Hospitals in Medicare ACOs reduced readmissions faster
PROVIDENCE, R.I. [Brown University] — Hospitals in Medicare Accountable Care Organizations outpaced non-ACO hospitals in reducing the rate at which patients discharged to skilled nursing facilities (SNFs) needed to return to the hospital, according to a new study in the January issue of Health Affairs.
When a hospital sends a patient to a nursing home, everyone's hope is that the patient won't have to resume acute care back at the hospital, though until recently about 1 in 4 did. To drive the rate down, Medicare has implemented measures including penalties for excess readmissions and encouraging hospitals to join ACOs in which they share an overall "bundled" payment with SNFs. This model gives providers the incentive that they can keep more profit if they reduce expensive extra hospital care.
The new study suggests that while all hospitals reduced their readmission rates following the adoption of readmission penalties, ACO hospitals reduced them at a faster rate.
"This is about understanding whether ACOs work or not," said Momotazur Rahman, a study co-author and assistant professor at the Brown University School of Public Health. "It seems like they are doing slightly but significantly better than non-ACO hospitals."
Faster dropping rates
To perform the study, the researchers gathered information on discharges and readmissions between 2007 and 2013 from 220 ACO-affiliated hospitals and more than 1,840 non-ACO hospitals in metropolitan areas around the country. They also had anonymized medical and demographic data on all the patients involved so they could statistically control for such factors in comparing ACO vs. non-ACO readmission rates.
They chose the study timeframe to straddle passage of the Affordable Care Act in 2010 and implementation of two variants of ACOs in 2012, "Pioneer" and "Shared Savings." In this way they could not only compare the readmission rates of different hospital types to each other, but also how each hospital type's readmission rate changed before and after Medicare's measures took effect.
They found that both ACO and non-ACO hospitals have succeeded on average in substantially reducing readmission rates, but that ACO hospitals have been measurably better. Over the entire period, relative to their readmission rate in 2007, non-ACO hospitals reduced readmissions by 13.1 percent, while hospitals in Pioneer ACOs reduced them by 14.9 percent and Shared Savings ACOs reduced them by 17.7 percent.
To gain further insights, the team analyzed the data in a few other ways. For example, they separated out quick readmissions (within the first three days after discharge) and longer-term ones (four to 30 days after). The reason, Rahman said, was that quick readmissions might reflect lapses in hospital care while longer-term ones might reflect lapses in SNF care. The analysis showed that readmissions declined significantly in both periods (not only overall but especially in ACOs vs. non-ACOs comparisons), suggesting that both hospitals and SNFs are improving their readmission prevention and that ACO-affiliated institutions are doing so more.
Another analysis showed that the pace of readmission reductions quickened further after the implementation of ACOs in 2012, suggesting that the policy really did make a difference.
Exactly how that difference occurred is not clear. The team did not see evidence in the data that hospitals became more selective about discharging patients to more capable SNFs, Rahman said. The data can't say how hospitals and SNFs improved their care quality, or whether the two types of institutions improved communication about patients.
But the prospect of not only avoiding reimbursement penalties but also keeping more of a bundled payment appears to have a real effect on reducing readmissions, the authors wrote.
"There is an additional incentive for ACO hospitals to coordinate the care of patients more efficiently," Rahman said.
The study's lead author is Ulrika Winblad, an associate professor at Uppsala University in Sweden who was a Harkness Fellow at Brown during the 2014-15 academic year. In addition to Winblad and Rahman, the paper's other authors are Brown Professor Vince Mor and Columbia University Assistant Professor John McHugh, a Brown University graduate alumnus.
The National Institute of Aging (grant: P01AG027296) and the Commonwealth Fund (20150004 funded the study).