This paper describes a modelling study of a multidisciplinary review panel which is responsible for matching levels of long-term care to the needs of older people. The study aims to understand the decision making process of the review panel and to predict placement decisions based on an applicant's attributes. Data were collected from cases notes presented to the London Borough of Merton review panel. A model predicting placement of an individual to residential home, nursing home or long-stay nursing care was built using logistic regression. and correctly predicts 78% of placement decisions. The model can be used as a means of checking the consistency of the review panel's placement decisions.