This article discusses the principal methodological problems related to hospital mortality as an indicator of clinical performance. Hospital mortality rates variation are due to various factors associated with patients' characteristics and to the specific diseases they are suffering. Socio-demographic variables, presence of comorbidity and severity may define case-mixes were chances of dying are not associated to technology deployed or quality of care. Relevant methodological aspects for calculating the rates include the quality of the source of data, time period and aggregation criteria. Various models that exist both for classifying severity of cases and for risk adjustment are presented and discussed. Explanatory models for mortality rates variation are also examined. The authors conclude that outcome indicators can be used as tool for health care service evaluation. For those conditions which death is not a rare event hospital mortality rates may constitute an useful tool for indicating services with low than expected quality of care.
Quality; Health Services Evaluation; Outcome Indicators; Hospital Mortality; Rate Adjustment