are you thinking of ”little [dot]” versus “big” dot measures? For those who may be less familiar, what are little versus big dot measures, and as we shift to paying for value, how can this help us as we think about quality measurement and equity?


DGS: The idea of big dot measures was a concept born out of some work that I was involved at the Institute of Medicine on a committee that was developing recommendations around a core, parsimonious measure set for the U.S. to use in evaluating health care quality and its impact. And then later in some work that I was involved with, with the Healthcare Payment Learning and Action Network (LAN).


What it really refers to is, if you think about the measures that comprise NQF’s endorsed measure set today, those are measures that have largely grown up out of fee-for-service payment, and as a result of that, are largely process-oriented measures. Process measures are a natural artifact of fee-for-service payment and fee-for-service thinking: you do a thing, you bill for that thing, we measure that thing.


Value-based payment really shifts us into a mindset around outcomes. It’s outcomes that we’re really talking about when we think of big dot measures. You and listeners are probably familiar with the idea of the Triple Aim introduced by CMS [Centers for Medicare and Medicaid Services] and The Center for Medicare and Medicaid Innovation, when that was first created. That Triple Aim is about cost, quality, and patient experience.


If we could have a big dot measure for each of those, that would really be the ultimate in big dot measures. And for cost, we do. In value-based payment contracts, we have a measure of total cost of care for a population – that is a quintessential big dot measure. We don’t have anything like that for health, or for quality or experience. So, we go to the next best thing and begin to develop measures that are outcomes for different areas of care. You may think of them as procedural outcomes or outcomes for patients with particular conditions. What are the outcomes that we’re trying to achieve for patients with substance use disorder? What are the outcomes that matter to patients who have congestive heart failure?

Those become the basis for big dot measures. And then the little dot measures, the process measures, are things that don’t necessarily have to be in those value-based contracts anymore, but rather used by the organizations who accept those contracts as a way to understand the inputs to the outcomes they’re trying to achieve. They might track all those process measures on dashboards along with the outcomes to see how they’re doing. But one of the great benefits of outcome measures – well two I’ll name – one is that they help us get to more parsimonious measure sets, and that has been something that has been hotly demanded for a long time especially with value-based care contracts. When we’re talking about population health, if you’re going to measure based on process and you’re going to have accountability for everything from prenatal care to end of life care – on a process basis, that’s a lot of measures. But if you’re going to measure on big dots or the outcome measures, it can be more parsimonious.


The second thing that’s so important about it is it cannot be so prescriptive. It can allow those who are providing care to decide how are they going to achieve those outcomes, not just checking the box of a series of process measures but, delivering on the outcomes in the way that they think based on their clinical judgement, and based on their understanding of patient behaviors, and patients’ individual life circumstances, how to get there.


JL: You’ve touched on something else that I think keeps coming back in my work and conversation with others: we want to know the mechanisms for which we achieve the outcomes we want; but also clinics and organizations want flexibility, they don’t want it to be prescriptive. I believe process measures have a role and it probably looks different than how they’ve been used historically.


Do you think that in a world that has more fleshed out and rigorous big dot measures, that little dot measures then are not used in assessing performance? Or do you think even in a more value-based world, we’ll have a mix of little and big dot measures together because of this tension around flexibility but also knowing mechanism?


DGS: It's a very thoughtful question, Josh. I can share from my own experience, and I’m drawing less on my role leading NQF and more on the role that I’ve had in the private sector both on the Executive Team at Blue Cross Mass [Blue Cross Blue Shield of Massachusetts] as one of the architects of a payment model that shaped the ACO [Accountable Care Organization] movement, and my role subsequent to that at Haven on the purchaser side of the table.

What I would say there is that where we have process measures that correspond directly to an outcome, we should be retiring those process measures from our contracts and using the outcomes. We don’t need to use both. We don’t need to be prescriptive about how to get to that outcome and the process measures. If they’re still useful to those who are in the contract to monitor their progress and understand what are the most important drivers of the outcomes that they’re accountable for, that’s fantastic.


Does that mean that we will have contracts that entirely exclude process measures? I don’t think so. Because, for example, preventive services, things like colorectal cancer screening, mammography screening, vaccinations for children. These are things where if we were to measure the outcomes, they’re so far out; the things that we’re preventing are so far out that the outcome measures just don’t have a close enough correspondence with the process measures. And the process measures in many cases, including the ones that I named, are so important to the health of the population that they should remain in our view.


JL: The next thing I wanted to ask you about – and you’ve kind of addressed it in this little-big dot framework – is what to retire and what not to do. I think we as a health care community often think about what do we need to do more; we’ve been inadequate in addressing equity when we think about quality and value, so we need to do x, y or z.


What are the things you think we need to undo, or to remove, to the extent that progress requires us to stop doing certain things?


DGS: I think that this period of time with the pandemic has really showed us that we can free ourselves from the tyranny of the office visit and the idea that health care is something that happens inside of the four walls of the provider system. We’ve known for decades that that’s a very provider-centric way of delivering care. But we really didn’t start freeing ourselves from that until we had to because of COVID.


I think that continuing to do that is a very, very important part of how we have to reimagine care. And to tie it back to our topic around equity, part of why I am an advocate – a passionate champion – for value-based payment and for payment reform is that I think that as we move to more outcomes-oriented accountability, that it forces those who are providing care to think outside the box of the clinical practice setting; the literal and figurative box of the clinical practice setting, to where patients live and work. It causes us to have to think about each individual patient and what stands in the way when they walk out the door, or disconnect virtually from a televisit, what stands between them and success at keeping their blood pressure under control; avoiding the onset of diabetes if they’re prediabetic; having a good outcome for their child if they’re pregnant? What are the barriers for that human being, that individual?


And I would say in terms of things we have to undo, it’s a one-size-fits-all way of thinking about care. Thinking that if we do the same thing for everyone, that’s equity. That’s not equity. You’ve probably seen some of the wonderful visuals that show equality versus equity. Equity really requires us to think about each individual human being and what would be required for that person to get them to success on that outcome measure – to get their blood pressure under good control, their hemoglobin A1c under good control, and so forth. So, I think that’s some of what we have to undo: the tyranny of the office visit and the thinking about a one-size-fits-all as the way that we’re going to get to good outcomes.


JL: There’s a term that a prior guest has used around this idea, that we’re a “chimera” – we know a lot about processes and steps with inputs into health care, but there’s a lot that we don’t know about what patients experience, what their goals are, how to get those measures you described just now under control, but also what that means for them in their lives.


I’m curious of your reaction to this idea that maybe we need to pare down and retire certain process measures we’ve used historically to allow a proliferation of measures in the areas of experience and equity. Your thoughts on that?


DGS: I think the measures that are the most missing from our active measure sets today are measures on patient-reported outcomes. Here again I’ll differentiate, as I did in introducing myself, between patient-reported measures of health versus patient-reported measures of care. Patient-reported measures of care and care experience, at this point, have become very well integrated into our national approach to measuring health care quality. Lots of folks have a desire to get to a next generation of patient-reported measures of care.


But what we have very little of, and is so critically important, is patient-reported measures of health. These are the measures that tell us how patients are feeling and how they’re functioning, and they let us monitor how that’s changing over time. Nothing could be more profoundly valuable to us in terms of developing the evidence base behind medicine than that – tracking whether the things that we are doing are showing up in terms of what patients tell us about how they’re feeling and how they’re functioning. Did we solve their problem? Did we make it better?


We’ve had those measures for decades. They are a central part of clinical trials; when drugs are developed, when devices are developed, it’s almost always the case that you need some of those measures in order to know whether your drug is having its desired effect, or whether your device is having its desired effect. But those measures have never really found their way to day-to-day clinical practice – save for a few visionary organizations nationwide that do use patient-reported outcome measures, or sometimes called PROMS, systematically for care.


And those organizations are more patient-centered, actually know what they’re doing that’s helping patients or not helping patients, and can also use this information to understand, for a given patient and their circumstances, what treatment, what intervention is likely to help or not, and therefore to get the treatment choices right the first time, more often, and to avoid wasteful or harmful interventions.


JL: In 10 years – in the context of payment and policy, and with an eye towards equity – what will quality measurement look like? And what will have changed?


DGS: I can’t say that I know what things will look like in 10 years, but I’ll share a little bit about what I hope. And that is that we will have achieved a health care system that is delivering more equitable outcomes.


We will not have maternal mortality rates that are three-fold higher for Black women in America than for White women in America. We will not have the inequities in measure after measure after measure, by race, ethnicity, age, gender. We will have really achieved greater equity. How will we achieve that is the question.


Personally, I think that one of the really key things to getting to that place that I think is a spot on the wall that we can all agree on, is that we need to be investing in health equity, not adjusting away differences in our performance measures. What I mean by that is rather than adjusting for differences in performance based on things like race or ethnicity or social drivers of health of any sort, that we rather would want to see investment. Those provider organizations that care for populations that disproportionally have variables that make it harder to achieve good outcomes should receive either more funding in recognition that it’ll take something more and different to achieve good results for that population, and/or should receive a higher level of reward for achieving a given level of outcome. So, either a multiplier on the reward or some upfront funding that recognizes that you have to do different things in order to get the same results for certain population subgroups, or both of those things together.


But I think that by coming to grips with the fact that we don’t know all of the little dots that add up to good outcomes, and that many of those little dots are things that have to do with the environment that patients are living in, that doesn’t excuse us as care providers from attending to those things. Just because they’re outside of our four walls doesn’t mean that we’re absolved form any need to ask about them and do what we can to attend to them. But it does mean that we need to recognize that the little dots that will add up to good outcomes will be different based on individual circumstances. And that sometimes what’s required to address those little dots is going to require a different level of investment in those who are providing care to those patient subgroups.


So where I hope we’re getting to through our focus on health equity, is health equity. How we’re going to get there: I hope we’ll get there by some investment in health equity rather than by adjusting for those differences in statistical ways.


JL: Dr. Dana Gelb Safran, thank you again for joining us on today’s conversation.


DGS: Thanks so much for having me.

bottom of page