Objectives Rehabilitation in the home (RITH) services increasingly provide hospital substitution services. This study examines clinical outcomes in a large metropolitan RITH service in Western Australia. Methods The 2010 database of Fremantle Hospital RITH service was interrogated to identify the clinical profile of cases, length of stay (LOS) and clinical outcomes. Negative outcomes included death or unexpected hospital readmission. Multiple logistic regression modelling was used to explore associations with negative outcomes. This study was reviewed by the Institutional Review Board which deemed it not to require ethics approval. Results There were 1348 cases managed by RITH: 70.6% were aged≥65 years; elective joint replacement (29.7%), medical conditions (20%), stroke (13%), hip fractures (10%) were major contributors. The majority (93.3%) were discharged after a median of 9 days. Negative outcomes occurred in 90 cases (6.7%), including five deaths (0.4%) and 85 readmissions (6.3%). Independent associations with negative outcomes included older age (odds ratio (OR) (95% CI); 1.02, P≤0.006), orthopaedic conditions (OR 1.91, P≤0.004) and longer inpatient LOS (OR 1.96, P≤0.003). Age above 80 years was independently associated with risk of negative outcome (OR 2.99, P≤0.004). Conclusions RITH had a low rate of negative outcomes. The database proved useful for monitoring quality of service provision. What is known about the topic? Rehabilitation in the home environment has proven cost effective for multiple conditions, particularly stroke and elective joint surgery, among others, facilitating better quality of life, with reduced rates of delirium and mortality. Overall there are few negative outcomes and death is rare. What does this paper add? Although RITH services are widely utilised as bed substitution services, there is scant literature on clinical outcomes while within the service. This study focuses on frequency of good and poor clinical outcomes in a well-established RITH service in Western Australia, suggesting pattern recognition of an at-risk cohort by identifying potentially useful predictors of poor outcome. What are the implications for practitioners? RITH services are a safe alternative for many, including older people. Health administration databases are useful tools to monitor clinical outcomes. Clinical indicators such as older age, long hospital stay and orthopaedic diagnoses may be useful predictors of poor outcomes in such services.