Skip Navigation

How Increased Competition from Generic Drugs Has Affected Prices and Returns in the Pharmaceutical Industry July 1998

This study draws on several different sets of data that cover sales revenues, prices, and quantities for prescription drugs sold in the United States (see Table A-1 for an overview). The data come from two private companies that collect and sell information about the pharmaceutical industry (Scott-Levin and IMS America), from three government agencies (the Food and Drug Administration, the Patent and Trademark Office, and the Health Care Financing Administration), and from Henry Grabowski, an economist at Duke University.
 

Retail Pharmacy Data Set

Many of the estimates in Chapter 3 rely on a set of retail pharmacy data purchased from Scott-Levin. That data set covers the number of prescriptions dispensed at retail pharmacies in 1993 and 1994 for all formulations of all prescription drugs in 66 narrowly defined therapeutic classes, as well as the revenues from sales of those drugs, valued at retail prices. (Those retail prices are the average of the actual retail transaction prices charged by pharmacies.) The total value of sales revenues in the data set equals approximately 70 percent of the total sales revenues of retail pharmacies in the United States from prescription drugs. The data set is based on Scott-Levin's Source Prescription Audit, which covers more than 34,000 U.S. retail pharmacies. Scott-Levin projects the sales data upward to reflect sales through all pharmacies in the United States (which numbered 67,939 in 1995).⁽¹⁾ Since retail pharmacies distribute roughly half of the value of prescription drugs, this data set represents approximately 35 percent of the value of all prescription drug sales in the nation.

The data are broken down by each dosage form of each drug in the 66 therapeutic classes. For example, if a multiple-source drug comes in both 50 milligram and 100 milligram tablets, the data set includes the sales revenues and number of prescriptions for each brand-name and generic manufacturer (if there are any) of both of those dosage forms. The set contains 454 different prescription drugs (or chemical entities), 177 of which are multiple source. Expanding that by the different dosage forms for each drug--many of which are produced by several manufacturers--brings the number of individual observations in the data set to 11,665. The Congressional Budget Office (CBO) added the chemical names of the brand-name drugs (using the reference book Drug Facts and Comparisons) and coded each observation so the generic drugs could be matched with their brand-name counterparts.⁽²⁾

CBO used that data set to estimate the total savings on prescriptions at retail pharmacies from generic substitution, to compare retail pharmacy sales of generic and brand-name drugs, and to analyze generic competition. Portions of the data set were also used to examine market concentration at the level of the therapeutic class for brand-name drugs and at the level of a single multiple-source drug for generics.

One drawback of the data set is that prescriptions are not the best measure of the quantity of sales. When comparing the prices of two drugs, the best comparison is one based on the price of an average daily dose, not the price of a prescription. Because prescriptions for a drug are typically dispensed in a variety of sizes (the quantity of dosage units, such as pills, varies), comparisons between them are potentially misleading. The variability in prescription sizes may be more of a problem for chronic drugs--which are taken over a long period of time--than for acute drugs. In the case of chronic drugs, whether a pharmacist dispenses a prescription that will last one month or four months may be arbitrary. However, since the data set covers such a large number of prescriptions, it seems reasonable to assume (where relevant) that the average quantity dispensed per prescription for one type of drug will be roughly equivalent to the average quantity dispensed for a close competitor. Moreover, such an assumption is necessary for carrying out any quantitative analysis because of the lack of better data.

CBO used prices per prescription to evaluate the reduction in prescription drug spending from generic substitution and the relative prices of brand-name and generic drugs. Those data were also used to evaluate the decline in the average prescription price as the number of generic manufacturers rises. The measurement error inherent in using a prescription as the unit of quantity could cause the estimated price difference between a brand-name drug and its generic counterpart to be either too high or too low--depending on whether generic prescriptions are smaller or larger, on average, than their brand-name counterparts. Consequently, the estimates of average prescription prices and of the savings to consumers from generic substitution should be viewed as rough figures, not exact ones.

All of the estimates based on average prescription prices cover only tablet and capsule dosage forms, which constitute 87 percent of all sales (or 91 percent of generic sales) in the data set. The average prescription price for those dosage forms appears more reliable than the average price when injectable and liquid dosage forms are included.
 

Total U.S. Sales at Average Invoice Prices

CBO purchased data on the total U.S. sales of 350 prescription drugs from IMS America. That data set covers all channels of distribution except mail-order pharmacies. The sales revenues are valued at the average prices charged on invoices to hospitals, pharmacies, and other purchasers. IMS America also calculated the difference in average invoice prices paid by different channels of distribution for 100 top-selling drugs that were largely distributed through retail pharmacies.

As discussed in Chapter 3, the average invoice price does not include rebates and some discounts that manufacturers give purchasers. As a result, the average invoice price slightly overstates the final price paid. Pharmaceutical Research and Manufacturers of America estimates that discounts and rebates (not including Medicaid rebates) amounted to about $3.5 billion in 1994. Assuming that none of those discounts and rebates were included on an invoice, that figure would equal 5.5 percent of total pharmaceutical sales valued at invoice prices. Although the excluded discounts and rebates are small overall, they could substantially alter the price dispersion figures in Chapter 3 if they were disproportionately received by a particular type of purchaser.

The calculation of the change in returns from marketing a new drug was based on data provided by Henry Grabowski for the average annual U.S. sales revenues of 67 brand-name drugs (valued at invoice prices). Those drugs were introduced between 1980 and 1984. The sales data cover the 1980-1991 period and thus capture the first eight to 12 years that those drugs were on the market. For drugs with only eight to 10 years of actual data, CBO relied on sales projections by Grabowski and John Vernon to determine average annual sales revenues through year 11 for all 67 drugs.
 

Pricing Data from the Medicaid Rebate Program

CBO obtained data from the Health Care Financing Administration (HCFA) on the average price that manufacturers charge wholesalers for drugs that are then distributed through retail pharmacies, as well as on the lowest price charged to any private purchaser (known as the best price). Manufacturers are required by the Medicaid rebate program to report those prices to HCFA for all brand-name drugs that Medicaid beneficiaries buy at retail pharmacies. CBO also obtained data on total Medicaid sales by prescription drug (valued at the price at which states reimburse pharmacies for purchases through Medicaid). Those data were used to assess the differences in price increases between 1991 and 1994 for multiple-source and single-source brand-name drugs.

Those prices reported to HCFA are among the best available (although they are not publicly available) to assess price changes for drugs channeled through retail pharmacies. They represent actual transaction prices, since all discounts and rebates to wholesalers and retail pharmacies are included. Both the average price to pharmacies and the best price are reported by dosage units, such as price per 50 milligram tablet. The average price to pharmacies of a particular dosage form of a drug is calculated by dividing the value of its total sales to wholesalers or chain pharmacies by the number of dosage units sold.⁽³⁾
 

Data from the Patent and Trademark Office and the FDA

The Patent and Trademark Office provided data on all drugs approved through 1995 that have received an extension under the Hatch-Waxman Act. The Food and Drug Administration (FDA) provided overlapping data on the average length of time those drugs spent in clinical testing and in having their new drug applications approved. Those data were used to calculate the average length of a Hatch-Waxman extension and the average time a drug spends in the FDA approval process.

1. National Association of Boards of Pharmacy, Survey of Pharmacy Law: 1995-1996 (Park Ridge, Ill.: National Association of Boards of Pharmacy, 1995), p. 90.

2. Facts and Comparisons, Drug Facts and Comparisons (St. Louis: Facts and Comparisons, 1995).

3. Details on how the best price and average manufacturer price are calculated can be found on HCFA's Web site at http://cms.hhs.gov/medicaid/default.asp.

---