Accounting for Observation Stays Shrunk Hospital Readmission Reductions

The decrease in readmission rates associated with the Hospital Readmissions Reduction Program (HRRP) was smaller after accounting for observation stays, indicating that the value-based program may be underperforming, according to a JAMA Network Open study.

The HRRP aims to decrease hospital readmissions by reducing Medicare payments by up to three percent for hospitals with higher-than-expected 30-day readmission rates for specific conditions and procedures.

Since the program’s implementation in 2010, studies have suggested that it has been achieving its goal of reducing readmission rates. However, the prevalence of observation stays has increased since the HRRP began, which are not counted as index hospitalizations or readmissions when calculating readmission rates.

Researchers used Medicare Part A and Part B claims to determine if the HRRP was associated with readmission reductions after accounting for observation stays as index hospital discharges and readmissions.

They compared readmission changes across three periods; the baseline period was from January 1, 2009, to March 31, 2010, the intervening period stretched from April 1, 2010, to September 30, 2012, and the post-penalty period was from October 1, 2012, to December 31, 2015.

The study looked at hospitalizations for three conditions targeted by the HRRP—acute myocardial infarction, heart failure, and pneumonia—and a comparison group of nontargeted conditions.

The final sample included almost 9 million index hospitalizations, 1.4 million of which were for one of the three targeted conditions. Only 3.3 percent of hospital discharges among the target conditions were observation stays, while 17.9 percent were for nontarget conditions.

The proportion of total hospitalizations that were observation stays increased throughout the study period for both targeted and nontargeted conditions.

Researchers found that readmissions for target conditions decreased faster after the HRRP’s announcement but returned to baseline in the post-penalty period.

Before accounting for observation stays, the readmission rate for target conditions fell from 22.14 percent in the baseline period to 20.65 percent in the post-penalty period, for a 1.48 percentage point reduction. For nontarget conditions, the readmission rate declined from 18.24 percent to 17.11 percent (-1.13 percentage points).

When researchers accounted for observation stays, the reduction in readmissions rates was smaller for target and nontarget conditions.

For example, the readmission rate for target conditions decreased from 23.32 percent in the baseline period to 22.66 percent in the post-penalty period, indicating a 0.66 percentage point reduction. This finding indicates that more than half of the decreases in readmission rates for target conditions was attributable to observation stays.

Similarly, the readmission rate for nontarget conditions fell from 18.58 percent to 17.82 percent, for a 0.76 percentage point decrease.

The study results suggest that the HRRP was less effective in reducing hospital readmissions when it accounted for observation stays as hospitalizations.

There have been concerns that hospitals may attempt to avoid readmission penalties by placing patients in observation when they are readmitted. Ignoring the growth of observation stays could make it more challenging to estimate the potential outcomes associated with the HRRP since they are not counted in the program, the study pointed out.

“Our results suggest that an increasingly larger share of hospital care will be invisible to quality metrics if shifts in observation stay practices are not accounted for in readmissions algorithms,” researchers wrote.

“The resulting risks of incorrect assumptions and program ineffectiveness extend beyond the HRRP to other quality programs, particularly given broader trends to both measure readmissions under value-based payment models and shift more conditions and procedures to outpatient management.”

The study also noted that readmission reductions following the implementation of the HRRP might be attributed to other factors, such as advances in clinical care delivery, greater use of home healthcare, and better diagnostic tests.

Past research has found that the HRRP could be improved, particularly for safety-net hospitals, by leveraging social risk adjustment and stratifying facilities based on patient population.