Measuring and accounting for the Hawthorne effect during a direct overt observational study of intensive care unit nurses

Abstract

Because suspecting nurses could alter hand hygiene (HH) behavior when observed, the goal of this article was to describe how the Hawthorne effect (HE) was measured and accounted for in a direct observational prospective study. Observations were made 8 h/d for 3-5 days in 5 intensive care units (ICUs) (4 hospitals) on a convenience sample of 64 ICU nurses in Texas. The HE was measured so if hand hygiene adherence rates of the first 2 hours were 20% higher than the last 6 hours, the first 2 hours would be dropped and an additional 2 hours would be added at the end of the observation period. Hourly rates were recorded during the observation period, using room entry and room exit. The difference between aggregated rates of the first 2 hours and last 6 hours was 0.56% (range, 0.02%-15.74%) and not significant. On 12 observation days, higher rates were observed during the first 2 hours. On 6 days, higher rates were observed in the last 6 hours, with difference in rates of 1.43% (day 1), 2.97% (day 2), and 1.42% (day 3). The attempt at measuring and accounting for the HE showed little difference in HH rates throughout the observation period. Based on these results, necessity of the observer moving locations during HH surveillance after 10-20 minutes, because of a feared HE, might not be necessary.