There is considerable disagreement over how hospital mortality rates should be measured. There are concerns that hospitals with a high number of geriatric beds and long-stay patients are bound to appear to do badly in terms of mortality rates. Analysis of death rates in English trusts shows mortality rates do vary for length of stay. After standardisation for age group and length of stay, the adjusted mortality rates show no bias against hospitals with more patients staying more than 28 days, or against hospitals with more geriatric beds.