Risk Adjustment

Show Me the HCCs — Final Notice by the Numbers

Written By
Meleah Bridgeford
Director of Risk Adjustment, Payer Solutions

The announcement for 2020 PY saw some major changes, both expected and not so expected. First, the expected. CMS has changed the model once again. They’ve chosen the Alternate Payment Count (APC) model. Back in our initial analysis of the model changes, we noted the decrease to risk scores between the 2017 and the 2020 models was 3.6%. However, this decrease was before the coding intensity and the normalization factors were taken into consideration.

Now that CMS has released the normalization factors, we can see how the average risk score decreases by 7% between the two models (using the 2018 normalization factor). The writing has been on the wall, so to speak, as both the 2017 model for RAPS and the 2020 model for EDPS have seen an increase in normalization factor. The remaining 50/50 blend between RAPS and EDPS notwithstanding, an overall decrease in the risk scores is coming.

This increase of the normalization factor shows that CMS is expecting an increase in the HCCs captured by MA organizations. This number, though, is not guaranteed to be reflective of your member population. In fact, it is guaranteed to only account for an “average” plan population. Some codes captured will come from the new HCCs, but this will only account for a portion of the decrease—we’ve estimated average reimbursements to fall by a full 3% more than CMS has accounted for if new HCCs are not captured. Premiums are going up, but coding trends overall will need to be increased at least by that new deficit.

This increase in coding trends will have an impact on RAPS submissions as well, which received its own increase in normalization factor. In fact, the RAPS 2017 Model blend saw a 5% increase to the normalization factor between the 2018 PY and the 2020 PY, with a 3.2% jump between the 2019 PY and the 2020 PY. So along with the decrease in the EDPS model—and utilization of the APC—there is a decrease on the RAPS side of the model as well. The 2017 model being utilized for the RAPS risk score does not have the benefit of the new HCCs added in 2019 and 2020 models, thus making any increase in risk scores reliant on an overall efficiency in HCC capture rates.

Perhaps most interesting of all is the impact to specific RAF Type Codes. The ESRD/Dialysis members saw the most significant impact with the switch over to the 2019 model—the difference between 2019 and 2020 models appears fairly minor as a result. In our initial analysis, we observed an impact to duals members being significant and this still appears to be the case as the impact of the normalization factor now decreases the average risk scores for this population even more. The prolonged use of the 2017 RAPS model will supplement this population most in light of risk score changes.

All in all, in an industry always looking for the next best breakthrough in suspecting analytics, chart retrieval rates, and coding capture results, the drive will be all the more frantic as individuals begin to realize the impact of the new model. Our suggestion is to stay ever mindful of your analytics. Know your data and know your results. Plan right now for the 2020 model changes. You will see the impact of focusing on different HCC capture programs. A diverse plan to find codes is about to be more necessary than ever before as we transition into the new models.

Indeed, it’s only through the capture of new HCCs that the lowering of the risk score can be counteracted— whether from a newly added HCC or one previously uncaptured.

Of course, this claim is meaningless without an example. So we took a sample member to see what happened when we found one of these new HCCs and not only does the risk score increase, but it does so at a rate of 1.9% between 2018 and 2020 and 0.4% between 2019 and 2020, showing the benefit of the APC model’s condition count risk scores. This also benefits the EDPS risk scores, which could be lower than the RAPS risk scores in some cases. Capturing the new HCCs will help reduce this variance, of course this also requires the monitoring of EDPS errors, which everyone should be doing with the new 50/50 blend.

In the future, Episource will be presenting on specific ways the new models could drive reductions in RAF and proactive methods to maintain accurate scores. There are many options for your organization to utilize, and we look forward to expanding on these topics moving forward.


We simplify the management of member programs by making it radically more efficient, and increasing value to healthcare organizations and their members

Presentation @ RISE VBC: Effective Collaboration Strategies Between Payer & Provider