New sequence types of Legionella pneumophila circulating in northern Italy and comparison with other regions of the world.

significant day-to-day change in the trend, and after the intervention there was no significant change in the day-to-day trend. In phase 2, individual compliance ranged from 72% to 100% with a median compliance of 92%, and a mean compliance of 91%. Nearly half (47%) of the nurses had compliance rates 95%. Compliance on room entry was 90% and on room exit 94%. Therewas no significant difference in compliance rates between work shifts. Using the alcohol sensor badges, we determined that there were 10 HHO per nurse-hour. A post-study questionnaire was completed by 14 nurses. Twelve felt that the alcohol sensor badge improved their compliance, one felt it had no effect, and one felt that it did not improve compliance. Ten respondents felt that all healthcare workers should wear the badge. Our study demonstrated easy use of an alcohol sensing badge, with rapid and significant improvement in hand hygiene compliance. Our results may underestimate the impact of this technology since performance was not reviewed by the nurse supervisor and no feedback was given. Moreover, we did not involve patients in the study. Patients could be instructed to observe the colour of the badge light before contact with the healthcare worker and request that he/she perform hand hygiene if the badge light is red, even though most patients are not comfortable asking caregivers to perform hand hygiene.6,7 We were also able to provide estimates of HHO per nursing hour. This is an underestimate since the badge only measures compliance on room entry and exit, and misses opportunities that may occur inside the patient room. The limitations of the study are its brevity and its performance in one ward of a single medical centre. In addition, compliance could only be assessed when the nurse performed hand hygiene using an alcohol product. This limitation could be overcome by adding a marker to liquid soaps used with water. Lastly, the quasi-experimental study design lacks randomisation, and may be complicated by temporal confounders and regression to the mean. Segmented regression analysis was performed to minimise these effects. To our knowledge, these results represent the most complete performance data and the highest level of compliance reported. A solution to monitoring hand hygiene compliance may be within reach.