What your patients need to know about the Medicare Rx data release


With the Centers for Medicare & Medicaid Services (CMS) release of Medicare Part D prescription drug data Thursday, your patients may have questions about your prescribing practices. Here are several points you might need to clarify in conversations with them.

The data shows which medications were prescribed to Medicare Part D beneficiaries, primarily broken down by individual practitioners or practices. The data release contains information for approximately $103 billion in drugs and supplies prescribed by more than one million health care professionals and paid under the Part D program in 2013. The data set shows physicians’ individual prescribing patterns, including the total number of prescriptions dispensed and the total drug cost paid by beneficiaries, Part D plans and other sources.

“The AMA is committed to transparency and the availability of reliable information for patients to make informed decisions about their medical care,” AMA President Robert M. Wah, MD, said in a written statement. But the data set is more complex than meets the eye.

“The limitations of [the data set] should be more comprehensively listed and highlighted more prominently so that patients can clearly understand them,” Dr. Wah said. “We are also troubled by the lack of context provided with the data that could help explain physician prescribing practices and pharmacy filling practices before conclusions are drawn.”

“In addition to improving transparency with the public, we are also calling on CMS to provide accurate, timely and actionable data to physicians that will support the implementation of new delivery and payment models that can improve patient care,” he said.

8 facts patients should know about the data

Here are eight issues patients should understand when looking at the data:

  • Dosage variations are unaccounted for. The data set doesn’t account for varying medication strengths or dosage levels or the spectrum of patient needs, so it isn’t wise to draw conclusions about treatments from this data. For instance, depending on the conditions and patients they treat, one physician might typically prescribe one 50 mg tablet of a particular drug while another might more commonly prescribe two 20 mg tablets. This could lead to erroneous comparisons of the drug costs and utilization they generated.
  • Differences in physicians’ typical patient mix are not adequately recognized. The specialty descriptions and practice types in the data are not detailed enough to identify physicians who routinely treat patients whose conditions generate higher drug costs. In some cases, physicians who appear to have the same specialty can serve very different types of patients. For example, physicians who work in a hospice or palliative care setting could look like outliers in their prescribing of opioids. 
  • The data shouldn’t be used to evaluate quality of care. The data solely focuses on payment and utilization, so it should not be used to evaluate care provided. The utilization part of the data may not be accurate if a patient had poor medication adherence or an adverse reaction to a pharmaceutical that required a prescription for alternative treatment.
  • The data does not provide a full picture of a physician’s practice. When there were fewer than 10 claims for a certain treatment by a particular prescriber, data isn’t included. In addition, some treatments could be listed under the physician’s organization instead of their individual National Provider Identifier, making their prescription levels seem lower than they actually are. The data set also is a limited view of the patients a physician cares for. It does not include treatments paid for by private insurance plans, for patients not covered under Medicare Part D or for Medicaid beneficiaries. It also fails to take into account patients’ health, socioeconomic status or medication adherence—all of which impact prescribing practices.  
  • Information about generic substitutions isn’t provided. Some physicians may look like outliers in how frequently they prescribe brand-name medications because they treat patients for conditions that are best treated by medications still on patent with no generic equivalent. In addition, comparing branded prescriptions versus generic alternatives to other parts of the country is difficult because state laws governing generic substitutions vary.
  • Charges and payments are different.  The cost information does not include manufacturers’ coupons or rebates that often help lessen patients’ out-of-pocket costs. Depending on where a patient gets his or her prescription, there also could be price differences. 
  • This data shouldn’t be compared with other data from CMS. There are significant differences between this data set and other CMS data, including last year’s physician Medicare claims data and financial data reported under the Physician Payments Sunshine Act. These releases have different timeframes and Medicare populations, and they also have identified physicians in multiple ways.
  • Physicians don’t have a way to correct the information reported. If data reported is inaccurate, physicians aren’t able to report it or update the data.

For more information, get the AMA’s guide to media reporting on the data set.