• Attention was drawn to the safety of patients in acute care hospitals in the early 1990s when studies found large numbers of potentially preventable deaths. Errors were initially ascribed to individual doctors and nurses, but later it was recognised that errors were mainly related to failure of systems rather than individuals.
• Mortality is not necessarily a good measure of hospital safety. It depends more on the nature of the patient's underlying clinical state and the type of intervention than on the safety of the hospital, and its prevention (as a measure of patient safety) contributes to the failure of hospitals to recognise and appropriately manage patients who are naturally at the end of life.
• It is difficult to find agreement on the best ways to measure patient safety in hospitals and, as a result of the enormous resources devoted to improving and studying safety, it is difficult to show that patient safety has improved. However, the concept of safety is beginning to include post-hospital outcomes, such as quality of life.
• A rapid response system is an organisation-wide patient safety system which recognises the deterioration of a patient's condition and provides urgent and appropriate care. Evaluating the impact of a rapid response system can provide information on hospital safety, including potentially preventable deaths and cardiac arrests.