The relationship of bed size and hospital type (private or public) was studied using Hospital-Wide Medical Indicator data on nosocomial infections submitted to the Australian Council on Healthcare Standards Care Evaluation Program by hospitals presenting voluntarily for accreditation in 1993. The aim was to determine if this process could simplify the establishment of hospital peer groups for comparison of risk in the absence of knowledge of patient illness severity indices. After adjusting for potential confounders in a logistic model, hospital type was found to be a significant predictor for the occurrence of infection in clean and contaminated wounds. Bed size was a significant predictor for the occurrence of hospital-acquired bacteraemia in private and public hospitals. The increase in the risk of developing hospital acquired bacteraemia with increasing number of beds was significant as a trend (P < 0.0001) in private as well as public hospitals. The results suggest that hospital type and bed size are initial indices for 'flagging' peer group variation and prompting a more detailed internal review.