A note on the history of the calculation of hospital statistics.

There is a continuing debate in the field of hospital infectious disease epidemiology about the correct way to calculate nosocomial infectious disease morbidity and mortality. On the one hand, there are the traditionalists who stick to common use and ease of understanding and divide the number of infected patients by either the number admitted or the number discharged during the period of observation (1). However, others, generally more mathematically inclined, never seem to tire of pointing out the shortcomings of this type of denominator (there is little quarrel with the numerator). They offer alternatives, the main one being to calculate the rates per patienttime of stay, i.e., per patient-day or patientweek (2, 3). Another alternative, especially in the context of clinical studies, is to calculate the rates per patients at risk (i.e., from time of admission).