A reader e-mailed me about the following in the recent Health Affairs article on health care spending in 2009.
Per enrollee, fee-for-service [Medicare] spending growth accelerated from 4.7 percent in 2008 to 6.9 percent in 2009. […]
Per enrollee, Medicare Advantage spending increased only 2.3 percent in2009—3.3 percentage points slower than the comparable figure for 2008.
Why did FFS Medicare growth accelerate and MA spending growth moderate in 2009? The honest answer is, I don’t know. It’d take a careful study to find out. But I can guess and I can ask a colleague (Steve Pizer, in this case). Here’s what we came up with:
Though enrollment in MA plans continued to grow in 2009, it probably grew at a slower rate than in 2008. In particular, new enrollment in private fee-for-service (PFFS) plans was likely lower in 2009 than 2008. Those are the plans that impose the highest per-beneficiary cost on the program. This enrollment moderation hypothesis is backed up by figures produced by the Kaiser Family Foundation.
Also, growth in payment rates to MA plans may have slowed. If payments did moderate, then plans would not have continued to enhance benefits at the same rate, which would also lead to a moderation in spending increases by the plans. Actually, what matters here is payments relative to costs. If by costs we take FFS costs, then payment relative to cost did not come down in 2009 (see MedPAC). But FFS costs are not MA costs and nobody knows MA costs except the plans themselves.
There may also be a selection effect. As MA enrollment increases, only the sickest individuals remain in FFS, increasing per beneficiary costs for that arm of Medicare.
That’s a sufficient amount of speculation on this bit of trivia. Suffice it to say, I would not believe a word anybody says about it (including me) until a detailed study is done. In particular, do not take the results as definitive evidence that MA all of a sudden ratcheted up utilization management and negotiated better prices with providers. That could be part of it, but one can’t know given the data at hand.