• Qualitative research
• Quality improvement
• Prehospital care

Interest in the utility of measuring preventable hospital deaths to drive improvement dates back to Florence Nightingale's first intimation that variations in mortality between London hospitals might reflect differences in quality of care.1 In 1999, the US Institute of Medicine's report ‘To Err is Human’ published the frequently quoted estimate of 44 000–98 000 preventable deaths annually in US hospitals, claiming that medical error represented the eighth most common cause of death in the country.2

This claim has fuelled ongoing and vigorous debate over actual numbers across many countries. Following well-publicised failures at Bristol Royal Infirmary3 and Mid Staffordshire NHS Foundation Trust,4 ,5 in England, the problem of preventable deaths has come to the wider attention of politicians and the public alike. Of late, politicians in England have developed a myopic focus on tackling preventable deaths as the key to raising performance across the NHS and look to single out discrete measures for bench-marking purposes, despite clearly not representing the complexity of modern-day hospitals.6

Acknowledgement of the existence of preventable hospital deaths is helpful in raising interest in the scale and burden of healthcare-related harm and encouraging commitment to improvement among clinicians and hospital managers. However, using preventable deaths as a comparative measure of quality between hospitals, if measures are not robust and fair, may overestimate the size of the problem and the risk to patients inducing unjustified levels of anxiety and fear and have a powerful stigmatising effect on hospitals identified as ‘high death rate’ outliers. Conversely, underestimation may lead to complacency and failure to acknowledge ongoing risks to patients. A thorough understanding of problems associated with both the concept and different approaches to measurement is needed to determine the role of preventable deaths in quality improvement.