On the use of multi-state multi-census techniques for modelling the survival of elderly people in institutional long-term care

^** Corresponding author. Email: chausst{at}wmin.ac.uk A good estimate of the length of stay (LOS) of elderly peoplein institutional long-term care (ILTC) is essential to bothcare providers (care home owners) and purchasers (local authoritiesin the UK) in order to forecast their resource availabilityand future requirements. Existing methods usually rely on cohortdata, which needs to be collected on a continuous basis forseveral years and are often not available electronically. Inthis paper we develop an approach based on multi-census datato fit a Markov model in continuous-time for the LOS of elderlypeople in ILTC (residential and nursing care). We use 3 yearplacement data provided by London Borough of Merton Social Servicesto compare the estimates of the model parameters obtained usingmulti-census data with those derived from cohort data, henceto validate the multi-census approach in this context. The implicationsof the proposed multi-census approach and its advantages inthe context of long-term care planning are discussed.