Page 67

67

information obtained from the HOS or HOS-M in one year to calculate frailty scores for the following year by applying the frailty factors that correspond to the ADL information gathered from the HOS or HOS-M data. Comment: Commenters expressed concerns regarding a number of aspects of frailty adjustment they believe result in underpayment for beneficiaries with the highest need including general concerns about the decline in frailty scores, the disproportionate impacts of lower frailty scores on vulnerable and high-need populations (e.g., dually eligible enrollees) due to payment decreases, concerns about low response rates and HOS/HOS-M survey administration. These commenters made a variety of recommendations, including: • Survey protocol modifications. One commenter recommended modifications to the current survey protocol such as CMS allowing FIDE SNPs to survey only enrollees who are at a nursing home level of care so that the level of frailty of enrollees being surveyed are on par with PACE participants. • Application to expanded beneficiaries. One commenter requested that CMS apply the frailty adjustment to D-SNP enrollees in Puerto Rico. • Monitoring and evaluation. Another commenter requested CMS continue to monitor payment to FIDE SNPs to assess the impact that reduced frailty factors will have on dually eligible enrollees and any impact this may have on risk adjusted payments and access to benefits. • Collaboration with stakeholders. A commenter requested that CMS collaborate with stakeholders on additional approaches to better measure frailty as well as how the HOS and HOS-M surveys are applied. Response: CMS acknowledges the concerns regarding several aspects of the frailty payment process. The HOS and HOS-M surveys are sufficient for a frailty adjustment to payment at the plan level because ADL data are collected to calculate frailty scores in the same manner that data are collected and used to calculate frailty factors for model calibration (i.e., limitations in activities of daily living collected from self-reported surveys). In addition, data are collected consistently across respondents, such that frailty scores are calculated using data collected in the same manner across plans, thereby allowing survey results to be compared across plans and relative to PACE (a requirement for determining whether FIDE SNPs receive a frailty adjustment in payment) and thus resulting in frailty payments that are comparable. Regarding the requests to apply the frailty adjustment to certain populations, CMS is only authorized by statute to apply frailty adjustment to payments to PACE organizations and is permitted to apply frailty to the payments to certain FIDE SNPs. CMS has explored ways of incorporating frailty into the risk adjustment model in order to account for frailty when making risk adjusted payments to all plans and found challenges with a number of approaches. We also note that, because the frailty factors are calculated using the residual of the CMS-HCC risk adjustment model (the difference between the predicted expenditure amounts and the actual