OBJECTIVE: "Readmission risk score", a 20-point, 4-dimensional tool, is generated from the electronic medical record. This study was performed to evaluate the ability of the readmission risk score to predict 30-day readmissions among older hospitalized patients.

METHODS: A retrospective study was conducted utilizing data from the electronic medical record. Using a cutoff value of 7, the readmission score sensitivity was 61%, specificity was 22%, positive predictive value 12%, negative predictive value 77%. The positive and negative likelihood ratios were 0.8 and 1.8, respectively.

CONCLUSION: The readmission risk score was associated with 30-day readmissions (median score of readmitted vs not readmitted patients was 8 vs. 5; P = 0.001), and it may be better at identifying those who are not at risk for readmission.