• Some questions and answers on quality bonus cost projections

    I’ve been getting some questions today about why my cost projection for the new Medicare Advantage quality bonus program differs so much from CMS’ figures.

    Although I don’t have access to CMS’ methods, I think the biggest single reason is because CMS is reducing projected spending in the first three years due to the phase-in of lower payment rates under ACA.  ACA gradually reduces Medicare Advantage payment rates from an average of about 13% over traditional Medicare to an average of about 1% over traditional Medicare.  This almost achieves the “level playing field” advocated by MedPAC.  The quality bonuses will not be paid in full to plans that are still in the phase-in period because they are still receiving high payment rates.  After a few years, most plans have phased to the new payment rates, but the CMS cost estimate stops after 3 years, so they never calculate the full impact of the bonuses.  I calculate the impact over 10 years, assuming the bonuses will stay in place, and I don’t adjust downward for the phase-in period because I think that adjustment doesn’t give a true picture of the ongoing size of the bonus program.

    I’ve also been asked why my projection changed so much when I received details of the proposed demonstration.

    There are three reasons why the calculations changed. First, I had assumed that the new program would have the same features as the old one.  I found out when I got documentation from CMS that middle-quality plans won’t get the entire 5% bonus, although the weighted average bonus will still be 3.9%– so pretty close.  Second, as Biles and Arnold explain, the original bonus program gives double bonuses to certain plans, accounting for about half the beneficiaries in bonus-qualifying plans.  The new program doesn’t extend double bonuses at all, so that makes a huge difference.  Finally, the original program pays bonus-qualifying plans a larger share of the difference between their bids and the local payment rates (known as benchmarks).  This is another feature that was not extended under the proposed demonstration and that reduces the cost estimate as well.

    Comments closed