Readmissions—the Same Old Thing?

1University of Michigan Medical School, Ann Arbor, MI

Meeting: Hospital Medicine 2011, May 10-13, Dallas, Texas.

Abstract number: 159


Hospitalized patients are frequently readmitted within 30 days of their discharge. It is believed that many patients are readmitted because of a recurrence or worsening of the same condition for which they were originally hospitalized. Such readmissions might be preventable if related to inadequate treatment or poor care transitions. However, the degree to which patient readmissions occur for the same reason remains uncertain, as most studies thus far have relied on administrative data.


To determine (1) the accuracy of billing data compared with clinical chart review in determining the reason for readmissions and (2) whether patients were readmitted for the same reason as the original admission based on clinical assessment.


All patients who were readmitted in November 2009 within 30 days of their discharge from the inpatient general medicine service were included. Sixty‐seven unique patients were identified. Administrative data regarding patient identification, reason for initial and subsequent admissions, demographic data, and length of stay were obtained. Only the first readmission was evaluated if more than 1 readmission occurred. Data elements in the chart review of the electronic medical record included discharge summaries, laboratory data, radiographic imaging, and inpatient progress notes from the primary and consulting services. Two reviewers evaluated the data independently looking at (1) the primary reasons (up to 4) for initial admission, (2) primary reasons for readmission, (3) whether these reasons matched the primary billing diagnosis‐related group (DRG) code listed for admission and readmis‐sion, and (4) categorizing the readmission into 1 of 4 nominal categories: same–planned, same–unplanned, different–planned, and different–unplanned. Cohen's kappa coefficient was used to calculate the degree of interrater agreement. The reasons for both admission and readmission in the administrative billing record matched the findings of the clinical reviewers in only 34 cases (51%). Only 8 readmissions (12%) were planned. Of the unplanned readmissions, 31 patients (46%) were readmitted for the same reasons and 28 (42%) for different reasons from their index hospitalization. The K coefficient was 0.63 between observers.


The majority of patients in this sample were readmitted in an unplanned fashion. Of these, only half were for reasons similar to their index hospitalization. Moreover, discrepancies in the reason for admission between billing DRG and clinical chart review were found in a substantial number of patients, highlighting the limited accuracy of using administrative databases for this purpose. These results are limited by sample size, single‐center experience, and the retrospective nature of the study.


R. Chang ‐ none; S. Nichani ‐ none; C. Kim ‐ none

To cite this abstract:

Chang R, Nichani S, Kim C. Readmissions—the Same Old Thing?. Abstract published at Hospital Medicine 2011, May 10-13, Dallas, Texas. Abstract 159. Journal of Hospital Medicine. 2011; 6 (suppl 2). Accessed March 30, 2020.

« Back to Hospital Medicine 2011, May 10-13, Dallas, Texas.