Today, CMS launched its Open Payments System. The website went live at 2pm. Open Payments gives users the option of (1) downloading gigantic excel files of data or (2) using CMS’s “data explorer tool.”
We took advantage of both options, hoping to analyze the raw data using Excel and also to get a feel for what patients would experience if they wanted to use Open Payments to find their doctor.
CMS announced that the Sunshine reports revealed $3.5 billion in payments for the last five months of 2013, there were approximately 4.4 million payments. Payments were made to 546,000 physicians and 1,360 teaching hospitals.
CMS broke down their data in several ways, including general payments and research payments, as promised. However, they also separated the data into “identifiable” at the physician level or “de-identified.” This made it necessary to combine all of these data sources to get an idea of the total amount of payments. It also demonstrated that research payment totals far exceed general payments. Furthermore, a full 90 percent of research data is de-identified.
CMS says such data will be identified in 2015 following the submission of corrected data and physician and teaching hospital opportunity to review/dispute. Further, “data that were disputed and not resolved by the end of the September 11 review period have not been published and will be updated at a later date.” CMS puts the current de-identified data figure at 40 percent which is lower than what the data shows.
We found the following figures:
General Payments: $976,743,814
This figure includes all meals (a huge portion of the data, many of which hover at or below $10); speaker and consulting fees; travel and lodging; educational materials; entertainment; and gifts. There were 4,283,132 total payments in the general database.
- The identifiable general payment reports 2,720,099 payments for a total of $669,561,563 in payments or transfers of value
- The de-identified data reports 1,563,033 payments for a total of $307,182,251
Research Payments: $1,486,242,674
Pharmaceutical companies spent vastly more on their research endeavors than on general payments. Notably, only about 10 percent of the research payments are identifiable on the covered recipient level in the Open Payments.
- The identifiable research payment spreadsheet has 23,226 lines of data equaling $155,815,828
- The de-identified data has 199,887 lines equaling $1,330,426,846
The best analysis of the data thus far is available at ProPublica which includes a break out of nature of payment categories.
In the ProPublica report they mention that companies reported by operating unit. The way the final rule was written, companies are penalized for mistakes on an operating unit basis and interpreted the requirement to report that way.
We will provide a detailed analysis and breakdown of the research payments category tomorrow.
How Will the Average Patient Use This Data?
In its press release, CMS announced that they were releasing the Open Payments data "to help consumers understand the financial relationships between the health care industry, and physicians and teaching hospitals." While we suspect the majority of users today were from the press, we attempted to use the system as an average patient would.
Instead of downloading an immense Excel file, we expect patients will use a page enabling them to “Explore the Data” in order to search for their doctor, if they pursue Open Payments in the first place. When the site went live, we were able to access this database after a process of trying to figure out which choice was the one you could find your doctor in. It was a slow, rocky ride. A partner at a leading data firm called us and expressed that they personally found the process difficult to use, so we can only imagine how a patient would feel.
I first went to the "General Payments" category with identifying recipient information. Not knowing my particular doctor's identification number, I searched by last name. The data eventually filtered down to hundreds of doctors and thousands of payment transactions. I located my doctor, wrote down her identification number, and then organized the data set by identification numbers. Each scroll was met with a processing signal, but eventually I found my doctor's 15-20 meal payments one after the other in the list. I calculated about $200 worth of value.
A fundamental issue in the usability of the site right now is how many categories there are for each payment. This will require patients to wade through a fair amount of confusing information before landing on each particular transfer of value, which they will then have to calculate up on their own. CMS has included a number of filtering options that can provide fairly advanced analysis of the payment data. We think that the press and/or the government may be taking more advantage of those analytic tools than patients.
Starting around 3pm, the website stopped working, despite numerous retries. As of this evening, the website is refreshing more quickly and seems to be in semi-working order. We will provide updates about the system as we continue to work through the databases.