How Can We Improve the Use of Administrative Codes to Measure Quality?

By | October 18, 2022

Historically, we have designated administrative codes as methods for reimbursement. However, with the evolution of value-based care, these same codes are now applied for risk stratification for adjustments to value-based payments. This variation of fees is necessary because the degree of illness and complexity drives the clinical needs and, thus, the cost.

However, using this methodology has created unintended dynamics. To receive the correct payment, clinicians must apply the administrative codes correctly. Consequently, the industry has focused on and emphasized the importance of accurate coding, resulting in a whole industry created to aid clinicians in ensuring this happens. The function allocated to the healthcare professional of selecting which codes to apply, has led to concerns and lawsuits concerning gaming the system by upcoding when inappropriate.

Currently, periodic audits are used as the system for checks and balances, and there are steep penalties if caught. Additionally, CMS has capped the increase in payments that occur with increased severity coding for many of its value-based programs. Neither method makes sense since, currently, there is no way to “police” all coding, and capping prevents the true intent from occurring and may penalize appropriate behaviors.

An additional dynamic is in play when we move more toward capitated payments. These payments rely heavily on severity adjustments for correct calculations. However, how do we adjust payments if one uses only the submitted administrative codes and then isn’t required to submit a claim for compensation with the needed administrative codes for risk adjustment? Currently, this requires one to manually update such information to the payer, which is incredibly cumbersome. Payers also need this information in the Medicare Advantage arena as the same information adjusts their payments from the government.

Considering all of this, it is important to periodically step back and contemplate ways to improve our present model.

Risk adjustment will never go away as the natural progression of diseases, and the human body always worsens over time, thus requiring more intense services. However, potential solutions do exist or, at least, enhancements to the present model. The first involves ensuring the robustness and rigor of the administrative data. And it’s important to remember that CMS has approved qualified clinical data registries developed for numerous clinical specialties and diseases. These registries will allow for better auditing. Unfortunately, this doesn’t review in real time, it is a  look-back model.

Also, somewhat of a look-back model, but much more accurate, is utilizing technology to better understand the severity of illness. Embedded within our electronic medical records exists information on diagnostic testing and results and descriptive information in our notes. Furthermore, we have the ability to “look” into those records directly for supportive information. One can think about this model prospectively at some future state. Of course, one can argue that currently, such information is what is driving the codes now. Still, we can be more accurate by removing the step of reviewing copious amounts of data to create the “right” administrative code and doing it automatically. The audit then goes in the other direction, ensuring the data is correct and actionable. At present, we gravitate towards utilizing technology to “edit” what humans are doing; however, there will be a time when the opposite should occur.

The most significant point is the imperativeness of considering all these dynamics and continuously enhancing our ability to better implement our administrative function of payment for services. There is no one right way, as it genuinely requires polarity management; how do we best optimize the complexity of the current question?