Identifying heterogeneity in the Hawthorne effect on hand hygiene observation: a cohort study of overtly and covertly observed results

Table 2 Differences in the Hawthorne effect among various departments, occupations, observation times, observation locations, number of indications per hand hygiene opportunity, and hand hygiene indications

Variable	Pair no.	Hand hygiene compliance (%)		Hawthorne effect (percentage point difference)	p value^a
Variable	Pair no.	Overt observation	Covert observation	Hawthorne effect (percentage point difference)	p value^a
Department (N = 6094)					0.255
Medicine	1212	76.6	53.3	23.3
Surgery	1258	76.1	48.5	27.6
Others	577	86.5	70.2	16.3
Occupation (N = 6094)					< 0.001
Nurse	2105	84.2	54.2	30.0
Physician	619	67.9	57.4	10.5
Caregiver	106	50.0	42.5	7.5
Others	217	64.1	57.1	7.0
Time^b (N = 6094)					1.000
Day shift	2493	78.5	55.9	22.6
Early night shift	516	77.1	50.4	26.7
Late night shift	38	73.7	28.9	44.8
Location (N = 6094)					< 0.001
Ward	2034	78.1	50.3	27.8
Intensive care unit	880	80.6	69.2	11.4
Outpatient department	133	64.7	24.1	40.6
Numbers of indications per hand hygiene opportunity^c (N = 4686)					0.541
1	2301	76.7	53.2	23.5
2	42	85.7	73.8	11.9
Hand hygiene indication^d (N = 1300)					0.130
Before touching a patient	270	72.2	47.4	24.8
Before clean/aseptic procedures	19	78.9	52.6	26.3
After body fluid exposure risk	18	77.8	50.0	27.8
After touching a patient	287	83.3	59.9	23.4
After touching patient surroundings	56	55.4	53.6	1.8
Overall (N = 6094)	3047	78.2	54.6	23.6

^aGeneralized estimating equations were used to test whether there were significant differences in the Hawthorne effect among different categories of each variable
^bThe day shift was 0800–1600 h, the early night shift 1600–2400 h, and the late night shift 2400–0800 h
^cOnly the matched pairs with the same number of indications for both overt and covert observation methods were included
^dOnly the matched pairs with the same HH indication for both overt and covert observation methods were included

ISSN: 1471-2334