«Abstract Advances in medical technology have long been believed to be a major driver of health care costs, but surprisingly little is known about the ...»
Estimates of Medical Device Spending in the United States
Roland “Guy” King, F.S.A., M.A.A.A.
Gerald F. Donahoe1
Advances in medical technology have long been believed to be a major driver of health
care costs, but surprisingly little is known about the aggregate value and prices of one of
the major types of medical technology, medical devices and diagnostics. This data note
analyzes spending on medical devices for 1989-2004. Prices for medical devices
consistently grew more slowly than either the Medical Consumer Price Index or the Consumer Price Index as a whole.
Background The role of medical technology in health care costs has long been a source of debate. It has been widely asserted that healthcare technology is a major driver of health cost increases.1 Other findings have suggested that returns on spending on medical technologies can far exceed their costs, particularly when longer term benefits are measured in terms of productivity and reduced disability.2 Literature on the systemic effects of advances in medical technology generally uses the term in a broad sense to include such factors as: (1) the development and use of new and improved medical devices and diagnostics; (2) development of new medical procedures;
(3) improvements in existing procedures; (4) increases in the number of procedures performed because of increased safety, effectiveness, or convenience, (5) development of new pharmaceutical products. Surprisingly little research has been done on the separate impact of medical devices, and there is virtually no published analysis on basic questions about medical spending on the devices themselves and on the role of price versus volume change.3 In part, this is because devices are rarely purchased separately from hospital, other institutional, or physician services and therefore do not appear as a separate line item in the national health spending accounts.
1 Roland “Guy” King is an independent Consulting Actuary and an industry leader in public sector health financing programs. Prior to 1995, he served as Chief Actuary for the Health Care Financing Administration (HCFA) for 16 years. Gerald F. Donahoe is a retired economist who worked as a statistician at the Bureau of the Census from 1962-1971 and with the Bureau of Economic Analysis, U.S. Department of Commerce from 1971-1997. The research for this paper was supported by AdvaMed.
1 This data note attempts to fill this gap by analyzing expenditures on medical devices as a share of overall national health expenditures and of growth in medical costs. In addition, our study analyzes the growth in the prices of medical devices and separates out price effects from other factors contributing to the growth in spending on medical devices and diagnostics.
Findings In this study, we estimated the cost of medical devices and diagnostics and the contribution of these products to increases in national health expenditures. We attempted to use the same methodological rigor in estimating medical device spending as is used by CMS (The Centers for Medicare and Medicaid Services) in compiling estimates of the major categories of national health care spending. The findings were interesting and, to
some degree unexpected:
• Medical devices are a relatively small share of total national health expenditures
• Over the fifteen year period studied, medical devices remained fairly constant as a share of national health expenditures
• Prices for medical devices grew more slowly than either the CPI for medical services or the CPI overall.
Discussion As described in the methodological appendix, we began our analysis by selecting categories from the North American Industry Coding System (NAICS). We then measured expenditures as manufacturers’ shipments plus imports minus exports and added margins for wholesale and retail trade, using Economic Census data and annual survey data. Price changes were measured using appropriate Producer Price Indexes and margin rates and incorporating a Fisher Index formula.
Our first finding was that medical devices make up a relatively small and constant share of national health expenditures. In 2004, the latest year studied, spending on medical devices and in vitro diagnostics totaled $111.7 billion, or 6.0% of total national health expenditures (Figure 1).2 2 See Table 1 in the Appendix on page 21 for the data underlying each of the following figures (1-4).
100,000 80,000 60,000 40,000 20,000 0 19 9 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 04 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 19 Throughout the fifteen-year study period (1989-2004), spending on devices remained relatively constant as a share of total national health expenditures. While spending as a share of expenditures did vary by year, it never fell below 5.4% and never rose above 6.1% (Figure 2).
5.5 5.0 4.5 4.0 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20
8.2 8.0 8.1 7.8 7.6 7.8 7.4 7.4 7.4 7.2
While medical device spending has grown at about the same rate as national health expenditures overall, prices for medical devices have actually grown far more slowly than the Medical Consumer Price Index or even the overall Consumer Price Index (Figure 4). During the fifteen-year study period, medical device prices have increased at an average annual rate of only 1.2%, compared to 5.0% for the MCPI and 2.8% for the CPI. This relatively slow rate of price increase suggests that the industry is highly competitive.
During much of this fifteen year period, a significant driver of changed medical practice has been the development of new medical devices from stents to implantable defibrillators to artificial hips and knees to new imaging modalities to new diagnostic tests to new surgical tools. In view of the conventional wisdom about the role of medical technology in driving up costs, it is surprising that the cost of medical devices has remained constant as a share of total national health expenditures. It is also striking that,
8.0 6.0 5.0 4.4 4.0 2.7 2.0 0.5 0.0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
5 To further determine the “medical boundary” we used manufacturing categories in NAICS (the North American Industry Coding System) because the data from which the estimates were developed are from the federal government statistical system, and that system is currently based on NAICS for industry data. The medical boundary narrows the economic activity universe to the eight categories shown below with their NAICS codes.
334510—Electromedical and electrotherapeutic apparatus 334517—Irradiation apparatus 339111—Laboratory apparatus and furniture* 339112—Surgical and medical instruments 339113—Surgical apparatus and supplies 339114—Dental equipment and supplies* 339115—Ophthalmic goods 339116—Dental laboratories* * These categories are not included in the study as discussed below.
Devices such as computers and autos that are used by the health services industry as well as by many other industries were not included.
Dental equipment and supplies (NAICS 339114) and dental laboratories (NAICS339116) were excluded, either because complete corresponding data were unavailable for all elements of the analysis (in the case of dental laboratories), or because dental care and related expenses are typically financed through different healthcare insurance mechanisms than the other products considered in the analysis.
We decided to drop Laboratory apparatus and furniture (NAICS 339111) because the apparatus portion was largely non-medical and no data were available to allocate the total. In 2002 there were over 700 thousand medical establishments in the United States, but most were offices of doctors and other practitioners, and these offices generally did not contain labs. Only 11 thousand of the 700 thousand establishments were medical and diagnostic labs, but there were 23 thousand food-processing establishments, many with quality assurance labs, and about four thousand institutions of higher learning, many of which have labs.
We believe that some types of hospital furniture should be classified as medical devices, to the extent they are regulated by the FDA. Operating room furniture and hospital beds 6 appear to fit both the dictionary and regulatory definitions of medical devices.
Unfortunately, lack of separate data prevented us from including this category.
Shipments of hospital beds are available for the entire period covered by the estimates, but separate codes are not available for imports and exports.
We further narrowed the scope by dissecting some of the remaining categories into medical and non-medical components, and where possible (with sufficient statistical accuracy), removed the non-medical portion to improve the estimates. The categories removed were: irradiation equipment used for non-medical uses; personal industrial safety devices and protective clothing (from Surgical appliances and supplies); and antiglare glasses and related goods (such as non-prescription reading glasses) from Ophthalmic goods.
Finally, we had intended to eliminate some double counting in the manufacturers’ shipments data caused by recording a shipment when shipped by a parts manufacturer and then recording the value a second time when embodied in the shipment of an assembled device. Unfortunately, sufficient data were not available to implement this plan. For example, Census data indicate that for 2004, X-ray tubes valued at $199 million were sold separately. But some of these sales (perhaps most of them) may be used as replacement tubes for existing machinery. Also, the Input-Output Tables prepared by the Bureau of Economic Analysis (BEA) for 1997 indicate that $119 million of the shipments of electromedical and electrotherapeutic apparatus was purchased and used by that same industry. This made up only about 1 percent of the industry’s output.
Appendix Methodology The general methodology involved measuring implied consumption (or expenditures) as manufacturers’ shipments plus imports minus exports. This is sometimes known as a “commodity-flow” procedure.
Shipments The most detailed shipments data are available from the Economic Censuses conducted by the Census Bureau in years ending in “2” and “7.” Somewhat less detail is available from the Annual Survey of Manufacturers (ASM) for other years. Shipments data used are “product shipments” in contrast to “industry shipments.” Product shipments are recorded on a “wherever made basis.” In other words, they include products made in industries primarily engaged in a specific activity as well as the same products made in industries primarily engaged in other types of manufacturing.
The Economic Census data were not used directly in deriving the estimates but were consulted as a control on the ASM data; much more detail is available for Census years.
7 The Census Bureau website provides bridge tables linking the NAICS codes with the Standard Industrial Classification (SIC) codes. The 1997 and 2002 Economic Censuses were tabulated using NAICS, and the earlier Censuses were tabulated using the SIC.
NAICS changed between 1997 and 2002, but no manufacturing classifications changed.
(The Economic Census for manufacturing did have some changes made by Census, however. They classified hospital beds in a different NAICS category.) Data on non-medical irradiation equipment are from the Census Bureau’s Current Industrial Reports, Series MA334S, “Electromedical and Irradiation Equipment.” This publication ended in 2004, but the products covered will appear in another of the CIR series reports in the future. Personal industrial safety devices and protective clothing are available in the ASM. Anti-glare glasses and related goods were interpolated between Census years (and extrapolated to 2004) using a residual nsk (not separately known) category in the ASM.
Imports and exports Imports and exports are tabulated by the Census Bureau from Customs and other documents, and were pulled from the website maintained by the United States International Trade Commission (ITC).
Import values used are the C.I.F. (cost, insurance, and freight) values. This represents the landed value of the merchandise at the first port of arrival in the United States. For this study we used “General imports” rather than “Imports for consumption.” Imports for consumption exclude imports that enter free trade zones and bonded warehouses, and they include merchandise that leave free trade zones and custom warehouses. However, Census Bureau studies have shown that the values leaving these entities can be severely misstated because of rules governing duties. As a practical matter the differences are not large for the categories included in this study. Separate data on import duties were not available.
Exports are valued at the F.A.S. (free alongside ship) value. This is the value of exports at the U.S. seaport, airport, or border port of export, based on the transaction price, including inland freight, insurance, and other charges incurred in placing the merchandise alongside the carrier at the U.S. port of exportation. The value, as defined, excludes the cost of loading the merchandise aboard the exporting carrier and also excludes freight, insurance, and any charges or transportation costs beyond the port of exportation.
“Total exports” rather than “Domestic exports” were used for this study. Total exports include “re-exports;” we decided to include these amounts because the re-exports are also reflected in the import data. The differences between total and domestic exports were 8 significant. For 2004, total exports for NAICS category 339112 were $7.8 billion, compared with domestic exports of 4.8 billion.
Imports and exports were tabulated on the basis of both NAICS categories (for 1996 forward) and SIC categories (for earlier years). In addition, a number of codes from the “Harmonized Tariff System” (HST) were tabulated in order to develop estimates needed to reconcile NAICS with the SIC and to remove non-medical portions of the broad categories discussed above.