This brief is the result of a collaboration between the Kaiser Family Foundation and the United States Bureau of Economic Analysis. Cynthia Cox is with the Kaiser Family Foundation; Abe Dunn, Lindsey Rittmueller, and Bryn Whitmire are with the Bureau of Economic Analysis.
Recently released health spending estimates from the Bureau of Economic Analysis (BEA) offer a new perspective on national health spending. These data, which show spending on major disease categories, can be used to gain insight into recent trends such as the health spending slowdown, how spending has changed since the passage of the Affordable Care Act (ACA), and some of the driving forces behind these trends. Researchers can now look at how health spending has grown across each disease category (such as cancers and circulatory diseases), and see whether growth was primarily driven by changes in prices (the cost of treating the disease) or the number of treated cases (treated prevalence).There was a significant increase in use of preventive services corresponding to the ACA provision Click To Tweet
In this brief, we use disease-based spending estimates to look specifically at how spending changed from 2010 to 2012 — the years following the economic downturn and the passage of the ACA — and we contrast these years with the 2005 – 2010 period. Starting in the mid-2000s, health spending began to grow at low levels, sparking researchers to investigate the causes. Several studies have described this slowdown as a result of the Great Recession – with lagged effects playing out over a number of years – in combination with structural changes in the health system. In a recent Health Affairs article, the slowdown in health care spending was studied by looking at disease-based trends over the 2005 – 2010 period and contrasting this with spending changes from 2000 – 2005. The updated disease-based spending estimates from BEA allow us to continue this analysis through 2012 and also examine spending changes that may have resulted from the ACA and the so-called “patent cliff,” when a number of commonly used drugs lost patent protection.
Overall, the findings in this brief show slower growth in the cost of treating diseases (price) and a simultaneous rebounding of growth in the number of treated cases in 2011 and 2012, relative to the slowdown period. We find that the Great Recession, the ACA, and the patent cliff are all influential factors in 2011 and 2012 spending trends. Of particular note, we find a significant increase in the use of preventive services in 2011 and 2012, which corresponds to the timing of the ACA’s preventive services provision that health plans cover preventive services without cost sharing.
Trends in Spending, Prices, and Treated Prevalence using Disease-Based Measures
The primary aim of the health system is to improve health through the treatment and prevention of illness and injuries, but, until recently, there has been limited information about how much we spend in the U.S. to treat various diseases. Using the official National Health Expenditure Account, for example, researchers can examine spending on various types of medical services and products (such as physician visits or prescription drugs), but the data do not allow users to view trends in spending for certain disease categories like cancers or circulatory conditions.
In January 2015, the BEA published its Health Care Satellite Account (HCSA), an experimental approach to measuring health spending, prices, and treated prevalence by disease category. The satellite account offers a new perspective on national spending by redefining the output for the health care sector. Whereas the traditional approach to measuring spending on health care treats the output of the health system as types of services (such as hospital stays), the satellite account defines the output as treatment of medical conditions (for example, circulatory conditions or cancers).
The satellite account can also be used to track changes in prices and number of treated cases of health care. In this brief, we refer to the cost of treating a particular disease as the “price,” but we note that the disease-based price index differs from the traditional concept of price indexes. While the traditional price index measures the price of specific services (such as the price of a physician office visit), the satellite account tracks the cost of disease treatment (for example, the average cost of treating a case of colon cancer). Price indexes in the HCSA, therefore, differ from official price indexes in that they are not only influenced by the price of a given treatment, but also by greater treatment intensity per visit, the number of visits and procedures, and shifts from lower-cost to higher-cost treatments.
Using the HCSA, we find that health spending growth in 2011 and 2012 was comparable to the average growth rate during the slowdown period (2005 – 2010), hovering around 4 percent per capita, but that cost per case (the disease-based price index) grew at even slower rates than had been the case during the slowdown period. In fact, 2011 and 2012 saw the slowest annual growth rates in cost per case (at 2.3 and 2.1 percentage points, respectively) since the beginning of the HCSA time series in 2000.
The chart below takes a look at growth in per capita spending, prices for disease treatment, and number of treated cases from 2000 to 2012. The difference between growths in per capita spending and disease-based price can be attributed to growth in treated prevalence per capita. Following the Great Recession, there appears to have been a rebound in the number of treated cases.
In 2011 and 2012, the number of treated cases rebounded and the cost per case grew at its lowest levels since 2000
The relatively stable growth in per capita spending in 2011 and 2012 compared to the slowdown period (2005-2010) may be viewed as the result of the combined effect of a slowing of growth in cost per case offset by a rise in the number of treated cases. In other words, growth in health spending per person appears to have leveled off in 2011 and 2012, but during this period prices grew more slowly, while the number of treated cases picked up.
Recent Trends in Spending by Disease Category
Additional patterns emerge when looking at spending on specific disease categories. The chart below shows average annual per capita spending growth rates in 2010 – 2012 compared to those seen in 2005 – 2010 across all condition categories, as well as the top seven spending categories. In the aggregate, annual growth in per capita spending slowed by only 0.5 percentage points to 3.9 percent annually between 2010 and 2012.
Cost growth varied by disease both during and after the initial health spending slowdown years
Turning to the disease-based trends, though, we find variation in the magnitude of growth rates. The routine care category, which mainly consists of preventive care services, maintained a high per capita spending growth rate of 6.1 percent annually. Meanwhile, growth for the circulatory conditions category continued to slow in 2010-2012. In 2012, circulatory conditions were overtaken by routine care as the largest overall spending category; the first time since at least 2000 that circulatory conditions had not ranked first in overall spending at the disease level.
With the exception of routine care and respiratory conditions, the remaining top spending categories saw slower per capita spending growth from 2010 – 2012 as opposed to the slowdown years, with differences ranging anywhere from 1.1 percentage points for circulatory conditions to 2.8 percentage points for neoplasms (cancers and tumors).
Factors of Spending and Price Growth
As previously alluded to, one of the benefits of using disease-based statistics is the ability to decompose per capita spending growth into contributions from disease-based prices (costs per case) and treated prevalence. Using the new HCSA statistics, we are able to utilize this decomposition at various levels in order to analyze spending trends for 2011 and 2012. In the remainder of this brief, three factors are highlighted which may have influenced the trends during this period: the Great Recession, the Affordable Care Act (ACA) and the patent cliff.
The recent extension of the HCSA’s time series to include data through 2012 allows for the break out of spending during the recessionary period (2008 – 2010) and the post-recessionary period (2010 – 2012).
Previous studies have found that the overall economy may have a lingering effect on health care spending of roughly 5 or more years. While there were certainly some immediate effects of the Great Recession observed from 2008-2010 (when treated prevalence growth slowed to zero), we found evidence of this lagged effect in the decomposition of per capita spending (also seen in disease-based prices).
Following the Great Recession, there was a rebound in treated prevalence and cost per case grew more slowly
Across all disease categories (in aggregate), per capita spending for health care grew slower during the recession than it had during the pre-recession period. During the initial years following the recession (2010 – 2012), per capita spending growth remained low and similar to the level during the recession period.
Spending, price index, and treated prevalence, by disease category
Per Capita Spending
Annual growth in per capita spending
Annual growth in disease-based price index
Annual growth in treated prevalence
|Routine care, signs and symptoms**||$788||6.4%||5.5%||6.1%||4.5%||4.3%||2.1%||1.9%||1.2%||4.0%|
|Endocrine system conditions||$440||6.1%||3.1%||3.2%||2.8%||1.0%||1.0%||3.2%||2.1%||2.2%|
|Nervous system conditions||$424||7.2%||5.6%||4.4%||5.1%||5.4%||3.0%||2.0%||0.2%||1.3%|
|Injury and poisoning||$375||5.8%||2.7%||2.8%||6.2%||4.0%||1.7%||-0.3%||-1.2%||1.1%|
|Source: Authors’ analysis of data from the Blended Account of the Bureau of Economic Analysis (BEA), which combines data from the Medical Expenditure Panel Survey and large claims databases. See appendix for more information on disease categories.
*Corresponds to the BEA’s “medical services by disease.”
**Corresponds to the BEA’s “symptoms,” which include preventive care, allergies, and flu-like symptoms.
***Blood diseases, perinatal conditions, congenital anomalies, and unclassified diagnoses.
Treated prevalence slowed down to zero annual growth during 2008-2010 then rebounded back to 1.6 percent growth post-recession. During the Great Recession, drops in employment resulted in people shifting from private insurance to either Medicaid or becoming uninsured, both of which are correlated with a reduction in the use of health care services. The rebound in treated prevalence that occurred post-recession in the aggregate was reflected across all condition categories, but there was variation in the magnitude of growth. For example, the number of patients seeking treatment for routine care nearly tripled compared to the recessionary period, accelerating by 2.8 percentage points to 4.0 percentage points in 2011 and 2012; meanwhile treatment prevalence of endocrine system conditions accelerated by just 0.1 percentage points.
Conversely, growth in cost per case during the recession period maintained similar rates as those pre-recession. It was not until the post-recession, 2010-2012 years, that a slowdown in cost per case from 3.7 to 2.2 percent annual growth was noticed. This trend extended to the disease category level (where prices slowed across the board and were 2.4 percentage points slower on average than the 2008-2010 period). Although economy wide inflation (as measured by BEA’s personal consumption expenditure (PCE) index) was relatively constant between 2005-2010 (2.0 percent) and 2010-2012 (2.2 percent), there was a noticeable slowing of the growth rate in the underlying service prices for the health care sector between these two periods from 2.9 percent to 2.0 percent (as measured by the PCE health by function index). That is, of the 1.5 percent slowdown in cost per case growth, about 0.9 percent is attributable to lower service prices (e.g., lower negotiated rates between insurers and providers).
Overall, the rebound in treated prevalence was completely offset by the slowdown in cost per case growth in 2011 and 2012, and led to aggregate per capita growth that remained relatively flat post-recession compared to growth during the Great Recession.
Affordable Care Act
While many of the key provisions of the ACA did not take effect until 2014, several were implemented as early as 2010. One, for example, being that Medicare reduced its provider reimbursements to physicians and hospitals in both 2011 and 2012, which aided in keeping costs per case contained across all illnesses and injuries.
At the disease-level, we began to see evidence of some effects of the ACA specifically in the routine care category. A provision of the ACA expanded coverage of preventive care services beginning in September 2010 for the privately insured and January 2011 for those insured by Medicare. Not only did the provision mandate coverage of a list of preventive care services, but these services became “free” at the point of service. That is, under this provision, no out-of-pocket costs were charged to the consumers.
Many of the preventive care treatments (such as routine medical exams, laboratory exams, and pap smears) are classified under the more disaggregate Clinical Classification Software (CCS) condition category, “exam or evaluation” condition, which is within the routine care category. Trends for this disaggregate CCS condition category can be seen in the chart below. As expected, there was a sharp uptick in the number of treated cases from 2010 to 2011 coinciding with the implementation of free preventive care services and continuing through 2012. Meanwhile the growth in cost per case remained relatively flat.
A sharp uptick in use of preventive care coincided with the Affordable Care Act’s preventive services provision
Between 2010 and 2012, there were a large number of brand-name drugs whose patents were set to expire, a phenomenon now known as the patent cliff. Also during this time an influx of available generic versions came to market. This allowed for patients to substitute away from costly brand-name prescriptions to more affordable generic versions and, therefore, helped to hold down growth in cost per case.
The chart below displays growth rates for two CCS condition categories within the mental illness category that were affected by generic substitution for brand-name prescriptions. Depression, which is classified under the CCS condition category mood disorders, had a pharmaceutical treatment called Effexor XR come off patent protection in July 2010. As can be seen in the chart, the growth in cost per case dropped to -3.1 percent annually while there was a rise in treated prevalence to 3.6 percent annually. This indicates that the cost to treat mood disorders fell, while simultaneously more people were being treated. A similar though less dramatic example of the effect of the patent cliff can be seen in the case of Adderall, a drug used to treat attention deficit and hyperactivity disorders (ADD/ADHD) that lost patent protection in May 2009. After Adderall came off patent, the cost to treat ADD/ADHD grew slower than it had in previous years, while treated prevalence grew by an average of 10.5% in 2010 – 2012 (compared to 5.0% annually in the previous five year period).
Price growth fell and treated prevalence increased for mood disorders after introduction of generic drugs
In addition to those prescription drugs used to treat mental illness, there were several other notable brand-name drugs coming off patent protection during this time that influenced low price growth in specific disease categories while consumers switched to generic alternatives. Within the circulatory conditions category, Cozaar and Hyzaar (angiotensin receptor blockers used to treat high blood pressure) lost patent protection in April 2010. Plavix, a drug used to help prevent circulatory conditions and typically given to patients who have suffered a heart attack, lost protection in May 2012. In other cases, Lipitor, a treatment for high cholesterol (November 2011), as well as Singulair, a popular treatment for asthma (August 2012), lost patent protection affecting the growth rates within the endocrine system and respiratory condition categories, respectively.
The BEA’s Health Care Satellite Account allows data users to examine trends in national health spending by disease, offering a new perspective of recent spending trends. In particular, the 2011 and 2012 updated estimates allow for the comparison of spending during the years following the recession and passage of the ACA to that of prior years. On the whole, these data show that spending continued to grow at similarly slow rates during 2011 and 2012 as had been the case during the 2005 – 2010 slowdown period. But while aggregate per capita spending growth was relatively constant, more subtle trends could be found when looking at the decomposed HCSA data. Specifically, underlying the per capita growth was a simultaneous slowdown in cost per case growth and an increase in the number of treated cases.
While many of these aggregate trends were reflected in trends at the disease category level, we also found some interesting disease-specific trends related to the Great Recession, the ACA and the patent cliff. Our analysis of this new spending data reaffirms a couple of findings from previous studies of recent health spending trends. As has been described in previous studies, we find evidence that the economic downturn played a role in the health spending slowdown and that it continued to play a role in health spending growth during the post-recession period. Following a slowdown in treated prevalence to zero annual growth during the recession years, there was a rebound in disease treatment during the post-recession period that can be seen across each major disease category. However, the magnitude of the uptick in treated prevalence varied quite a bit by disease, suggesting that disease-specific mechanisms and structural factors were also at play.
Additionally, our analysis reaffirms prior findings that the patent cliff played a role in health spending from 2010 – 2012. The HCSA’s disease-based spending estimates allow us to examine prices and treated prevalence for conditions that were affected by commonly used drugs going off patent. Across a number of these conditions, we found that the loss of patent protection corresponded to a lower growth in the price of treatment.
Lastly, using this new data, we find that the ACA’s preventive services provision likely had an upward effect on the number of patients treated for preventive care. Although most of the coverage expansions of the ACA did not take effect until 2014, the provision requiring that most plans cover preventive services without cost sharing went into effect in late 2010 and early 2011. Our analysis finds that in 2011 and 2012, coinciding with the implementation of the preventive services provision of the law, there was a sharp uptick in of the number of patients receiving routine exams and evaluations, many of which are services mandated to be covered without cost-sharing under the ACA. At the same time, prices held relatively steady for these services. Our use of the HCSA allows – for the first time following implementation of the ACA – examination of national trends in spending, price, and number of treated cases by condition category. A significant advantage of the HCSA is the ability to examine prices and number of treated cases of specific conditions categories – such as routine care, mood disorders and ADD/ADHD – that are the focus of policy changes and other developments.