Observation status, ED visits create illusion of fewer readmissions, quality gains

Even as the government claims preventable hospital readmissions have dropped, a dramatic increase in hospitals' use of observation status and emergency care in lieu of readmission suggests many providers are gaming the quality standards, according to a Health Affairs blog post.

Hospitals' use of observation status has been a source of controversy in recent years. Critics say the designation confuses patients; states such as Virginia have passed legislation requiring hospitals to inform patients when they are designated under observation status.

Observation stays jumped 96 percent between 2006 and 2013, accounting for more than half of the decline in Medicare admissions, according to MedPac data. Moreover, about one in 10 hospital stays within 30 days of an earlier discharge are classified as "observation," according to the Centers for Medicare & Medicaid Services (CMS).

Further analysis by Health Affairs shows the rise of observation status accounts for 36 percent of the claimed reductions in readmissions.

The authors--David Himmelstein and Steffie Woolhandler, both physicians and professors of public health at the City University of New York--flat out call it "cheating." 

Increased emergency department (ED) treatment also accounts for some of the readmission reductions, they write. After controlling for the 0.4 percent increase in ED use within a month of discharge, readmissions for patients with urgent medical problems are only down 0.3 percent over the last three years.

CMS, meanwhile, claims a figure more than three times as much.

Upcoding is corrupting data, too: Readmissions have actually increased for heart attack patients after controlling for ED visits and observation status, according to the authors.

These problems, the post states, come down largely to the fact that many Medicare pay-for-performance measures are unproven. It "risks quality failure on a monumental scale [and] pressures hospitals to cheat."

To learn more:
- here's the blog post
- read the MedPac data (.pdf)
- here's the CMS data (.pdf)