OBJECTIVES
(1) To determine whether all State Medicaid agencies collect drug manufacturer rebates for all physician-administered drugs.
(2) To estimate the potential savings that would result if all State Medicaid agencies collected drug manufacturer rebates for physician-administered drugs.
BACKGROUND
The Medicaid program, established under Title XIX of the Social Security Act, is administered by States and financed with State and Federal funds. Medicaid pays for medical and health-related assistance for certain vulnerable and needy individuals and families. All 50 States and the District of Columbia provide coverage for prescription drugs under the Medicaid program.
From 1997 to 2000, Medicaid spending for outpatient prescription drugs grew twice as fast as total Medicaid spending. The total payments for outpatient prescription drugs in fiscal year 2000 were $21 billion.
Drug Manufacturer Rebates
In fiscal year 2000, drug manufacturers paid the Medicaid program $4 billion in rebates. This reduced Medicaid drug expenditures by 19 percent (total payments before rebates, $21 billion; after rebates, $17 billion).
State Medicaid agencies receive manufacturer rebates for drugs under the Medicaid Drug Rebate Program. This program was established by Federal law in 1990 (see section 1927 of the Social Security Act). States cover all prescription drugs produced by manufacturers who have entered into rebate agreements under this program. Physician-administered drugs (drugs that a medical professional administers to a patient in a physician’s office) are among the drugs covered by the rebate program. Only two States, Arizona and Tennessee, did not participate in the rebate program at the time of our review.
Under the rebate program, manufacturers are required to provide a rebate on drugs paid for by a State Medicaid agency. In order to collect a rebate, the State is required to identify the drugs by their national drug codes and provide quarterly units-paid data to the manufacturer. The manufacturer then pays the State a rebate, based on the unit rebate amount per drug multiplied by the number of drug units for which the State made payments.
Codes Used to Bill Physician-Administered Drugs
Self-administered drugs are typically billed by pharmacies in pharmacy claim formats using national drug codes to identify specific drug products. Physician-administered drugs, on the other hand, are more often billed by medical providers on professional service claims. (An example of a physician- administered drug is a prescription drug given by injection in a doctor’s office.) Professional service claims identify services, medical equipment, and physician-administered drugs by procedure codes (i.e., the Healthcare Common Procedure Codes).
While it is possible to use either a national drug code or a procedure code to bill for a physician-administered drug, these codes identify different things. The national drug code is an 11-digit numeric code, which is divided into 3 segments identifying (1) the firm that manufactures, distributes, or repacks the drug product; (2) the specific strength, dosage form, and formulation of the product for a particular firm; and (3) the product’s package size. The procedure code is a 5-digit alpha-numeric code that identifies a drug by its generic name; route of administration (e.g., oral or injection); and identifies the number of drug units allowed per reimbursement amount for that code.
As mentioned above, States must identify drugs by their national drug code in order to collect rebates. Therefore, if a State requires the use of national drug codes for physician-administered drugs, identifying the drug by its national drug code for rebate collection is not a problem. If a State requires the use of procedure codes for physician-administered drugs, identifying the drug by its drug code can be difficult. For example, if the drug reimbursed by Medicaid has more than one manufacturer, the drug would have more than one national drug code. Therefore, the State would not be able to determine which national drug code matched the procedure code.
Past OIG Work Found that Few States Collected Rebates for Physician-
Administered Drugs
In 1996, the Office of Inspector General (OIG) conducted the study, Appropriateness of Medicare Prescription Drug Allowances, (OEI-03-95-00420). It examined Medicare allowances for prescription drugs through a comparison with Medicaid reimbursement mechanisms and Medicaid drug rebates. During the study, OIG found that while most State Medicaid agencies participated in the rebate program, only six were collecting rebates on physician-administered drugs. States’ ability to collect rebates for physician-administered drugs was contingent upon their ability to identify the drugs by their national drug codes.
Federal Effort to Encourage States to Collect Rebates for Physician-Administered Drugs
In 2002, the Centers for Medicare & Medicaid Services (CMS) informally gathered information regarding State collection of rebates for physician-administered drugs. This was done through the Pharmacy Technical Advisory Group ? a group composed of CMS and State representatives. From this data collection, it appeared to CMS that States could gain considerable savings if they collected rebates on these drugs. CMS wrote to State Medicaid Directors in March 2003, encouraging them to look into system conversions that would facilitate a crosswalk of procedure codes to national drug codes and make rebate collection possible for these drugs. A crosswalk is the identification of national drug codes for drugs represented by procedure codes.
Availability of Medicare Crosswalk
States that use procedure codes for physician-administered drugs need a crosswalk to national drug codes in order to collect rebates on these drugs. A Medicare crosswalk is available on the Internet. It contains drugs paid by Medicare in a particular quarter and it is updated quarterly. The Internet address is: http://www.cms.hhs.gov/home/medicaid.asp.
METHODOLOGY
Preliminary Research
We conducted background research on the Medicaid Drug Rebate Program and code requirements for drug claims, including a review of Federal laws, regulations, manuals, and program memoranda. We reviewed Medicaid expenditures for prescription drugs, and we also reviewed public and private studies addressing the Medicaid prescription drug program and cost controls.
State Medicaid Data
In January 2003, we began collecting a formal set of data from the 49 States (i.e., 48 States and the District of Columbia) that participate in the Medicaid Drug Rebate Program. For the sake of brevity, we use the word “State” in this report as a synonym for both “State Medicaid agency” and “State Medicaid agency representative.” We refer to the District of Columbia as a State for the same reason. We sent our data request to State Medicaid directors.
We asked each State whether they collect rebates for physician-administered drugs, the year they began collecting rebates and for what types of drugs, the type of code they use to bill physician-administered drugs, the types of changes they made to their systems in order to collect rebates for physician-administered drugs, and the estimated costs of changing systems for this purpose. We also asked each State to provide us a set of financial data for physician-administered drugs in calendar year 2001, namely, total payments per code, total units paid per code, rebate dollars requested of manufacturers per code, and rebate dollars received from manufacturers per code.
Forty-eight States answered the questions we sent them, and most States provided all or some of the financial data we requested. The State of Nevada provided financial data but did not respond to the questions. We did not verify the data States sent us. Table 1 in the Appendix includes the State financial data.
CMS Data
We used CMS’s 2001 list of national Healthcare Common Procedure Codes with complete definitions to identify the universe of codes that represented physician-administered drugs.
We used calendar year 2001 data from CMS’s Medicaid Statistical Information System (MSIS) to identify payments and units paid per code for physician-administered drugs for States that did not send us this information (see Table 1 in Appendix A). Five States did not provide payment and unit data, and an additional two States did not provide unit data. We aggregated the State payment data received directly from States with data retrieved from MSIS to determine total Medicaid payments and utilization for physician-administered drugs.
CMS’s Medicaid Drug Rebate Initiative database (hereinafter called rebate database) was used to identify drug unit rebate amounts in calendar year 2001. Since unit rebate amounts are established quarterly, we took the unit rebate amount for each quarter and calculated the average. The average unit rebate amount was then used in our calculation to determine potential savings.
We used the January 1, 2002 update of the Part B Drug Calculation File from CMS’s Single Drug Price contractor. The CMS contractor uses this file to establish a national Medicare payment allowance for procedure codes. The file is organized by procedure code. Data in this file for each of the 485 procedure codes for physician-administered drugs included the procedure code description and a list of associated drug products with their national drug codes. The CMS contractor used the 2001 Red Book, a national drug-pricing reference, to identify average wholesale prices for drugs in this file.
We used the Part B Drug Calculation File to determine which procedure codes represented single-source drugs and which represented multiple-source drugs on the market in 2001. A single-source drug is a brand name drug that is manufactured under a patent and has no competing products. A multiple-source drug is a drug whose patent has expired and is manufactured by a number of different drug companies under different names. Multiple-source drugs may include generic drugs and brand name drugs.
Single-source drugs are more likely to have only one national drug code and are more easily matched to a procedure code than multiple-source drugs. If the Part B Drug Calculation File showed only one drug manufacturer for a procedure code, we took that to be an indication the procedure code represented a single-source drug. We then looked up that drug in the October 2001 issue of Red Book ™ for Windows® to confirm whether it was indeed a single-source drug. Drugs in the Part B Drug Calculation File that had multiple manufacturers or were not confirmed as single-source drugs in Red Book were put in the multiple-source drug category.
We also used the Part B Drug Calculation File to identify the national drug codes for single-source and multiple-source drugs that corresponded to each procedure code.
Analysis of Payments for Physician-Administered Drugs
Using State financial data, we identified payments for single-source and multiple-source, physician-administered drugs per State. Single-source drug payments for 46 States ranged from $43,000 to $20 million, for a national total of $148 million. Multiple-source drug payments for 46 States ranged from $12,000 to $18 million, for a national total of $152 million. Because California and Nevada use local codes (codes used by the individual States), and Massachusetts uses non-specific codes for physician-administered drugs, we did not identify payments for single-source and multiple-source drugs for these three States. We did, however, aggregate total payments for these three States with total payments for other States. Total national payments for physician-administered drugs in 2001 were $364 million. A summary of payments is provided in
Table 1 of Appendix A.
Potential Savings Analysis Drugs Used to Calculate Savings. We reviewed all physician-administered, single-source drugs and 40 multiple-source drugs for which States made payments in 2001 but did not request a rebate, and for which a unit rebate amount was available in the rebate database. The number of single-source drugs reimbursed by each State ranged from 17 to 149. Of the 40 multiple-source drugs reviewed, the number reimbursed by each State ranged from 5 to 39. We reviewed drugs billed with national procedure codes only. We did not review any drugs billed with local codes. Nor did we review vaccines or immunizations because these products are not covered under the rebate program.
We selected the 40 multiple-source drugs in the following way. We took the State-reported payment data for physician-administered drugs billed with procedure codes and removed payments for codes that received rebates in 2001. We then aggregated the payments nationally by code. We arrayed the aggregated payments from high to low and selected 40 codes for multiple-source drugs having the highest payments. These 40 drugs represented $71 million in total Medicaid payments. Table 2 in Appendix A lists the 40 drugs.
In order to separate single-source drugs from multiple-source drugs, we matched the State’s procedure codes with procedure codes in our crosswalk of single-source drugs. If the procedure code paid by the State did not match a procedure code in the single-source crosswalk, we included it in the multiple-source category.
For States that collected rebates on single-source, physician-administered drugs, we removed those codes from our analysis of the State’s potential savings.
Identifying Unit Rebate Amounts. Unit rebate amounts for drugs are stored in the rebate database under the national drug code. Having completed a crosswalk for single-source drugs and selected multiple-source drugs, we compiled a list of national drug codes for the drugs under review in the single-source and multiple-source categories. We then obtained the unit rebate amount from the rebate database for each national drug code and calculated the average of the unit rebate amount for all quarters in 2001 for each national drug code. When more than one drug product was on the market for a particular procedure code, and therefore, had more than one national drug code and unit rebate amount in the rebate database, we identified the median unit rebate amount.
In instances where the strength of the national drug code did not exactly match the strength of the procedure code, we applied a conversion factor to calculate the correct unit rebate amount for the procedure code. The unit rebate amount per procedure code was then used to calculate potential savings for each procedure code in each State.
Potential Savings Calculation. We used the following steps to determine potential savings:
Step 1. For each procedure code, we multiplied the total number of drug units paid for the procedure code in 2001 by the unit rebate amount per procedure code. The product of this calculation was the potential savings for the procedure code.
Step 2. We calculated each State’s total potential savings by adding the potential savings for each procedure code.
Step 3. In order to determine total Medicaid potential savings, we summed the potential savings from each of the States.
We noticed that in Step 1, savings for some procedure codes exceeded payments. This might have been the result of a State’s units-paid data not accurately representing paid claim units. In any State where potential savings for a procedure code exceeded the payments for that code, we removed the code from our savings calculations.
Reporting the Potential Savings. Table 3 in Appendix A contains potential Federal and State savings per State. These savings were calculated on physician-administered drug payments of $169 million. This was the subset of payments for single-source and 40 multiple-source drugs for which the States did not request rebates in 2001. In addition, these payments do not include vaccines or immunizations, drugs billed with local codes or non-specific codes, and codes deleted because savings calculations exceeded payments.
____________________________
This study was conducted in accordance with the Quality Standards for Inspections issued by the President’s Council on Integrity and Efficiency.