The Recent Slowdown in Health Care Spending: Explanations and Predictions

Featured in print Bulletin on Aging & Health

Health care expenditures in the U.S. have grown faster than GDP for many decades, leading to concerns about the long-term sustainability of public and private insurance programs. Yet over the past decade, the growth in health care spending has moderated, leading to much speculation about whether the era of rapidly rising costs may finally be coming to an end or whether this is merely a short-term phenomenon driven by the recent recession.

In Is This Time Different? The Slowdown in Health Care Spending (NBER Working Paper 19700), researchers Amitabh Chandra, Jonathan Holmes, and Jonathan Skinner explore the causes of the recent slowdown and ask whether it is likely to continue.

The authors begin by noting that the recent slowdown in health care spending and utilization is not the first such episode, as health care costs grew more slowly than GDP in the early 1990s before resuming more rapid growth later that decade. They also point out that the U.S. health care system is an aggregation of three different systems — private insurance, Medicare, and Medicaid. These systems differ in their means of controlling costs, and the authors argue that it is necessary to look at trends in utilization, prices, and technological development within each system in order to understand past and future spending growth.

What role has the recession played in the spending slowdown? Studies estimating the relationship between GDP growth and health care spending have yielded a wide range of results. One explanation for this, the authors suggest, is that different components of health care spending are affected differently by an economic downturn. Medicare is largely insulated from recession-related budget cuts, and the authors show that there is no empirical relationship between Medicare expenditures and GDP growth. Private health insurance spending, by contrast, is strongly associated with GDP growth. Medicaid spending is in between, as policy makers tend to cut provider reimbursements and limit covered services when state government budgets are tight, but economic downturns also make more people eligible for Medicaid coverage. A study cited by the authors suggests that there remains a roughly 1 percentage point drop in health care costs even after accounting for the effect of the weak economy.

Another leading explanation for the slowdown in spending is the Affordable Care Act (ACA). As the slowdown predates the ACA's passage by several years, the ACA is unlikely to be the only reason that cost growth has moderated, but the law may have been a factor over the past few years and its effect could potentially grow in the future. One change that is already being phased in is lower payments to Medicare Advantage plans and providers and to hospitals with poor quality indicators. However, other provisions of the ACA, such as the extension of insurance to dependents up to age 26, and the expansion of Medicaid coverage, are likely to increase costs. In the future, the law's support of Accountable Care Organizations (ACOs), which offer a new way of organizing and reimbursing health care, may also save money, but this is as yet largely untested. In sum, more time is needed before the ACA's effect on costs can be definitively determined.

Moving away from these usual suspects, the authors identify three other explanations as leading contributors to the cost slowdown. The first is higher patient cost-sharing, as more employers offer high-deductible insurance plans and the share of the working-age population that is uninsured continues to climb. The second is state-level efforts to curb Medicaid cost growth by cutting provider fees and limiting access to specialists and other high-cost services. The third is a decrease in pace of technological growth, particularly for the Medicare population. Drug spending moderated and even fell in 2012, as generic drugs claimed a larger share of the market and relatively few new blockbuster drugs were introduced. Among Medicare patients, the use of some popular treatments including heart stents, coronary bypass surgery, and hip and knee replacements declined. The authors estimate that cuts to Medicaid providers reduced the annual growth in health care spending by 0.5%, while decreased utilization due to higher cost-sharing for the privately insured and access restrictions for Medicaid patients reduced spending by a further 0.4%.

Will health care costs continue to grow at a similar rate or revert to long-term trends? The authors use three approaches to explore this question. The first is to examine what new treatments are in the technology pipeline. Historically, technological innovation has been the largest contributor to cost growth, thus future innovations will play a key role in future cost growth. While drug development has been slow in recent years, there are new treatments for heart disease and cancer that are likely to add substantially to costs in the coming years. Some, such as proton beam therapy for prostate cancer patients, have not been proven to be effective yet are very expensive; each new proton beam facility costs hundreds of millions of dollars. The authors suggest that, compared to other countries, the U.S. is more likely to adopt high cost treatments with little proven value, as Medicare is legislated to pay for any treatments that do not cause harm and private insurers often follow suit.

Second, the authors look at trends in the value of health care stocks, and conclude there is little evidence that the market believes that the rate of technological progress is slowing. Third, the authors use their own estimates of future growth in prices and utilization as well as the effect of the aging population to predict an annual real growth in health care spending of 1.2 percent above GDP growth. This is lower than the historical growth rate and relatively similar to the current rate. At this rate, the health care sector would grow from 17.9 percent of GDP today to 23 percent in 2032.

As the authors note, this estimate "is not a cause for celebration," as growth at this rate would require some combination of tax increases or drastic cuts in non-health care spending to continue to be able to fund Medicare and Medicaid. But "more optimistically, we also recognize that the structure and balance of power among providers and insurers may be undergoing fundamental changes...Similarly, accountable care organizations in Medicare and the move towards bundled payments could encourage providers to switch from expensive and unproven therapies to cheaper ones." In the end, the authors caution that any policy solutions "must be concerned about the long-term technology pipeline that will continue to deliver new technology with large price tags but with the potential for very modest health benefits."

The authors acknowledge funding from the National Institute on Aging (P01-AG19873).