BACKGROUND: Crisis resolution teams (CRTs) provide a community alternative to psychiatric hospital admission for patients presenting in crisis. Little is known about the characteristics of patients admitted despite the availability of such teams. METHODS: Data were drawn from three investigations of the outcomes of CRTs in inner London. A literature review was used to identify candidate explanatory variables that may be associated with admission despite the availability of intensive home treatment. The main outcome variable was admission to hospital within 8 weeks of the initial crisis. Associations between this outcome and the candidate explanatory variables were tested using first univariate and then multivariate analysis. RESULTS: Patients who were uncooperative with initial assessment (OR 10.25 95% CI-4.20-24.97), at risk of self-neglect (OR 2.93 1.42-6.05), had a history of compulsory admission (OR 2.64 1.07-6.55), assessed outside usual office hours (OR 2.34 1.11-4.94) and/or were assessed in hospital casualty departments (OR 3.12 1.55-6.26), were more likely to be admitted. Other than age, no socio-demographic features or diagnostic variables were significantly associated with risk of admission. CONCLUSION: With the introduction of CRTs, inpatient wards face a significant challenge, as patients who cooperate little with treatment, neglect themselves, or have previously been compulsorily detained are especially likely to be admitted. The increased risk of admission associated with casualty department assessment may be remediable.