Back in January, PTs and OTs experienced a pretty dramatic shift in the way they coded for initial evaluations. As you probably—hopefully—recall, this change required therapists to select evaluation CPT codes based on the complexity of each individual eval. (If this doesn’t sound familiar—or if you want a refresher on the details of the new evaluation code set—then I’d recommend reading this post ASAP.)

This was a major change—one that thrust our entire profession into uncharted territory. And if you followed the chain of events leading up to this transition, you’ll remember that, originally, payments for these codes were going to be tiered according to complexity, with providers receiving different reimbursement amounts for low, moderate, and high complexity evaluations. The logic CMS developed to establish those differentiated payment rates was largely based on projection; after all, they had no real data to inform their code valuations. And that was part of the reason why, in the end, CMS decided against offering higher payments for more complex evaluations.

Fast-forward seven months: As the largest rehab therapy-specific EMR on the market, WebPT has the privilege of collecting a lot of industry data. A good portion of it is information that only our internal data analysts get excited about. But, in some cases, our data is incredibly insightful and relevant beyond our own four walls. This is one of those cases.

Since January 1, 2017, WebPT users have completed more than a half-million evaluative notes—which means we have a whole lot of evaluation CPT codes saved in our system. And in analyzing that coding data, we’ve identified some interesting trends, including the following:

1. The actual distribution of code usage does not match CMS’s projections.

As we explained in this post, CMS expected PTs to use:

  • the moderate complexity evaluation code 50% of the time,
  • the low complexity code 25% of the time, and
  • the high complexity code 25% of the time.

Furthermore, CMS used those projections to come up with tiered valuations that would, in theory, yield work neutrality. Still, as we explained in the same post, “there’s no way to guarantee that the actual distribution of reported codes would fall in line with those projections—which means there is no way to guarantee budget neutrality.”

This was a big part of the reason CMS hesitated to assign different reimbursement rates to different evaluation complexity levels. And as it turns out, that hesitation was not unfounded. According to our data (shown below), the distribution of low-complexity and moderate-complexity codes is nearly identical: 45.94% of billed evaluative codes fell into the low-complexity (level 1) category, and 45.20% fell into the moderate-complexity (level 2) category. However, only 8.86% of billed evaluative codes fell into the high-complexity (level 3) bucket.

Evaluation Code Distribution by Complexity

What does that mean for the future of code valuations? Well, looking at the overall data, it would appear that CMS could actually save money by implementing a tiered pricing structure based on the originally proposed work RVUs. But, when we dive deeper into the data, a different story emerges.

2. Compared to younger patients, elderly patients underwent a significantly higher number of high-complexity evaluations.

Looking at the overall age distribution for all levels of complexity (shown below), you can see that in general, the distribution skews toward the post-50-year-old population, with the largest spike occurring between ages 65 and 72. So, we can safely assume that, compared to beneficiaries with other insurance types, Medicare patients are undergoing more therapy evaluations.

Total Distribution by Age

That being said, there is a clear—albeit much smaller—spike in distribution around ages 14-18. This likely represents the high school sports population.

When we look at the age distribution for each individual level of complexity, however, things get a little more interesting. As you can see in the graph below, the distribution for moderate-complexity evaluations clearly skews toward an older population, and the distribution for high-complexity evaluations shows spikes in the early childhood and elderly age ranges. And considering that older patients are much more likely to have Medicare as their primary insurance, this theoretically means that, in addition to undergoing more therapy evaluations overall, Medicare patients are—compared to commercially insured patients—undergoing more high-complexity evaluations.

Complexity Distribution by Age

This is important, because it could affect Medicare’s willingness to implement a tiered pricing structure in the future. After all, why would Medicare pay more for high-complexity evaluations when the majority of patients undergoing those evaluations are, in fact, Medicare beneficiaries? And if Medicare doesn’t adopt differential pricing based on complexity, then there’s a good chance private payers won’t, either.

3. Compared to men, women undergo more evaluations and account for a greater proportion of high-complexity evaluations.

As shown below, women accounted for about 30% more evaluations than men.

Total Distribution by Gender

Additionally, men were less likely than women to undergo moderate-complexity or high-complexity evaluations.

Complexity Distribution by Gender    

Any number of factors could be responsible for these disparities. For example, multiple studies have shown that men are more reluctant—and less likely—to seek medical care for issues affecting their health and wellbeing. So, it would make sense that they would account for fewer initial PT and OT evaluations. Additionally, as this article points out, once men finally do schedule appointments with care providers, they are “less likely to be honest once they get there.” So, it would make sense that our society’s “cultural script about masculinity that tells men they need to be tough, brave, strong and self-reliant” would deter them from completely and accurately describing the full range of their issues and impairments—to, in essence, downplay the severity and complexity of their ailments. And that, in turn, could result in fewer moderate-complexity or high-complexity evaluations—even in cases where the evaluations should actually be more complex.

4. Evaluations for cases involving abnormalities of gait and mobility tended to be more complex than evaluations of patients with other types of diagnoses.

Of the four diagnosis categories shown in the graph below, the R26 category—which accounts for diagnoses related to abnormalities of gait and mobility—was associated with a greater relative number of moderate-complexity and high-complexity evaluations. (For reference, the other diagnosis categories analyzed were M25 [joint disorders], M54 [dorsalgia], and M79 [soft tissue disorders].) This may be due to the fact that gait and mobility issues often result from other concurrent injuries and conditions, and evaluations of patients with multiple diagnoses tend to be more complex.

Complexity Distribution by Diagnosis Category  

Now, it’s important to remember that these codes are still fairly new, and these trends could change as PT and OT providers get more comfortable with the code selection process. It’s also important to note that without any financial incentive to not only accurately code, but also accurately justify each coding choice, there’s a good chance that the coding data is not totally representative of the actual complexity of the evaluations completed since January 1, 2017. All of that being said, I want this post to serve as a reminder that data matters—it can, and will, influence future decisions and policies that will impact not only our payments, but also our reputation as doctorate-level practitioners with the clinical expertise necessary to accurately classify, categorize, and code for the services we provide.


Special thanks to Joe Dundas for performing all of the data analysis for this post.