Predictors of 30-day hospital readmission in patients with community-onset pneumonia
Methods: Retrospective evaluation of adults admitted to a 624-bed U.S. academic medical center with an ICD-9 diagnosis code of pneumonia from 1/2010 - 12/2011. All patients had a new or progressive infiltrate on chest imaging and at least two of: hypo/hyperthermia, WBC <4k or >12k, or respiratory symptoms (e.g. cough, shortness of breath). Chi-squared, Student’s t-test, and Mann-Whitney U were used for univariate analysis. Variables with a p<0.1 in univariate analysis were candidates for the stepwise multivariate logistic regression to identify factors independently associated with readmission.
Results: 508 patients included with 19.3% (98/508) experiencing 30-day readmission. Baseline characteristics at index hospitalization for the readmitted/non-readmitted groups were the following: median age: 66.5y/65y (p=0.77); CURB-65 score >2: 33.7%/12.2% (p<0.001); ICU admission: 7.1%/6.6% (p=0.85); WBC: 12.9K/12.3K (p=0.48); immunocompromised status: 19.4%/9% (p=0.003); MRSA colonization: 12.2%/3.7% (p=0.001); HCAP: 63.3%/44.9% (p=0.001); aspiration risk: 26.5%/18.8% (p=0.086) COPD: 33.7%/26.3% (p=0.145); heart failure: 21.4%/15.4% (p=0.15); diabetes 30.6%/26.8% (p=0.45); hemodialysis 12.2%/3.7% (p=0.001); antibiotic use in the last 90 days: 41.8%/28.9% (p=0.013); median number of hospitalized days in the last 180 days: 4/0 (p<0.001). In the final multivariate model, primary viral pneumonia was protective for 30d hospital readmission OR=0.115 [95%CI 0.01-0.96] and four variables were independently associated with readmission: MRSA colonization, OR=3.13 [95%CI 1.33-7.38]; hemodialysis, OR=2.56 [95%CI 1.07-6.13]; CURB-65 score >2 OR=3.58 [95%CI 2.05-6.24]; immunocompromised status, OR=2.71 [95%CI 1.39-5.32].
Conclusion: Independent predictors of 30-day readmission in multivariate analysis are typically not modifiable; preventive interventions should be focused on these four patient populations to decrease hospital readmission.
A. E. Gross,
K. M. Olsen, None
M. E. Rupp, 3M: Consultant and Grant Investigator, Consulting fee and Research grant
A. C. Kalil, None