Hello, fans of healthcare pricing (looks out to an arena of thousands of cheering actuaries...) Back in July, we did a live blog as we voraciously devoured payer price transparency data. Oddly enough, people seemed to like that.
So, Price Points will be a similar series of short, informal dives into the new payer data, except what's actually in it vs. how we got it (+ an absolutely adorable new name). We're just eager to show, at a high level, what's possible with this new data.
This Price Point is written by yours truly, Chris, CEO of Turquoise, and notably not an engineer (if I was able to throw this data into Tableau / Excel, anyone can).
BTW: these blogs assume you've completed Steps 1-99: Find, download, download again, buy Amazon servers, hire engineers, cry, download some more, parse, buy more servers...etc. Or, you've got the cleaned up data from Turquoise or another source.
Next step: Attach to a provider directory and geocode.
The payer data in the wild lacks even basic provider information (like name, address, or the provider's favorite bone in the body). There's just a numeric provider identifier, either NPI or Tax ID (aka TIN). Unfortunately, payers like to use TINs, which are harder to connect to other data sources for non-payers than NPIs. We've attached the data to a TIN-friendly provider directly, which lets us marry it up to cleaner NPI data and derive physical locations.
Here's what it looks like to map Cigna PPO rates across TX surgery centers for facility fees, CPT 29881 (a type of knee arthroscopy). Challenge: can you spot what looks off? (I hope so)
Next step: Attach provider specialties to service categories.
Unlike the hospital data, where we can generally assume that a hospital will only list a billing code for a service it provides, payers are less savvy in knowing exactly what services each provider...uh, provides. In theory, payers could reason this out with claims data, but that wasn't part of the assignment from CMS.
So, data folks like us have to attach the provider's specialty (such as a taxonomy code, or another classification of your choice) to the service category the code falls under. I.e, an eye surgery center does not mess with knees.
The process laid out above still doesn't tell us for certain if the provider offers a specific service. To use this data for patients at scale, there are a host of other considerations that will require elbow grease: we'll save those for another blog.
However, let's have a look at how the data cleans up once we get specialty into the mix, and we remove a few closed / outdated providers.
What does this leave us with? It looks like there's a fee schedule rate that applies to all of Texas Health's surgery centers. With Baylor Scott & White, the $1,549 rate looks relatively low. This is either: 1) a great deal in Dallas, 2) a general surgery center that does not specialize predominantly in orthopedics, or 3) an issue in Cigna's PPO rates file. Regardless, it's interesting.
A final step - let's browse around TX!
Every region in the Lone Star state has different prominent health systems driving the market. For kicks, let's put Dallas next to Austin and Houston and compare Cigna PPO's facility rate of CPT of 29881:
Spoiler alert: MBC Ambulatory Surgery Center doesn't appear to be super real per Google, unfortunately. Otherwise, we can see that Memorial Hermann surgery centers generally may have more favorable rates than their Texas Health counterparts in Dallas. Cigna PPO members with achy knees, take heed.
Going a bit further: exploring other payers in TX, or more
In future Price Points (this is fun just to say!), there are myriad of routes to continue exploring. Here are a few things we can try next:
- Compare to the hospital disclosed negotiated rates data for Cigna (from the hospital files)
- Compare against other payers in TX (after all, Cigna has low single digits market share in TX - UHC or BCBS could be interesting comparisons)
- Compare against other products within Cigna (OAP, eg), or other products at other payers
- Call every provider in the graph and ask them, definitively, what their favorite bone in the body is (the ulna, maybe?)
Until next time!