This study aimed to develop predictive models to assist clinicians working in acute care to determine which stroke and lower limb orthopaedic patients should be discharged to nursing homes, inpatient rehabilitation, or home. The subjects were 223 patients with stroke and 125 patients with lower limb orthopaedic problems, aged 60 years and over, and the 22 acute care and rehabilitation clinicians who cared for these patients. Patient status was measured within 72 hours of discharge using the Adult FIM(SM) (Guide for the Uniform Data Set for Medical Rehabilitation, 1993) and ten additional items from either the RICFAS (Rehabilitation Institute of Chicago, 1987) or developed for this research. Four mathematical models were developed to predict discharge destination using stepwise discriminant function analysis. Using between three and seven items from the FIM(SM) alone or FIM(SM) plus additional variables, these models correctly classified between 74.9% and 80.5% of patients. In conclusion, clinicians are offered a selection of models to guide their discharge destination decisions for two large groups of patients. While these models cannot replace clinical judgement, they increase the transparency of decisions. The use of objective models to guide clinical decisions are essential, particularly given increasing pressure to justify patient access to costly rehabilitation services.