Readmissions to hospital are important outcomes since they occur frequently, haw substantial implications for patient morbidity, account for a large portion of hospital costs, and may be a marker for poor quality care. There is substantial interest in using readmissions to gauge the effectiveness of medical interventions. Yet to date there has been little work to identify the sorts of readmissions that may specifically indicate problems with the quality or effectiveness of care or to validate this outcome. This investigation is designed to determine whether readmission outcomes, properly defined, can be used to assess the quality of hospital care received by patients. Because not all readmissions may indicate poor quality care, we will solicit expert medical opinion to correctly specify Related Adverse Readmissions (RARs), defined as readmissions potentially related to less than optimal care during the initial hospitalization. This will be done for 2 prevalent conditions, congestive heart failure and pneumonia. We will then identify 1800 patients in four northeastern states who were either readmitted with an RAR, readmitted within 31 days for other reasons, or not readmitted. will be abstracted by nurses to assess severity on admission and explicit measures of quality. Implicit chart reviews by physicians to implicitly assess quality will also be performed. Our analytic goal is to test whether RARs can be used as a valid and reliable method for identifying patients who may have received lower quality care. In addition we will whether readmission to a different hospital is associated with lower quality care and whether subsets of RARs are associated with particularly poor quality of care. If this study finds no association between RARs and other measures of quality of care it will suggest that readmission outcomes for medical conditions cannot be used to indicate problems with the quality of hospital care during the index admission. On the other hand, if the findings point to an important quality-readmission relationship, we will have demonstrated an important prototype methodology that can be employed nationally with little modification. Potentially our methods can be extended to other conditions in a system of quality assurance in which examination of the care received by patients who have RARs leads to identification of prevalent quality problems and steps to their resolution.
