What Accounts For The Lower Growth In Health Care Spending?


“This post is from Healthaffairs.org. It is an easy to understand summary of the lowering of the growth of health care spending comparing 2000-2005 with 2005-2010.” Bill Chesnut, MD

What Accounts For The Lower Growth In Health Care Spending?  Kenneth Thorpe_February 18, 2016


Several recent studies have examined the factors accounting for the recent slowdown in growth in real per capita health care spending. The most recent entrant to this literature is the newly published work ofDunn, Rittmueller, and Whitmire. Earlier work has identified two large categories of change as contributors to slower growth: the 2007-2009 recession, and other structural changes in the delivery and payment of health care. Understanding the relative role of these two factors is important in determining whether the slowdown is temporary or will be sustained.

The new paper by Dunn and colleagues builds on a framework that I developed with my colleagues at Emory in 2004-2005, that decomposes the growth in spending to changes in the prevalence of treated disease, spending per case treated, and population growth by medical condition. We have published several updates to this initial paper over the past ten years.

We found that the roles played by changes in treated prevalence and cost per case since the mid-1980s differ by time period and by payer. Rising treated prevalence has consistently been the major driver of rising Medicare spending over time, while prevalence plays a smaller role in growth in private health insurance spending. Understanding the factors that account for these variations over time may provide useful information in projecting trends in health care spending.

Dunn and colleagues examine data from 2000-2005 and 2005-2010 using the aforementioned decomposition. They conclude that the spending slowdown observed between 2005 and 2010 was traced largely to lower growth in cost per case. The reductions differed, however, by medical condition. One of their conclusions is that reduced growth in cost per case is linked, in part, to blockbuster drugs, such as Norvasc and Toprol XL to treat hypertension, which came off patent during the 2005 to 2010 time period.

Indeed, our tabulations show that nominal spending on hypertension during this period actually declined by $2 billion. Patent expirations have assumed an even more important role more recently, as $84 billion worth of branded drugs came off patent between 2010 and 2013 alone. Tracking trends in drugs scheduled to come off patent will provide important information regarding future spending trends.

Several other factors could also underlie the slowdown in the growth in treated prevalence and spending per case treated going forward. On the spending side, delivery system reforms and the migration away from fee-for-service payment are structural changes that will continue to put downward pressure on growth in spending per case. These reforms include bundled payments that provide incentives for more efficient placement of patients in post-acute care settings and reductions in readmissions and the costs of the initial hospitalization. Understanding the role of these shifts in the site of treatment will be helpful in projecting future spending trends.

Several other factors are likely to affect trends in treated prevalence. Most important may the recent slowdown in the growth in obesity rates, which will affect both the incidence and prevalence of several chronic conditions. There is already some evidence that the incidence and prevalence of diabetes has leveled off between 2008 and 2012. Tracking the role of changes in chronic disease incidence and prevalence overall and by payer will provide useful information in determining whether the slower growth in spending is short lived or may continue.

Adding these epidemiologic prevalence trends to the more traditional national health account projections methodology may enable more precise health care spending projections overall and by source of treatment.