Unpacking the Triple Aim Model

By David A. Kindig, MD, PhD

Many thanks to those who offered some great blog ideas for 2011 – reader suggestions are always welcome. Jay Gold’s question about the Triple Aim caught my eye because it’s something I’ve been giving a lot of thought to recently: “How in your view does population health tie in with the other two elements (quality of care, cost-effectiveness) of Berwick's Triple Aim?”

This is a relevant and timely question, since the Triple Aim is getting much more attention since Don Berwick became the Administrator of the Center for Medicare and Medicaid Services. In a blog interview with John Whittington last fall, we explored the origins of Triple Aim within the Institute for Healthcare Improvement and their emerging regional strategy beyond individual health systems or plans.

I am a big fan of Triple Aim and have been privileged to present at several of their recent meetings. My appreciation for this work is based in the belief that IHI/Triple Aim is one of the leading forces in the U.S. to motivate healthcare systems to stretch beyond traditional clinical roles and work towards broad population health improvement. In fact, population health is prominently featured at the top of the Triple Aim triangle:

It’s great to see population health front and center in a model. But there’s a communication problem here that needs to be acknowledged: what does “population health” in the Triple Aim model actually represent?

My concern is that people looking at the Triple Aim model won’t necessarily know what to make of the “population health” leg of the triangle. As a result, they may focus on the other two more concrete goals of improving the experience of healthcare and reducing per capita costs of healthcare. In no way do I want to diminish the importance the model’s clinical goals, which may represent our best short-term strategy. However, the reality is that even major progress in these two areas over the next decade will not help us achieve our goals related to healthy life expectancy and disparity reduction.

To achieve our broad population health goals we need to understand and intervene across the whole spectrum of determinants, not just health care. Triple Aim leaders understand and appreciate this, but I’m not sure this is clearly communicated by the Triple Aim model so I’ve drafted an alternate view that combines the Triple Aim model with our model of population health.

As the figure shows, two of the legs of the Triple Aim stool relate only to a single determinant, healthcare. I’d like to see the population health part of the Triple Aim model more clearly defined to make it clear that population health outcomes (among populations or individuals) are influenced by multiple determinants, most of which are beyond health care (note the reddish boxes in the figure above).

I have discussed this constructive criticism openly with Triple Aim leaders and will continue to do so. They have a powerful brand which is serving us well in many regards. But when Berwick and others talk about “better health” as a shorthand for “population health” in the Triple Aim model, let’s be certain that everyone understands this means actions beyond just health care. Several Triple Aim sites have developed impressive population health agendas and are moving forward with broad and ambitious health goals. But I can’t help but wonder if a revamped model might enhance existing efforts and spur new dialogue and action in AND out of the health care system to improve health.

David A. Kindig, MD, PhD, is Emeritus Professor of Population Health Sciences and Emeritus Vice-Chancellor for Health Sciences at the University of Wisconsin School of Medicine and Public Health.

Comments

[Greg Randolph]

David, your point is right on about the murkiness of the population health component of the TA model. It is good to see you attempting to address this with TA leaders, and indeed there is an opportunity, and even obligation, for other public health leaders to shed light on the population health component and what it really means. I often hear a very health care centric understanding of population health, that is along the lines of "the population of patients we care for". Your idea to breakout the components of population health is a very good one.

[Matt Stiefel]

Hi David. I strongly agree with your point that population health requires actions well beyond healthcare. As Triple Aim measurement faculty along with Kevin Nolan, my principal focus is population health and this is a point central to all our work and presentations, and I use a driver diagram very similar to yours to illustrate that point. A prerequisite and difficult task for organizations new to the Triple Aim is to clearly define a population denominator, essential to both the population health and per capita cost aims. For many US healthcare systems, however, this is a new and challenging perspective, since many US healthcare systems take care of "patients", not populations. Defining a population, even if it starts with a small population, is the key first step toward the Triple Aim, as well as a broader conception of population health.

[David Kindig]

Great points, Greg and Matt. Thanks for joining the conversation.

[Ken House]

I've actually been quite impressed with the description of "population health" in the Triple Aim material. The message has always seemed quite clear that an effective Triple Aim initiative must look (way) beyond the medical community. Do people out there really not understand Triple Aim's definition of "population health" ...or is it something else?

[Ed Donovan, James M. Anderson Center for Health Systems Excellence]

There are possible advantages to using geographic boundaries to define the target population in efforts that use Triple Aim principles - rather than plan-based or provider based definitions. These latter definitions necessarily exclude individuals who might otherwise benefit from being included in the target population. Furthermore, geographically defined target populations, as the denominators for per capita spending estimates, provide opportunities for resource reallocation considerations that do not arise when the "population" is plan- or provider-based. Because many of the same social determinants of health are shared among different plan and provider patient populations, sharing, ie not duplicating, interventions directed toward mitigation of negative social determinants of health may increase efficiency.

[Geof Swain]

Presumably we benefit both as individuals and as populations when any of the modifiable determinants of health (MATCH model) are improved.

Thus, if the two bottom points of the Triple Aim model represent "interventions at the level of individual medical care" (improving the experience of medical care, and improving the per-capita cost of medical care), then perhaps the top point of the Triple Aim model should represent "interventions" at the population level rather than "outcomes" at the population level. Such population-level interventions would necessarily include improvements in physical environment, health behaviors, and socioeconomic determinants of health.

Alternatively, perhaps the Triple-Aim model could be replaced by a competing Quadruple-Aim Model - - something that would probably look suspiciously like the MATCH model of modifiable determinants of health [i.e.: a) physical environment, b) medical care; c) health behaviors, and d) socioeconomic factors]. If folks want to add "experience of care" and "per capita costs of care" as sub-points under the medical care portion of such a Quadruple-Aim model, I'm sure that would be fine.