— Janet Weiner (@weinerja) April 18, 2016
This has to be fraud, this has to be scandal?
The study at the NBER argues that it has a good first step of an overbilling screening test:
Our idea is very simple. Every provider has a fixed number of hours in any given period; and most of the service codes that are submitted for reimbursement require that the provider spends certain amount of time with the patient. If
the hours worked implied from the service codes a provider submits to CMS are implausibly long, the provider is suspicious for overbilling.
I am very skeptical that this will have practical impact for a couple of reasons.
The biggest problem is how Medicare handles Master level clinician billing. Their data method does not address this problem.
Providers are uniquely identified by their National Provider Identifier (NPI) and characterized by a limited set of basic information (e.g. address, individual or organization indicator, gender and specialty). Places are categorized into office settings and facility (such as hospitals) settings, and reflect where the provider furnished a service. Services are identified by a 5-digit alpha-numeric code specified in the Healthcare Common Procedure Coding System (HCPCS).
The problem is simple. Some states allow master level clinicians to work to the full scope of their license and to bill under their own authority. Other states do not allow master level clinicians to operate independently. The certified nurse practicioners (CRNPs), nurse anthesiologists, physician assistants, physical therapists and other master level clinicians operate under the supervision of an MD or a DO. All of their billing rolls up to the MD or DO. Specialty clinics will often be set up so that one or two doctors will have half a dozen master level clinicians underneath them. Those master level clinicians will do most of the work but the doctor in an eight hour day could conceivably bill for twenty four to thirty hours worth of work. This study rolls up master level clinicians as providers in the states where they are allowed to bill independently.
The second issue is the time value of the Relative Value Unit.
I have a good auto mechanic. He is trustworthy in that he has told me that he could do $1,200 worth of work on my old Blue Beast and get her past inspection but by the end of the year, he would probably need to do another $1,000 worth of work. And that it would be a good idea to get a new car. He has called me and said that he could do the work on the exhaust system, but another garage could do the work at better quality and a better price because they are really good welders and he is just competent. He is a good mechanic. He also routinely bills 10 hours or more for eight hours in the garage itself. He bills off of a national time study book where each task is assigned a calculated time. When he is working he usually beats that time by a good margin.
RVU’s have the same problem. They are a skill and time estimate. Some providers out of a very large sample will naturally be more efficient than others. Some providers will be able to complete a task a bit faster than others. And this assumes the time value attributed to a particular procedure code is an honest mean time. There is a good argument that RVU’s, especially for specialty codes, overestimate the time needed to perform tasks. So there is a combination of providers who are naturally a bit more efficient interacting with codes that are unrealistically overtimed.
When Medicare started to release billing data, I was skeptical of simple division studies:
This means we can’t do a simple time management bullshit detection study based solely on “This provider is claiming he is doing 17 Medicare Part B procedures a day. Each of these procedures takes 30 minutes… IMPOSSIBLE”. That type of first level analysis might identify odd situations, but most will be explained
I’m still skeptical.