Nicholas Bagley at the Incidental Economist passes along a clear case of fraud, waste and abuse at a massage clinic. He then laments the oversight process failing to catch the fraud at a low level and writes a simple, elegant paen to informatics that is completely useless:
What’s demoralizing is the familiarity of this story. Back in 1997, the D.C. Circuit issued another opinion involving a Washington, D.C. physician who claimed he worked more than 24 hours in a day. Any halfway-sophisticated computer system should have uncovered fraud this blatant. Well over a decade later, however, Wheeler was able to do exactly the same thing. Even then, it took an informant’s “tip” to clue officials into the fraud.
A good computer program can investigate oddities and identify outliers. It can’t prove fraud or even get a 2 sigma estimate of fraud.
There is something that he is missing. Claims payment and claims processing rarely reflects reality, especially when services are performed by mid-level clinicians. We saw this with the recent release of CMS Medicare claims data. Some of the highest paid individuals were truly performing all the services that they were getting paid for. However, others were medical directors of large studies or claims rolling up to a single NPI or Medicare ID:
-
Claims rolling up to a provider’s NPI or Medicare ID. Non-MD/non-D.O. clinicians such as Certified Nurse Practicioners, Physician Assistants, Master and Doctorate level Physical Therapists etc. often will roll their billing up to a doctor’s Medicare billing number. 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 by seeing three or four CRNPs/PAs doing most of the work that the doctor than bills for.
The story that Bagley flags in his post is for a massage clinic. Assuming they are billing for therapeutic and rehabilitative massages (a legitimate billing category), most of the providers who are actually performing the massages are certified massage therapists or atheletic trainers who are loosely suprevised by a single doc. It is quite possible that five, six, or seventy people are billing under a single claim number so seventy hours of compensated care can be provided in a single day. I know when I tweaked my ankle and needed physical therapy, the billing was done by an orthopedic surgeon who I saw once for eight minutes and then I worked with a great therapist for the next eight weeks. Every session I had, I saw at least four therapists on the clock with patients.
An automated outlier detection system is useful. But it won’t be the end all and be all of fraud detection and prevention. Medical billing is too byzantine and convoluted for that.
Someguy
Fraud at a massage clinic? That’s terrible.
I was expecting a happy ending to that story.
Richard Mayhew
@Someguy: Try the beef and remember to tip your waitress
askew
As someone who works with healthcare fraud, this entire post is so true. Trying to identify a unique provider based on the number of different types of IDs and various claims systems is what I do all day. It’s kind of a nightmare.
RobertB
@Someguy: http://instantrimshot.com/
gwangung
Fruit of the free market, doncha know….
Richard Mayhew
@askew: I’m sorry —that job sucks.
I’ve spent too much time in my career trying to match up addresses across half a dozen domains, so I don’t want to think about matching NPI to SSN to TIn to local propiertary ID to taxonomy to birthmarks
askew
@Richard Mayhew:
Addresses are just as bad as matching IDs. I weep when the project I am working on involves getting unique providers with 1 address.
max
An automated outlier detection system is useful.
The Wire:
max
[‘Well, *I* certainly feel safer already.’]
KG
as a lawyer, it’s not uncommon to bill more hours than you actually work. part of it is rounding (lawyers typically bill in 6 minute increments – so 15 minutes becomes 18 minutes), part of it is the ability to double bill items (such as, I spend all day at a mediation, while I’m waiting for the mediator, I draft a letter on another case), part of it is having a firm policy for billing a certain amount for particular things (one firm I worked at had a rule that you never bill less than .3 hours on something). so, I could see how a doctor could do the same thing. but 24 hours in a day is just stupid.
Richard Mayhew
@askew: yeah. I’m spoiled. Our address data internally is well curated, highly structured, and we own the data so when formatting changes, we can do it ourselves. We have some control and knowledge about our data set. however, when we have to play nicely with our state regulators, their data set is not controlled by them; it is controlled by the docs so the following addresses are all considered “valid”
123 Main Street
123 Main St
123 E. Main St
123 East Main Street
Suite 100 123 East Main Street
1st Floor 123 East Main Street
Main Street and Subsidiary Avenue
500 Big Town Commons
Same fucking location, all on the same file without any coherent crosswalk to a shared identifier.
Richard Mayhew
@KG: But to use the law example — if you have three paralegals working for you, they roll their billing up to your fictitious Lawyer ID Number so 30 billable hours in a day is not hard to achieve under a single fictitious Lawyer ID number
japa21
The key to whether or not fraud is involved, in those case you talk about Richard, is the use of modifiers.
If the codes being billed are for full MD services, then there is an issue. Generally speaking, various mid-level providers are paid at a level less than full MDs.
If PAs, CNPs, PTs and OTs are performing the services but the claims are not coming in that way, then fraud is involved.
Additionally, claims are to show the name of the specific provider as well as the billing provider. Again, if services are being rendered by a mid-level but the name is not included, there is the potential for fraud.
As you point out, a practice, be it a massage clinic, a PT office, a large medical group or a hospital will have an NPI and TIN exclusive to that practice, and if the reimbursement goes to the practice, those are the only numbers that count. Each of the individual providers will have their own NPI, but that is somewhat irrelevant.
jl
Why are the claims rolled up into one MD’s ID?
Shouldn’t the MD bill for whatever low level on call or consulting stuff, and the actual providers bill?
Is it just easier to do it that way? Something related to an MD’s order required for reimbursement? No reimbursement codes for non-MD providers? What?
scav
@Richard Mayhew: Lord, you haven’t even waded into the non-standard part of the address involving offices in larger buildings. This and general ID DB cross-walk building was sooooo much of my life . . . . ZIPS to FIPS! Cass certified!
eta oh, you did at the bottom, front-loaded which is the worst place too!
japa21
@jl: Actually, in many of these cases, it is the practice doing the billing, not the providers. Now, the practice may well be under the physician’s name, usually with an SC, LLC, or some such after the name. But it is the practice billing for the services rendered by the lower level provider.
Trollhattan
O/T but while amusing myself googling “Alabama man” I found this doozy.
http://www.msnbc.com/msnbc/alabama-gop-offer-money-to-find-voter-fraud
But, that’s not the only right being oppressed today in Alabama.
FlipYrWhig
Shouldn’t they, like, not do that? Or have the practice number be the one everyone uses, and a “primary responsibility” code for everyone who does a specific task?
Calouste
@Trollhattan:
It always surprises me that these supposedly Christian ammosexuals (there is a pretty big overlap) think they should be allowed to carry a gun in a church. Swords to Plowshares, the Prince of Peace, Turn the other Cheek, seems like half the content has been expunged from the Bible they read.
Mnemosyne
@FlipYrWhig:
G sometimes has to deal with medical billing. It’s screwed up far beyond what you would believe. How anyone gets paid, I have no idea.
ranchandsyrup
NONONONO the existence of any fraud, waste, or abuse immediately demands the immediate termination of the program. Because reasons.
Villago Delenda Est
In order to fight the fraud at low level, you need to have lots of paid inspectors whose job it is to detect the fraud and bring it to the attention of those with the legal means to deal with it.
To do that, you need funding for the inspectors.
To do that you need to have non shitheads in legislatures and congress who will approve the funding.
To do that, you need to defeat the shitheads and the voters who support the shitheads in elections.
Villago Delenda Est
@Calouste: These people are not Christians. They are heretical worshipers of Mammon and Morloch.
piratedan
@Calouste: pretty sure that they stopped somewhere in the Old Testament because if they ever got around to what that Jesus guy was supposedly practicin and preachin, well it’s patently obvious that guy was a freeloading hippie.
Bill Arnold
@japa21:
(and Richard Mayhew, and askew)
What’s the current bleeding edge in medical insurance billing fraud detection?
Is there any machine learning involved (supervised or unsupervised or both), or are the detection methods mostly hand-crafted?
KG
@Richard Mayhew: ah, that’s where law and medicine are different. paralegals, clerks, and other staff are typically billed at a lower level than lawyers (who depending on their rank may also be billed differently). so, you might have 20 hours for an associate on a case, 15 for a paralegal, and 5 for a partner, but they’d all be billed separately. Still, I get where you’re coming from.
Ryan G
IANAD but my shifts are all coded on the date they start… if I work 12 hours on June 2nd dayshift and then 12 hours + 2 overtime on June 2nd nightshift, that’s 26 hours in a day. Nevermind that I’m not done until 9AM on the third.
Doesn’t apply to the specific case here but I’ll bet there are thousands of ER docs – and especially residents – who run into this sort of thing.
soprano2
As a person who works with HIS maps regularly I know what a PITA addresses can be. I remember how shocked I was the first time I was in a meeting where the subject of addresses came up. I thought they were simple, they’re anything but.
Richard Mayhew
@soprano2: Yeah, every couple of years, we get a wet behind the ears fresh from the management intern who wants to do something AWESOME and Six SIGMA-Y with address data and it takes a good three months to demonstrate that this project/idea is a great idea but it will take much more than three months to accomplish a minimal efficiency gain, so if you really want to see a three basis point increase in the claims auto-adjudication rate and can find thirty man years to build the project and 2 FTEs to maintain it, we can do it, BUT…..