We are increasingly aware of the tremendous geographical variation in health care—in utilization, prices, and the growing role of local market power that arises from consolidation. The idea behind the Healthy Marketplace Index, supported by the Robert Wood Johnson Foundation (RWJF), was to develop a concise way to compare health care markets in different places, somewhat in the spirit of the County Health Rankings and Roadmaps program, also funded by the RWJF. The goal was to come up with timely measures that reflected important attributes of health care markets—such as prices, efficiency, and provider consolidation.
The thought was that this would be interesting to employers and other stakeholders, such as local civic and political leaders, and would hopefully contribute to efforts to improve competitiveness in local health care markets. We knew when embarking on this project that employers consider population health and health care costs to be very important aspects when deciding where to locate their businesses. The results demonstrated how complex an endeavor managing health and costs can be because of the complexities of health care markets.
The danger inherent in any project that attempts to create an index or ranking system is that there is a tendency to sacrifice accuracy for simplicity, resulting in measures that either minimize differences or overemphasize them, thereby distorting true comparisons in the name of creating something that is easy to comprehend or perhaps grabs headlines.
Our partnership with the Health Care Cost Institute (HCCI) was a good fit because, in addition to its access to commercial health insurance claims data, it had both the willingness and the ability to put a process in place that would result in a credible set of measures.
The RWJF’s initial planning grant was very modest and included support for creation of a technical advisory committee, comprised of representatives from academia, government, and the private sector, but definitely with the heaviest representation from academia. Unlike many technical advisory committees, which often play a largely symbolic role, this group worked extremely hard for a period of almost two years at creating the methodology for the Healthy Marketplace Index.
The challenges were many. One of the earliest was choosing the appropriate geographic level. States were too large, counties too small, and neither really met the definition of a health care market. It was decided to base the index on Core Business Statistical Areas (CBSAs), which comprise an urban area and surrounding suburbs.
A second challenge concerned adequacy of data. The main data source for the index was HCCI’s database of commercial insurance claims, which captures roughly one-quarter of those in the United States with employer-sponsored insurance. Yet despite the massiveness of this data set, there was not sufficient penetration in all markets. The Healthy Marketplace Index project this year included forty-one CBSAs where at least 25 percent of the commercially insured (nonelderly adult) population was captured in the HCCI data. As the HCCI holdings grow, hopefully future versions of the Healthy Marketplace Index will include more CBSAs, and it will be possible to analyze subpopulations with specific conditions.
The Healthy Marketplace Index measures price, utilization, resource use, and competition among providers. In the case of price and utilization, separate inpatient and outpatient measures were created. The price index was based on a “marketbasket” of “allowed amounts,” or negotiated prices of services from the HCCI claims data. The utilization measure captures use of services, adjusted by price. The attempt to measure resource use or productivity—that is, a way of relating use of health care services to population health—was perhaps the most challenging. The project measured health using two different measures, one from the RWJF County Health Rankings data, and the other from diagnosis codes in the claims data. Then, a ratio of resource use to health was constructed from the claims data.
The competition measure was based on consolidation in hospital markets, in part because of a lack of comprehensive data on outpatient facilities. We would like to be able to measure competition in markets for outpatient care in a future version of the index.
What did we learn? Health care markets are complicated. One of the challenges of reporting the results was that there are no simple overarching takeaways, nor is there the “shock and awe” effect that we have seen in the many mainstream exposés on unexplained variation in health care.
The distribution of high and low prices did not fit any particular geographic pattern, and there were markets where inpatient prices were high and outpatient prices were low, and also markets where the reverse was true. Further, we found no correlation between utilization and prices, which may reflect the fact that benefit designs may vary geographically in ways that counteract underlying differences in negotiated prices. Another measure looked at the relationship between resource use and population health and found examples of markets that seemed to under- or overconsume health resources relative to need.
While it is possible to identify better- and worse-performing markets from these data, it was quickly evident that there was no way to “rank” health markets on a single scale, as performance in the different domains did not fit together neatly. As project director Eric Barrette observed, “A big challenge was producing metrics that can be understood by a broad audience and were usable but were also accurate. We tried to avoid any methodology that was too technical or required a series of assumptions to interpret. I think we succeeded with this for some of the measures, e.g., the price index, but for some of the other measures we will probably need to make revisions over time.”
The Healthy Marketplace Index measures on price, utilization, and competition were easiest to understand, while the productivity or resource use measures were the most challenging.
Future plans include refining some measures, such as the productivity index, and hopefully increasing both its geographic scale and timeliness. Larger sample sizes will allow us to address more specific questions related to subpopulations. Being able to relate these measures to changes in benefit design and payment method would also be extremely interesting—to see how these changes in the way that providers are paid are reflected in care delivered and in prices.
So far, most of the interest in the Healthy Marketplace Index project has come from journalists and the research community. There are many research questions that can be addressed with these data, but we also hope that future versions of the Healthy Marketplace Index will attract a broader range of interest from employers and local stakeholders, who will use the measures to assist them in policy development and contract negotiations that will help their health care market deliver higher-value care.
