Abstract

Hospital mortality rates could be useful indicators of quality of care, but careful statistical analysis is required to avoid erroneously attributing variation in mortality to differences in health care when it is actually due to differences in case mix. The summary hospital mortality indicator is currently used by the English National Health Service (NHS). It adjusts mortality rates up to 30 days after discharge for patient age, sex, type of admission, year of discharge, comorbidity, deprivation and diagnosis. Such risk-adjustment methods have been used to identify poor performance, most notably at mid-Staffordshire NHS Foundation Trust, but their use is subject to a number of limitations. Studies exploring whether variation in risk-adjusted mortality can be explained by variation in healthcare have reached conflicting conclusions. Furthermore, concerns have been raised that the proportion of preventable deaths among hospital admissions is too small to produce a reliable ‘signal’ in risk-adjusted mortality rates. This provides hospital managers, regulators and clinicians with a considerable dilemma. Variation in mortality rates cannot be ignored, as they might indicate unacceptable variation in healthcare and avoidable mortality, but they also cannot be reliably used to judge the quality of healthcare, based on current evidence.