Universal precautions (UPs) for all health care workers (HCWs) have been recommended by the Centers for Disease Control and mandated by the Occupational Safety and Health Administration in 1992. Evaluation of the outcomes of UPs is difficult, since the testing of all HCWs for bloodborne pathogen infection has not been readily accepted. An alternative to evaluation of outcomes, i.e., the presence or absence of bloodborne pathogen infection, is to measure the use of UPs itself and assume that decreased exposures to bloodborne pathogens will result in decreased infections with these pathogens. The purpose of this study was to determine UPs compliance by direct observation and by self-report questionnaire in a random sample of critical care nurses, a group with a high daily index of exposure to blood and body fluids. Additionally, knowledge, attitudes, and supply and equipment variables that might impact on the rate of compliance were also examined. Data were collected on a random sample of 25 nurses on two critical care units in a military medical center. Each nurse was observed performing at least two procedures that would expose him/her to blood or body fluids, using the Universal Precautions Assessment Tool (UPAT) to score compliance rates. The same sample of nurses was then asked to complete an 85 item questionnaire that produced a score for knowledge of UPs, attitudes toward UPs, and their opinions of the quality, fit, availability, and accessibility of the supplies and equipment provided for their use. The 25 nurses were observed performing 73 procedures on 59 patients. The overall observed compliance score for all nurses was 67%, with a range of 25%-100%. Nurses who reported any prior blood or body fluid exposures to patients with a known or suspected diagnosis of HIV or hepatitis had higher compliance with UPs than those who did not report such exposure. Nurses who reported lower numbers of exposures to the blood and body fluids of any patient also had higher UPs compliance scores. Self-reported compliance scores ranged from 8% for eye protection to 93% for glove use. Pearson r correlation coefficients were obtained for the subscales of each nurse's questionnaire and the observed compliance score. No statistically significant correlations were obtained between knowledge of UPs, attitudes towards UPs, or the quality, fit, availability, or accessibility of supplies and equipment and the observed compliance rates. Power analyses revealed that the sample size was too small to reveal significant findings. It is possible that a larger sample size would demonstrate that these factors are indeed associated with UPs use. It is also possible that the UPAT is not an appropriate tool to use for correlations with associated factors. These findings also suggest that UPs is an effective strategy to prevent exposure to blood and body fluids.