Medicare Advantage Payments

CMS itself has acknowledged there may be issues with whether it is adequately adjusting payments to Medicare Advantage plans to account for the costs of seriously ill patients.  An Avalere analysis indicates the significance of the problem.   (Avalere Report)   Medicare does its risk adjustment primarily through Hierarchical Condition Categories modeling.  The model has been revised, most recently in 2014, to add, delete or modify conditions.  Avalere compared actual fee-for-service expenditures for beneficiaries with various chronic conditions to the costs projected by the CMS model.  The researchers concluded that the model under-predicted costs for these beneficiaries by a total of 3.3% or $2.6 billion annually.  On the other hand, the model appears to over-predict spending for people with no chronic conditions.  In particular, the model forecast spending that was too low for rheumatoid arthritis by 15.3%, osteoarthritis by 12%, chronic kidney disease by 4% and dementia, which is not even included in the model, although it is one of the highest cost and fastest growing diseases.  The model was even lower than actual spending for very high utilizers within each chronic condition.  Some of the conditions with over-prediction of spending included heart attack by 11% and transplants by 15%.  Some of the mismatch on under-prediction would seem to me to be driven by higher specialty drug spending for arthritis, spending that often is picked up in Part B.  An interesting side fact of the analysis is that 57% of the Medicare population is said to have three or more chronic conditions.  CMS has recently indicated that it will change the model and pay more with people who are dually-eligible for Medicare and Medicaid.  This subset of patients tends to have more chronic conditions and higher costs.  Whether these changes fully address the model inaccuracies is uncertain, but they likely erase at least some of the dissonance between model and reality.  CMS has a difficult task with its Medicare Advantage rate-setting.  It wants to not pay plans more than it would be paying for the same beneficiary in the fee-for-service arm of the program, but it also wants to pay plans enough to entice them to enter or stay in the program.  The risk-scoring model unfortunately is the main mechanism for that balancing and appears to have some weaknesses.