CHICAGO (Feb. 11, 2015)—A new editorial published in JAMA by Northwestern Medicine experts strongly questions the use of several hospital quality measures by the Centers for Medicare & Medicaid Services (CMS) in its pay-for-performance programs. These eight measures, collectively known as the Patient Safety Indicator 90 (PSI-90) composite measure, account for about one-third of CMS's Hospital-Acquired Condition (HAC) Reduction Program and Hospital Value-Based Purchasing (VBP) Program scores, which financially penalize hospitals with the poorest scores. The editorial argues that flaws in the PSI-90 measure and its use in CMS's programs inaccurately identify problems in hospital care and cause hospitals to be unfairly penalized.

"The PSI-90 measures were originally created to help hospitals measure adverse events and address their quality improvement efforts," said lead author Karl Y. Bilimoria, MD, MS, vice-chair for quality for Northwestern Medicine and director of the Surgical Outcomes and Quality Improvement Center at Northwestern University Feinberg School of Medicine. "These measures were not designed to be representative of a hospital's overall quality, let alone used by an organization such as CMS as a determining factor in how much they reimburse or penalize hospitals for the quality of care they provide."

The PSI-90 component measures track the rate of eight different adverse events including blood clots, pressure ulcers, hip fracture and infections. First used by CMS in its pay-for-performance programs in 2014, these eight PSI-90 measures now account for 35 percent of CMS's overall HAC Reduction Program score and 30 percent of their Hospital VBP program.

One of the main problems identified by the editorial, "Concerns About Using the Patient Safety Indicator-90 Composite in Pay-for-Performance Programs," is that PSI-90 contains flawed measures due to surveillance bias. A recent JAMA study looked at PSI-12, which tracks the rate of patients diagnosed with blood clots following surgery, and found that hospitals with higher rates of postoperative bloods clots were often the hospitals that were most vigilant in screening patients for them. Thus, high-quality hospitals often paradoxically appeared to be providing poor care. The PSI-12 measure and other PSI-90 components, such as PSI-03 which monitors rates of pressure ulcers, were found to potentially penalize hospitals that are proactively screening patients for surgical complications.

"Ensuring patients following surgery are appropriately screened for blood clots and pressure ulcers plays an important role in providing safe, high-quality care for patients," said Gary A. Noskin, MD, senior vice president and chief medical officer at Northwestern Memorial and professor of medicine at the Feinberg School. "Pay-for-performance policies should be rewarding hospitals that do an excellent job of identifying these conditions before they cause harm to patients, but currently these measures do the opposite."

The authors also argue that CMS' use of PSI-90 measures in pay-for-performance programs causes other problems including:

  • Individual complications being redundantly counted by both the Hospital VBP program and HAC Reduction Program
  • Failure to accurately distinguish harmful complications from those that resulted in little to no harm
  • Inadequate risk adjustment for hospital-to-hospital differences in patient populations.

"Beyond the flaws that many of the measures have independently, one the major concerns we raise is that CMS' pay-for-performance programs weigh the importance of each PSI-90 measure based how often their complications occur, not on their clinical implications," said Bilimoria. "This leaves out consideration of differences in a complication's increased length of stay, cost, morbidity and risk of mortality, all of which vary widely between the PSI-90 composite measure's components."

Even with the potential flaws discussed by the authors, they conclude that the measures can still be effective in pay-for-performance programs if they are properly constructed.

"The complications that these measures monitor are important and should be areas of focus for quality improvement in hospitals, but issues like surveillance bias, redundancy and lack clinical relevancy need to be corrected to encourage real healthcare improvement," added Bilimoria.

The JAMA editorial was published online February 5, 2015, and is available on the JAMA website.