"Background: The selection of patients for rehabilitation, and the timing of transfer from acute care, are important clinical decisions that impact on care quality and patient flow. This paper reports utilization review data on inpatients in acute care with stroke, hip fracture or elective joint replacement, and other inpatients referred for rehabilitation. It examines reasons why acute level of care criteria are not met and explores differences in decision making between acute care and rehabilitation teams around patient appropriateness and readiness for transfer. Methods: Cohort study of patients in a large acute referral hospital in Australia followed with the InterQual utilization review tool, modified to also include reasons why utilization criteria are not met. Additional data on team decision making about appropriateness for rehabilitation, and readiness for transfer, were collected on a subset of patients. Results: There were 696 episodes of care (7189 bed days). Days meeting acute level of care criteria were 56% (stroke, hip fracture and joint replacement patients) and 33% (other patients, from the time of referral). Most inappropriate days in acute care were due to delays in processes/scheduling (45%) or being more appropriate for rehabilitation or lower level of care (30%). On the subset of patients, the acute care team and the utilization review tool deemed patients ready for rehabilitation transfer earlier than the rehabilitation team (means of 1.4, 1.3 and 4.0 days from the date of referral, respectively). From when deemed medically stable for transfer by the acute care team, 28% of patients became unstable. From when deemed stable by the rehabilitation team or utilization review, 9% and 11%, respectively, became unstable. Conclusions: A high proportion of patient days did not meet acute level of care criteria, due predominantly to inefficiencies in care processes, or to patients being more appropriate for an alternative level of care, including rehabilitation. The rehabilitation team was the most accurate in determining ongoing medical stability, but at the cost of a longer acute stay. To avoid inpatients remaining in acute care in a state of 'terra nullius', clinical models which provide rehabilitation within acute care, and more efficient movement to a rehabilitation setting, is required. Utilization review could have a decision support role in the determination of medical stability."