Benchmarking Evaluation & Management Codes

With the advent of EMRs, doctors are now able to track, monitor and visualize data that was not as easy to obtain as before.

One data set that EMR systems feature, are E/M codes. E/M codes, if you are not familiar with the term, stand for Evaluation and Management. E/Ms are represented by CPT codes 99211-99215.

Evaluating your practice’s CPT data is highly beneficial because the data reveals insights into the practice. For example, gathering E/M codes allows you to analyze each code’s distribution (e.g., how many 99213 is the practice ‘or a provider’ coding in comparison to 99214).

But how does one know if the distribution is a proper distribution? In other words, if a practice’s 99213 is two times greater than 99214s, is that good or bad?

My go-to guy for this sort of questions is Chip Hart. Chip has access to large data sets over the span of many years thanks to his company’s customers. And when it comes to this sort practice management analysis stuff, Chip is the biggest nerd I know.

Coincidentally, Chip posted on SOAPM a response that addressed this very question of E/M code distribution and what is an appropriate benchmark for pediatric practices.

I took Chip’s response (with his permission) and adapted it for the blog post. Enjoy:


B: You’ve done some work on E/M distribution, have you not?

C: I’ve done a lot of work on the semi-mythical E/M distribution topic.

B: OK, let’s get to it. What is the E/M distribution benchmark practices should use?

C: I want to mention a few things first.

  1. We are talking about Pediatrics data specifically. What Family Practice does, for example, is interesting or helpful in an argument, but really doesn’t pertain to pediatrics.
  2. I’m assuming that you really want to know what practices are DOING, not in fact what they SHOULD be doing. There is a difference.

B: Anything else we should know before you share the data?

C: Yes, let me remind you that what other people are doing should only act as a mild guidepost…just because a practice’s distribution is different doesn’t mean the practice is safer or losing money or whatever.

B: This sounds like an important point to highlight.

C: I know practices who do a great job with 60% 99214s and I know practices who should be in jail for their 15%.

B: Give me the bottom line then.

C: The bottom line is this, providers should chart what they did [in an exam] and code what they chart. Nothing more, nothing less.

B: Noted. Now, let’s get to the data.

C: Here’s some real pediatric data from millions of pediatric visits.

Screen Shot 2015-11-08 at 4.36.35 PM
2010 – 24% | 2011 – 25% | 2012 – 27% | 2013 – 29% | 2014 – 30% | 2015 – 31%


B: I’m confused by the numbers… I thought you were going to share with us a bell curve. What did you do here?

C: To provide a single, simple number, I just add the 99214s+99215s and divide it by the total 99212-99215 set.

B: What about the 99211?

C: In this benchmark, I am only looking at 99212s through 99215s.

B: Why is that? The 99211 are part of the EM codes.

C: Pediatricians shouldn’t be doing 99211s, and the “normal” curve does imply that pediatricians do as many 99214s as 99213s, for example.

B: Your “simple” single number changes the bell curve.

C: The peak [of the bell curve] is between the 99213 and 99214, not the 99213.

B: Let me see if I understand. To calculate my practice’s E/M distribution, I add all the 99214s and 99215; then I add all the 99212 thru 99215; and finally, I divide the total of 99214 and 99215 by the total of 99212-99215. Correct?

C: Yes.

B: Walk us through the interpretation of the result.

C: Looking at the chart above, for 2015, our clients bill a 99214 or 99215 31% of the time they do an E/M.

B: This way of calculating and benchmarking E/M distribution is different. E/M distribution charts traditionally show the percent for each code for a specific time.

C: I think this data is a lot better than the MGMA data for a variety of reasons (namely sample distribution).

B: What about wellness codes that were billed with an E/M code, do you factor them in?

C: This [data] does not include 9921X codes done during a well visit (i.e., 99213-25).

B: Are we talking new and established E/M coded or just established?

C: The data I shared above does not include NEW 9920X codes;

It is important to reiterate that while Chip’s data represents millions of pediatric claims, you should use Chip’s data as a reference among many.

In other words, just like a sailor uses multiple navigation tools and visual aids to determine its position, speed, and course – instead of a single reference point – you should use additional data points to determine how well your practice is doing.

Don’t forget to visit Chip’s blog Confessions of a Pediatric Practice Management Consultant