Service Use Variation In MA and FFS Medicare

Some of the early work on geographic variation in health spending and utilization was done using Medicare data.  Although the Medicare Advantage program now comprises one-third of Medicare enrollment, research on utilization in that arm has been difficult because the data has been episodically and incompletely collected and made available by CMS.  MA encounter data is now required to be filed by the plans and CMS is making it public.  So we can expect to see more research such as that recently published in Health Affairs, which compared use, and variation in use, of hospitals, home health services and skilled nursing facilities in the Medicare Advantage population.   (HA Article)   You might think that given tighter utilization management controls in the MA plans, there would less variation.  Not so, according to the study, which covered the period 2007 to 2013 and analyzed data on over 50 million beneficiaries.  Using hospital referral regions as the geographic unit, variation was compared only for the year 2013.  A variety of factors were used to adjust the analyses.  Compared to fee-for-service Medicare beneficiaries, MA enrollees are more likely to be female, older, a minority and less likely to be dually-eligible.

On an adjusted basis, there was more use of home health and skilled nursing care in FFS Medicare, but hospital care was roughly comparable between the arms.  In 2013, for example, SNF use was 2327 days per year per 1000 beneficiaries for Medicare Advantage and 2902 days for fee-for-service.  In the same year, home health care was 4712 days in MA and 7257 in FFS.  Hospital days were 1228 in Medicare Advantage and 1305 in FFS.  While there were fewer days in MA, the difference did not reach statistical significance.  In both parts of the Medicare program, geographic variation in utilization was highest for home health services, followed by skilled nursing care and then hospital use.  The Medicare Advantage program actually had greater geographic variation in SNF care and hospital use than did the traditional Medicare program, which had more variation in regard to home health care. The variation was often significant, over a factor of two in some cases.  Utilization was fairly correlated for all three services in both parts of the program in a specific region, suggesting that broader practice patterns in a community are a factor.  While Medicare Advantage generally reduces utilization in some services, it does not appear to do so by application of some national standard of care patterns, since wide geographic variation in utilization remains even on an adjusted basis.  Doctors gonna practice how doctors gonna practice.