Health Affairs, 23, no. 5 (2004): 217-225 doi: 10.1377/hlthaff.23.5.217 © 2004 by Project HOPE	New Online   Senate Health Reform Bill   Rewarding Providers   Public Option Policy Brief   Health Reform & Abortion   Delivery System Reform

This Article Abstract Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (12) Citing Articles via Google Scholar Google Scholar Articles by Moeller, J. F. Articles by Banthin, J. S. Search for Related Content PubMed PubMed Citation Articles by Moeller, J. F. Articles by Banthin, J. S. Related Collections Pharmaceuticals Health Spending Consumer Issues Elderly

Health Tracking

TRENDS

Looking Inside The Nation’s Medicine Cabinet: Trends In Outpatient Drug Spending By Medicare Beneficiaries, 1997 And 2001

Abstract

 
We examine trends in outpatient prescription drug spending by the Medicare civilian, noninstitutionalized population in 1997 and 2001 using nationally representative data from the Medical Expenditure Panel Survey. We find that the 72 percent increase in drug spending over this period, in excess of price inflation for all goods and services, is primarily attributable to increases in the number of prescriptions per drug user and in the price per prescription. We also find, however, that an increase in the number of users is the primary reason for growth in a number of the fastest-growing subclasses of drugs.

These questions are even more pressing for the Medicare population, since beneficiaries use more prescription drugs than the rest of the population. With the enactment of the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003, which adds coverage for outpatient prescription drugs, policy interest in spending patterns is high. Several recent studies examine trends in outpatient drug spending using private insurance claims databases.² However, data in those studies are limited to people with private group coverage and are not representative of the Medicare population, many of whom do not have private drug coverage.

In contrast to previous studies, this paper uses data that are nationally representative of the noninstitutionalized Medicare population. Our sample includes Medicare beneficiaries of all ages with private and public supplemental benefits as well as those without any prescription drug coverage. We present unique population estimates of the numbers of beneficiaries purchasing prescription drugs, overall and by therapeutic class and subclass, for two recent years, 1997 and 2001. We also analyze trends in spending, average prices, and percentage generic drugs used, overall and by therapeutic class and subclass.

The impact of a new drug is best seen within its therapeutic class, grouped with other drugs used to treat the same conditions, or within the more specific therapeutic subclass, grouped with other drugs that use the same mechanism of action. Furthermore, within therapeutic classes it is possible to measure how many people switch from old to new drugs and whether the new drugs increase the number of users in the therapeutic class.

The analysis of spending patterns by treatment category also could be useful in designing drug benefits, particularly if coverage is designed to vary by treatment category or type of drug. For example, drug formularies and reference pricing (RP) schemes are often based on therapeutic categories. Such provisions can be effective in curbing spending for high-price drugs that fall into the same treatment category as other, lower-price substitutes.

Data And Methods

Top Data And Methods Study Results Discussion Editor's Notes NOTES

 
We used data on a nationally representative sample of prescription drug purchases, collected as part of the Medical Expenditure Panel Survey (MEPS) Household Component (HC) for 1997 and 2001. MEPS, sponsored by the Agency for Healthcare Research and Quality (AHRQ), yields national estimates of medical spending and use, sources of payment, insurance coverage, medical conditions, and health status for the civilian, noninstitutionalized U.S. population.³ MEPS HC collects information on the medication name, number of purchases, and the condition for which the drug was prescribed for each medication purchased during the year for each member of the sampled household. Signed permission forms were required to contact pharmacy providers listed by the household in the MEPS Pharmacy Component (PC) to obtain detailed information about annual household drug purchases, including the National Drug Code (NDC); purchase dates; and the medication name, strength, quantity dispensed, total charge, and payment sources and amounts. A matching program was developed to link the PC drug purchases and related information to the HC drug mentions.⁴ The 1997 and 2001 MEPS HC data yield an unweighted sample of 4,471 and 4,308 beneficiaries and 94,754 and 111,235 drug purchases in each respective year.

Each drug purchase was assigned a therapeutic class, subclass, and active ingredient and brand names using the NDC to link the MEPS prescribed medicine files to the Multum Lexicon database, a product of Cerner Multum Inc. The Multum therapeutic classification system categorizes drugs in a manner designed to replicate the types of organizational schemes used in practice by physicians and pharmacists.⁵ We also used the active ingredient(s) for each drug to merge our files with data from the Food and Drug Administration’s (FDA’s) New Drug Applications (NDA) list to determine the year in which the newest active ingredient in each drug was first approved.

Drug spending for 1997 is expressed in constant dollars by inflating it to 2001 U.S. dollars using the Consumer Price Index for all items averaged across all U.S. cities (CPI-U). All standard errors are adjusted for the complex survey design of MEPS as well as for correlations in observations of individuals across years. We define the Medicare population as all beneficiaries of any age within the civilian, noninstitutionalized U.S. population. We include all people in the MEPS sample with at least six months of Medicare coverage.

Study Results

Top Data And Methods Study Results Discussion Editor's Notes NOTES

 
Average annual use and spending. Medicare beneficiaries’ aggregate drug spending increased from $31.5 billion in 1997 to $54.0 billion in 2001, an increase of 71.6 percent in excess of price inflation for all goods and services (Exhibit 1).⁶ This large increase was driven by a 9.9 percent increase from 1997 to 2001 in the population with use (p.< 10); a 23.6 percent increase in prescriptions per user; and a 26.3 percent increase in the average prescription price.⁷ This increase reflects price increases above general price inflation as well as changes in drug purchases, including the introduction, or expanded use, of higher-price new drugs. More than half of beneficiaries’ drug spending was financed by out-of-pocket payments in both years.⁸

Differences among types of drugs. The rate of spending growth and the relative importance of its three components vary greatly for different types of drugs. The following sections highlight results from Exhibits 2–4 that document these trends for selected therapeutic classes and subclasses.⁹

Hormones Total spending for hormones ranked second (behind cardiovascular drugs) in 1997 and 2001 and more than doubled over this time period (Exhibit 2). The large increase in spending for hormones was largely driven by a 114.4 percent increase in spending for antidiabetic drugs. The most important component in increased spending for antidiabetic drugs was a 74.0 percent increase in the average prescription price (Exhibit 4). Increased purchases of metformin (Glucophage), approved by the FDA in 1995, and the introduction of pioglitazone (Actos), approved in 1999, contributed to this increase.

Gastrointestinal agents. Between 1997 and 2001, spending for gastrointestinal (GI) agents grew at about the same rate as aggregate spending for all drugs. Within the GI class, however, there was a huge shift between subclasses. Spending on proton-pump inhibitors rose by 175.5 percent, while spending on histamine-2 (H2) antagonists showed no growth. Both types of drugs are used to treat the same conditions, including heartburn and ulcers, but proton-pump inhibitors are regarded as superior to H2 antagonists.¹⁰

The introduction of proton-pump inhibitors also appears to have increased the number of users of GI drugs. The number of beneficiaries using GI agents rose 24.9 percent from 1997 to 2001. Users of proton-pump inhibitors rose 141.7 percent, while users of H2 antagonists declined by 24.4 percent (Exhibit 3). By 2001, omeprazole (Prilosec), approved by the FDA in 1989, and lansoprazole (Prevacid), approved in 1995, ranked third and sixth, respectively, in total sales among the Medicare population (Exhibit 5). The expanded use of the higher-price proton-pump inhibitors also contributed to an 11.9 percent increase in the average price per prescription for all GI agents (Exhibit 4).

Psychotherapeutic drugs Spending for psychotherapeutic drugs among Medicare beneficiaries more than doubled from 1997 to 2001 (Exhibit 2). This 118.4 percent increase resulted from a 23.8 percent increase in beneficiaries who used psychotherapeutic drugs, a 24 percent increase in prescriptions per user, and a 42.2 percent ($19) increase in the average prescription price (Exhibits 3 and 4). Increased sales of antidepressants and antipsychotics accounted for nearly 90 percent of this increased spending.

The most important component of increased spending for antipsychotics, which grew by 249.9 percent, was a 108.5 percent increase in the average prescription price (Exhibit 4). Sales exceeding $1.3 billion of the single-source antipsychotic drugs olanzpine (Zyprexa) and risperidone (Risperdal) contributed to the higher prices for psychotherapeutic drugs (data not shown).

The 101.4 percent increase in spending for antidepressants resulted from a 38.6 percent increase in the number of users, a 23.1 percent increase in prescriptions per user, and an 18.0 percent increase in the average price per prescription.¹¹ Over the same period, there was also a slight decline in the percentage of all antidepressant prescriptions that were dispensed as generics (Exhibits 2–4).

Analgesics. Nearly all of the 125.6 percent increase in spending for analgesics was attributable to the emergence of cyclooxygenase-2 (COX-2) inhibitors, which were not approved by the FDA until after the beginning of our study. Celecoxib (Celebrix) and rofecoxib (Vioxx) were approved by the FDA in 1998 and 1999, respectively (Exhibit 5). COX-2 inhibitors grew from no users and no sales in 1997 to 5.5 million users and $2.0 billion in sales in 2001 (Exhibits 2 and 3).

The rapid increase in users of COX-2 inhibitors coincided with an overall increase in the number of beneficiaries who used analgesics and a reduction in the number who used non-steroidal anti-inflammatory agents (NSAIDs) (Exhibit 3). NSAIDs and COX-2 inhibitors are both used to treat pain associated with osteoarthritis, but COX-2 inhibitors are potentially safer than NSAIDs, which may cause adverse GI effects including bleeding and ulcers.¹² The shift from NSAIDs, which are available as generics, to COX-2 inhibitors also contributed to a 50.5 percent increase in the average prescription price for analgesics (Exhibit 4).

Discussion

Top Data And Methods Study Results Discussion Editor's Notes NOTES

 
Over the period 1997 to 2001, stocking the medicine cabinet of the nation’s noninstitutionalized Medicare population became much more costly. The 71.6 percent increase in aggregate drug spending in excess of inflation resulted in large part from increases in the numbers of beneficiaries using certain therapeutic classes and subclasses of drugs. Several of the largest percentage increases in users occurred among subclasses of cardiovascular drugs including anticholesterol drugs, beta-blockers, antihypertensive combinations, and ACE inhibitors. While other subclasses of drugs such as sex hormones, thyroid drugs, antidepressants, and coagulation modifiers had large increases, there was a dramatic increase in the number of users of proton-pump inhibitors and COX-2 inhibitors. Although NSAIDs and H2 antagonists had reductions in the population with use, this reflected a switch from one subclass to another. Much of the increase in users at the class and subclass levels was among beneficiaries who were already using at least one prescription medicine.

Diuretics, beta-blockers, and antidepressants were the only subclasses with statistically significant increases in prescriptions per user. There was, however, significant growth in the number of prescriptions per user at the therapeutic-class level, including cardiovascular drugs, hormones, GI agents, psychotherapeutic drugs, analgesics, and respiratory agents. At this higher level of aggregation, the increases in prescriptions per user reflect, in part, an increase in the number of unique drugs (and the number of subclasses of drugs) used by each beneficiary.

The 26.3 percent growth in average prescription price over all drugs is driven by changes in prices within classes or subclasses and by shifts in drug use across classes and subclasses. Trends in prices within classes and subclasses were mixed, but there were no statistically significant declines in average prices. Antihyperlipidemic agents and proton-pump inhibitors, which had no within-subclass price growth but have high prices relative to the overall average, illustrate the potential for shifts in use across classes to have large effects on aggregate prices.

The therapeutic subclasses with the most rapid spending growth were antihyperlipidemic agents, antidiabetic agents, proton-pump inhibitors, COX-2 inhibitors, anti-psychotics, and antidepressants. In each of these subclasses, spending more than doubled in constant dollars from 1997 to 2001, and by 2001 relatively new, high-price single-source drugs dominated the markets with no or declining competition from generic drugs.

Our estimates of drug use and spending and the changing patterns of use and spending at the therapeutic-class level from the 1997 to the 2001 MEPS are comparable to other estimates from the literature over this period.¹³ Using nationally representative data, our estimates reinforce knowledge of pharmaceuticals’ growing role in health care costs. Examination of trends by therapeutic class and subclass underscores the extent to which prescription drug spending is affected when new drugs are introduced. As heavy users of prescription drugs, who pay out of pocket for a large portion of their drug spending, Medicare beneficiaries are especially sensitive to both the financial costs and the potential benefits of these new drugs.

Our study also points to the need for more information on the relative cost-effectiveness of drugs within clinically significant therapeutic categories. Medicare beneficiaries will want to continue to control their out-of-pocket drug spending even though a drug benefit has been added to their Medicare coverage. Sensible choices among competing drugs require unbiased information on their relative medical effectiveness and cost-effectiveness. Such information should be readily available to both consumers and physicians.

Editor's Notes

Top Data And Methods Study Results Discussion Editor's Notes NOTES

 
John Moeller (JMOELLER6{at}wi.rr.com), formerly director, Division of Modeling and Simulation, Center for Financing, Access, and Cost Trends, Agency for Healthcare Research and Quality, is currently a consulting economist. Ed Miller and Jessica Banthin are in that division.

The views expressed here are the authors’ and do not reflect the views of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services. The authors thank Nancy Kieffer, Fred Rohde, and Jun Tian of Social and Scientific Systems for their invaluable programming support.

NOTES

Top Data And Methods Study Results Discussion Editor's Notes NOTES

 

1. K. Levit et al. "Inflation Spurs Health Spending in 2000," Health Affairs 21, no. 1 (2002): 172–181.[Free Full Text]
2. C.P. Thomas, G. Ritter, and S. Wallack, "Growth in Prescription Drug Spending among Insured Elders," Health Affairs 20, no. 5 (2001): 265–277[Abstract/Free Full Text]; R. Dubois et al., "Explaining Drug Spending Trends: Does Perception Match Reality?" Health Affairs 19, no. 2 (2000): 231–239[Abstract]; and M. Merlis, Explaining the Growth in Prescription Drug Spending: A Review of Recent Studies, Report for USDHHS Conference on Pharmaceutical Pricing Practices, Utilization, and Costs (Washington: Institute for Health Policy Solutions, 2000).
3. S.B. Cohen, Sample Design of the 1996 Medical Expenditure Panel Survey Household Component, MEPS Methodology Report no. 2, Pub. no. 97-0027 (Rockville, Md.: Agency for Healthcare Research and Quality, 1997).
4. J.F. Moeller et al., Outpatient Prescription Drugs: Data Collection and Editing in the 1996 Medical Expenditure Panel Survey, Pub. no. 01-0002 (Rockville, Md.: AHRQ, June 2001).
5. Approximately 4.5 percent of drug purchases linked to more than one therapeutic class and were assigned to a unique class using a combination of condition information and random assignment. About 1 percent of drug purchases could not be linked to Multum.
6. The percentage growth in spending is equal to the growth in the population with use times the growth in prescriptions per user times the growth in the average prescription price. For example, for aggregate drug expenditures: 1.716 = (1.099) x (1.236) x (1.263).
7. All other results mentioned in the text are significant at the .05 level unless otherwise noted. The price measure is technically not a pure measure of price because prescription prices are not standardized for days supply and strength. See Merlis, Explaining the Growth.
8. We do not present results on out-of-pocket payments by therapeutic class. We find that for less (more) costly drug classes, a relatively high (low) percentage is financed by out-of-pocket payments. This is probably attributable to the presence of fixed copayments for drugs. We also find that the percentage paid out of pocket within therapeutic classes and subclasses was stable from 1997 to 2001.
9. Therapeutic classes that accounted for less than 3 percent of spending were not included in the exhibits. The fifteen subclasses were all in the top ten in purchases or spending, or both, in at least one year, 1997 or 2001.
10. B.T. Vanderhoff and R.M. Tahboub, "Proton Pump Inhibitors: An Update," American Family Physician 66, no. 2 (2002): 273–280.[Web of Science][Medline]
11. We find evidence of considerable off-label use of psychotherapeutic drugs among Medicare beneficiaries (for example, use of valium to treat a sleep disorder or pain, rather than to treat anxiety).
12. J.A. Doshi, N. Brandt, and B. Stuart, "The Impact of Drug Coverage on COX-2 Inhibitor Use in Medicare," Health Affairs, 18 February 2004, content.healthaffairs.org/cgi/content/abstract/hlthaff.w4.94 (3 June 2004).
13. See, for example, S. Foote and L. Etheredge, "Increasing Use of New Prescription Drugs," Health Affairs 19, no. 4 (2000): 165–170[Medline]; Merlis, Explaining the Growth; Thomas et al., "Growth in Prescription Drug Spending"; and C.W. Burt, "National Trends in Use of Medications in Office-Based Practice, 1985–1999," Health Affairs 20, no. 4 (2002): 206–214.[CrossRef]

                      What's this?

This article has been cited by other articles:

				D. J. Cobaugh, E. Angner, C. I. Kiefe, M. N. Ray, C. L. LaCivita, N. W. Weissman, K. G. Saag, and J. J. Allison Effect of racial differences on ability to afford prescription medications Am. J. Health Syst. Pharm., November 15, 2008; 65(22): 2137 - 2143. [Abstract] [Full Text] [PDF]

				W. Wei, A. Akincigil, S. Crystal, and U. Sambamoorthi Gender Differences in Out-of-Pocket Prescription Drug Expenditures Among the Elderly Research on Aging, July 1, 2006; 28(4): 427 - 453. [Abstract] [PDF]