Q&A: Which US counties deliver the best—and worst—health care?

New research from IHME evaluates health care systems across 3,110 US counties to understand which are performing the best and which are not. This research uses the Triple Aim framework, an established goal for health care systems which focuses on three aspects: improving health outcomes, providing a positive patient experience, and maintaining a low cost per person.

Co-author of this study and Managing Research Scientist at IHME, Haley Lescinsky, explains the importance of these findings as a guide for health systems in US counties.

Video transcript

This transcript has been lightly edited for clarity

What is the Triple Aim that your research examined, and why is monitoring it important?

The goal of our study was to look at how well the health care system is performing. And there are many different ways that you can do that. And so, for this study, we use something called the Triple Aim framework to guide our approach. The Triple Aim framework was introduced in 2012 by the Institute for Healthcare Improvement, and it’s a widely recognized model that’s often used to set goals for health care systems.

The Triple Aim says that a strong health care system should aim to do three things simultaneously. Firstly, it should improve health outcomes. Secondly, it should provide a positive patient experience. And then, thirdly, it should keep per person health care spending low. We think that tracking progress on all three gives a pretty good picture of how the health system is performing overall. And so we built our study around those goals.

What resources did you utilize in order to carry out this research?

One of the main data sources we used in this project is called the DEX dataset, which stands for Disease Expenditure. It’s a new dataset that came out in February, and it provides county-level estimates of health care spending and utilization. It’s actually a really exciting dataset because it opens the door for a lot of new research, including analyses at the county level.

For this project, we use the DEX dataset to link county-level spending data from DEX with health outcomes data, specifically data on life expectancy, from another workstream at IHME, the US Health Disparities project. So, together, the DEX data and the US health disparities data cover the first two parts of the Triple Aim, spending and health outcomes. And then for the third part, which is patient experience, we built a composite score using a number of variables that we extracted from national survey data.

What is the difference between adjusted and unadjusted scores, and why do we use both?

One of the more important and also more complex parts of this work is the difference between the unadjusted and adjusted performance scores. So, first, the unadjusted score is just the geometric mean of the three Triple Aim measures, so the geometric mean of life expectancy, spending, and patient experience. What we found is that this unadjusted score tends to be higher in counties that already have things like better baseline health.

So those counties have lower smoking rates and lower obesity rates. And then they also generally have more privilege – so they have higher income levels and education rates. In other words, the unadjusted score reflects a lot of population-level disparities. And our goal was not just to assess that, but also to assess how effective the health system actually is in serving its unique population.

And so for that reason, we created the adjusted score, which controls for those underlying factors. And so that means that the adjusted score, which is the one that we use for our main results, tells us how well the health care system is performing given the context that it operates in.

What were the key findings of this research?

We observed a lot of variation in performance across both counties and states in the United States. When we weight populous counties more, so the bigger counties count more in the state-level values, some states really stand out in terms of having strong adjusted performance. Those include Rhode Island, Iowa, Idaho, Virginia, and Hawaii. And then on the flip side, the states that struggled the most were the District of Columbia, Alaska, Delaware, Mississippi, and Oklahoma.

What does the research point to as the optimal path to improving Triple Aim performance?

This research wasn’t designed to test causal relationships, but we did find some really interesting associations with higher-performing health systems under the Triple Aim framework. So, first, we saw that there was better performance in states that had expanded Medicaid and had more generous eligibility limits for Medicaid enrollment.

Secondly, we found that there was stronger performance in places that had more value-based care models. So, this means adoption of managed care: We saw that for both Medicare Advantage and for Medicaid managed care. And then finally, performance tended to be worse in areas that had really high market concentrations, both for insurers and for hospitals. And so that indicates that competition in the market for insurers and hospitals is better for a strong health system performance.