CMS is proposing major changes to overhaul E&M services.
As I read through more and more articles and white papers on the new proposed 2019 evaluation and management (E&M) changes, I got more rather than less confused.
It feels like the Centers for Medicare & Medicaid Services (CMS) had this epiphany one day in its sleep and then decided to change the landscape of the entire E&M model the next day. While I agree with the concept (simplification), and I am excited that CMS has been listening to the needs of providers, it feels like they are going about this in the wrong way.
First of all, any change such as this, which may result in huge financial and utilization impacts for practices, should be rolled out slowly. For example, how about we pick 1,000 or even 10,000 providers at random and subject them, as a test group, to the new proposed changes? In that way, we will have the opportunity to see what the impact will involve. For example, what if the overall payments are significantly more than estimated as a result of changes in physician behavior? Or what if the payments are lower than expected, imposing a more onerous financial burden on doctors?
Since this is about efficiency, I would like to see whether this really improves that. If it really does give the provider back three weeks each year to be more productive, then maybe a reduction in per-unit payment may be balanced by the ability to process more units (i.e. see more patients). But then, from an accounting standpoint, the practice would need to model around variable expenses, which normally don’t change based on volume. So a reduction in payment per unit when the underlying expense is the same may lead to a lower profit margin. Obviously, this won’t be the case for everyone, but I believe that it has to be considered.
Another concern I have is based on standards. For example, in the past, like with the budget neutrality reduction factor on the work relative value unit (RVU) or the conversion factors, private payors, like lemmings, seem to fall right in line with CMS changes. What’s good for the goose is good for the gander. But what if this isn’t good for the goose?
I ran an analysis of what this would have looked like if it were to have been in place in 2016, and the result was a net increase in payments of around a quarter of a billion dollars. When I ran the 2016 utilization data against the difference using the 2018 RVU data, the net impact was an increase in payments of nearly half a billion dollars. If this holds true, and it was based on some simple math and complex algorithms, then the private payors would likely reject this change, meaning that each practice would now have two standards for submitting claims for office visits. Well, at least two standards. Add in the totally confusing and administratively complex Medicaid systems, and it just adds to the confusion.
And speaking of this massive increase in payments, how will CMS handle that? My understanding of the Budget Neutrality Act (BNA) is that Medicare cannot expand or contract by more than $20 million from the year before. So if it’s true that this may add some half a billion dollars to payments, then CMS is going to have to find a lot of places to cut over $400 million from other procedures and services (which, in and of itself may cause more damage than the benefits associated to some providers for the increases).
I have also been wondering how the change in the RVU is going to comport with current analytical models. The Resource-Based Relative Value Scale (RBRVS) Update Committee (RUC), which is owned and maintained by the American Medical Association (AMA), pretty much has a monopoly on the assignment of RVUs. For sure, the work RVU is tied almost completely to the RUC time recommendations. And what may be a not-very-well-known fact is that the practice expense RVU is also tied to those time recommendations. In fact, if you run the work RVU against the RUC time data, you get a correlation coefficient that is very close to 1 (0.956). I am not sure how the four RVU values will be computed (two for new office visits and two for established office visits). I might just take the weighted average for the two lowest codes and assign that to the low level, plus a weighted average for the top three levels for the high level, and hope that the math works out. But no matter how it is done, just like non-mathematical manipulations have occurred in the past, I believe that CMS is going to have to play around with the final values. That means that, as with the practice expense RVUs of the past, the new RVUs may be tied to budget restrictions rather than meaningful analytics.
Finally, how is this going to impact those providers that are engaged with productivity-based compensation? Whether it’s just a bonus model or actually tied to base salary, this radical a change in the current model may create more disruptions than we would like to see. I mean, the reality is that those physicians who may experience a reduction due to RVU values that, on average, are lower than what is currently being reported may very well increase the number of patients they see, spending less time with each, but still getting paid the same per visit. Don’t get me wrong; I’m not saying that this will negatively impact quality. I’m saying that it might impact quality, and whenever quality of care is called into question, research should always follow. In the past, CMS has imposed rules and regulations based on fear of changes in physician behavior, such as the elimination of physician-owned specialty hospitals. We have also seen changes to RVU values and conversion factors based on CMS concerns over changes in physician coding and billing patterns. If there really is that concern, then now may be a time to challenge that assumption.
In conclusion, kudos to CMS and its administrative staff. This appears to be a Herculean effort to improve efficiency, reduce administratively complex and medically unnecessary rules and regulations, and make life generally better for the average provider. My opinion is that in any change of this magnitude, testing should be done first. When possible, we should simulate, but in this case, it has be done in vivo. We need real-life living guinea pigs to get a better handle on the potential consensus, both good and bad and both big and small. And that means selecting a sample of physicians and transitioning them to the new model over some reasonable time frame. Then and only then will we be able to estimate the full impact these changes might have, giving us the opportunity to tweak the model so that we can optimize the results across the entire provider spectrum.
And that’s the world according to Frank.