Background: The benefit of critical care unit (CCU) admission in acute heart failure (AHF) remains unclear. The risk factors for in-hospital events best managed in a CCU including the need for CCU restricted therapies (including invasive and non-invasive mechanical ventilation, mechanical circulatory support devices, and intravenous vasopressors or vasodilators) have not yet been formulated. The purpose of this study was to develop a clinical prediction model for adverse outcomes or CCU restricted therapies in patients with AHF. Methods: Using data from the ASCEND-HF trial, patients with AHF who did not require critical care related therapies within the preceding 12 hours of randomization were selected. The primary outcome was an in-hospital composite of the requirement of CCU specific therapies, death, myocardial infarction, cardiogenic shock, resuscitated sudden cardiac death, or ventricular arrhythmias requiring intervention. Model discrimination and calibration were evaluated using the c-index and the Hosmer-Lemeshow test, respectively. Results: The study cohort included 4772 patients and the primary composite outcome occurred in 547 (11.5%) patients. A total of 11 variables were independent predictors of the primary composite outcome as follows (Figure): chronic respiratory disease, prior ACE inhibitor, angiotensin receptor blocker or aldosterone antagonist use, Asian race, body mass index, diastolic blood pressure, respiratory rate, resting dyspnea, hemoglobin, sodium, and blood urea nitrogen. The simplified clinical prediction model demonstrated modest discrimination (c-index= 0.66) and good calibration (Hosmer-Lemeshow Goodness of Fit=7.017, p=0.535). Conclusions: In an international dataset of patients with AHF, we derived a clinical prediction describing patients who are likely to need a CCU. This model may be useful as a triage tool to identify patients with AHF who may benefit from admission to a higher acuity CCU.