Bundled payments is another one of CMS’ initiatives to reign in Medicare spending and other payers are adopting them as well. An evaluation of the CMS bundled payment program is conducted and issued every year and the latest report was released recently by Lewin Associates. (Lewin Report) The structure of the program, conditions covered and participants have changed over time, so the evaluation becomes complex. The three models evaluated are one covering a hospitalization and readmissions in 30 days for the same condition; one covering post-acute care and one covering both. This evaluation included episodes initiated up to September 2015, patient surveys and interviews with providers. The comprehensive model was the most popular and accounted for 85% of all episodes. The post-acute care accounted for most of the rest, as the inpatient-only episodes drew little interest. The comprehensive model also experienced few withdrawals, only 6% of participants, while 13% withdrew from the post-acute care model and over 50% from the acute only model. The most popular episode to participate in was joint replacement, followed by congestive heart failure. The participating providers tended to be larger, urban and not-for-profit. They also had higher average pre-bundled program payments than did non-participants, suggesting they knew there were costs they could reduce.
Only one condition resulted in statistically significant lower spending for Medicare and that was joint replacement. Under the comprehensive model, spending was 4.5% less and under the post-acute care one it was 7.1% lower. In both models this appears to be due to less use of post-acute care facilities, with a concomitant increase in home health care. Across most conditions there were few statistically significant changes in quality. For a few there was an increase in ER use and unplanned admissions. Patient satisfaction generally showed slight improvement, and for joint replacement there was a larger gain, while for sepsis, which had longer hospital stays under bundled payments, there was a decline. This is consistent with the notion in both cases that being institutionalized reduces patient satisfaction. For unclear reasons, unplanned readmissions were somewhat higher for a number of episodes. May be the effect of too-quick hospital discharges. There were some indications that providers may have selected bundled episodes involving healthier patients. Not surprisingly, the providers who achieved the greatest savings had the highest initial costs and healthier patients. It is early in the process, so it is understandable that the results are somewhat inconclusive. Providers clearly are figuring out where they may be able to make money and probably avoiding other episodes. The impact on Medicare to date is negligible.