How We Analyzed Medicare’s Drug Data
ProPublica obtained Medicare Part D data from the Centers for Medicare and Medicaid Services (CMS) under the Freedom of Information Act. Here follows more information about the data and how we analyzed it.
ProPublica obtained Medicare Part D data from the Centers for Medicare and Medicaid Services (CMS) under the Freedom of Information Act. The data includes information from 2007 through 2011 about prescribing by physicians and other providers under Medicare’s drug benefit program. Before ProPublica first obtained this data in 2012, Medicare had not released prescribing information with provider identities. No patient information was disclosed.
Prescribers include any health professional who wrote a prescription that was filled by a beneficiary in Medicare Part D. In addition to doctors, nurses, physician assistants, dentists and others with prescribing authority are included in the data.
In 2011, more than 1.6 million providers wrote nearly 1.2 billion prescriptions, including refills. The data does not include prescriptions that were written during hospital or short skilled nursing home stays because those are paid for under different parts of Medicare.
CMS provided ProPublica with two different prescribing files for each year, allowing reporters to analyze the data in different ways. One file was broken down by provider ID and drug name. The other file was grouped by the first 9 digits of a standard code, called the National Drug Code, which is used to classify medications, and indicates the dosage strength of the drugs. For each provider and drug, the data included the total number of claims, including refills dispensed, the retail cost of the drug, the days of supply and the number of units (i.e. pills or ounces).
In cases where a provider wrote 10 or fewer prescriptions for a specific drug, CMS removed some information to protect patient privacy.
We also received similar files listing only those prescriptions written for patients 65 and older in 2011. Because Medicare covers seniors as well as the non-elderly disabled, this file allowed us to identify physicians who were prescribing drugs that are considered particularly risk for older patients. The American Geriatrics Society compiles what’s called the Beers list of drugs that may be inappropriate for seniors; often safer alternatives exist.
CMS also provided data on the top 500 drugs prescribed nationally and in each state.
To understand how best to analyze and categorize the drugs, we consulted dozens of experts. They included pharmacologists, academics, government officials who work with prescribing data and specialists in geriatrics and psychiatry.
We used data from First Databank, a company that sells and analyzes health information, to classify drugs by category, such as narcotics or antipsychotics. Within each category, we identified the highest prescribers and those whose drug choices were significantly different from other providers. We also examined prescribing patterns for many individual drugs.
The Medicare Part D data includes codes, not names, to identify individual prescribers. Most have a unique federal health care ID called an NPI, or National Provider Identifier. We used NPI data that is freely available for download online from CMS to find the names, addresses and other information about individual prescribers in the Part D data. The NPI data also indicates what providers reported as their primary specialties. (Some also classified themselves as a group practice.) Our analysis used NPI data downloaded in July 2013.
We also obtained a database of U.S. Drug Enforcement Administration registrations from the National Technical Information Service to identify providers who did not have an NPI but did have a DEA registration number. Those providers do not have primary specialties in our Prescriber Checkup news application.
In some cases, we were unable to identify prescribers because they did not have an NPI number or DEA number or because pharmacies used “dummy” ID codes that don’t match back to an individual. These represent about 1 percent of all claims.
The data could not tell us everything. We interviewed many high-volume prescribers to better understand their patients and their practices. Some told us their numbers were high because they were credited with prescriptions by others working in the same practice. In addition, providers who primarily work in long-term care facilities or busy clinics with many patients naturally may write more prescriptions.
In addition, the type of patients some doctors see may affect their rate of prescribing name-brand drugs. Some of the physicians who prescribe name-brands at far higher rates than their peers specialize in treating HIV/AIDS. The drugs for these patients are expensive and there aren’t comparable generics.
When a provider writes a prescription, the pharmacy filling it enters the information into a computer system. That is transmitted to a health plan for payment and then relayed to Medicare. Medicare compiles all of the prescriptions into a massive database. (Prescriptions written by doctors but not filled by patients wouldn’t be included.)
Prescriptions can vary in length, from days to months. Some researchers who analyze prescribing patterns adjust the number of prescriptions to a 30-day standard to allow for easier comparisons between providers. This app does not. Rather, it notes when the average length of a provider’s prescriptions for a given drug is higher or lower than normal. In some cases, a provider may write shorter prescriptions while adjusting a patient’s medication regimen, for example. That could mean that the provider’s prescription count may appear higher than his or her peers. On the other end, if providers write prescriptions for more days than average, they may have fewer prescriptions. Prescriber Checkup indicates that a prescriber’s average prescription length was significantly higher or lower from his or her peers if it was more than two standard deviations from providers in the same specialty and state for that drug.
Prescriber Checkup features a chart comparing prescribers to others in their state and specialty. Itclusters prescribers based on their drug preferences and volume and represents them visually under the heading “Another View.” A provider who appears far to the right has drug preferences and prescribing volume that are markedly differ from others.
Illinois psychiatrist Michael Reinstein, for example, appears far to the right for his state and specialty. Not only are his overall prescribing numbers higher than most of his peers, his pattern is different. The rankings table shows that Reinstein’s No. 1 drug, clozapine, is the No. 7 drug among his peers. His No. 3 drug, FazaClo, is No. 26 among other Illinois psychiatrists.
If a provider’s specialty is not known, or if there are 10 or fewer providers in a given specialty and state, comparisons will not be shown.
Medicare’s failure to monitor what doctors are prescribing has wasted billions of taxpayer dollars on excessive use of brand-name medication and exposed the elderly and disabled to drugs they should avoid.
Latest Stories in this Project
- Even After Doctors Are Sanctioned or Arrested, Medicare Keeps Paying
- The Perils of Problematic Prescribing: A Double Dose of Warnings
- Medicare’s Drug Program Needs Stronger Protections Against Fraud, Watchdog Says
- Using Prescriber Checkup: A Quick Guide
- Some Predictions on How Medicare Will Release Physician Payment Data
Our Hottest Stories
- Segregation Now
- Long After Sandy, Red Cross Post-Storm Spending Still a Black Box
- Even After Doctors Are Sanctioned or Arrested, Medicare Keeps Paying
- How the Labor Department Has Let Companies Off the Hook for Unpaid Internships
- Shake-Up Inside Forensic Credentialing Org
- The U.S. Government: Paying to Undermine Internet Security, Not to Fix It
- Meet the Doctor Who Gave $1 Million of His Own Money to Keep His Gun Research Going
- Republicans Say No to CDC Gun Violence Research
- What Newly Released Docs Tell Us About the IRS and How It Handles Dark Money Groups
- Medicaid Programs Drowning in Backlog