What’s Driving Health Care Costs?


Value-based payment (VBP) models are an effort to rein in the growth of health care costs and improve quality. However, it’s unclear what overall impact VBP models are having on health care costs. Even though health care is provided at the local level, most evaluations examine health care spending at the national level. To address this disconnect, we conducted quantitative and qualitative market-level assessments. Our goals were to examine the impact of population-based, value-based care within a market; identify what measurable factors were associated with differing costs; and understand how business leaders are thinking about value-based care and cost reduction.

Leavitt Partners, the Healthcare Financial Management Association (HFMA), and McManis Consulting, with participation from Mark McClellan at Duke University, conducted three mixed-methods studies:

  1. Growth of Population-Based Payments Is Not Associated with a Decrease in Market-Level Cost Growth, Yet” examined the impact of population-based VBP on per-beneficiary-per-year (PBPY) health care spending and quality of care. The study used growth curve modeling and fixed-effects regression analyses of Medicare and commercial claims data.
  2. Market Factors Associated with Medicare Costs and Cost Growth” examined which market factors are correlated with PBPY health care costs and cost growth within a market using growth curve modeling. The study used and aggregated multiple data sets from public and private sources.
  3. What Is Driving Total Cost of Care? An Analysis of Factors Influencing Total Cost of Care in U.S. Health Care Markets” combined qualitative interviews conducted during site visits of nine markets and the quantitative findings from the studies above to understand factors that may be influencing total cost of care in US health care markets.

Key findings from the studies include:

  • Based on data from 2015, there was no association between an increase in population-based VBP and slowing of health care costs in a given market. Our study did not include episode-based payments.
  • Health care leaders across markets believe further changes to payment and delivery models are coming. Less clear is what, or who, will be the catalyst to push further change.
  • Some stakeholders expressed stronger support for other types of VBP models, including episode-based models and models that address the needs of specific patient groups.
  • The question of “what type of competition” in a market may be more important than “how much” competition. Lower-cost markets featured competition among a few health systems with well-aligned physician practices and geographic coverage across their market.
  • Lower-cost markets appear to benefit from organized mechanisms, including state-sponsored or endorsed reporting agencies, for more transparent sharing of information on provider quality and costs.Based on quantitative and qualitative evidence, the studies contribute to our understanding of the dynamics of competition, integration, and transparency on health care costs in a market. Below, we summarize findings from the three mixed-method studies and provide some policy implications.

Population-Based VBP Models Are Not Lowering Market-Level Health Care Costs … Yet

VBP dates back to 2005 with the Physician Group Practice Demonstration. The Affordable Care Act (ACA) significantly accelerated the proliferation of VBP models with the creation of the Medicare Shared Savings Program(MSSP) and the Center for Medicare and Medicaid Innovation, which was tasked with developing and testing innovative new models. Commercial VBP arrangements have also taken hold in the years since the ACA’s passage.

Given the growth of VBP, we wanted to examine whether, in the first few years following the ACA, these models were influencing the total cost of care. We used Medicare data from 2012 to 2015 and commercial data from 2012 to 2014 to assess the early impact of these models. We restricted our study to population-based VBPs, which included models with upside risk only (shared savings), both upside and downside risk, and global budgets, but excluded episode-based (bundled) payments.

We did not find a statistical relationship between the level of penetration of population-based VBPs in a market and a decline in health care costs for Medicare or commercial payers. Nor did we find an improvement in quality. When we limited our analysis to just those markets with higher levels of population-based VBP penetration (at least 30 percent), our results suggested a very modest, not statistically significant, market-level decrease in cost growth. Despite this null finding, our results provide an important baseline for future research.

Possible Explanations

There are several potential explanations for the null findings. For one, our study period (2012–15) may simply have been too early to see signs of population-based VBP lowering health care costs. Although today 561 MSSP accountable care organizations (ACOs) (the largest of Medicare’s ACO programs) cover 10.5 million beneficiaries, at the beginning of our study period in 2012 and 2013, only 220 MSSP ACOs covered 3.2 million beneficiaries. Many interviewees told us not enough lives were covered under VBP. Indeed, in some markets, less than 1 percent of lives were part of a VBP arrangement.

Second, although participation in population-based VBP models is growing, few models involve the provider taking on downside risk. As of 2018, the majority (82 percent) of MSSP ACOs were in the non-risk-bearing Track 1, which means they share in savings if they spend less money than their assigned benchmark, but they will not incur financial losses if they spend more than the benchmark. Our site visits found that although different markets had varying levels of population-based VBP activity, no market had significant numbers of providers participating in downside risk. Several interviewees stressed the need to take incremental steps to more risk.

Fee-for-service payment remains quite profitable for many providers and health systems. Even for those that have begun to take on risk-based contracts, fee-for-service payment represents the majority of total revenue. As long as the status quo remains lucrative, it’s difficult to make the business case for why a provider should undertake the effort to switch to a value-based focus that may lead to a reduction in use and total revenue.

Still, several interviewees said they believed the move toward paying for value would continue, even if there’s some uncertainty over whether Medicare or private payers will lead the movement. It’s possible that when VBP models outweigh fee-for-service payments in a market, we’ll reach a “tipping point” and health care cost growth will decline. Many interviewees expressed enthusiasm for other VBP models, such as those based on episodes of care (bundled payments) and those designed for specific populations (for example, the frail elderly). These models may make more sense for specialty providers who perform a certain type of procedure or care for a certain type of patient.

Other Market Factors

If these initial population-based VBPs results don’t show a relationship to health care cost growth, then which market-level factors do correlate? For our second quantitative analysis, we used a variety of public and private data sources to examine the relationship among several market-level factors beyond value-based payment and Medicare costs and cost growth between 2007 and 2015. All the factors together explained 82 percent of variation in baseline Medicare costs (Exhibit 1). 

The prevalence of chronic diseases was the most influential predictor of market costs, accounting for 41.5 percent of the variance. Hospital quality metrics, market socioeconomic status, and the concentration of hospitals and insurers also helped explain market-level costs.

Using these same factors to predict Medicare cost growth was less fruitful, explaining only 27 percent of the variation in Medicare cost growth—substantially less than the 82 percent of baseline costs. As Exhibit 2 shows, a much weaker association exists between chronic disease prevalence and Medicare cost growth. Significant additional research should be done to identify factors that predict cost growth.

These findings matter for several reasons. First, they reinforce efforts currently underway to contain costs, including strategies to prevent and better manage chronic conditions, reduce hospital readmissions, and reduce the number of individuals without insurance. Second, although we know less about what drives health care cost growth in a market, meaningfully reducing spending in a market relies on developing strategies that target cost growth, instead of baseline costs. More research that focuses on what’s driving cost growth is needed.

The Role Of Competition And Transparency On Costs

The interviews we conducted add insights into these market-level findings. We identified two distinguishing characteristics of higher- and lower-cost markets: type of competition in the market and degree of transparency in the market. We recognize that while there are some common lessons, health care markets differ significantly and their approaches to care, costs, and VBP models will vary.


We know competition can help drive down costs and increase quality in health care markets. However, how much competition, and what type, seems to make a difference. For example, we found that the lower-cost markets in our nine site visits had at least one integrated delivery system. Consolidation in these markets had resulted in two to four health systems with geographic coverage across the market. In these markets, physicians were generally employed by the health system or worked in close alignment with it. Health plan competition matters as well, particularly with respect to innovation in new payment and care delivery models. Portland, Oregon, and Minneapolis-St. Paul, Minnesota, two of the lowest-cost markets, both had competitive health plan landscapes.

Conversely, the markets we visited with less integration and seemingly more provider competition actually had higher costs. These included Los Angeles, California (which had higher Medicare costs only), Baton Rouge, Louisiana, and Oklahoma City, Oklahoma. One reason for this may be that there is less focus on addressing unnecessary use in these markets.


Transparency is often cited as a strategy that will help contain costs. Similar to competition, the type of transparency effort matters. We found that some lower-cost markets seemed to benefit from organized transparency mechanisms, including state-sponsored or endorsed reporting agencies and employer coalitions that made information on provider quality and costs publicly available. For example, in 2005, the Minnesota Medical Association and health plans in the state together formed MN Community Measure, a nonprofit organization tasked with the collection and dissemination of data on the quality and cost of providers across the state. Today, providers are required to submit data to the organization. Our interviewees expressed optimism but acknowledged more work is needed to optimize consumer-oriented transparency tools, which research has so far shown to have had only minimal use.

Policy Recommendations

Our research led us to three primary policy recommendations to help improve health care quality and lower costs (for additional ones, see the fullstudies).

  1. Continue movement toward payment models that increase financial incentives to manage total cost of care and closely monitor the impact of doing so because our findings show that the majority of payments in a market continue to flow through fee-for-service, instead of value-based arrangements. Experiments should continue with population-based VBP models but should not be confined exclusively to these models. Episode-based payment models, for example, may be better suited to certain types of providers who perform a certain procedure (for example, a knee replacement) instead of care for a general population of patients.
  2. Balance the benefits of competition with the benefits of integration. The lower-cost markets we studied had competition among two and four systems with well-organized provider networks that had been developed through vertical integration or strong alignment of physician practices. Most of the lower-cost markets also had an integrated delivery system—with vertically integrated health plan, hospital, and physician capabilities—as a competitor in the market.
  3. Support more transparent sharing of information on health care cost and quality within markets. Lower-cost markets in the qualitative study had organized mechanisms for the sharing of information on health care cost and quality, whether through employer coalitions, statewide reporting agencies, or both.

Although differences exist among each health care market, all markets can act to improve quality and reduce costs. Our studies suggest several actions different stakeholders in each market can take to improve care for their populations.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.