Fact Sheets

Updated prescriber-level Medicare data

Updated prescriber-level Medicare data
Data serve as a rich resource into Medicare Part D costs, utilization, and trends

The Centers for Medicare & Medicaid Services (CMS) has made available its second annual release of data that details information on the prescription drugs that were prescribed by individual physicians and other health care providers and paid for under the Medicare Part D Prescription Drug Program.

The new 2014 dataset describes the specific medications paid for under the Medicare Part D program and statistics on their utilization and costs for 38 million beneficiaries enrolled in a Medicare Part D plan, who represent 70 percent of all Medicare beneficiaries. It provides data on more than one million distinct health care providers who collectively prescribed $121 billion in prescription drugs under the Part D program. New in the 2014 data are distinct beneficiary counts, prescription drug event counts and total drug costs aggregated by drug category for opioids, antibiotics, antipsychotics, and high-risk medications among the elderly. In addition, a prescriber enrollment status field has been added to indicate whether the prescriber is enrolled, not enrolled or opted out of the Medicare program. The data are posted on the CMS website at

CMS created the Part D Prescriber Utilization and Payment public use files (Part D Prescriber PUF) using information from Prescription Drug Event data submitted by Medicare Advantage Prescription Drug (MA-PD) plans and by stand-alone Prescription Drug Plans (PDPs). The new dataset identifies providers using their National Provider Identifier (NPI) and presents the specific prescriptions dispensed at their direction, listed by brand name (if applicable) and generic name as designated by First Databank.  

For each prescriber and drug, the dataset includes the total number of prescriptions that were dispensed (including original prescriptions and any refills), and the total drug cost. The total drug cost includes the ingredient cost of the medication, dispensing fees, sales tax, and any applicable administration fees. It’s based on the amounts paid by the Part D plan, Medicare beneficiary, other government subsidies, and any other third-party payers (such as employers and liability insurers). Total drug costs do not reflect any manufacturer rebates paid to Part D plan sponsors through direct and indirect remuneration or point-of sale rebates.  In order to protect beneficiary privacy, CMS did not include information in cases where ten or fewer prescriptions were dispensed.

The Part D Prescriber PUF provides key information to consumers, providers, researchers, and other stakeholders to help drive transformation of the health care delivery system. These data enable a wide range of analyses on the type of prescription drugs paid for under the Medicare Part D program, and on prescription drug utilization and spending generally. 

The new dataset can be used to examine the relative rank of drugs by utilization. Below are the top ten drugs in 2014 by claim count[1] (Table 1). All of the top ten drugs are generic drugs, and the top nine drugs were among the drugs with the highest claim counts in 2013. The claim counts for these drugs ranged from 22.1 to 38.3 million claims and the total drug costs for each drug ranged from $136 million to $748 million.

Table 1. Top Ten Drugs by Claim Count, 2014


Drug Name

Total Claim Count

Beneficiary Count

Prescriber Count

Total Drug Cost






Levothyroxine Sodium





Amlodipine Besylate




















Atorvastatin Calcium










Metformin HCl










Table 2 below shows the 2014 top ten drugs by total drug cost. These drugs are all brand name drugs with relatively fewer claims than the top drugs by claim seen in Table 1. In 2014, Solvaldi had the highest total drug costs at $3.1 billion, but  the costs for each of the top 10 drugs were all more than $1 billion.

Table 2. Top Ten Drugs by Costs, 2014

Drug Name

Total Drug Cost

Beneficiary Count

Prescriber Count

Total Claim Count





















Advair Diskus










Lantus SoloSTAR

(insulin pen)




















As this is the second year that CMS has released this Part D prescribing data, the datasets can also be used to compare 2013 and 2014 data. Table 3 presents total number of Part D claims and total drug costs for 2013 and 2014. From 2013 to 2014, the total number of claims increased from 1.37 billion to 1.42 billion, a 3 percent increase from 2013 to 2014. However, total drug costs increased from $104 billion to $121 billion, reflecting a 17 percent increase from 2013 to 2014.

Table 3: Overall Claim Count and Total Drug Cost, 2013 and 2014




Percent increase 2013-2014

Total Claim Count




Total Drug Costs




The two years of available data can also be used to investigate growth trends for specific drug products. Figure 1 below shows the percentage change between 2013 and 2014 in total drug costs for the top ten drugs with the overall highest 2014 total drug costs. Lantus Solostar and Lantus insulin products had the highest growth in total drug costs between 2013 and 2014 with growth rates of 47 percent and 32 percent, respectively. Abilify, Januvia, and Revlimid also had high growth rates of 20 percent or higher.  Advair Discus had a very low growth in total drug costs of only 1 percent.  Part D total drug costs are affected by both the volume of prescriptions filled as well as the unit prices of the individual products, and the cost trends shown here do not reflect any manufacturer rebates or discounts, which may also vary from year to year. 

Figure 1.  Percentage Change in Total Drug Cost, 2013 vs. 2014

Figure 1 displays a bar chart of percentage change in total drug cost for the top ten drugs with highest drug cost in 2014. In order of highest percentage change in drug cost, Lantus Solostar had the highest growth rate between 2013 and 2014 at 47 percent. Growth rate for Lantus insulin products was 32 percent, growth rate for Revlimid was 24 percent, growth rate for Januvia was 22 percent, growth rate for Abilify was 20 percent, growth rate for Crestor was 15 percent, growth rate for Spiriva was 10 percent, growth rate for Nexium was 5 percent and growth rate for Advair Diskus was 1 percent.  No growth rate was reported for Sovaldi.  Since Sovaldi was not available until the end of 2013, any attempt at comparison across years would be invalid.

* Sovaldi was not available until the end of 2013, so comparisons across years are not valid.

New Data Added for 2014

The new 2014 dataset also can be used to examine patterns of antibiotic prescribing in the Medicare program. For example, these data can inform where high rates of antibiotic prescribing are occurring across the U.S. This map (Figure 2) shows that states in the South and Midwest have rates of antibiotic prescribing that are higher than the national average of 1.39 fills per beneficiary.

Figure 2. Antibiotic Fills per Capita, 2014

Figure 2 shows a U.S. map of antibiotic prescribing rates by State in 2014.  The map shows that states in the South and Midwest have the highest rates. It shows that Nebraska, Kansas, Arkansas, Louisiana, Mississippi, Alabama, Tennessee, Kentucky, Indiana, and West Virginia have claim per capita rates between 1.57 and 1.85, higher than the national average of 1.39. The lowest antibiotic prescribing rates are in the Northeast and West, particularly, Main, New Hampshire, Vermont, New Mexico, Montana, Washington, Oregon and California. In these states there were an average of 1.02 – 1.16 Antibiotics prescribed per beneficiary.


Use of high-risk medications in the elderly is a measure that assesses medication management in order to prevent the harms associated with certain medications for this population. CMS used the list of medications maintained by the Pharmacy Quality Alliance for this measure, which includes select prescription drugs recommended by the American Geriatrics Society to be avoided in persons aged 65 years and older because of a high risk of serious side effects, when there may be safer drug choices.[2] The map below (Figure 3) depicts the average number of high risk medications prescribed to elderly beneficiaries across the U.S. The national average for prescribing high risk medications is 0.86 claims per elderly beneficiary. The lowest rates of high risk medication prescribing are found in some Western states, Midwestern states, such as Minnesota, Wisconsin, and Michigan, as well as states in New England. States with the highest rates of high risk medication prescribing, which range from 1.09 to 1.31 claims per elderly beneficiary, are concentrated in the South.

Figure 3. High Risk Medication (HRM) Fills per Capita among Elderly Beneficiaries, 2014

Figure 3 shows a U.S. map of average number of high-risk medications prescribed to elderly beneficiaries in the U.S. The map shows that the south and Midwest have the highest rate of prescribing high-risk medications, particularly, Oklahoma, Arkansas, Louisiana, Mississippi, Alabama, Georgia, South Carolina, North Carolina, Tennessee and Kentucky. These states saw an average of 1.09 to 1.31 high-risk medications per each elderly beneficiary.  The lowest rates were found in Vermont, Delaware, Michigan, Wisconsin, Minnesota, New Mexico, Arizona, Nevada, and Oregon.  In the states with the lowest rates an average of 0.51 and 0.72 high-risk medications were prescribed per elderly beneficiary.


Although the Part D Prescriber PUF has a wealth of payment and utilization information about Medicare prescription drug events, the dataset also has a number of limitations that are worth noting. First, the information presented in this file does not indicate the quality of care provided by individual clinicians.  Second, given that the data contain information only from Medicare beneficiaries with Part D coverage, clinicians typically treat many other patients who do not have that form of coverage, the data in the Part D Prescriber PUF may not be representative of a prescriber’s entire prescribing pattern nor be fully inclusive of all prescriptions written by the provider. Additionally, the data in this file are limited to medications covered by the Part D program and drugs statutorily excluded by the Part D program, which may be covered by individual Part D prescription drug plans through supplemental coverage. Since not all Part D plans have supplemental coverage for excluded products, utilization and cost statistics presented in the data likely underestimates the true use of these products in this population.

The total drug costs included in these data reflect the prescription drug costs incurred by Medicare Part D beneficiaries, including costs that are paid by Medicare, by beneficiaries, and by third-party payers. The Part D prescription drug program is administered by private Part D plan insurers. Medicare pays Part D plans a monthly, risk-adjusted capitation payment for each enrollee. Beneficiaries also pay a monthly premium. In addition, Medicare pays Part D plans additional subsidies to cover reduced cost-sharing for low-income beneficiaries and a portion of the costs for beneficiaries whose drug costs are very high. Following each benefit year, CMS shares risk with plans by reconciling the capitation and various subsidy payments to actual drug cost expenditures determined from Prescription Drug Event records, and any manufacturer rebates or other direct and indirect remunerations received by the plan. Therefore, because the drug expenditures derived from the Prescription Drug Event data comprise only a piece of the payment process, it is not possible to directly attribute total drug costs at the prescriber or drug level to payments from the Medicare Trust Fund. Furthermore, these total drug costs do not reflect any manufacture rebates.

Visit to view the new prescriber dataset.  

To read the press release, visit:  


[1] Claims count represents prescription drug fills which can cover varying lengths of time. 
[2] High risk medications are defined based upon the Pharmacy Quality Alliance drug list used for CMS’ Part D Star Ratings display measure. For additional information see: