This represents one of the first studies to characterize differences in hand hygiene practices not only by profession but also by level of patient contact. Additionally this study demonstrates that staff members in the NICU using an alcohol-based product were significantly more likely to touch neonates with clean hands than staff on the NICU using an antiseptic detergent. The findings of this study are consistent with previous reports of suboptimal hand hygiene,9-13
as well as reports of differences in hand hygiene behavior between different disciplines of hospital staff5, 14, 15
or different hand hygiene regimens.16-18
More than one-half of the touches made directly to neonates were with unwashed ungloved hands, providing a means of contaminating the immediate environment of each neonate and a potential reservoir for patient-to-patient spread of potential pathogens. Although adherence to appropriate hand hygiene by every profession did increase as the level of contact with the neonate increased (i.e. staff were least likely to have uncleaned, ungloved hands when they directly touched an infant), unchanged gloves were still worn for 15.9% of direct contacts with neonates, and cleaned, gloved hands were used in only 17% of direct contacts.
A recent study reported that gender and profession may interact, because it was shown that women and nurses of both genders tended to wash more often than men and physicians.19
However, in this study we did not find that nurses demonstrated greater adherence to hand hygiene protocols than other health care professionals. We observed that nurses wore used gloves for 22.3% of direct contacts and no gloves with uncleaned hands for almost one-third of direct neonatal contacts. Nurses also had cleaned gloved hands for direct contact with neonates less frequently than any other professional group. These findings are concerning, considering that more than one-half of all contacts with neonates and their equipment and environment were made by nurses.
This study had strengths and limitations. The large number of observations allowed for good statistical power and there was excellent interrater reliability between the two observers. Because the observations were conducted openly by research assistants involved in the larger clinical trial, the NICU nurses and physicians were probably more aware of their own hand hygiene practices and of the fact that observations were being made. However, workers who came to the unit less frequently may not have been aware of the study or the purpose of the research assistants. The groups of workers least likely to be aware of the observations (i.e. other health care workers such as phlebotomists, respiratory therapists or radiology technicians) had the highest rates of appropriate hand hygiene when directly touching a neonate.
The data collection was limited to a single month, and observations were only made during the day shift. Hence we were unable to compare the hand hygiene practices of staff and visitors during other shifts. Because our primary aim was to examine the overall types and numbers of contacts to neonates in the NICU while protecting the anonymity of individual staff members, we did not identify hand hygiene practices of specific members of the staff. Therefore it was not possible to analyze the data for gender differences in practice or to count the number of different individuals contacting each neonate. Further we could not determine whether only certain staff members practiced poor hand hygiene repeatedly whereas others habitually practiced good hand hygiene. Finally the observation instrument was developed for the purposes of this study and has not been tested in other studies.
The hand hygiene practice differences between the two hospitals were substantial. NICU A has a smaller, older, more crowded unit whereas NICU B is a newer, spacious facility. We are uncertain why neonates were touched significantly more often in NICU B, but this could be a surrogate marker for care needs or for differences in practice patterns. The universal gloving policy of NICU A almost certainly influenced practice, but this was not the only explanation for practice variations because both gloved cleaned hands and ungloved but cleaned hands were used significantly more often in NICU A than NICU B. Also the difference in protocols does not explain the large difference in the rate of touches with unwashed ungloved hands.
Staff in the two NICUs were using two different hand hygiene products. Staff members on the unit using the alcohol product (NICU A) were significantly less likely to touch neonates with unclean or ungloved hands than staff on the unit using a traditional antimicrobial soap. The fact that the alcohol hand rubs were available in pocket size bottles as well as being placed at each bedside precludes the necessity to travel to a sink and facilitates hand hygiene. Others have found that such alcohol products are more convenient and have been associated with increased adherence to accepted hand hygiene standards.10, 20-22
Our data are consistent with this observation and with the recently published CDC Guideline for Hand Hygiene in Health Care Settings3
which recommends alcohol-based rubs over antimicrobial soaps for staff hand hygiene.
In previous research traditional interventions to change hand hygiene behavior such as staff education and feedback have not resulted in sustained improvements.23
One intervention study demonstrated a significant positive association between overt administrative commitment to hand hygiene and improved frequency of handwashing as well as reductions in rates of certain infections. The administrative support was demonstrated by communications in the hospital newsletter, posters, pay stubs and other formats encouraging handwashing and by requiring that each staff member complete a competency demonstration in handwashing during orientation.24
Hence it would seem that an intervention which includes the adoption of the newer alcohol-based products accompanied by strong administrative support has the potential to improve hand hygiene practices. Because staff often overestimate the frequency with which they practice hand hygiene,15, 25-27
observational studies such as our study can be used as baseline data to increase staff awareness of actual practices and to evaluate the effects of interventions to improve hand hygiene behavior.