Our What's New in the Patient Safety World columns for February 2015 “17% Fewer HAC’s: Progress or Propaganda?” and January 2016 “” discussed interim data sets from AHRQ which showed that there was a 17% reduction in hospital-acquired conditions (HAC’s) between 2010 and 2014. Over the 4-year period the biggest reductions in HAC’s percentage-wise were seen for CLABSI’s (-72%), CAUTI’s (-38%), and post-op venous thromboembolism (-43%).
AHRQ’s Chartbook on Patient Safety summarizes patient safety measures, including overall hospital-acquired conditions (HAC’s) and hospital-associated infections (HAI’s), highlighting the trends between 2010 and 2014 (AHRQ 2016). Similarly, CDC has reported substantial reductions in CLABSI’s, CAUTI’s, surgical site infections (SSI’s), hospital-onset C. difficile infections, and hospital-onset MRSA bacteremias over a roughly similar time frame (CDC 2016).
While we’ve had some degree of skepticism in interpretation of the data, overall we’ve felt comfortable that true progress is being made.
One of the many interventions cited as contributing to the apparent improvement in HAI and HAC rates is the financial penalty hospitals pay for poor performance in these rates (primarily for CMS/Medicare patients but also for some other insurers). We’ve always been concerned about how coding changes have obfuscated some quality parameters. For example, we’ve always been concerned about changes in sepsis coding may have artificially lowered mortality rates for both sepsis and pneumonia (see our
March 2016 What's New in the Patient Safety World column “”). Now a study from the Stanford Graduate School of Business questions whether coding practices have similarly impacted quality reporting for the HAI’s reported to CMS as well (Bastani 2015).
Bastani and colleagues note that CMS does not directly monitor the occurrence of the various HAI’s. Rather it collects administrative (billing) data from hospitals and does chart reviews of a small sampling to assess validity. The Stanford researchers used more sophisticated techniques to assess how rampant “upcoding” might be. In particular, one form of upcoding would be assigning a designation present-on-admission (POA) to an infection when it was, in fact, a hospital-acquired infection. Upcoding would be financially beneficial to hospitals either by increasing reimbursement or avoiding penalties.
They compared rates of HAI’s in states that require strict reporting of HAI’s to those in states that have weaker reporting requirements. Overall, they found hospitals in the more weakly regulated states reported lower rates of HAIs and higher rates of POA infections. They estimate there are more than 10,000 upcoded infections annually, resulting in an added costs of $200 million to CMS.
Bastani and colleagues are careful to not impute a motive to such “upcoding” While such could be intentional in attempt to avoid the CMS penalties, they also note it might reflect lack of clinical knowledge by “coders” or lack of communication between clinicians and coders (talk about being tactful and politically correct!).
They conclude that their findings suggest, contrary to widely-held beliefs, increasing financial penalties alone may not reduce HAI incidence and may even exacerbate the problem. They make several policy recommendations based on their results, including a new measure for targeted HAI auditing and suggestions for effective adverse event reporting systems.
AHRQ (Agency for Healthcare Quality and Research). Chartbook on Patient Safety. March 2016
CDC (Centers for Disease Control and Prevention). National and State Healthcare Associated Infections Progress Report. 2016
Bastani H, Goh J, Bayati M. Evidence of Strategic Behavior in Medicare Claims Reporting. Stanford Graduate School of Business 2015; Working Paper No. 3396; July 13,,2015