CMS Ends RAPS for Risk Adjustment - Are You Prepared?

Starting this year, diagnosis information will now be the main source of information used to risk-adjust payments to Medicare Advantage (MA) plans. Diagnosis codes are used by the Centers for Medicare & Medicaid Services (CMS) as part of their risk adjustment processing system to calculate risk scores for each member enrolled in MA. In the past, CMS has used diagnosis submitted into its Risk Adjustment Processing System (RAPS) by MA plans. Starting in 2015, CMS began transferring the main origin of data to diagnosis codes on encounters submitted as part of the Encounter Data System (EDS). In this post we’ll look at the differences between encounter and RAPS data. We’ll also cover what challenges these adjustments pose to MA plans.    

One of the major differences between a RAPS submission and an encounter submission is that the encounter submission includes more information from the patient encounter. The data presented in an encounter submission is more closely aligned with a traditional Medicare fee-for-service (FFS) claim, which includes data points from the ANSI 837 v5010 claim format. Whereas a RAPS data submission only includes the date of service, the diagnoses, provider type and beneficiary ID. 

Another difference of note is, under RAPS it’s up to the MA plans consideration as to what diagnoses get submitted for risk adjustment. It’s left up to the plans to determine which diagnosis best represent the patient’s conditions, and it is not required of the plan to submit each individual diagnosis from each patient encounter. However, under CFR 42 422.310(d)(1), the new system, MA plans are now required to submit records for every patient encounter. In turn, CMS now reviews the encounter data and decides which diagnoses will be entered for risk adjustment, not the MA plan. 

CMS has established a screening logic that’s based on procedure codes using a subset of Current Procedural Terminology (CPT) and Healthcare Common Procedure Coding System (HCPCS) codes to assess diagnoses on encounter submissions for risk adjustment. 

What this means is that in order for a record to be deemed as an acceptable source of diagnoses for risk adjustment, the encounter submission must be affiliated with one of CMS’s screening procedure codes. CMS established the procedure code list to certify that diagnoses specifically from eligible healthcare providers and visits are accepted. For example, procedures from technicians are typically excluded, while in person visits between provider and patient are usually accepted.   

There are other data points from the encounter record that are also important. The dates of service must occur within a certain time frame for the service to be eligible for risk adjustment. More information on the development of the procedure code list can be found here. 

Lastly, the timeframe in which the data must be reported is also different. RAPS data is required to be reported quarterly, but encounter data is required to be reported monthly at the very least. 

Between years 2015 and 2021, the calculation of risk score was established using both RAPS and encounter data. What was known as a blended risk score. That blend by design has by slowly increasing the weight of the encounter data. When you look back at 2021 the risk score was 75 percent encounter data and 25 percent RAPS. The increased weight of encounter data in risk score calculation will no longer allow plans to supplement encounter data with RAPS. Such was the case from 2019 to 2021, when plans could include diagnosis data from inpatient records submitted to RAPS.  

The shift to encounter data and the implications for risk adjustment were set in motion by requirements in the 21st Century Cures Act and are expected to be completed this year (2022). The changes include the diagnoses acceptable for risk adjustment and other modifications that will affect the overall value of diagnoses used in risk score calculations. See our explainer on condition counts.    

The shift to encounter data has raised some concern with health plans, specifically around the quality of the data. Concerns that encounter data is often incomplete or prone to inaccuracies have plans worried that it could equate to lower risk scores, causing lower reimbursements. America’s Health Insurance Plans (AHIP), one of the nation’s largest associations, indicates research that suggests risk scores from encounter data were four to 16 percent lower on average than that of scores using RAPS.  

Government agencies like the Department of Health and Humans Services Office of Inspector General (OIG) and the U.S. Government Accountability Office have raised concerns about the accuracy of encounter data and its reliability related to program payment. The concerns raised brought forth the decision to slowly phase in the use of encounter data over time. The slower implementation was designed to let CMS and plans take extra time to ensure the quality of the encounter data itself.  

With all the concerns brought forth by various organizations it will be essential that plans and the federal government continue to evaluate the quality of the encounter data. Plans will no longer be required to support two processes for submitting risk adjustment data, it is yet to be seen if the costs of maintaining only one system will equal decreased costs and oversight. It may happen that an increase in validation costs may occur compared to only using RAPS data, time will tell. 

The increased volume of encounter data and the frequency of that data could represent a new set of additional costs. In addition, some plans might choose not to sunset their RAPS processes until they know the quality of diagnosis reporting is 100 percent accurate. Or perhaps until they streamline the process of capturing encounter data altogether.

The transfer of management over which diagnoses are entered for risk adjustment from MA plans to CMS increases the challenge of identifying documentation to support the diagnoses submitted.

The range of distinguishing documents to support the diagnoses is now broadened to every submitted encounter and previous diagnosis recognized for risk adjustment. The significance of this from an audit standpoint is of note since requirements state insurers are to refund CMS within 60 days of becoming aware of an overpayment.

Submissions accepted for risk adjustment and the disposition of the submitted encounter are provided to plans via EDS return reports. It is still left up to the plan to determine which of the accepted encounters are actually supported by documentation in the patient record. 

This burden can potentially be mitigated by acquiring software that can rationalize patient data using artificial intelligence. The latest software from ForeSee Medical can rationalize encounter diagnoses with all available supporting medical record documentation and pinpoint missed revenue opportunities. For example, diagnoses supported in medical record documentation, but may have been unsubmitted on an encounter. Accurate and complete diagnosis coding during the encounter is required for providers to perfect their Medicare risk adjustment scoring and increase RAF value capture. 

There are major changes rolling out to the modeling and data used for risk adjustment this year and plans will need to be prepared. These changes made it even more vital than ever to have methodical systems in place for data validation. As it goes with all risk adjustment changes, the implications will be varying for MA plans depending on enrollee population and available means for submitting and maintaining the accuracy of patient data.   

Although there are many unanswered questions, it’s likely they will be answered as things progress. One question, since the filtering logic is focused around procedure codes, CMS may choose to have those procedure codes validated down the road. Another question, could at some point encounter data be used to regulate the CMS-HCC risk adjustment model and replace the traditional Medicare fee-for-service claim as a data source for regulation? A decision like that would need to be implemented in the long term, since CMS would need to outline policy and tech issues related to implementation. 

In the shorter term, improved encounter data could make it possible to have more full spectrum analysis of the actual value of care provided in a MA program. This more robust analysis could bring the added patient information that encounter data could bring as opposed to the traditional RAPS system. Additionally, the updated frequency of encounter data may encourage a more prospective approach to coding patient conditions and provide CMS with more real time insights on utilization trends.

For healthcare organizations looking to succeed in the evolution to value-based care delivery models, ForeSee Medical is a specialized software platform designed to increase the profitability of Medicare Advantage risk contracts. Using A.I. including NLP technology and machine learning, ForeSee Medical perfects HCC risk adjustment scores, empowering providers to positively influence health outcomes.

Blog by: The ForeSee Medical Team