Coding and Outcomes Measurements

New research published in the Journal of the American Medical Association suggests that coding changes may be responsible for apparent movement in results on some outcome or other quality measurements.   (JAMA Article)   The researchers examined trends in pneumonia morbidity and mortality from 2003 to 2009.  Pneumonia accounts for one million annual hospital admissions and over $10 billion in health spending and has been a focus for various clinical care guideline and pay-for-performance programs.  Over the course of these programs, it appears that the care guidelines were being adhered to more frequently and that mortality was improving.  The authors suspected that coding changes might be responsible for some of the apparent improvement in mortality.  Hospital billing and clinical departments have a fair amount of latitude in assigning disease and diagnosis codes.  In particular, the researchers looked at situations where pneumonia might have been switched to a secondary diagnosis.

Over the 2003 to 2009 study period, the number of hospitalizations with pneumonia as a principal diagnosis declined 27%, but cases where pneumonia was a secondary diagnosis increased rapidly.  Although mortality rates appeared to have decreased for principal diagnosis of pneumonia, on a combined basis it actually increased in absolute terms, but did decline slightly after adjustment for comorbidities (which by the way, might be subject to coding changes over this time period as well).  The authors’ analysis suggests that moving principal diagnoses away from pneumonia accounts for most of the perceived improvement in mortality rates over this time period.  Hospitals in essence changed the number of and which cases ended up being measured for reporting and pay-for-performance programs, without necessarily actually improving care.  Hospitals, aided by a number of vendors, have become very creative in gaming coding systems both for reimbursement and quality improvement purposes.  CMS and other payers will need to be equally creative to sort out what is really happening in terms of meaningful patient outcomes.