Researchers looking at hospital data for patients admitted with acute MI, heart failure and pneumonia have found little evidence of an association between risk-adjusted rates of mortality and readmission. The findings were published online Feb. 13 in the Journal of the American Medical Association.
In 2007, the Centers for Medicare & Medicaid Services (CMS) began publicly reporting all-cause risk-standardized hospital mortality rates (RSMR) for acute MI and heart failure (HF). It started reporting RSMR for pneumonia in 2008, and in 2009 began reporting 30-day, all-cause risk-standardized readmission rates (RSRR) for the three conditions.
Harlan M. Krumholz, MD, of Yale University School of Medicine in New Haven, Conn., and colleagues attempted to quantify an overall correlation between RSMR and RSRR for MI, HF and pneumonia, and within subgroups based on hospital characteristics.
The study authors suggested that there is a perception that decreased mortality may lead to increased readmission rates, as sicker patients are discharged. They also noted the possibility that strategies that lead to decreased mortality also suppress readmission rates, reflecting high standards of care, thus rendering the reporting of both measures redundant.
The researchers identified Medicare patients 65 years old and older who were hospitalized with a diagnosis of acute MI, HF or pneumonia between July 1, 2005, and June 30, 2008, and used the Medicare Enrollment Database to establish mortality.
Using hierarchical logistic regression models, the researchers estimated 30-day all-cause RSMRs and RSRRs at all nonfederal acute care hospitals. They stratified the samples according to region, safety-net status, rural or urban. For each disease condition they ranked the hospitals’ RSMRs and RSRRs into quartiles, with the lowest quartiles encompassing the hospitals with the lowest rates of mortality and readmission.
The median RSMR for acute MI was 16.57 percent, for HF 11.06 percent and for pneumonia 11.46 percent. The median RSRR for acute MI was 19.87 percent, for HF 24.42 percent and for pneumonia 18.09 percent. The researchers found no meaningful correlation between RSRR and RMRR except a weak association (-0.17) for HF, which they explained was most prominent among the hospitals with an RSMR for HF of less than 11 percent. Subgroup analysis did not reveal any differences based on hospital characteristics.
The researchers noted that many hospitals were placed in the quartiles with the lowest rates for both RMSR and RSRR: 8.5 percent for acute MI, 5.4 percent for HF and 6.4 percent for pneumonia. Similarly, many hospitals were placed in the quartiles with the highest rates for both measures: 6.7 percent for acute MI, 5.3 percent for HF and 7.2 percent for pneumonia.
These results “should allay concerns that institutions with good performance on RSMRs will necessarily be identified as poor performers on their RSRRs,” Krumholz et al wrote. “For AMI and HF, there was no discernible relationship, and for HF, the relationship was only modest and not throughout the entire range of performance.”
The researchers noted that they assessed patterns and that at the individual hospital level there may be a correlation between RSMRs and RSRRs, and that their use of hierarchical modeling may obscure a relationship because low-volume hospitals were included; they cited these as limitations of their study. They also pointed out that their study did not assess the value of RSMR and RSRR as measures of quality.