Introduction: Inhospital mortality has been used as an outcome quality indicator in the USA and in England to compare and benchmark hospital performance. It is now also possible to measure this outcome indicator in Switzerland, but it is important to highlight limitations and precautions to its use.
Methods: We collected administrative data from acute care community hospitals in the Canton of Valais, Switzerland, for the year 2001. We assessed rates of global and disease specific inhospital mortality and calculated crude and adjusted relative risks of inhospital mortality, specific to each hospital.
Results: The crude rates of the global inhospital mortality varied from 1.25% to 1.80% between hospitals. The variation for disease specific mortality rates was low. After adjustment, differences between relative risks were almost never statistically significant.
Discussion: The use of inhospital mortality as an quality indicator, need to be done with cautions, in particular adjustment for the case-mix, exclusion of patients in palliative care and analysis of disease specific rates.