Medicare Part D and prescription drug prices.

Berndt, Ernst R. ; Frank, Richard G.

CONGRESS passed initial versions of the Medicare Prescription Drug, Improvement, and Modernization Act in late June 2003. This landmark legislation provided for a prescription drug benefit for all Medicare beneficiaries over age 65 and for individuals under age 65 who have certain disabilities. This new prescription drug benefit is called Medicare Part D. The House and Senate versions of the bill differed, and after considerable negotiations and maneuvering, the House passed a unified version of the bill by a 220-215 vote on November 22, 2003. On the next day, the Senate passed the legislation by a 54-44 vote. On December 8, 2003, President George W. Bush signed the final conference committee version into law. The Medicare Part D prescription drug benefit was fully implemented on January 1, 2006.

The congressional and public debate on the merits of this legislation was extensive and heated. Controversy surrounded issues such as what would the effects of moral hazard on prescription drug demand and prices be? How should the Federal Government exercise its considerable buying power? How restrictive or broad should formularies be? How much competition should there be among private plans offering benefits? How high would monthly premiums be, and how would they vary with benefit design? And of course, how much would this new program cost?

Medicare Part D has been with us now for over a year. What has happened? In terms of assessing its impact on prescription drug prices, there are at least three important considerations on which we focus in this paper. First, how has the Bureau of Labor Statistics (BLS), the source of official government price statistics, monitored and measured prices paid by consumers (the Consumer Price Index (CPI)), as well as prices received by manufacturers from sales to the first point in the distribution chain (Producer Price Indexes (PPIs) subsequent to the implementation of Medicare Part D? Specifically, what measurement changes and assumptions were required in order to assess the impact of part D on consumers' and producers' prices? Second, given provisions of the part D legislation and the BLS procedures for measuring prices, what do we as economists expect regarding the impact of part D on consumers' and producers' prices? And third, what price changes have been observed by the CPI and the PPIs leading up to and then following full implementation of the part D legislation on January 1, 2006?

Background history and literature

Over the years, as U.S. public policy has lead to expanding health insurance coverage, policy analysts have evaluated not only government and elderly out-of-pocket expenditures on health care but also the price and quantity components of these expenditures.

Related concerns have focused attention on the overall price inflation experienced by the elderly versus the nonelderly and more specifically on relative prices paid by the elderly, versus the nonelderly for health care goods and services.

For at least seven decades, the BLS Medical CPI (MCPI) has grown about half again as fast as the overall CPI; between 1927 and 1996, for example, the MCPI rose at an average annual growth rate of 4.59 percent, compared with 3.24 percent for the CPI (Berndt and others 1998a and 1998b). In the 11 years since then, between January 1996 and January 2007, these average annual growth rates were 3.91 percent for the MCPI and 2.49 percent for the CPI. Congressional concern over these differential rates of inflation has involved a number of initiatives.

Prior to the introduction of Medicare in July 1966, the Social Security Administration anticipated that the existence of the new insurance might have an impact on medical care prices. Therefore, in the summer of 1965, the administration arranged with BLS to collect supplementary prices for three surgical procedures and two in-hospital medical services that were particularly prevalent among the elderly though not necessarily limited to them. The three surgical procedures were cholecystectomy (removal of the gall bladder), prostatectomy (removal of the prostrate gland), and fractured neck of femur (hip surgery), and the two in-hospital services were acute myocardial infarction (treatment of heart attack) and cerebral hemorrhage (stroke). Among the major results of this study, as stated in a report to the President and summarized by Dorothy E Rice and Loucele A. Horowitz, was the finding that the index of the five in-hospital surgical and medical procedures that were particularly significant for the aged did not increase as rapidly during 1966 as the combined index for physicians' fees regularly priced for the CPI (Rice and Horowitz 1967, 28; U.S. Department of Health, Education and Welfare 1967). (1)

Several decades later, in response to a mandate contained in the 1987 amendments to the Older Americans Act of 1965, the BLS created an experimental price index for elderly consumers (CPI-E). The CPI-E employs differential expenditure weights for the elderly (defined as households headed by persons aged 62 and older) and the nonelderly based on data from the Consumer Expenditure Survey (CES), but the CPI-E assumes that within each category weight, the distribution of prices, the outlets in which consumers buy, the use of coupons, and the availability of discounts, as well as the quality of the items purchased, are the same for the elderly and as for the nonelderly (U.S. Department of Labor, Bureau of Labor Statistics no date). From 1982 through 1996, the CPI-E for the elderly grew 67.9 percent, while the CPI rose 62.5 percent, implying that over that 15-year period, the average annual growth rate of the CPI-E, at 3.77 percent, was slightly greater than the 3.53-percent growth rate of the overall CPI (Berndt and others 1998a and 1998b). In the 11 years since then, between January 1996 and January 2007, the averages have been 2.68 percent and 2.49 percent, respectively. The larger health care expenditure weights for the elderly, along with greater measured medical price inflation, account almost entirely for the difference in the growth rates between these two series. In this context, one qualifying note emphasized by the Boskin Commission was that medical care prices are likely to have overstated inflation by not fully accounting for improvements in quality (U.S. Senate Finance Committee 1996). If this is correct, then as Moulton and Stewart have noted "A reduced rate of inflation for medical care would mitigate and perhaps eliminate any difference between the CPI-E and the official CPI" (Moulton and Stewart 1997, 21). (2)

Relatively little research has focused on price differentials between the elderly and the nonelderly for health care goods or services. (3) Among various medical care goods and services, pharmaceuticals have become an increasingly important component of the medical care armamentarium. Moreover, prescription drugs are likely to be one case in which within stratum consumption patterns of the elderly likely differ substantially from those of the nonelderly.

Berndt and others (1998a and 1998b) have examined whether prescription drug price inflation in the 1990s differed between the elderly and the nonelderly, when age-related substrata variations in consumption were taken into account. They examined prices at three alternative points in the distribution chain and reported three sets of findings.

First, at the initial point in the distribution chain involving manufacturers' sales to wholesalers, retailers, and hospitals--transactions that are monitored and reported by various BLS PPIs--there is essentially no age-related aggregate price differential despite very significant differences in the baskets of drugs ultimately destined for use by the elderly and the nonelderly. Specifically, using prescription drug data from the National Disease and Therapeutic Index survey, maintained by IMS Health, to record elderly and nonelderly number of prescriptions by therapeutic class and applying these proportions to the BLS PPI weights by therapeutic class, the authors found that the PPI for pharmaceuticals destined for ultimate use by the elderly increased from 1.000 in 1990 to 1.331 in 1996, while that for the nonelderly rose a virtually identical amount, from 1.000 to 1.329, over the same 6-year period.

A second finding focused on an intermediate point in the distribution chain involving acquisition prices of retail pharmacies for purchases primarily from wholesalers as measured by the IMS retail prescription audit; these retail sell-in transactions take place at a point in the distribution chain that is in between the PPI and CPI and is not monitored by BLS price measurement programs. The authors focused on three therapeutic areas--antidepressants (used twice as intensively by the nonelderly, at 4.69 percent, as by the elderly, at 2.35 percent), broad and medium-spectrum antibiotics (also used about twice as intensively by the nonelderly, at 15.79 percent, as by the elderly, at 7.44 percent), and calcium channel blockers (for hypertension, used about three times more intensively by the elderly, at 6.18 percent, as by the nonelderly, at 2.01 percent). The authors found that between 1990 and 1996, retail acquisition price inflation for antidepressants destined for use by the elderly, at 7.02 percent, was less than that for ultimate use by the nonelderly, at 10.9 percent. Further research revealed that the elderly disproportionately used older generic drugs whose prices rose less rapidly than branded drugs during this time period. For antibiotics, however, especially from 1992 to 1996, the reverse occurred--the antibiotics price index for the elderly increased 7.74 percent, while that for the nonelderly rose only 2.40 percent. Additional research suggested that the greater elderly price inflation since 1992 appeared to reflect the more rapid growth in the elderly's use of the newest, branded drugs for which bacterial resistance was generally less likely. Finally, for the calcium channel blockers, there was essentially no difference in price inflation between 1990 and 1996-10.0 percent for the nonelderly and 11.1 percent for the elderly.

Data constraints prevented Berndt and others (1998a and 1998b) from undertaking a comparable analysis of retail sell-out prices across various therapeutic classes. Instead, the authors confined their analysis to sales by retail pharmacies to consumers and other payors (monitored by the IMS method-of-payment survey) to the antidepressant therapeutic class. Over all age groups, between 1991 and 1996, gross margins for antidepressants sold by retail pharmacies (sell-out prices relative to sell-in prices) fell about 3.5 percent, in part because of the growth of managed care and pharmaceutical benefit manager firms during that timeframe. Additional research found that young consumers appeared to have enjoyed most of the benefits of the increased buying power of managed care, for gross margins on the antidepressants they purchased fell by 3.8 percent. In contrast, for the antidepressants purchased by the elderly who are disproportionately large users of generic drugs, retail margins actually increased slightly.

These results suggest that no general age-related pattern of price inflation differentials for prescription pharmaceuticals is likely to emerge. Instead, the empirical significance of brand versus generic consumption, use of new versus old drugs, and various age-related quality attributes (once-a-day versus multiple daily dosages, extent of adverse interactions with other drugs, and seriousness of side effects and adverse reactions) must most likely be examined on a class-by-class basis before any general conclusions can be reached. (4) Moreover, even these class-specific variations may change with time, particularly when major institutional and market changes take place.

An example of such a major legislative development is the Medicare Prescription Drug, Improvement, and Modernization Act, which was passed by the U.S. Congress in 2003 and which mandated a Medicare Part D prescription drug benefit for the elderly and disabled, beginning on January 1, 2006.

Medicare Part D: Timelines, essential features, and BLS price measurement

Legislative history and essential features

The Medicare Prescription Drug, Improvement and Modernization Act (Medicare Modernization Act) was introduced into the House of Representatives on June 25, 2003, sponsored by Speaker Dennis Hastert. After an initial electronic vote failed, several Republicans changed their vote, and early on the morning of June 27, 2003, it passed by a 216-215 vote. The Senate passed its version of the bill by a 76-21 vote on June 26, 2003. The bills were then unified in a conference committee and came back to the House for approval on November 21, 2003. After various legislative maneuvers and vote changes by congressional representatives, around 5:30 a.m. on November 22, 2003, the House passed the unified bill by a 220-215 vote. The Senate's consideration of the conference report was less heated but still controversial, and the bill finally passed the Senate by a 54-44 vote on November 23, 2003. President Bush signed the bill into law on December 8, 2003 (Wikipedia 2006).

Under provisions of the Medicare Modernization Act of 2003, a prescription drug benefit was created as Part D of Medicare, to become available beginning January 1, 2006, whereby Medicare beneficiaries (including those disabled and under age 65) would receive a statutorily defined standard prescription drug benefits after a $250 annual deductible, would pay 25 percent of costs up to $2,250, 100 percent of costs between $2,250 and $5,100 (a gap of $2,850, commonly referred to as the "donut hole"), and 5 percent of costs above $5,100. Plans were granted freedom to construct alternative benefit designs that were actuarially equivalent to the standard benefit, such as no deductibles and tiered copayments rather than 25 percent coinsurance (Cubanski and Neuman 2006). Expected monthly premiums were estimated to be about $37, with variations depending on copayment structures, formulary design, and retail pharmacy network benefit provisions.

As a temporary and transitional step to assist beneficiaries more immediately with their prescription drug purchases, the Medicare Modernization Act of 2003 also created a program whereby Medicare-approved discount cards were issued to beneficiaries for use beginning on June 1, 2004. These cards were to help seniors purchase prescription drugs at reduced prices until the full part D benefit was implemented in January 2006. The discount cards did not provide actual insurance benefits but instead were cards issued by Medicare-approved private-sector entities (pharmacies, pharmacy benefit management firms, insurers), giving Medicare beneficiaries approximately a 15-20 percent discount on out-of-pocket cash prices for prescription drugs; discounts were on the steeper end for generic drug purchases (U.S. Department of Labor, Bureau of Labor Statistics 2006). Subsidies were also made available to some low-income beneficiaries. Other important dates were October 1, 2005, the first day for private companies to release details of their individual plans, and November 15, 2005, the first day that individuals could enroll in a part D prescription drug plan.

One other significant aspect of the Medicare Modernization Act of 2003 concerned those individuals over age 65 who had been receiving prescription drug benefits under state Medicaid programs and those under age 65 with certain disabilities. These "dually eligible" beneficiaries saw responsibility for purchasing their prescription drugs transferred from Medicaid to the Medicare Part D program, effective January 1, 2006. It is estimated that these dually eligible individuals accounted for about 29 percent of all part D enrollees (Cubanski and Neuman 2006, exhibit 5, page w8). Under the Medicaid "most-favored-nation" rules, manufacturers have been required to offer Medicaid the lower of the "best" price they sell to the private sector or a discount of 15.1 percent below the average manufacturer price for branded drugs, whichever is lower. (5) Under Medicare Part D, however, pharmaceutical manufacturers instead negotiated prices with private prescription drug plans (PDPs) (Frank and Newhouse 2007). Manufacturers' prices charged to PDPs were exempt from the "most-favored-nation" pricing calculations.

Medicare Part D price monitoring by the BLS

Given the substantial lead time between initial legislative approval in June 2003 and final full implementation of Medicare Part D in January 2006, the various BLS price measurement programs had considerable time to adapt their data collection and aggregation procedures as necessary to reflect changing prices associated with implementation of Medicare Part D.

Since the PPI measures prices only at the first point in the distribution chain (for pharmaceuticals, most commonly from manufacturers' sales to wholesalers and large retail chains), price changes directly realized by Medicare Part D beneficiaries are out of scope--the PPI does not identify and monitor prices paid by final purchasers, such as the elderly at retail or mail order. For the PPI, therefore, implementation of Medicare Part D required no significant changes in the data gathering protocols. Instead, the PPI continued to introduce new branded and generic drugs as supplemental samples into its sample of price quotes on an annual basis. (6)

In contrast to the PPI program, the BLS CPI program faced a number of serious challenges in adapting its price measurement protocols to capture price changes resulting from the introduction of the transitional Medicare discount card and then the launch of the full Medicare Part D program. Because the Centers for Medicare and Medicaid Services (CMS) Web site contained a pricing utility set up explicitly for beneficiaries to determine how the various discount card plans compared with each other in terms of drugs covered and their prices, beginning in October 2004, the CPI flipped a portion of its existing sample--the senior cash-discounted portion that had been receiving about a 10-percent discount--from discounted cash to Medicare discount card, where the sample recorded an average additional discount of 15 percent off retail and mail order cash prices; these quotes were then employed in the aggregate index calculations (U.S. Department of Labor, Bureau of Labor Statistics 2006). As of December 2004, the BLS had been collecting 1,111 price quotes for prescription drugs (U.S. Department of Labor, Bureau of Labor Statistics 2005). Since CMS ceased supporting the pricing utility that yielded the Medicare discount card price quotes in November 2005, for November and December 2005, the BLS estimated these price quotes as being approximately 25 percent off the full cash price quotes they continued to collect. (7)

To account for the introduction of Medicare Part D in January 2006, the BLS CPI program employed a variant of the directed substitution rule by which the product characteristics of the new item were already known and determined (rather than going through the entire disaggregation process). In particular, the CPI recorded the price changes that occurred for the same prescription as it switched from being paid with a Medicare-approved discount card (December 2005) to the full Medicare Part D benefit price (January 2006). The latter was calculated by taking quotes from a single nationally offered private prescription drug benefit plan that conveniently allowed direct pricing via an online pricing utility. (8) In cases where the national part D plan only offered the generic equivalent of a brand drug covered by the discount card plan, the CPI recorded the price change between the brand discount card and the generic part D price. Note that only the changes from the discount card to part D were captured by the BLS CPI and that the quoted changes are those based on a single national plan. (9) In particular, the CPI program has not attempted to capture price quotes of formerly uninsured cash, or partly insured, customers who subsequently obtained part D coverage. Similarly, since direct substitution procedures were employed, any switches from retail to mail order that occurred because of part D private prescription drug plan benefit design were also not captured by the CPI.

Because a portion of the Medicare-approved discount cards that came into the CPI sample in 2004 was rotated out of the sample and was not adequately re placed through rotation, BLS augmented its Medicare-approved discount card sample to match CMS' estimate that approximately 3.7 percent of the U.S. population had been issued such cards. This was accomplished by the BLS randomly assigning part D quotes to their existing sample. As a result, the part D sample may not mirror a market snapshot that would have emerged had the BLS initiated the part D drugs from the pharmacy based on their traditional "last 20" prescription method. We note in passing that in the future, when BLS initiates a new sample frame, it will finally be able to measure and directly compare prescription drug prices paid by the elderly through Part D with purchase prices paid by the nonelderly. These new data could yield some very interesting research findings and in principle, could be incorporated into the CPI-E.

Coincidentally, the BLS CPI program has been wrestling with how to incorporate prescription-only to over-the-counter (Rx-to-OTC) switches into its medical care CPI, which includes both types of drugs. Two very prominent recent Rx-to-OTC switches have involved Claritin for the treatment of allergies (switch approved November 27, 2002) and Prilosec OTC for the treatment of frequent heartburn (approved June 20, 2003) (U.S. Food and Drug Administration 2003, 2002). Conversations with BLS CPI personnel reveal that when there is an Rx-to-OTC switch, the BLS treats the initial price of the OTC variant as the final price of the Rx version, and then it treats subsequent OTC price changes as only affecting the OTC price index. Note that since the BLS CPI is based on a Laspeyres aggregation framework, which has the property of reproducible aggregation, the Laspeyres aggregate of an Rx price index and an OTC price index is numerically equivalent to a Laspeyres index aggregated simultaneously over all Rx and OTC products. (10) A related pilot project is under way at the BLS CPI program, involving the creation of separate brand and generic CPIs for prescription pharmaceuticals. Currently, the BLS only publishes an aggregate of prescription pharmaceuticals.

Expectations regarding impact of Medicare Part D on BLS price measures

As we have written elsewhere, we believe the BLS faces enormous challenges in reliably measuring price inflation for health care goods and services, including prescription drugs. (11) The introduction of Medicare Part D benefits likely increases these challenges and difficulties for the BLS. What are reasonable expectations regarding how the introduction of Medicare Part D affected price inflation as measured and reported by the pharmaceutical CPI and PPI? Four points are worth noting.

First, prior to the implementation of Medicare Part D, about 25 percent of the elderly had been paying cash prices for prescription drugs for the entire year. (12) As of January 1, 2006, these individuals became eligible to enroll in Medicare Part D and benefit from the lower prices negotiated on their behalf by private prescription drug plans (Frank and Newhouse 2007). Because undoubtedly, not all of those who were eligible actually enrolled (estimates are that slightly more than 90 percent of those eligible obtained creditable coverage (Cubanski and Neuman 2006)), as we have seen, the price declines experienced by those individuals who did enroll will not have been captured by the CPI. In this sense, to the extent such transaction types are not being captured, growth in the prescription drug CPI has been overstated. Looking to the future, although some Medicare Part D transactions will have been uncovered by the Consumer Expenditure Survey (CES) data (none from 2005, but presumably those from the 2006 CES), the resulting new CES weights will be set as of December 2007 for use beginning finally with the January 2008 CPI.

Second, we expect the introduction of new or additional insurance to increase demand due to moral hazard. Danzon and Pauly (2002) have estimated that between 25 percent and 50 percent of the total growth in U.S. prescription drug spending between 1987 and 1996 can be attributed to increased drug insurance coverage by employers and Medicaid. On the other hand, since as noted earlier, a substantial portion (between 25 percent and 40 percent) of new Medicare Part D beneficiaries had previously been paying cash prices, branded manufacturers now faced a reduced demand from the cash-paying segment of consumers. Which of these two effects dominates--increased demand from moral hazard versus reduced number of cash-paying customers--is not obviously a priori. Whether the combined demand function over cash-paying and new Medicare Part D insured individuals shifted outward or inward is in theory ambiguous and is therefore an empirical matter. Also unclear are expectations regarding the timing of any price changes. Specifically, whether price increases occurred on or after the time of the implementation of Medicare Part D or in anticipation of it depends on numerous factors beyond the scope of this paper.

Third, as noted above, switching dually eligible individuals from Medicaid coverage, which entailed "most-favored-nation" pricing to Medicare private prescription drug plans (PDPs), which are not subject to the Medicaid mandatory rebates, provided the PDPs with less bargaining power than the state and Federal Medicaid purchasers had previously been able to exercise. Recall that it is estimated that 29 percent of the Medicare Part D enrollees had previously been dually eligible (Cubanski and Neuman 2006, exhibit 5, page w8). To the extent that this has occurred, we might expect prices of drugs disproportionately used by the previously dually eligible individuals to increase more rapidly than other drugs, at least as measured by the PPI. Below we comment on the therapeutic drug classes that are likely to be more intensively utilized by previous dually eligible individuals.

Fourth and finally, in their negotiations with CMS regarding formulary design, the PDPs were constrained by CMS to include a minimal number of (often at least two) drugs with preferred status in each therapeutic class and in some cases, such as the antidepressants, all drugs (Huskamp, et al. 200; Huskamp 2003). Since payers' buying power relative to manufacturers stems in large part from payers' ability to either exclude drugs entirely from their formulary or at least banish them to the third tier with the highest copayment, this broad formulary policy constrained the buying power of the PDPs, and may have led to reduced rebates and increased prices.

Together, these four considerations suggest that potentially offsetting impacts on prices are associated with the passage and then the implementation of Medicare Part D legislation. The net effect of these various impacts is in theory ambiguous, and is therefore an empirical matter. Moreover, given the 30-month timespan between the June 2003 initial passage of the legislation and its full implementation in January 2006, it is also unclear what to expect in terms of the timing of any price changes--price changes in anticipation of the full implementation of the Part D benefit could be larger or smaller than those following its implementation. However, what is clear is that we expect PPIs in therapeutic classes, including drugs disproportionately used by previous dually eligible individuals, to increase more rapidly than PPIs for drugs in other classes.

Results: Trends in BLS measures of pharmaceutical CPI and PPI price inflation

We now move on to a discussion of trends in BLS measured price inflation, with a particular focus on dates surrounding developments in Medicare coverage of prescription pharmaceuticals. We begin with the CPI and focus on five time periods over the last 11 years.

The first two periods are (1) January 1996-January 2000 (the early history) and (2) January 2000-June 2003 (June 2003 was the month in which initial House and Senate versions of the Medicare Prescription Drug, Improvement, and Modernization Act were passed). We then divide the following 30-month time period until the January 1, 2006, implementation of Medicare Part D into two equal 15-month time intervals: (3) June 2003-September 2004 and (4) September 2004-December 2005. We then focus on the year following the implementation of the Medicare Part D program: (5) December 2005-December 2006. For each of these time periods, we compute average annual growth rates.

Results: The CPIs

As we noted earlier, the set of price quotes interpreted as reflecting Medicare Part D transactions is based in part on the BLS' flipping Medicare discount card quotes on to Medicare Part D, based on online price quotes from a single national private prescription drug plan' Web site and in part on randomly taking certain existing price quotes and converting them to a part D comparison over time. The latter set of quotes may, however, have not originally been those of elderly individuals, and thus the composition of prescriptions in the part D subsample may not be representative of that for the overall elderly population enrolled in part D.

In table 1, we compare the distribution of prescriptions by therapeutic drug class in the overall sample of prescription drug CPI quotes with that in the part D subsample over the January-October 2006 timeframe. There are six therapeutic classes in which there are zero part D quotes--the prescription shares of these classes except for anesthetics (at 9.67 percent) in the overall sample are quite small, and together, the six zero-share part D classes account for 12.94 percent of the overall sample prescriptions. Not surprisingly, in the cardiovascular and metabolics/nutrients classes, the elderly part D share is considerably larger than in the overall sample; in contrast, for central nervous system and analgesics, the elderly part D share is smaller than in the overall sample.

Average annual growth rates of various CPIs are presented in table 2 over the five time intervals discussed above. In the first row, we provide average annual growth rates of the "all items-urban" CPI, and in the second row, the experimental or elderly CPI (E-CPI) for "all items-urban." The E-CPI grows slightly more rapidly than the "all items" CPI, with the differential ranging from about 0.10 percent to 0.22 percent and having no distinct time trend. Previous literature has attributed this differential to the larger share of medical care expenditures for seniors along with above-average inflation for medical care.

In the third row of table 2, we show average annual growth rates for the overall medical care CPI, and in the fourth row, the medical care E-CPI, which differ to the extent that the elderly and nonelderly shares of the components (medical care commodities, medical care services, hospital and related services, and health insurance) of overall medical care differ, and these components experience varying rates of inflation. In three of the five time intervals, the medical CPI-E grows slightly less rapidly than the overall medical CPI, and the reverse occurs in two time periods. Over the 11-year timeframe between January 1996 and January 2007, the medical CPI-E grew at an average annual growth rate of 3.905 percent, virtually identical to the overall medical CPI, at 3.913 percent.

Rows five and six provide average annual growth rates separately for medical care services and medical care commodities; the BLS does not compute experimental CPI-Es at this level of aggregation, only overall CPIs. In each of the five time intervals, average annual growth rates of medical care services (which includes physicians,' dental, hospital and nursing home and adult day care services) are greater than those of medical care commodities (prescription and OTC drugs and medical supplies), with the differential since 2000 ranging between 1.0 percent and 2.3 percent and tending to become larger in more recent times.

Finally, in the last row of table 2, we provide average annual growth rates for prescription drugs, which include medical supplies. Between 1996 and 2005, annualized price inflation for prescription drugs ranged from about 3.6 percent to 4.3 percent, but in 2006 following the implementation of Medicare Part D, it fell to about half its previous rate, to 1.856 percent. In summary, in recent times, there appears to have been a substantial decline in the rate of growth of the CPI for prescription drugs, particularly following the implementation of the Medicare Part D benefit in January 2006.

Results: The PPIs

We now turn to a consideration of the PPIs for pharmaceuticals. Recall that the PPI monitors prices received by the manufacturer (net of discounts and prompt payment price reductions) from sales to the first point in the distribution chain, which for pharmaceuticals is usually either wholesalers or large retail chains. Participation by manufacturers in reporting to the BLS is voluntary; participation rates have been around 65 percent. Although considerable pharmaceutical manufacturing takes place in Puerto Rico, from the vantage of the BLS PPI program, Puerto Rico is not part of the United States. (13)

The BLS PPI for pharmaceuticals includes both prescription and OTC products. Medicaid purchases are explicitly out of scope for the CPI (because they are government purchases), but for the PPI, the identity of the ultimate consumer is irrelevant; thus, the PPI will incorporate prices paid by among others, Medicaid purchasers (that is, state governments and the CMS). In principle, the pharmaceutical PPI also tracks changes in prices that occurred when Medicare-Medicaid dually eligible individuals switched to the Medicare Part D program in January 2006, although the types of transactions are defined quite narrowly and at best, changes in weights occur only at annual intervals.

We report average annual growth rates for various pharmaceutical PPIs in table 3 for five time intervals: (1) June 2001-June 2003 (because some price series did not begin until June 2001); (2) June 2003-September 2004 (the first 15 months after initial passage of the Medicare Part D legislation); (3) September 2004--December 2005 (the final 15 months before the implementation of Medicare Part D in January 2006); (4) December 2005-December 2006 (to monitor changes associated with the first year of the implementation of Medicare Part D); and (5) January 2000-December 2006 (for some price series, data from the beginning of this decade). We remind readers that the PPI is a sample of products selected using probabilities proportional to sales; while we mention particular brand products in various therapeutic classes below, we have no information regarding whether those specific brands are in the PPI sample.

The first row in table 3 indicates that the overall pharmaceutical PPI has grown at about 4.1 percent annually since 2000, with slightly larger annual growth at 5.4 percent in the 15 months leading up to the implementation of Medicare Part D; (14) since December 2005, growth has returned to just under 4 percent. There is considerable heterogeneity in average annual growth rates, both across time intervals and among therapeutic classes. Prices of prescription analgesics (pain medicines), for example, only grew at a 1.4-percent annual rate in the 15 months leading up to the implementation of Medicare Part D, but then they grew at a much larger 4.9-percent annual rate following its implementation. (15) By contrast, prices of anticoagulants grew at a 4.7-percent annual rate between June 2003 and September 2004, but since then, they have grown at 0.1-0.2 percent annually. (16)

The antispasmodic/antisecretory market class includes drugs for the treatment of heartburn (such as the [H.sub.2]-antagonists and proton pump inhibitors--brands like Zantac, Prilosec and Nexium). This category has experienced particularly volatile price growth--averaging around 3.5 percent annually from June 2001 to September 2004, then grew at a very high annual rate of 22.8 percent up through December 2005, (17) and continued to grow at a 5.1-percent annual rate since then. (18) In table 1, this class of drugs would be in the gastrointestinal category, and data there suggest that the prescription drug share of gastrointestinal drugs is approximately the same for the elderly and nonelderly. We know of no data on whether this class of drugs is consumed disproportionately by the previous dually eligible individuals.

Returning to table 3, we see that cancer therapy products (where utilization might be expected to be disproportionately by the elderly, though typically covered by Medicare Part B for many years) had an average annual growth rate of about 4.4 percent over the entire January 2000-December 2006 timeframe. In the 15 months leading up to the January 2006 implementation, prices rose at an annual rate of 3.7 percent and since then, at a slightly smaller rate of 2.5 percent. By contrast, the class entitled "other neoplasms, endocrine system and metabolic diseases, including hormones" includes a number of antiosteoperosis drugs for postmenopausal women, and thus its utilization is likely to be disproportionately by the elderly. (19) As seen in table 3, over the entire January 2000--December 2006, price growth has been relatively high in this class, averaging 8.6 percent annually; between September 2004 and December 2005, it increased at an average annual growth rate of 11.1 percent, and most recently, it continued at a relative high average annual growth rate of 11.7 percent.

Of particular interest in the context of Medicare-Medicaid dually eligible individuals are psychotherapeutic drugs, which are used disproportionately by Medicaid beneficiaries. (20) For the entire class of psychotherapeutic drugs, price growth accelerated from about 6 percent annually between June 2001 and September 2004, to about 8 percent annually since then. (21) The next row in table 3 indicates that this price acceleration was particularly marked in the antidepressant subclass of psychotherapeutic drugs. For antidepressants, the average annual growth rate between September 2004 and December 2005 was 14.6 percent, more than twice that during the previous 15 months at 6.3 percent; (22) this average annual growth rate has fallen since the implementation of Medicare Part D, but it is still substantial at 10.1 percent in 2006. (23) Interestingly, average annual growth rates are lower, albeit still considerable in the subclass of psychotherapeutics designated as "other psychotropics, including tranquilizers," which includes the second generation atypical antipsychotic drugs for treatment of schizophrenia and bipolar mania disorder. In recent years, the medical literature has identified several medications within this class as being associated with side effects of weight gain and diabetes, and their cost-effectiveness over earlier less costly products has been called into question. (24) For this aggregate class of other psychotropic drug, prices grew at an average annual growth rate of around 6.0 percent between June 2003 and September 2004, they grew at a slower annual rate of 3.9 percent in the 15 months leading up to the implementation of part D, and since then, they have grown at an annual rate of 5.5 percent.

In summary, therefore, although there is considerable heterogeneity over time intervals and among therapeutic classes, there is evidence based on PPI trends suggesting that some prescription drugs likely disproportionately used by the elderly (for example, the antiosteoporosis drugs for postmenopausal women) and by the Medicaid-Medicare dually eligible individuals that are now covered by Medicare Part D (such as various types of psychotherapeutic drugs) have experienced very considerable price growth leading up to and following the implementation of the new Medicare Part D benefit. A common, but clearly not uniform, pattern is that price increases in the 15 months leading up to the implementation of the part D benefit in January 2006 were greater than those observed since its full implementation in January 2006. Although at a much higher level of aggregation, this PPI evidence is consistent with preliminary findings from Frank and Newhouse (2007) that are based on more detailed brand data, which are discussed below. However, there is also substantial PPI price growth during these time periods for the antispasmodic/antisecretory class of drugs--drugs that are not likely to be used disproportionately by the elderly. More research will be needed to clarify these early findings.

Results from an additional data source

We have explored additional heterogeneity in the price response to passage of the Medicare Modernization Act by examining price movements among branded prescription drug products in the top 50 in U.S. sales, based on detailed research where these drugs have been stratified by the age composition of their purchasers. (25) We have constructed pharmaceutical PPIs (Laspeyres and Fisher indexes) for this entire sample of drugs and for various subsets. Using IMS Health data that track sales of prescription drugs from manufacturers and wholesalers to drug stores, we selected brand name drug products from among the top 50 in U.S. sales that had no generic competition. From among these, we identified two cohorts of drugs that together included eighteen products. The first consists of a set of drugs where 55 percent or more of the sales of the drugs were likely to have been to people over age 65 (the sales shares by age are based on data on physician drug mentions provided from surveys of physician office visits conducted by IMS Health). (26) The second group is made up of drugs where less than 35 percent of the sales are likely to have been to people age 65 or more. (27) From these data, we calculated monthly prices and quantity of sales based on extended units. The period observed begins in June 2003 and extends through June 2006.

Using these data, we constructed six price indexes that are analogous to PPIs but that are at a much more disaggregated level. Specifically, we calculated fixed-weight Laspeyres and chained Fisher indexes for each of the two cohorts defined by the age of the purchasers, as well as an overall index for all 50 drugs. This yields six price index series. The six indexes are displayed in chart 1. The fixed-weight Laspeyres indexes--L-elderly, L-nonelderly, and L-all drugs--refer to the drugs disproportionately used by the elderly, the nonelderly, and the entire set of 50 drugs, respectively; the corresponding chained Fisher indexes are designated F-elderly, F-nonelderly, and F-all drugs, respectively.

[GRAPHIC OMITTED]

Chart 1 reveals that the two PPIs calculated for the drugs in the nonelderly purchasers cohort grew at lower rates than the cohort of drugs where the majority of purchasers were over age 65. Thus, by June 2006 there was a 5.3 percentage point difference in the final value of the Fisher index for the elderly and the nonelderly drugs (F-elderly and F-nonelderly). The index for the elderly cohort ended between 3 and 4 percentage points higher, depending on the index, than the corresponding index for all 50 drugs. Together, these data suggest that prices of prescription drugs likely used to treat people over 65 years of age, and thus are more likely to have been influenced by the passage and implementation of Medicare Part D legislation increased more rapidly than did drug prices for prescription drugs likely used to treat the general population.

Concluding remarks

The implementation in January 2006 of the Medicare Modernization Act that provided for Medicare Part D prescription drug benefits for the elderly created monitoring challenges for Government statistical agencies, such as the BLS. It has also created the opportunity for the BLS eventually to assess any differences in prices paid by the elderly and by the nonelderly for the same branded or generic prescription drug. Although the implications of the Medicare Modernization Act for the PPI program were relatively minor, those for the CPI program were greater and more complex. The CPI program did not attempt to capture price quotes of formerly uninsured cash-paying or partly insured consumers who subsequently obtained part D coverage or for those switching from retail to mail order because of part D. Hence, it is likely that the CPI for prescription drugs overstated actual inflation between 2005 and 2006. Nonetheless, it is notable that the CPI for prescription drugs grew only by 1.9 percent between December 2005 and December 2006, roughly half the annualized 3.8-percent rate in the previous 15 months.

With respect to the various pharmaceutical PPIs, theoretical predictions regarding the price impacts of Medicare Part D are generally ambiguous, since the moral hazard increase in demand could be offset by the reduction in the number of cash-paying consumers. There is some evidence suggesting that drugs disproportionately used by the elderly (for example, antiosteoporosis drugs for postmenopausal women) and by the Medicaid-Medicare dually eligible individuals subsequently covered by Medicare Part D (for example, psychotropic drugs) experienced considerable price growth leading up to and following the implementation of Medicare Part D. Although the evidence is not uniform, a common observed trend is for price increases in the 15 months leading up to the implementation of Medicare Part D to be greater than in the previous 15 months following initial passage of the enabling legislation, and in the year following full implementation.

Using data from a different source, IMS Health, on the 50 top selling brands stratified by age of purchaser, we report evidence consistent with the notion that between June 2003 and June 2006, price increases for drugs likely used primarily by the elderly were larger than were those for prescription drugs likely used primarily by the nonelderly. (28)

The implications of changes in purchasing arrangements for drugs used by Medicare beneficiaries and the resulting price impacts stemming from the implementation of part D are just now beginning to be observed. A great deal of new data will soon be emerging, which will facilitate research on the impacts of institutional changes on both out-of-pocket prices paid by consumers and on revenues received by prescription drug manufacturers (analogous to CPIs and PPIs for prescription drugs). This new learning is likely to be important for the interpretation of the continued evolution of health care price indexes and for the evaluation of public policies.

Acknowledgments

The authors thank Frank Congelio, Dan Ginsburg, and Francisco Velez at the Bureau of Labor Statistics for helpful discussions and data support and Samuel Kina for research assistance. Professor Frank acknowledges research support from the National Institute of Mental Health, Grant ROIMH069721. A previous draft of this benefited from the comments of Jack E. Triplett. Any opinions expressed in this paper are those of the authors and do not necessarily correspond with those of the institutions with which they are affiliated or the research sponsors.

References

Berndt, Ernst R., Zvi Griliches, and Joshua G. Rosett. 1993. "Auditing the Producer Price Index: Micro Evidence from Prescription Pharmaceutical Preparations." Journal of Business and Economic Statistics 11, no. 3 (July): 251-264.

Berndt, Ernst R., Iain M. Cockburn, Douglas L. Cocks, Arnold M. Epstein, and Zvi Griliches. 1998a. "Is Price Inflation Different for the Elderly? An Empirical Analysis of Prescription Drugs." In Frontiers in Health Policy Research, edited by Alan Garber, 33-75. Cambridge, MA: MIT Press for the National Bureau of Economic Research.

Berndt, Ernst R., Iain M. Cockburn, Douglas L. Cocks, Arnold M. Epstein, and Zvi Griliches. 1998b. "Prescription Drug Prices for the Elderly." Monthly Labor Review 121 (September): 23-34.

Berndt, Ernst R., David M. Cutler, Richard G. Frank, Zvi Griliches, Joseph E Newhouse, and Jack E. Triplett. 2000. "Medical Care Prices and Output." In Handbook of Health Economics, edited by Joseph P. Newhouse and Anthony C. Culyer, 119-180. Amsterdam: Elsevier Science B.V.

Berndt, Ernst R., David M. Cutler, Richard G. Frank, Zvi Griliches, Joseph P. Newhouse, and Jack E. Triplett. 2001. "Price Indexes for Medical Care Goods and Services: An Overview of Measurement Issues." In Medical Care Output and Productivity, edited by David M. Cutler and Ernst R. Berndt, 141-198. Chicago: University of Chicago Press for the National Bureau of Economic Research.

Cubanski, Juliette, and Patricia Neuman. 2006. "Status Report on Medicare Part D Enrollment 2006: Analysis of Plan Specific Market Share and Coverage." Health Affairs Web exclusive (November): w1-w12.

Danzon, Patricia, and Mark V. Pauly. 2002. "Health Insurance and the Growth in Pharmaceutical Expenditures." Journal of Law and Economics 45: 587-613.

Duggan, Mark. 2005. "Do New Prescription Drugs Pay for Themselves? The Case of Second-Generation Antipsychotics". Journal of Health Economics 24 (January): 1-31.

Frank, Richard G. 2001. "Prescription Drug Prices: Why Do Some Pay More Than Others Do?" Health Affairs 20 (March/April): 115-128.

Frank, Richard G., and Joseph P. Newhouse. 2007. "Mending Medicare's Prescription Drug Benefit: Improving Consumer Choices and Restructuring Purchasing." Hamilton Project discussion paper i sponsored by the Brookings Institution. Harvard Medical School, Department of Health Care Policy Management, April 10.

Frank, Richard G., Ernst R. Berndt, Alisa Busch, and Anthony E Lehman. 2004. "Quality-Constant Prices for the Ongoing Treatment of Schizophrenia: An Exploratory Study." Quarterly Review of Economics and Finance 44, no. 3 (July): 390-409.

Huskamp, Haiden A. 2003. "Managing Psychotropic Drug Costs: Will Formularies Work?" Health Affairs 22 (September/October): 84-96.

Huskamp, Haiden A., Meredith B. Rosenthal, Richard G. Frank, and Joseph P. Newhouse. 2000. "The Medicare Prescription Drug Benefit: How Will The Game Be Played?" Health Affairs 19 (March/April): 8-23.

Lieberman, Jeffrey A., T. Scott Troup, Joseph E McEvoy, Marvin S. Swartz, Robert A. Rosenheck, Diana O. Perkins, Richard S.E. Keefe, Sonia M. Davis, Clarence E. Davis, Barry D. Lebowitz, Joanne Severe, and John K. Hsiao. 2005. "Effectiveness of Antipsychotic Drugs in Patients with Chronic Schizophrenia." New England Journal of Medicine 353, no. 12 (September 22): 1,209-1,223.

Morton, Fiona Scott. 1997. "The Strategic Response by Pharmaceutical Firms to the Medicaid Most-Favored Customer Rules." RAND Journal of Economics, (Summer): 269-290.

Moulton, Brent R., and Kenneth J. Stewart. 1997. "An Overview of Experimental U.S. Consumer Price Indexes." Washington, DC: U.S. Bureau of Labor Statistics, March.

Newhouse, Joseph P. 2004. "How Much Should Medicare Pay For Drugs?" Health Affairs 23 (January/ February): 89-102.

Polsky, Daniel, Jalpa A. Doshi, Mark S. Bauer, and Henry A. Glick. 2006. "Clinical Trial-Based Cost-Effectiveness Analyses of Antipsychotic Use." American Journal of Psychiatry 163 (December): 2,047-2,056.

Rice, Dorothy P., and Loucele A. Horowitz. 1967. "Trends in Medical Care Prices." Social Security Bulletin 30 (July): 13-28.

Rosenheck, Robert A., Douglas L. Leslie, Jody Sindelar, Edward A. Miller, Haiqun Lin, T. Scott Stroup, Joseph McEvoy, Sonia M. Davis, Richard S. E. Keefe, Marvin Swartz, Diana O. Perkins, John K. Hsiao, and Jeffrey Lieberman. 2006. "Cost-Effectiveness of Second-Generation Antipsychotics and Perphenazine in a Randomized Trial of Treatment for Chronic Schizophrenia." American Journal of Psychiatry 173 (December): 2,080-2,089.

Sharpe, Andrew, ed. 2006. "Symposium on the Boskin Commission Report After a Decade." International Productivity Monitor 12 (Spring): 1-83.

Triplett, Jack E. 1999. "Accounting for Health Care: Integrating Price Index and Cost-Effectiveness Research." In Measuring the Prices of Medical Treatments. Edited by Jack E. Triplett, 220-250. Washington DC: The Brookings Institution.

U.S. Department of Health, Education, and Welfare. 1967. Report to the President on Medical Care Prices. Washington DC: U.S. Government Printing Office.

U.S. Department of Labor, Bureau of Labor Statistics. 2006. "Introduction of Medicare Part D in to the CPI Sample." Electronic manuscript. Washington, DC, January 13.

U.S. Department of Labor, Bureau of Labor Statistics. 2005. "Measuring Price Changes for Medical Care in the CPI." Electronic manuscript. Washington, DC.

U.S. Department of Labor, Bureau of Labor Statistics. No date. "The Limitations of the Unpublished Experimental Consumer Price Index for Americans 62 Years of Age and Older." Electronic manuscript. Washington, DC.

U.S. Food and Drug Administration. 2002. "FDA Approves OTC Claritin." Online at www.fda.gov/bbs/topics/news/2003/NEW00916.html (accessed December 21, 2006).

U.S. Food and Drug Administration. 2003. "FDA Approves Prilosec OTC to Treat Frequent Heartburn." Online at www.fda.gov/bbs/topics/news/2003/NEW00916.html (accessed December 21, 2006).

U.S. Senate Finance Committee. 1996. Final Report From the Advisory Committee To Study The Consumer Price Index. Washington DC: U.S. Government Printing Office.

Wikipedia. 2006. "Medicare Prescription Drug, Improvement, and Modernization Act" Online at <en.wikipedia.org/wiki/Medicare_Prescription_Drug_Improvement,_and_Moderniz> (accessed December 20, 2006).

(1.) Rice and Horowitz report that the December 1965-December 1966 average annual growth rates ranged from 2.5 percent for cholecystectomy to 6.9 percent for prostatectomy, and the combined index for physicians' fees regularly priced for the CPI rose 7.8 percent (Rice and Horowitz 1967, 25).

(2.) For additional discussion, see the various articles in Sharpe (2006).

(3.) In this context it is worth noting that because of Medicare reimbursement policies to physicians and hospitals, the elderly purchase much of their health care under administered prices.

(4.) The cost-effectiveness of medications in certain classes likely varies by patient age. Triplett (1999) links price indexes to cost-effectiveness analyses.

(5.) For details, see Morton 1997; Frank 2001.

(6.) For a discussion of supplemental sampling and other details on the PPI, see Berndt, Griliches, and Rosett 1993; Berndt and others 2000, 2001. We have benefited from correspondence with Frank Congelio in the BLS PPI program regarding recent supplemental sample introductions.

(7.) We are not aware of any emprical analyses substantiating the average 25-percent discount off of full cash price for these consumers.

(8.) Cubanski and Neuman (2006) report that 10 organizations captured 72 percent of the part D enrollment, primarily in low premium plans and those associated with name recognition. Two organizations--UHC-Pacific (United) and Humana--dominated, together accounting for 45 percent of part D enrollment.

(9.) We are unaware how the CPI program deals with varying copayments, deductibles, and rebates.

(10.) This assumes of course that the OTC and Rx weights are adjusted appropriately in month two after the switch.

(11.) For example, see Berndt and others 2000; Berndt and others 2001.

(12.) If beneficiaries that paid cash prices for part of the year are counted this figure may be as high as 40 percent (Frank and Newhouse 2007).

(13.) For further discussion on details regarding the pharmaceutical PPI, see Berndt, Grilliches, and Rosett (1993).

(14.) About a third of this inflation occurred between June and July 2005.

(15.) About half of this increase occurred between June and July 2006.

(16.) A 41-percent decline in this PPI occurred between December 2000 and January 2001. BLS officials indicate this was due to entry by generic drugs.

(17.) About half of this increase occurred between April and May 2005.

(18.) Most of this increase took place between June and July 2006.

(19.) From table 1, we see that the class of "hormones" has roughly an equal share of around 10 percent for both the elderly and the nonelderly. The hormones class also includes contraceptives, however, which are not generally used by the elderly. Clearly, the hormone class is heterogeneous.

(20.) For example, see Newhouse (2004); Duggan (2005); Frank, et al. (2004).

(21.) Almost all of the September 2004-December 2005 inflation took place between June and July 2005.

(22.) The PPI for antidepressants increased by 19.1 percent between June and July 2005.

(23.) The antidepressant price growth is somewhat surprising. Prozac, the leading selling antidepressant, lost patent protection and experienced generic entry beginning August 2, 2001; yet from table 3, we see that between June 2001 and June 2003, prices in this subclass grew at an average annual rate of almost 11 percent. Similarly, the branded antidepressant Zoloft lost patent protection and experienced generic entry beginning June 30, 2006.

(24.) For example, see Freedman, et al. (2006); Lieberman, et al. (2005); Polsky, et al. (2006); Rosenheck, et al. (2006).

(25.) This research has previously been discussed in greater detail in Frank and Newhouse (2007).

(26.) Included in this group were the branded drugs Aricept, Flomax, Xalatan, Forteo, Coreg, Plavix, Fosomax, Actonel, Norvasc, and Evista.

(27.) This group includes Advair, Prevacid, Nexium, Singulair, Acipbex, Zoloft, Effexor, and Wellbutrin XL (this last drug was dropped from most analyses since a generic version of the molecule was also on the market).

(28.) See Frank and Newhouse (2007) for further details.

Ernst R. Berndt is a professor at the Sloan School of Management at the Massachusetts Institute of Technology. He is also director of the National Bureau of Economic Research Program on technological progress and productivity measurement. Richard G. Frank is a professor at Harvard Medical School and a research associate with the National Bureau of Economic Research.

Table 1. Distribution of Prescriptions by Therapeutic
Class in the Overall and Medicare Part D Samples,
January-October 2006
[Percent]

 Prescription share

 Overall Part D
Therapeutic class sample sample

 Analgesics 8.10 14.63
 Anesthetics 9.67 0.00
 Antidotes 1.16 0.00
 Antimicrobials 9.88 9.76
 Cardiovascular 14.30 17.07
 Central nervous system 11.99 7.32
 Gastrointestinals 5.26 4.88
 Hematologics 1.79 2.44
 Hormones 10.20 9.76
 Immunologics 0.11 0.00
 Metabolics/nutrients 9.57 14.63
 Neurologics 3.47 4.88
 Oncolytics 0.32 0.00
 Ophthalmics 1.47 0.00
 Otics 0.21 0.00
 Respiratory tract 9.04 9.76
 Skin/mucous membrane 2.00 2.44
 Unclassified/miscellaneous 1.47 2.44
Total 100.00 100.01

Table 2. Annual Average Growth Rates
of Alternative Consumer Price Indexes (CPIs)
[Percent]

 Jan. 1996- Jan. 2000- June 2003-
CPI Jan. 2000 June 2003 Sept. 2004

All items-urban 2.250 2.450 2.690
CPI-E-all items-urban 2.404 2.674 2.910
Medical care 3.206 4.324 4.297
CPI-E-medical care 3.158 4.468 4.380
Medical care services 3.201 4.675 4.821
Medical care
 commodities 3.157 3.142 2.677
Prescription drugs 4.132 4.254 3.602

 Sept. 2004- Dec. 2005-
CPI Dec. 2005 Dec. 2006

All items-urban 2.900 2.540
CPI-E-all items-urban 3.003 2.687
Medical care 4.103 3.563
CPI-E-medical care 3.893 3.297
Medical care services 4.498 4.094
Medical care
 commodities 2.913 1.816
Prescription drugs 3.751 1.856

Table 3. Average Annual Growth Rates of
Alternative Producer Price Indexes (PPIs)
[Percent]

 June 2001 June 2003 Sept. 2004
 through through through
PPI June 2003 Sept. 2004 Dec. 2005

All pharmaceuticals 4.23 4.38 5.42
Analgesics--prescription 3.49 3.59 1.36
Antibiotics--broad
 and medium spectrum 3.31 5.31 5.35
Anticoagulants 2.05 4.70 0.22
Antispasmodic/antisecretory 3.75 3.43 22.84
 Other digestive or
 genito-urinary preps 3.73 2.74 2.44
Bronchial therapy 6.23 4.10 1.94
 Other prescription
 respiratory preparations 7.98 6.20 6.94
Cancer therapy products 5.46 0.30 3.71
Other neoplasms,
 endocrine system,
 and metabolic diseases,
 including hormones 10.92 7.29 11.06
Cardiovascular 3.95 4.60 4.04
 ACE inhibitors 1.78 1.70 0.38
 Other cardiovascular 5.74 6.32 6.90
Insulin/antidiabetes 6.09 9.47 6.99
Multivitamins-prescription
 and over the counter (OTC) 0.80 0.81 -0.2
 Other prescription
 vitamins and nutrients 3.88 2.81 3.28
Psychotherapeutics 5.79 6.13 8.69
 Antidepressants 10.99 6.26 14.59
 Other psychotropics,
 including tranquilizers 2.81 6.01 3.85
 Other central nervous
 system and sense organs -5.13 5.74 5.76
Skin prescription preparations 4.19 13.32 3.97

 Dec. 2005 Jan. 2000
 through through
PPI Dec. 2006 Dec. 2006

All pharmaceuticals 3.90 4.12
Analgesics--prescription 4.92 4.41
Antibiotics--broad
 and medium spectrum 3.71 4.63
Anticoagulants 0.09 -5.47
Antispasmodic/antisecretory 5.13 7.33
 Other digestive or
 genito-urinary preps 4.19 n.a.
Bronchial therapy -0.9 3.52
 Other prescription
 respiratory preparations 4.54 n.a.
Cancer therapy products 2.53 4.39
Other neoplasms,
 endocrine system,
 and metabolic diseases,
 including hormones 11.66 8.63
Cardiovascular 3.88 3.90
 ACE inhibitors 0.00 n.a.
 Other cardiovascular 6.63 n.a.
Insulin/antidiabetes -0.99 4.28
Multivitamins-prescription
 and over the counter (OTC) 1.63 1.30
 Other prescription
 vitamins and nutrients 2.90 n.a.
Psychotherapeutics 7.66 5.89
 Antidepressants 10.09 10.21
 Other psychotropics,
 including tranquilizers 5.45 n.a.
 Other central nervous
 system and sense organs 2.61 n.a.
Skin prescription preparations 6.84 n.a.

n.a. Not applicable because the BLS series begins in June 2001.

ACE Angiotensin-converting enzymes