Potential Savings from Substituting Generic Drugs for Brand-Name Drugs: Medical Expenditure Panel Survey, 1997–2000 FREE

The cost of prescription drugs is of great concern to Americans. The substitution of cheaper generic drugs for more expensive brand-name drugs might reduce prescription drug costs.

Using data from the 1997–2000 Medical Expenditure Panel Survey Household Component, the researchers estimated that substitution of a generic for a brand-name drug whenever available would have saved approximately $46 per year for adults younger than 65 years of age and approximately $78 per year for older adults.

While the per capita savings of generic substitution appear modest, national savings would be substantial: about $6 billion for adults younger than age 65 years and about $3 billion for older adults.

–The Editors

Prescription drug spending is increasing at a rate of over 10% per year and currently represents 11% of all health care expenditures (1). In 2001, expenditures for prescription drugs in the United States were $141 billion (1). The passage of a Medicare prescription drug benefit has resulted in much debate about the cost of this coverage. The program is designed to offer $410 billion in new drug benefits over a 10-year period (2). Containment of drug spending will be central to the stability of this benefit as well as health care expenditures in general.

Generic drugs are typically less expensive than brand-name drugs, and prices for generics have historically increased less than those for brand-name drugs (3). The U.S. Food and Drug Administration examines generic formulations and approves them as bioequivalent to brand-name drugs in safety, strength, and quality (4). Although the bioequivalence of some drugs is controversial (5–7), generic drugs are widely believed to provide the same therapeutic effects as their brand-name alternatives (4).

Use of a generic formulation instead of a brand name for multisource drugs (that is, those with ≥1 generic available) could be one mechanism for limiting drug expenditures. Since the 1980s, almost every state has enacted laws to allow and in some cases mandate generic substitution (8). Patients, physicians, and pharmacists may also influence whether a generic is dispensed for a multisource product (9–15). A study of Medicaid prescription drug spending in 2000 found that $229 million, approximately 1% of the total amount reimbursed for prescription drugs by Medicaid, could have been saved by wider use of generic drugs (16). Despite the importance of this issue, we know of no estimates of the potential savings associated with greater generic substitution in populations other than Medicaid beneficiaries.

The objectives of this study were to estimate the use of multisource drugs, specifically generic formulations, among a nationally representative sample of adults. We also estimated the potential savings associated with broad substitution of generic drugs for all multisource products. Because of their implications for the Medicare drug benefit, our findings are stratified by age.

This analysis is based on data from the 1997-2000 Medical Expenditure Panel Survey Household Component (MEPS-HC), which involved a nationally representative sample of the civilian, noninstitutionalized U.S. population and was conducted by the Agency for Healthcare Research and Quality (17). The MEPS-HC sample is drawn from respondents to the previous year's National Health Interview Survey (NHIS). The NHIS uses a stratified, multistage probability cluster sampling design to obtain a representative sample of the U.S. population and oversamples persons of African-American and Hispanic ethnicity (18). The MEPS-HC data include sampling weights that reflect the sampling frame and adjustments for household nonresponse. The MEPS-HC provides data on demographic characteristics, health status, insurance coverage, and the utilization of health care services for all individuals in sampled households. Household respondents provided information on the names of all outpatient medications used by each household member and the names and locations of the pharmacies where each medication was obtained (19). They were also asked for permission to request records from these pharmacies. Pharmacy providers were asked to provide the data necessary to assign a National Drug Code, which is specific for manufacturer, ingredients, strength, package size, quantity dispensed, total charge, and sources of payment. The linkage rate between information provided by respondents and data obtained from pharmacies was 67% in 1997, 79% in 1998, 79% in 1999, and 77% in 2000. The Agency for Healthcare Research and Quality performed detailed matching, imputation, consistency checks, sensitivity checks, and reconciliation algorithms (19).

Data for each drug from the First DataBank National Drug Data File were merged with the MEPS data by using the National Drug Code. The National Drug Data File indicates whether a drug is available only as a brand name or as a multisource product; generic status is assigned annually. The National Drug Data File also includes a categorization of the therapeutic class and the average wholesale price for each product. The average wholesale price is the manufacturer's suggested list price for a wholesaler to charge a pharmacy and is typically higher than a pharmacy's actual acquisition cost, particularly for brand-name drugs (3).

Adults who were older than 18 years of age and used at least 1 outpatient multisource product were included in this analysis. Brand-name drugs were included in the analysis only if a generic alternative was available in an identical strength and form. A median of 4 prescriptions per person (range, 1 to 60 prescriptions per person) was included in this sample.

Because retail prices for medications vary widely and are not uniformly available, we used data from MEPS about the total expenditures for each drug (that is, the amount spent by an individual out-of-pocket plus the amount paid by any insurance coverage). We then calculated the mean total cost per unit (that is, per tablet, tube, vial, or dropper bottle, as appropriate) for each of the 7056 products used by individuals in our sample (that is, we calculated the average per unit cost for all occurrences of each product). The availability of the actual drug expenditures for each product is an improvement over previous estimates of drug costs, which used a standard discount of the average wholesale price because actual expenditures were not available (for example, estimating that all generic drugs cost 75% of the average wholesale price) (20–21). We then estimated the annual savings that would result if each person in the sample switched from a brand-name drug to a corresponding generic formulation, using the following equation:

Annual savings = Σ (total units of brand-name product dispensed per year) × (average per unit expenditure for brand-name product – average per unit expenditure for identical generic product).

Using data from 2000, we calculated overall savings from a population perspective (for example, total saved for all people ≥65 years of age) and as median annual savings per person (for example, out-of-pocket savings plus savings to any insurance plan per person). Calculations were done with SAS, version 8 (SAS Institute, Inc., Cary, North Carolina); the sampling weights described earlier were used to account for differences in the probability of sample selection as a result of the clustered design, planned oversampling, and nonresponse. The weighted results therefore represent estimates for the noninstitutionalized U.S. population.

This work was supported by the Agency for Healthcare Research and Quality (P01 HS10771 and P01 HS 10856). The National Cancer Institute provided partial support (R01 CA 10184). The funding sources had no role in the design, conduct, or reporting of the study.

Fifty-six percent of all prescription medications (345  781 observations that were weighted to represent 3 686 700 000 prescriptions from 1997 to 2000) were multisource drugs, accounting for 41% of total prescription drug expenditures. Of these drugs, 61% were dispensed as a generic. Generic use increased over the course of the study, from 58% in 1997 to 64% in 2000. In 2000, total medication expenditures for adults in MEPS were $53 billion for those younger than age 65 years and $27 billion for those at least 65 years of age.

Among adults who were dispensed a multisource drug in MEPS, approximately 75% were younger than age 65 years and approximately 25% were at least 65 years of age (Table 1). Most individuals in both age groups were women and described their race or ethnicity as white. Among individuals younger than 65 years of age, approximately half did not receive education beyond high school; the median household income was $54 761. While most of these working-age adults had employer-sponsored insurance, 14.4% were uninsured and 7.6% were covered by Medicaid or another public program. Most did not report any comorbid conditions. The median annual number of prescriptions dispensed was 3.

Among individuals at least 65 years of age, 68.9% did not receive education beyond high school; the median household income was $26 274. Approximately 33% had only Medicare coverage, 57.1% had additional private coverage, and 9.5% had additional public coverage, including Medicaid. Almost half reported at least 1 comorbid condition. The median annual number of prescriptions dispensed was 8.

For individuals younger than 65 years of age, the most common therapeutic classes of multisource drugs dispensed were psychiatric, analgesic, and cardiac (Table 2). Among individuals at least 65 years of age, cardiac drugs were the most common drugs dispensed, followed by diuretics and psychiatric drugs. Overall, 60.7% of multisource prescriptions were dispensed as generics. Thyroid medications were most likely to be dispensed as generics, and contraceptives were least likely. Broad generic substitution of antihistamines and cardiac, gastrointestinal, and hypoglycemic agents were each associated with a potential annual savings of more than $1 billion.

For adults younger than age 65 years, the median annual per person savings associated with broad generic substitution was $45.89 (interquartile range, $10.35 to 158.06) (Table 3). This savings would be shared between an individual and his or her health plan. Median per person savings increased with age and number of chronic conditions, and men saved more than women. Median per person savings were largest for low-income individuals and varied by insurance status.

For adults at least 65 years of age, the median annual per person savings associated with broad generic substitution was $78.05 (interquartile range, $19.94 to $241.72). Median per person savings increased with the number of chronic conditions and varied by insurance status.

The estimated national savings associated with widespread generic substitution in 2000 was $5.9 billion (95% CI, $5.5 billion to $6.2 billion) per year for adults younger than 65 years of age (11.1% of all drug expenditures) and $2.9 billion (CI, $2.6 billion to $3.1 billion) per year among those at least 65 years of age (10.7% of drug expenditures). From a national perspective, the potential annual savings associated with generic substitution among adults younger than 65 years of age was $4.1 billion for those with employer-sponsored insurance and $388 million for those with Medicaid or public coverage. For adults at least 65 years of age who were dually eligible for Medicare and Medicaid, the potential savings was $1.7 billion per year.

These findings suggest that broad generic substitution of outpatient prescription drugs could save approximately $8.8 billion, or approximately 11% of drug expenditures for adults in this sample, in the United States each year. For individuals at least 65 years of age, who are expected to be eligible for the new Medicare prescription drug benefit, the annual savings is estimated to be $2.9 billion, or approximately 10% of the annualized cost of this coverage after accounting for inflation (2). Despite these considerable societal savings, absolute per person savings would be small.

Although broad generic substitution would only modestly reduce national drug expenditures, the absolute savings, particularly at a time when employers and public programs are struggling with drug expenditures, are significant. Broad dispensing of generic products would achieve savings without compromising safety. Generic drugs are believed to provide therapeutic effects similar to those of their brand-name alternatives (4). The standards and regulations for manufacturing generic products are the same as for brand-name drugs (22). Recent efforts to remove barriers to generic drug approval suggest that policymakers see an opportunity to control expenditures through greater use of generic drugs (4, 23). However, some have argued that greater generic substitution could limit drug development by pharmaceutical companies because of decreased revenue (24).

Using aggregate data for Medicaid drug payments from 48 states in 2000, Fischer and Avorn (16) estimated that approximately $229 million (approximately 1% of total drug expenditures) could have been saved through wider generic substitution. These data did not include information on the 56% of Medicaid beneficiaries who were covered by a managed care plan in 2000 (25). Our estimate of the savings for persons covered solely by Medicaid was higher, perhaps because our sample included individuals covered by a managed care plan. Differences in the estimates may also be related to differences in the calculations of the cost of prescription medications. Fischer and Avorn's estimate was based on the lowest generic price for each product paid by Medicaid in a particular state, whereas ours was based on the average total cost reported by individuals in MEPS nationwide. Our analysis also estimated savings for individuals with other types of health coverage. Medicaid beneficiaries are likely to use a different array of brand-name and multisource drugs than adults with other forms of health coverage.

Choosing a brand-name drug when a generic is available may be influenced by a variety of factors. State laws regulate generic substitution. While most states permit pharmacists to substitute a generic unless directed by the physician or patient, a minority of states mandate that a pharmacist substitute a generic unless overridden by a physician's order (8). Health plan policies may facilitate the use of generic drugs. However, although tiered formularies may encourage the use of generics, they may also lead to greater out-of-pocket expenditures and even to the discontinuation of long-term medications (26). Reference pricing may encourage greater use of generic drugs by offering reimbursement for the lowest-priced therapeutically equivalent drug (27). Some patients may believe that a brand-name formulation is superior (9–10), perhaps because direct-to-consumer advertising influences patient beliefs about medications (28). Previous research has shown that physicians and pharmacists play an important role in the decision to choose a brand-name or generic formulation of a drug (11–15).

Our analysis does not include the substitution of drugs within a class (that is, therapeutic substitution). We included only brand-name drugs with an identical generic. Therapeutic substitution of angiotensin-converting enzyme inhibitors has been shown to reduce expenditures without adverse clinical outcomes (29). Future research should more broadly examine the effect of therapeutic substitution on clinical outcomes and expenditures.

Our study has several additional limitations. Because retail drug pricing in the United States varies widely and is proprietary, there are no nationwide data to calculate actual costs or savings. For this reason, we used costs reported in MEPS. We do not have information about formularies or copayments required by health plans for specific products, both of which may influence the utilization of specific medications (26, 30). We also cannot calculate how much of the potential savings would benefit individuals or their health plans. Regardless of the distribution, however, we believe that the savings are important from a societal perspective because they result from curtailing unnecessary expenditures for drugs that have an identical generic available.

Our study examined the potential savings associated with generic substitution through 2000. In subsequent years, several “blockbuster” drugs have lost patent protection (31–32). Because detailed pharmacy data were available for 67% to 79% of the MEPS sample (depending on the year), our findings underestimate total drug expenditures in the United States. We do not know whether MEPS respondents who used a multisource drug were more or less likely to be matched with a pharmacy claim. Finally, we did not examine the potential savings associated with greater generic substitution among children.

These data provide per person and national estimates of the potential savings associated with broad generic substitution of outpatient prescription drugs for adults. Interventions to stimulate competition in the generic market, to reduce the approval times for generic drugs by the U.S. Food and Drug Administration, and to limit opportunities to extend the patent life of brand-name drugs could increase the potential savings (33–34). Greater use of generic medications could result in important health care savings in the United States while maintaining quality of care.