NEJM: Study on CHF readmission rates may inform policy
“Unplanned readmissions after hospitalization are costly and reflect suboptimal patient outcomes,” Arnold M. Epstein, MD, from the Harvard School of Public Health and Brigham and Women’s Hospital, both in Boston, and colleagues wrote. “Evidence of suboptimal care at hospital discharge and shortly thereafter has prompted clinical interventions to improve discharge planning, ensure timely follow-up and improve transitional care.”
While the focus has been to help thwart readmission with better disease management programs and better transitions of care, these types of clinical interventions may not be good enough, especially if they are not widespread.
Because there is a weak relationship between publicly reported measures of discharge planning and readmission rates, Epstein et al set out to assess the regional variation in readmission rates among patients hospitalized for congestive heart failure (CHF) or pneumonia to determine whether regional variations in readmission rates can be explained by variations in population-based rates of admission.
To do so, the researchers used 2008 Medicare data to calculate 30-day, 60-day and 90-day readmission rates for patients discharged with CHF or pneumonia for each referral region. Variations in hospital referral region rates were examined.
There were 306 hospital referral regions in the study. These referral regions included 4,432 hospitals and 234,477 hospital discharges between Jan. 1, 2008, and July 30, 2008, with a diagnosis of CHF and 4,497 hospitals and 237,025 discharges with a diagnosis of pneumonia.
The authors reported that readmission rates after an index hospitalization for CHF varied from 11 to 32 percent across referral regions. The rates for pneumonia varied from 8 to 27 percent across the referral regions. A hierarchical model showed that readmission rates ranged from 19 to 31 percent for CHF and from 13 to 24 percent for pneumonia.
The hospital referral regions in the highest quartile of readmission for CHF had an average readmission rate of 28 percent; those in the lowest quartile had an average readmission rate of 20 percent. Hospitals with higher readmission rates were more likely to be medium- or large-size hospitals, be located in the Northeast and be for-profit, private nonprofit, urban or academic hospitals. These hospitals also had higher numbers of female, black and Hispanic patients and lower rates of mortality.
Those with the highest rates of CHF readmission had low performance on the HCAHPS (Hospital Care Quality Information from the Consumers Perspective) discharge planning metrics.
“Case mix, the number of cardiologists per 100,000 population, and discharge planning also explained a notable proportion of the variation in univariate analyses, although their explanatory potential diminished substantially in multivariable models,” the authors wrote.
The researchers estimated that each year there were nearly 115,568 readmissions within 30 days of patients being discharged from a CHF hospitalization and 84,854 readmissions among those who first presented at the hospital with pneumonia.
If hospitals in the highest quartile for readmissions reduced utilization rates to the rates in hospitals in the lowest quintile, CHF readmission rates could drop from 24.6 to 21.2 percent, eliminating 14 percent of these readmissions.
“The quality of transitional care is in need of substantial improvement: many elderly patients fail to receive any ambulatory care before readmission, and only a minority of primary care physicians report receiving key discharge information about their recently hospitalized patients,” the authors noted.
Due to this fact, policy makers have focused on transitions of care between the hospital and ambulatory care settings. While these efforts can reduce some admissions, it will become more important to focus on general use of hospital care, the authors claimed.
“In summary, we found a substantial association between regional rates of rehospitalization and overall admission rates,” the authors concluded.