Two measures for evaluating hospital performance – 30-day mortality and 30-day readmission rate – were found to be fully independent of each other in the first comprehensive examination of the relationship between the two, according to a report in the Feb. 13 issue of JAMA.
Physicians, researchers, and policy analysts have raised concerns that hospital mortality rates and readmission rates could interact with each other in such a way that would impair their accuracy as measures of a hospital’s performance.
For example, hospitals with lower mortality might be more likely to have higher readmission rates because the very interventions that improve mortality might also create a higher-risk group of discharged patients. Conversely, hospitals with higher mortality might have had patients die before they could be readmitted, so their readmission rates would be artificially lower.
If such interactions occurred, hospitals might be forced to choose one performance measure over the other, said Dr. Harlan M. Krumholz of the section of cardiovascular medicine at Yale University, New Haven, Conn. and his associates.
Alternatively, if 30-day mortality and 30-day readmission rates were found to have a positive correlation, it could be inferred that they reflect similar processes and are redundant.
Dr. Krumholz and his colleagues assessed both measures in Medicare fee-for-service beneficiaries across the United States who were hospitalized for acute MI (590,809 deaths and 586,027 readmissions), heart failure (1,161,179 deaths and 1,430,030 readmissions), or pneumonia (1,225,366 deaths and 1,297,031 readmissions) over a 3-year period.
"We failed to find evidence that a hospital’s performance on the measure for 30-day risk-standardized mortality rate is strongly associated with performance on 30-day risk-standardized readmission rate. These findings should allay concerns that institutions with good performance on [mortality measures] will necessarily be identified as poor performers on their [readmission measures]," the investigators said.
At all levels of performance on the mortality measure, "we found both high and low performers on the readmission measures," they noted (JAMA 2013;309:587-93).
Their findings were consistent across all types of hospitals, regardless of size (bed capacity), teaching status, location (urban or rural), safety-net status, or type of funding (public or private).
The study results indicate that mortality and readmission measures reflect mutually distinct processes and are therefore not redundant, Dr. Krumholz said.
The study was supported by several agencies within the U.S. Department of Health and Human Services.