Dartmouth Atlas on Hospital Readmissions

Hospital readmissions certainly create additional health spending and are also promoted as a marker of poor quality.  CMS has an ambitious initiative underway to reduce inappropriate admissions, which it believes may save as much as $17 billion a year.  A new report from the Dartmouth Atlas authors looks at geographic variation and other factors related to readmissions.   (Dartmouth Report) Based on 2009 data from over 300 hospital referral regions and all the leading academic medical centers, the results show that average 30-day readmission rates for a medical admission for Medicare beneficiaries were 16.1%, with a low of 11.5% in Ogden, Utah and a high of 18.9% in Pontiac, Michigan.    For surgical admissions, the average was 12.7%, ranging from 7.5% in Rapid City, South Dakota to 19% in the Bronx.

Looking at six specific diagnoses, the coefficient of variation ranges from .07 for medical admissions to .20 for hip fracture.  Generally, regions had the same rates of readmissions for all six diagnoses.  There were also high rates of variation across the prestigious academic medical centers, and again, that variation existed in all categories of admission reason.  There was little change in readmission rates in the last five years, which the authors take as a sign that either substantial efforts aren’t being made to reduce them or that the efforts aren’t working.  Another likely explanation, given the attention readmissions have received, is that most simply aren’t avoidable.   The report specifically finds that a high percentage of patients don’t see a physician within 14 days of discharge, which it takes as evidence of poor care coordination.   There was tremendous variation in this percentage as well.

While the report contains valuable information, it ignores recent research suggesting that most hospital readmissions may be unavoidable; that early followup after discharge or any followup at all doesn’t seem to make a difference in readmission rates and it does not critique the CMS initiative which appears to be poorly designed if the goal is to truly discourage inappropriate readmissions.  The report also ignores the role of CMS’ hospital reimbursement methods in encouraging early discharges, which may lead to problems.  There are interventions that appear to improve care coordination and the overall longitudinal care of patients and those interventions should be widely adopted.  But programs that penalize hospitals for readmissions without any examination of the clinical circumstances of a particular case are unfair and will inevitably lead to unintended consequences for many patients which may worsen their care, just as other CMS reimbursement-related programs have.