As a teacher I give a grade of D- with a C+ for effort.

Yesterday I began to dig into the first public report by the Center for Medicare and Medicaid Services of their Open Payments program, also known as the Sunshine Act. Under this part of the Accountable Care Act (ACA), some (but not all) payments and transfers of value to physicians and teaching hospitals by pharmaceutical companies, medical device manufacturers, and group purchasing organizations must be reported. There is little doubt that such payments powerfully influence the practice of medicine for better, but mostly for worse. Initial reporting in the media finds individual physicians and organizations scrambling to justify or to correct the often very startling amounts received.

I still believe the program is a great idea. The public has no idea of the extent to which industry has captured the health policy market, including the professional activities of physicians and other practitioners. Much will be learned by focusing on the outliers on both ends of the curve are already emerging from the mass of data. Indeed, I hope to be able to contribute myself. Nonetheless, in fairness to all involved, it must be recognized that this was a deeply flawed implementation.

Born to be bad.
Of course anything that makes it way through the legislator-lobbyist gauntlet has a very high chance of being flawed to the point of ineffectiveness. Crippling laws you don’t like is just as good as passing laws that you do. Thus, payments to medical schools, or to the third-party middlemen of today’s continuing medical education programs need not be reported– even though the ultimate beneficiaries of the payments are essentially the same. Want to camouflage your new drug roll-out? Done deal!

Even if there had been bipartisan and professional agreement to make the Open Payment program work, the obstacles were many and large. Just getting all the computers of a myriad of companies and hospitals to talk to each other in the same language, or to be able to assemble and standardize information residing in different places is an administrative and computational nightmare. It obviously was not possible in the timeframe available. Right off the bat doctors were finding that payments being ascribed to them were erroneous. (We physicians are seeing the some same problems with our industry- and government- mandated electronic health records.)

Look now or forever hold your peace.
Although an interval before public release was allotted to allow physicians and hospitals to register and sign in to the CMS Open Payments Portal to review or challenge their data, this was not straightforward. Furthermore, the already short period allowed for review was further encroached by unavailability even in the days before the review period was closed.

Nonetheless, CMS decided to go ahead and publish what they had on schedule. To do so, they released the information in two separate files – one in which there was a reasonable probability that the correct recipient had been “identified,” and another in which for one reason or another the recipient was uncertain, termed “de-identified”. Just before the public release, it was being said that as many as 30% of payments would be left out. As it happens, they were not left out. Some were published separately, and some not at all.

Data Fact Sheets.
CMS provided a Open Payment Data Fact Sheet summarizing the data. On the face of it, it is abundantly clear to me that it is the information still hidden in the de-identified data, in the data being disputed, and the data being delayed that is critically most important to the goals of the program. If the rule is “follow the money, then it remains still in these currently non-disclosed vaults! I took the liberty of restating and editing the Fact Sheet to bolster my opinion.

Amazing numbers.
For example, 39% of the combined payment records were de-identified. However, these made up 63% of the dollar value of all reported payments and transfers. Most of the money was left under the table! What is more, the amount of each individual payment per record in the de-identified group was much higher – $1294 vs. $481. I wonder why?

24% of payments not published at all!
A stunner is evident in Table 4 where we learn that another $1.1 billion worth of payments was not published at all in either the identified or the de-identified file. These represented unresolved disputed payments or requests by industry for delays. The average value per payment of the disputed records was a hefty $57,111! The average payment related to drugs and devices that were newly approved by the FDA (and thus eligible for a request to delay publication for up to four years) was $2900. It is behind such new drugs and devices that industry puts its greatest marketing muscle, and therefore correspondingly should be of greatest value to a public interested in identifying conflicts of interest.  I am unaware of any prior disclosure by CMS or by the traditional news media of this “grey money.”

It gets even worse!
There has been almost no review by physicians and teaching hospitals of the amounts of money and transfers of value to them being reported by the host of outside organizations. Only 4.8% of physicians and 29.8% of teaching hospitals named in the report even registered to check what was being attributed to them. Only a minuscule 0.3% of payment records were disputed and of those, 71.5% remained as unresolved disputes at the end of the review period. Clearly, the accuracy of the vast majority of these records– those public now and those to be released later– has been unexamined. The large sums of money involved in the disputed claims suggests to me that some recipients had a good reason to review and make sure the transactions were accurate. If these large sums were reported in error, than many other physicians and hospitals will find that they too are being credited with money they did not receive.

What was the rush?
I will let CMS speak for itself. I am not the first to say that if this report had been presented to any journal of which I had been on the editorial board or served as a reviewer, I would have strongly recommended against publication. The research methodology is too weak and the data collection to incomplete to support any robust conclusions. Surely the professionals within or advising CMS would have known this. My first thought is, that facing a potential Republican capture of the US Senate, and a stated intent by this business-friendly political party to reverse as much of the ACA as possible, the Democratic administration wanted to get whatever it had out there while it could. Once out there where the public could get a glimpse of what was happening, it would be harder to kill. Does anyone else have a better explanation?

In closing.
Imagine the cost in time and money to keep track and insure the accuracy of some 4.5 million individual transactions, and to distill it into a single meaningful report. Against that cost must be weighed the many billions of dollars that are spent unwisely, unnecessarily, or worse in our non-system of medical care. The real tragedy is that examining such a tradeoff is a reasonable, even a necessary thing to do.

Peter Hasselbacher, MD
President, KHPI.
Emeritus Professor of Medicine, UofL
October 3, 2014