Poster 28 Predictors of Readmission to Acute Care and Mortality in the Inpatient Rehabilitation Setting Among Cardiovascular Patients. Heather K. Vincent, PhD, MS (University of Florida, Gainesville, FL); Sally Gamon, RN; Kevin R. Vincent, MD, PhD. Disclosure: H.K. Vincent, ARA Research Institute, AMRPA; S. Gamon, None; K.R. Vincent, None. Objective: To identify factors associated with readmission to acute care and mortality in the inpatient rehabilitation facility (IRF). Design: Retrospective study. Setting: A free-standing university-affiliated IRF. Participants: Cardiovascular patients who received acute care and subsequent inpatient rehabilitation (N 138; 64.1% men; age, 72.4 11.8y; 80.8% Medicare insured) from December 1, 2002, to March 31, 2006. Diagnoses included coronary artery disease, heart failure, peripheral vascular disease, aneurysms, acute myocardial infarction (MI), heart valve repair or replacement, heart transplant, and peripheral bypass graft. Interventions: Not applicable. Main Outcome Measures: Patient variables (interventions, diagnostic test results, care needs, FIM instrument scores), discharge disposition, readmissions to acute care, and mortality. Results: 15.4% of patients were readmitted to acute care for MI symptoms, failure to thrive, falls, mental status change, and infection. 2.4% of patients expired during rehabilitative care; causes of death included second infarctions, heart failure exacerbation, and blood loss. Patients readmitted to acute care had a longer IRF length of stay (LOS) than remaining patients (23.2 5.1d vs 13.8 7.8d, respectively; P .05) but total IRF charges did not differ statistically ($23,550 vs $19,567, P .279). Ageand sexadjusted regression models revealed that independent predictors of readmission to acute care included presence of dysrhythmias or foot ulcers, an indwelling catheter, angina, peripheral vascular disease, anxiety, lung atelectasis, development of infections, percentage of days active with therapies, and IRF LOS (all models, P .05). Independent predictors of mortality in the IRF included previous infarction prior to rehabilitation, angina, presence of diabetes, and specific admission low FIM subscores (bathing, verbal expression, problem solving, memory, auditory comprehension) (all P .05). Conclusions: Prediction of acute care readmission and mortality is multifaceted, and these preliminary findings can be used to start identifying which patients are less likely to succeed in the IRF.