Adjusting for Readmission Penalty Status

The CMS hospital readmission penalty program is sometimes judged to be a success because it seems to have lowered readmission rates.  But concerns have been raised because it appears to disproportionately affect hospitals serving poor and/or minority populations, even if they aren’t delivering worse care.  Congress got involved (always a positive) and mandated that CMS make some adjustment for socioeconomic status.  A new study in the Joint Commission Journal on Quality and Patient Safety suggests that adjusting for that factor may not solve the problem.   (JCJQPS Article)   What CMS is proposing to do is assign hospitals to one of five peer groups, within which they would be compared; so the researchers examined whether this would fix the perceived problem without eliminating the  effectiveness of the penalties in reducing inappropriate readmissions.  The authors hypothesize that it may be volume that drives quality, so maybe it is being a smaller hospital that is more a factor.  Having poorer quality because you are small is not a valid excuse in their judgment.  They used data from 2017 to conduct their analyses of what the penalties would look like if the new adjustment is implemented.  They divided hospitals into deciles based on socioeconomic status of patients and reviewed readmission rates by each of the six conditions used for penalty measurement.

They found some significant variation across conditions even in what hospitals treat more low socioeconomic status patients with that condition.  But in general, hospitals with more low socioeconomic status patients had higher readmission rates.  Somewhat confoundingly, for some conditions, joint replacement and coronary artery bypass surgery, high volume appears clearly linked to lower readmission rates, but for others, pneumonia, heart failure, heart attack and chronic obstructive pulmonary disease, low volume is associated with lower readmission rates.  The interactions of socioeconomic status and volume may inappropriately affected expected readmission rates, according to the authors, thereby affecting penalty calculations.  What is most clear is that there is something wrong with the design of the hospital readmission penalty program.  And it tracks back to trying to use overall statistics on patient treatment instead of doing the obvious thing, the thing that would have little possibility of error.  Examine each readmission to determine if it could have been avoided by actions of the hospital.  If so, attribute it to the hospital.  If not, don’t attribute it.  Pretty straightforward and much fairer.