Effect of ‘regression to the mean’ on excess readmissions

Editor's Note

Strong evidence suggests that most declines in excess readmissions--after implementation of Medicare’s Hospital Readmission Reduction Program (HRRP)--at poorly performing hospitals can be explained by a statistical phenomenon called “regression to the mean [RTM],” this study finds.

In RTM, entities farther away from the mean in one period are likely to be recorded closer to the mean in subsequent periods, simply by chance.

Of 3,258 hospitals analyzed, those with an excess readmission ratio (ERR) greater than 1.0 during the FY 2013 measurement window had decreases in ERRs in the subsequent 3 years, and hospitals with ERRs no greater than 1.0 had increases.

The same results were found when the analysis was done using an alternate measurement window that predated HRRP.

Comparing actual changes in ERRs with expected changes because of RTM, 74.3% to 86.5% of the improvement in ERRs for penalized hospitals was explained by RTM.

These findings call into question the notion that HRRP had an effect on readmissions, the researchers say.