Health Affairs, 23, no. 5 (2004): 226-232 doi: 10.1377/hlthaff.23.5.226 © 2004 by Project HOPE	New Online   How Would Obama, McCain Cover The Uninsured?   Debating Cost Of Uninsured   Try Medicare-For-All   HA Blog Top 10

This Article Abstract Figures Only Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when eLetters 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 ISI 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 Google Scholar Google Scholar Articles by Shaya, F. T. Articles by Mullins, C. D. Search for Related Content PubMed PubMed Citation Articles by Shaya, F. T. Articles by Mullins, C. D. Related Collections Business Of Health Consumer Issues Insurance Coverage - Children Pharmaceuticals Health Spending

Health Tracking

TRENDS

Prescription Drug Spending Trends For The Privately Insured In Maryland, 2000–2001

Abstract

 
Out-of-pocket drug spending per patient rose 25 percent in 2001 for the privately insured nonelderly in Maryland, outpacing the 10 percent increase in total drug spending per patient. Prescription drug use patterns changed little in response to a 30 percent rise in copayments, with the number of prescriptions per patient decreasing by 3 percent, and the proportion of generic drugs remaining steady. About 23 percent of patients spent less than $50 on drugs in 2001, while 21 percent spent more than $700. Findings suggest that a greater proportion of the economic burden of prescription drugs in coming years will be borne by those who use them.

The long-term effects of increased patient cost sharing on health care use, outcomes, and insurance coverage represent major health policy concerns. Policymakers and plan administrators recognize that if the economic burden for patients is too high, the resulting potential decrease in drug use may come at the expense of patients’ health and downstream health care costs.³ At the same time, there is increased interest in greater patient participation in health care decision making and financing.⁴

We analyzed the 2000 and 2001 drug usage and payments of privately insured people in Maryland to ascertain trends in cost sharing among the nonelderly. Maryland is one of the few states to require claims-level information on prescriptions from insurers. Plans and employers are not identified by the carrier providing the data; hence, proprietary concerns are minimized. Although not representative of the entire country, a unique broad cross-section of financial data from private insurers were available for this analysis, which may provide new insights about the privately insured.

Data And Methods

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

 
Study sample. We analyzed thirty million prescription drug claims filed by 1.5 million privately insured, nonelderly Marylanders in 2000 and 2001. Maryland requires insurers and managed care organizations to submit annually each prescription claim made on behalf of a Maryland resident.⁵

Data from plans where the employer contracts directly with a pharmaceutical benefit manager (PBM) are unavailable because the insurer does not manage the drug benefit. Plans where the insurer contracts with a PBM are included. Nationwide, it was estimated that 32 percent of workers were covered by such employer "carve-out" plans in 2001.⁶

Data elements. The financial data include both the amounts paid for a given prescription by the insurer and the patient’s out-of-pocket cost. A breakdown between patients’ copayments, coinsurance, and deductible amounts is not available. Patient information includes age, sex, and ZIP code. Drug information includes the fill date, formulation, and whether the product was generic or brand-name.

A cross-sectional analysis was conducted at both the prescription and patient-year levels. Study methods included an examination of time trends and characteristics of patients, expenditures, and patients’ out-of-pocket share, including totals, means, medians, and distributions.

A separate analysis was performed on drugs newly approved by the U.S. Food and Drug Administration (FDA) during 1997–2001.⁷ Furthermore, we identified sixteen different high-volume or high-cost drug therapies, by referencing utilization studies and consulting with health care plans, which accounted for 57.9 percent of claims and 58.4 percent of spending in our sample.⁸

Study Results

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

 
Study population. Prescription claim records provide information on 1.46 million insured people with at least one claim in 2000 and 1.42 million such people in 2001. The difference in numbers is likely attributable to changes in enrollment or missing data; a breakdown of patients by the thirty payers submitting data shows considerable turnover for each payer. Because personal identifiers could not be linked across years, and no data were available on the total number of insured people, we cannot determine whether the percentage of people using their drug benefit changed from one year to the next. Thus, our conclusions are limited to prescription drug users, hereafter referred to as "patients," and are presented on a per patient basis.

The total number of privately covered, non-elderly people in Maryland is estimated at 3.7 million in 2001.⁹ Applying the national percentage, 62 percent, of privately insured, nonelderly people who use their prescription benefit, and excluding the national percentage, 32 percent, of those who are covered by their employer through PBM carve-out plans, we would have expected records on about 1.6 million patients.¹⁰ This number is consistent with the 1.4 million patients observed in our sample.

Nearly 29 percent of patients in the sample were ages 45–64 in 2000, and nearly 24 percent were under age 18 (Exhibit 1). The comparable proportions found in a national random sample, the Medical Expenditure Panel Survey (MEPS), are similar: 32 percent and 26 percent, respectively, of nonelderly people with a drug expense.¹¹ Fifty-eight percent of our patients are female, which is also the proportion found in MEPS. The age and sex distribution was the same in both years. The distribution of prescriptions by therapeutic category was also stable, with the proportions differing by less than a percentage point from 2000 to 2001.

Copayments. The median copayment for brand-name drugs increased by 50 percent, from $10 in 2000 to $15 in 2001. Copayments per prescription for generic drugs rose an average of 18 percent, as some insurers raised these copayments from $5 to $7 or $8. Across both brand-name and generic drugs, the average copayment per prescription went up 29 percent, from $10.50 to $13.50. Meanwhile, the total costs per prescription, including insurers’ and patients’ payments, rose by 13 percent. Brand-name drug costs per prescription rose 12 percent (from $72 to $81); interestingly, generic drug costs per prescription increased by twice as much on a relative basis (from $15 to about $18).

Drug usage. There were few changes in prescription drug usage during 2000–2001 that might have contributed to the increase in total cost. In fact, the number of prescriptions per patient dropped by 3 percent, from a mean of 9.75 to 9.43 prescriptions per year. Use of generic drugs accounted for 36 percent of prescriptions in both years and did not increase in any one of the sixteen high-volume/cost therapeutic categories. Brand-name drugs accounted for 90 percent of total spending in 2000 and 2001; drugs on the market less than four years accounted for 9 percent of spending in both years.

Distribution of drug spending. Prescription drug spending went up dramatically with patients’ age, at a rate of 30–50 percent per decade (Exhibit 2). The magnitude of the growth rate is consistent with national random-sample data.¹² These increases were mainly associated with increases in use, but also with higher costs per prescription. For example, patients in their early fifties used twice as many drugs as patients in their early thirties; their cost per prescription was 17 percent higher as well. Six percent of their prescriptions were for newer drugs approved in the previous four years, while for patients in their early thirties, 4 percent were for newer drugs. Use of generic drugs (at 36 percent of prescriptions) did not generally vary with age. However, 45 percent of prescriptions for children under age five were for generics.

View larger version (19K): [in this window] [in a new window]   EXHIBIT 3 Distribution Of Maryland Nonelderly Patients’ Annual Out-Of-Pocket Spending For Prescription Drugs, 2001

The lowest-spending patients were generally paying a high percentage share of their expenses out of pocket. Patients with less than $50 in expenses paid a median of 86 percent (mean 73 percent) of their total bill. Most of these patients received just one or two inexpensive prescriptions, which frequently did not exceed the required copayment. At the other extreme, patients with $5,000–$10,000 in annual total payments were liable for a median 11 percent (mean 13 percent) of their total expenses. The median out-of-pocket share of the entire sample was 34 percent.

Discussion

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

 
Effect of increased copayments. From 2000 to 2001, the out-of-pocket prescription drug expenses of privately insured Marylanders rose 25 percent per patient as total drug spending increased 10 percent. This finding is not surprising, given recent payer efforts to increase consumer cost sharing. However, the magnitude of the rise is substantial. For comparison, national spending data for 2001 from the Centers for Medicare and Medicaid Services (CMS) show accelerated growth in out-of-pocket drug spending, from 11 percent in 2000 to 13 percent in 2001, with total drug spending rising by 16 percent.¹³ Because spending by the noninsured is included in the consumer out-of-pocket spending total, these data understate the effect of changes in copayments by the insured.

The 30 percent increase in average copayment per prescription was concurrent with a 3 percent decline in per patient use of both brand-name and generic drugs, while the share of generic drugs remained the same from 2000 to 2001. By comparison, a major PBM found a 6 percent increase in prescriptions per enrollee in 2001.¹⁴ The same per capita growth was found in an analysis of nationwide samples of retail and mail-order sales in 2001.¹⁵ Taken together, our data and the other cross-sectional aggregate data suggest that, overall, increasing copayments may not greatly slow utilization.

Demand might have been even higher without the increased copayments, considering other factors driving growth, such as more aggressive management of chronic disease and drug advertising.¹⁶ Also, data were not available to control for turnover in employers and plans or the possible movement of patients to plans most favorable to their utilization patterns (that is, selection bias). More detailed, longitudinal studies of cost sharing that control for such factors have been performed. For example, RAND retrospectively modeled the effects of copayments on nonelderly members of fifty-five large-employer plans in 1997–1999. Results showed that a doubling of copayments in either two- or three-tier plans reduced utilization by a third, after adjusting for such factors as members’ age, health, and copayments to physicians.¹⁷

There is evidence from a Canadian study showing that increased copayments were associated with a reduction in the use of essential drugs. In turn, this reduction led to greater use of hospital emergency departments or increased hospitalizations, thus potentially raising overall health care costs and making a negative impact on health outcomes.¹⁸ For nonelderly and higher-income patients, the limited number of studies to date have not found a direct link between drug cost sharing and health outcomes.¹⁹

Implications of spending distributions. A small portion of patients accounted for a large portion of total spending. The higher spenders are disproportionately older (the oldest 9 percent in our sample accounted for 20 percent of total expenses).

Patients’ out-of-pocket burden generally increases proportionately with their total expense, with interesting exceptions at the extremes. At the lower end, the half of the population buying the fewest and least expensive drugs paid 86 percent of their total expenses out of pocket, in large part because of the fixed copayment structure. At the higher end, patients who incurred the highest drug expenses usually paid a relatively low share of their total expenses out of pocket. Only about 1 percent of the patients in our study paid more than $1,000 out of pocket in 2001.

The skewness of the spending distribution and its strong correlation with age have implications for how drug plan designs or regulations should allow enrollees to choose among various levels of coverage. Introducing a $250 deductible, for instance, would effectively require 58 percent of the patients in our sample to pay all of their drug expenses in 2001. The deductible would not affect chronic users, who are already paying more than $250 out of pocket. The savings to the payer are not substantial—about 15 percent of total costs for our study population.²⁰ However, a deductible may cause many healthy employees to opt for another plan, their spouse’s coverage, or no coverage at all. This raises concerns about possible adverse selection.

Our findings run counter to expectations that recent increases in cost sharing will greatly slow growth in prescription drug use. Rather, our results suggest that a greater proportion of the economic burden of prescription drugs will be borne by those who use them.

Editor's Notes

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

 
Fadia Shaya (fshaya{at}rx.umaryland.edu) is an assistant professor in outcomes research and epidemiology and an associate director of the Center on Drugs and Public Policy at the University of Maryland (UM) School of Pharmacy. Steven Blume is a research analyst at the center. Daniel Mullins is chair of the Pharmaceutical Health Services Research Department, a professor of pharmacoeconomics, and an associate director of the center, at the UM School of Pharmacy.

This research was supported by a grant from the Maryland Health Care Commission to the Center on Drugs and Public Policy at the University of Maryland, Baltimore. The authors gratefully acknowledge Alan Lyles for his input throughout the study, Linda Bartnyska and Ben Steffen of the commission for their advice and insight, and several anonymous reviewers for their suggestions and comments. Preliminary results from this work were presented to the Maryland Health Care Commission, 17 July 2003, and the commission published highlights from this work in its instate, quarterly newsletter, "Spotlight on Maryland" (September 2003).

NOTES

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

 

1. Centers for Medicare and Medicaid Services, "Table 2: National Health Expenditures Aggregate Amounts and Average Annual Percent Change, by Type of Expenditure, Selected Calendar Years 1980–2002," www.cms.gov/statistics/nhe/historical (17 June 2004).
2. Henry J. Kaiser Family Foundation and Health Research and Educational Trust, Employer Health Benefits: 2003 Annual Survey (Menlo Park, Calif.: Kaiser Family Foundation, 2003).
3. D. Steinwachs, "Pharmacy Benefit Plans and Prescription Drug Spending," Journal of the American Medical Association 288, no. 14 (2002): 1773–1774[Free Full Text]; and U.E. Reinhardt, P.S. Hussey, and G.F. Anderson, "U.S. Health Care Spending in an International Context," Health Affairs 23, no. 3 (2004): 10–25.[Abstract/Free Full Text]
4. J.C. Robinson, "Renewed Emphasis on Consumer Cost Sharing in Health Insurance Benefit Design," Health Affairs, 20 March 2002, content.healthaffairs.org/cgi/content/abstract/hlthaff.w2.139 (17 June 2004); J.K. Iglehart, "Changing Health Insurance Trends," New England Journal of Medicine 347, no. 12 (2002): 956–962[Free Full Text]; and K. Levit et al., "Trends in U.S. Health Care Spending, 2001," Health Affairs 22, no. 1 (2003): 154–164.[Abstract/Free Full Text]
5. Code of Maryland Statute 10.25.06 (Baltimore: Maryland Health Care Commission, Analysis and Data Systems Division, 15 January 1996).
6. Kaiser/HRET, Employer Health Benefits: 2001 Annual Survey (Menlo Park, Calif.: Kaiser Family Foundation, 2001).
7. U.S. Food and Drug Administration, Center for Drug Evaluation and Research, "New Molecular Entity Reports," www.fda.gov/cder/rdmt (17 June 2004); and Center for Biologics Evaluation and Research, "Products: Approvals by Year since 1996," www.fda.gov/cber/products.htm (17 June 2004).
8. The sixteen categories were antibiotics, abortive agents for acute migraine, short-acting beta-2 agonists for asthma, systemic corticosteroids, cold and cough preparations, proton pump inhibitors, statins, angiotensin receptor blockers, selective serotonin reuptake inhibitors, sedative hypnotics, hormone replacement therapy and oral contraceptives, asthma controller drugs, HIV drugs, narcotic analgesics, nonsedating antihistamines, and nonsteroidal anti-inflammatory drugs. The Multum Drug Dictionary (Denver: Multum, 2002) was the source of active ingredient and drug category data.
9. Kaiser Family Foundation, "Maryland: Distribution of Nonelderly 0–64 by Insurance Status, State Data 2001–2002, U.S. 2002," State Health Facts Online, statehealthfacts.kff.org (17 June 2004).
10. Agency for Healthcare Research and Quality, MEPS, "Household Medical Expenditures, Table 2, Prescription Medicines 2000," July 2003, in MEPS Household Compendia of Tables, www.meps.ahrq.gov/CompendiumTables/TC_TOC.htm (17 June 2004). Of the privately insured nonelderly, 61.7 percent are reported to have a prescription expense. Express Scripts, Drug Trend Report 2001 (St. Louis: Express Scripts, 2002), 6. Of the insured whose prescription benefit was managed by Express Scripts as the PBM, 61.5 percent used their benefit in 2000, 62.9 percent in 2001.
11. Calculated from AHRQ data on age and sex distribution and percentage with an expense. Includes all insured (both public and private) and uninsured people.
12. Ibid.
13. CMS, "Table 3: National Health Expenditures, by Source of Funds and Type of Expenditure" (see Note 1).
14. Express Scripts, Drug Trend Report 2002 (St. Louis: Express Scripts, 2003), 12.
15. National Institute for Health Care Management, Prescription Drug Expenditures in 2001: Another Year of Escalating Costs (Washington: NIHCM, 6 May 2002).
16. Ibid.
17. G. Joyce et al., "Employer Drug Benefit Plans and Spending on Prescription Drugs," Journal of the American Medical Association 288, no. 14 (2002): 1733–1739[Abstract/Free Full Text]; B. Motheral and K.A. Fairman, "Effect of a Three-Tier Prescription Copay on Pharmaceutical and Other Medical Utilization," Medical Care 39, no. 12 (2001): 1293–1304[CrossRef][ISI][Medline]; and S. Kamal-Bahl and B. Briesacher, "How Do Incentive-Based Formularies Influence Drug Selection and Spending for Hypertension?" Health Affairs 23, no. 1 (2004): 227–236.[Abstract/Free Full Text]
18. R. Tamblyn et al., "Adverse Events Associated with Prescription Drug Cost-Sharing among Poor and Elderly Persons," Journal of the American Medical Association 285, no. 4 (2001): 421–429.[Abstract/Free Full Text]
19. S. Schneeweiss et al., "Outcomes of Reference Pricing for Angiotensin-Converting-Enzyme Inhibitors," New England Journal of Medicine 346, no. 11 (2002): 822–829.[Abstract/Free Full Text]
20. Assumes that the deductible is per patient, not per family. Also assumes that all patient liability in our 2001 data was made up of copayments or coinsurance, not deductibles.

                      What's this?