Dollars Per RVU Benchmark for Pediatricians

[Quick segue - come see the likes of me, Susanne Madden, Richard Lander, Rick Oken, and more speak at the VA AAP and Pediatric Alliance 4th Annual Meeting in March.]

Igor has been really, really busy lately and it's great!

He's off researching a concept for Dr. Schwartz (I can't remember what clever nickname I gave him here on the blog once) where the latter posits that part of the decline in visits for 2011 can be attributed to the higher HDHP plans and, as a result, they are seeing fewer "multi-patients-per-visit" episodes.  In other words, it used to work that a mom would bring in Sick Kid #1 and toss in one or two "oh, by the ways" because the copay was just $10 per or something.  Now, with the HDHPs hitting the deductible for each of those kids, the docs will stare a sick kid right in the eye and the mom'll say, "Oh, he's fine."

We're looking into it.  Igor and I have a plan.

Meanwhile, SOAPM folks discussed the concept of "Dollars Per RVU" recently as a benchmark.  When you discuss "Dollars CHARGED per RVU" in your office, it creates what's known as your "Frequency Adjusted Conversion Factor" and essentially tells you how much, on average and relative to Medicare, you charge.

When you look at it on a "Dollars PAID per RVU" you get a sense of how much you are being paid.  Sadly, as you'll see in a moment, our clients aren't getting paid much more than the 95-99% of Medicare they were getting a few years ago.  That's not to say they don't make more money, per se - simply coding properly can generate a lot more income, even at a steady or declining payment rate.

So, Igor and I took at a look at all the PCC clients from 11/09 through 10/10, cleaned up the appropriate outliers, removed all non-RVU-valid procedure data, and here's what we learned.

PCC's pediatric customers generate, on average, 97% of 2010 Medicare rates in payments.  It worked out to about $35.14-per-RVU.  If you want to figure out $$/Work RVU, just remember that Work RVUs are part of the 50/48 split, so you really just need to double it to figure it out.

This is essentially identical to the figure I calculated almost 10 years ago when I first looked at the data.

Playing with the numbers, what do I see?

The range for the figure is from $13.74 (38%) to $57.67 (160%).  That's right - we've got practices who average less than 40% in Medicare and practices who average over 160% of Medicare.  For the very same work.  95% of PCC's practices are between 76% and 118%.  About 10% are under 70% and 10% are over 120%.

Any correlations I could find?

• Practices with a Sick:Well visit ratio less than 2.0 hopped up to 103% and those over 2.5 drop to 91%.  However, many of our clients with the largest S:W ratios (>4) jump way up to 115% and beyond.  I suspect they could be coding differently.
• Our clients who perform a lot of 99214s and 99215s - as a group, they average over 75%, some over 90% - perform very poorly, with a 76% Medicare payment rate.  Most of these practices are extremely high Medicaid, but not all.
• Our clients who charge the most - over 200% of Medicare - do alright, as they collect 103% of Medicare.  Those who charge too little (< 140%), fare poorly at 86% of Medicare.  Those who charge much too little (< 120%) are at 76% of Medicare.
• There is clearly massive regional bias.  NJites bring in 86%.  NY brings 97% (to my surprise). And the state that Jon Caine loves to hate, MA, averages - you ready? - 122% of Medicare.  Our sample size may be suspect (about $25m in charges), but our MA clients clearly get paid well.  So does NH. Oklahoma - 112%.
• Gut instinct tells me that Medicaid volume affects this number, but I can't prove it much.  I don't trust the Medicaid percentages from our clients (not all of them configure it in such a way we know it's a Medicaid plan without manual intervention), but there looks like slight correlation.  The top 10% Medicaid practices come in at 92% while everyone who sees NO Medicaid, has 99%.  But take that one with a BIG grain of salt

How's that for a night's work?