Understanding Implications of the Medicare Provider Utilization and Payment Data

This week, Medicare’s public release of “Provider Utilization and Payment Data” has caused quite a stir. Yet, it has raised more questions than it has answered. From what I have read on Facebook and Twitter, there seems to be widespread confusion concerning the implications of this data. The goal for this post is to provide some context and to discuss some common comments I have heard about the release.

“It shows how overpaid physicians are.”

What you are “really” interested in is MD salaries. MDs may in fact be overpaid, but these data don’t address that issue at all. MDs salary is not always equivalent to “average reimbursement” from Medicare. First, the RNs, receptionists, valets, lab techs, custodians, etc. Since none of those occupations can bill Medicare, their salary must come from somewhere; presumably these people aren’t working for free. So either it its wholly cross-subsidized from non-Medicare patients (e.g. Blue Cross-Blue Shield) or it comes partially out of the professional fee and partially out of the facility fee. I don’t know how every hospital does it, and there is significant variation along service lines depending on the proportion of Medicare patients (e.g. pediatrics vs vascular surgery).

Second, most MDs see non-Medicare patients as well which means they (or the practice/hospital/health system) are getting other revenue streams. In brief, you have no idea as to the denominator without data on non-Medicare payers.

In the case of physicians who are employed by another business entity (e.g. group, practice, hospital), this money goes into the community pot for that entity. The nature of the community pot differs by individual organization. Here are some possible scenarios:

  • A small rural family medicine practice with 3 MDs. Reimbursement might be divided equally among MDs. Of course this is after all the other people who work in the office get their cut.
  • A large urban practice with 10 equity partners & 10 non-equity employees. Reimbursement goes to community pot which takes care of contractual obligations first (i.e. paying the RNs, the receptionist, non-equity employee-MDs). Anything extra goes to the partners. In theory MD salaries are in proportion to Medicare billing. Specifically, an individual MD’s salary is they are a function of the partner’s equity and available funds are whatever is left over after contractual obligations.
  • Salaried MDs of large hospitals/health systems (e.g. Cleveland Clinic). Goes in to the general fund. MDs salaries are whatever they can negotiate. If Medicare reimbursement goes up/down it doesn’t impact any individual MD until their 1-yr contract is up. Then presumably depending upon whether the hospital/department is in the red or black and whether or not the MD negotiated a bonus, he/she may get a  bonuses vs. get fired. Who gets fired is probably a function of productivity and political clout.

Be careful of what you read about the total payments or average payments. There is rather sizable skew in the data. Fortunately, @StuartBuck1 posted the histogram of payments on twitter (pic.twitter.com/TKknvfjkD8). The modal payment (most commonly received amount) was $10. In brief, relatively few MDs are “Medicare Millionaires” in terms of revenue generated, much less salary.

“Healthcare is bankrupting the country.”

  • You are probably right, but don’t forget that these data don’t even represent all of the money Medicare pays out. Medicare pays at least as much in facility payments as it does on provider payments. Not to mention there are all sorts of other expenditures, like GME payments, and durable medical equipment.
  • Don’t be shocked by a big number until you know what the denominator is.

“This misleading datadump causes more problems that it was worth.”

On a fundamental level I disagree with this idea. As a general rule, I am against the government having access to data to which citizens lack access. In some circumstances, claiming “national security” may be realistic, but Medicare provider payments are hardly an example of this exception to my general opinion regarding extreme transparency.

“These data are misleading.”

You can only be mislead when you have a specific question in mind. For example:

Question: “Did you have sex with that woman?”

Answer: “I did not have sexual relations with that woman.”

Truth: We in fact engaged in oral sex.

Ok, this situation is probably misleading to the general public (though not to many attorneys). Asking the right question and not expecting a certain answer go a long way to finding the “Truth”.

In contrast, the Medicare data was not released to answer any particular question. If the question you are actually trying to answer can’t be answered by this data release, I would argue the answer is for Medicare to release even more data. For example, say you want to know “how much is the average MD paid from Medicare”? You actually need to know how Medicare dollars are dispersed once it hits the hospital. Feel free to ask your local hospital how it pays its MDs.

The best solution to this who ting is to go and take a look at the data yourself. If you can’t do that, let me know and maybe I can help. If I can’t find someone who can. Here are some resources to get you started:

Methodology Readme for the Public Use File. http://go.cms.gov/1qmqOXn

Blog post from Policy Prescriptions. http://bit.ly/R80q7O
CMS (Medicare) website for the Public Use File, excel files for data. http://go.cms.gov/1nau6eA
Source: Stuart Buck / Twitter

Source: Stuart Buck / Twitter