Coding for Hospital-acquired Infections

One well-meaning attempt to improve quality in health care is reducing the incidence of health care-acquired conditions, which occur when the delivery of health care in a substandard manner results in a new medical problem for the patient.  One subset is hospital-acquired infections and CMS has had a policy for several years of not paying for care resulting from such infections and of reporting hospitals’ infection rates publicly.  More recently, an additional financial penalty has been added to reduce all Medicare reimbursements to hospitals with big HAI rates.  A paper from Stanford University finds that hospitals may be using coding tactics to avoid the penalties.  (Stanford Paper)   Hospitals basically self-report whether an infection was present on admission, and therefore not penalized, or occurred during the hospitalization.  So far, it appears that there has been little decrease in the rate of HAIs, and it may be because hospitals are using coding to avoid penalties, according to the researchers.  There are two ways a hospital could do this.  One is to code an infection as present-on-admission when it wasn’t and the second is to leave hospital-acquired infections off the diagnosis code list.

The authors used the fact that some states also have programs requiring reporting of hospital-acquired infections and that there is variance in the strength of these programs to detect improper coding to estimate miscoding for Medicare.  In states with “strong” regulations, hospitals are required to provide details on the events and causes.  The researchers found that in these states, hospitals generally provided higher quality according to a variety of measures.  Notwithstanding this, these hospitals in the strongly regulated states appeared to have lower present on admission and higher HAI rates, which is illogical, and strongly supports the idea that in less regulated states providers feel freer to miscode.  (It also suggests this miscoding is intentional.)  Using a large set of Medicare hospital claims and comparing HAI reporting across state regulation types, the authors conclude that for just two of the common HAIs, central line-associated blood stream infections and catheter-associated urinary tract infections, the researchers find that there are over 10,000 infections improperly coded as present on admission; infections which cost CMS $200 million a year.

It is estimated that on any day about 4% of hospitalized patients have a hospital-acquired infection and 75,000 patients a year die from one.  So this is a serious quality issue.  But as usual, CMS’ good intentions don’t turn into a well-designed, effective program to address the problem.  Shame on hospitals for spending effort trying to avoid the penalties by coding instead of stopping infections, but they are responding to the financial incentive they were given.  Hospitals that are truthful in their coding are being unfairly treated in this process as it looks like they have worse quality than their cheating peers and they get penalized financially while those cheating peers avoid penalties.  Manual reviews to compare medical records and coding on claims is expensive, but maybe CMS needs to do more of this on a random, surprise basis, especially in weakly-regulating states, and when it finds questionable practices, not just hospitals, but their managers, especially CEOs and heads of coding groups, should be penalized.  Penalizing individuals is important because otherwise it is easy for responsibility to be diffused across the “organization”, which is a legal fiction.