Author + information
- Rohan Khera,
- Yongfei Wang,
- Khurram Nasir,
- Zhenqiu Lin and
- Harlan M. Krumholz
We sought to evaluate concerns about US hospitals gaming readmission measures by deferring admissions for AMI and HF during the 30-day period to avoid readmission penalties.
For 2011-2017, we calculated each hospital's daily rates of readmission from post-discharge day 1 to 60 for elderly Medicare beneficiaries hospitalized with AMI/HF. In a quasi-experimental approach, we used non-parametric regression-discontinuity to evaluate whether there were unexpected inflections in readmissions rates around day 30.
At 3279 US hospitals, there were a median 29 (IQR 7, 79) AMI and 78 (IQR 33, 154) HF discharges/year. The median readmission rates in days 1-30 and 31-60 were 17% (IQR 11%, 23%) and 6% (0%, 10%) for MI, and 22% (18%, 26%) and 11% (8%, 14%) for HF. Readmission rates decreased exponentially across the 60 post-discharge days, with consistently lower rates in 2016-17 compared with 2011-12 for days 1-30 and 31-60 (P generalized additive models <.01 [AMI/HF]). There were no discontinuities in readmission for AMI/HF at the 30-day mark, with a >95% power to detect a 0.1% difference (Fig). There were no discontinuities in mortality rates overall, or readmission/mortality rates at hospitals incurring readmission penalties in the years after the penalties (P>0.05).
Readmissions for AMI and HF decreased without inflections around the 30-day mark, arguing against a systematic gaming of readmission measures by pushing readmissions beyond the measure evaluation period.
Lounge & Learn Pavilion
Sunday, March 17, 2019, 9:00 a.m.-9:15 a.m.
Session Title: Young Investigator Awards: Outcomes Research
Abstract Category: Outcomes Research
Presentation Number: 916-08
- 2019 American College of Cardiology Foundation