Hand washing for preventing diarrhoea

 

Characteristics of included studies [ordered by study ID]

Bartlett 1988

---

Methods	Cluster-randomized trial Method of allocation sequence: unclear Allocation concealment: unclear Blinding: assessor Inclusion of participants in the analysis: unclear Length of follow up: 12 months Cluster-adjustment method: not adjusted
Participants	Number: 26 day-care centres, with 374 children Inclusion criteria: not stated Exclusion criteria: not stated Age: < 3 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group meetings (directors and caregivers) 2. Provision of posters and handouts depicting the procedures taught Control: 3. Visited to review surveillance procedures, but no instruction in disease prevention or management provided
Outcomes	Diarrhoea rates
Notes	Location: Maricopa County, Arizona, USA Duration: October 1981 to September 1984

---

Black 1981

---

Methods	Cluster-randomized trial Method of allocation sequence: unclear Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 6 months Cluster-adjustment method: not adjusted
Participants	Number: 4 day-care centres, with 116 children Inclusion criteria: not stated Exclusion criteria: not stated Age: < 3 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group education Control: 2. No intervention
Outcomes	1. Diarrhoea rates Not used in this review: 2. Estimate of load of diarrhoea causative agent
Notes	Location: suburban Atlanta, Georgia, USA Duration: June 1976 to April 1977

---

Bowen 2007

---

Methods	Cluster-randomized trial Method of allocation sequence: random-number table Allocation concealment: unclear Blinding: none Inclusion of all participants in the analysis: 93% (3962/4256) agreed to participate Length of follow up: 2003 to 2004 school year Cluster-adjustment method: adjusted
Participants	Number: 87 schools (57 intervention; 30 Control); with 3962 children (2670 intervention; 1292 control) Inclusion criteria: public primary schools; at least 20 students in first grade year in 2003 to 2004; no overnight boarders; at least 1 running water tap for every 30 first grade students Exclusion criteria: no compulsory hand washing or provision of hand-cleaning products before school lunch; no commercial hand washing promotion programmes at school during previous 5 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Expanded programme: as standard programme plus continuous supply of Safeguard soap for school sinks; 1 student from each class recruited to assist peers with hand washing techniques, and remind them of key hand washing opportunities; teachers asked to encourage this student weekly but not instructed to enforce hand washing behaviour 2. Standard programme: Proctor and Gamble's 'Safeguard' promotion programme delivered in Chinese schools since 1999; teachers deliver programme to first grade children during single 40 minute classroom session; also single 2 h training session for each first grade teacher delivered by Proctor and Gamble staff; teacher's pack contains guidebook outlining hand washing, basic information on infectious disease transmission, 5 posters describing hand washing procedure, videotape, and 5 wall charts for classroom hygiene competition; student take-home pack includes hygiene board game, parent booklet on hand washing, and 50 g bar Safeguard soap Control: 3. All 3 groups received government hygiene educational programme consisting of a cursory statement manual about hand washing after using toilet and before eating
Outcomes	1. Diarrhoea rates Not used in this review: 2. School absences 3. Rates of other common illnesses
Notes	Location: 3 counties in Fujian province, South-East China

---

Butz 1990

---

Methods	Cluster-randomized trial Method of allocation sequence: unclear Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 12 months Cluster-adjustment method: not adjusted
Participants	Number: 24 family day-care centres, with 108 children Inclusion criteria: not stated Exclusion criteria: unclear Age: 1 month to 7 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group training (in-home instruction to day-care providers) Control: 2. No intervention
Outcomes	Incidence of infectious disease symptoms (diarrhoea)
Notes	Location: Baltimore, Maryland, USA Duration: 12 months

---

Carabin 1999

---

Methods	Cluster-randomized trial Allocation sequence: day-care centres were stratified by incidence of respiratory infections and block randomized by geographical areas Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 6 months Cluster-adjustment method: adjusted
Participants	Number: 52 day-care centres, with 1729 children Inclusion criteria: presence of at least 1 sandbox and 1 play area; at least 12 available toddler places Exclusion criteria: not stated Age: 18 months to 3 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group hygiene training (educators) 2. Handouts Control: 3. No intervention
Outcomes	Diarrhoea rates
Notes	Location: Quebec, Canada Duration: September 1996 to November 1997

---

Haggerty 1994

---

Methods	Cluster-randomized trial Allocation sequence: unclear Allocation concealment: unclear Blinding: assessor Inclusion of participants in the analysis: unclear Length of follow up: 6 months Cluster-adjustment method: adjusted and unadjusted results given
Participants	Number: 18 sites, with 1954 children Inclusion criteria: not stated Exclusion criteria: not stated Age: 3 months to 35 months
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group training Control: 2. No intervention
Outcomes	Diarrhoeal rates
Notes	Location: Kikwit, Bandundu Province, Zaire (Democratic Republic of Congo) Duration: October 1987 to December 1988

---

Han 1989

---

Methods	Cluster-randomized trial Allocation sequence: unclear Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 4 months Cluster-adjustment method: not adjusted
Participants	Number: 350 households (162 intervention and 188 control) with 494 children Inclusion criteria: households with 1 or more children between 6 and 59 months; those in which regular follow up was possible; not allergic to soap; gave informed consent Exclusion criteria: not stated Age: < 5 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Small group education (households) Control: 2. No intervention
Outcomes	1. Incidence of diarrhoea 2. Incidence of dysentery
Notes	Location: Nga-Kha ward of Thin-Gun-Kyun township, Rangoon, Burma (now Myanmar) Duration: June to November 1985

---

Huang 2007

---

Methods	Individually randomized trial Allocation sequence: unclear Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: 100% Length of follow up: 1 year
Participants	Number: 73 intervention, 75 control Inclusion criteria: patients with acquired immune deficiency syndrome (AIDS) at local human immunodeficiency virus (HIV) clinic; HIV-1 infection verified by both ELISA and Western Blot; AIDS by CD4 counts and plasma HIV RNA; been on highly active anti-retroviral therapy (HAART) for at least 6 weeks and without diarrhoea for at least 3 months
Interventions	Both groups: 3 dedicated study nurses educated participants on health problem associated with contaminated hands and provided specific hand washing instructions at enrolment; hand washing technique demonstrated by nurses, including wetting hands, lathering completely with soap, running together for at least 15 seconds, and drying hands with towels; all 148 participants then demonstrated adequate hand washing technique Intervention (see  Table 2 for detailed description): 1. Weekly telephone call from nurses to determine number of hand washing episodes per day, ensure compliance, answer questions, re-educate participants on importance, and go over instructions Control: 2. Weekly telephone calls but only to ascertain diarrhoea episodes
Outcomes	1. Incidence of diarrhoea 2. Hand washing behaviour Not used in this review: 3. Microbiological diagnosis of diarrhoea episodes
Notes	Location: USA (location unclear) Duration: 1 year (exact dates unclear)

---

Kotch 1994

---

Methods	Cluster-randomized trial Allocation sequence: unclear Allocation concealment unclear Blinding: assessor Inclusion of participants in the analysis: unclear Length of follow up: 7 months Cluster-adjustment method: adjusted
Participants	Number: 24 day-care centres, with 389 classrooms Inclusion criteria: children < 3 years; present in the day care at least 20 h per week; absence of chronic illness or medication that would predispose to infection; youngest of potentially eligible children in the same family; consenting English-speaking parents with access a telephone; intending to remain in day-care centre throughout study Exclusion criteria: not stated Age: < 3 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group training 2. Curriculum for caregivers Control: 3. No intervention
Outcomes	Diarrhoeal rates
Notes	Location: Cumberland County, North Carolina, USA Duration: October 1988 to May 1989

---

Ladegaard 1999

---

Methods	Cluster-randomized trial Allocation sequence: unclear Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 4 months Cluster-adjustment method: not adjusted
Participants	Number: 8 day-care centres, with 475 children (212 intervention, 263 control) Inclusion criteria: not stated Exclusion criteria: not stated Age: < 6 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Small group practical demonstration Control: 2. No intervention
Outcomes	Diarrhoeal rates
Notes	Location: Odense, Denmark Duration: 6 months

---

Luby 2004a

---

Methods	Cluster-randomized trial Allocation sequence: computer-generated Allocation concealment: serially numbered Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 12 months Cluster-adjustment method: adjusted
Participants	Number: 36 neighbourhoods (25 intervention, 11 control), with 4691 children Inclusion criteria: household located in the study area; have at least 2 children < 5 years; intention to reside in the house for the duration of study Exclusion criteria: not stated Age: < 15 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group training using slide shows, pamphlets, and video tapes Control: 2. No intervention
Outcomes	Diarrhoeal rates
Notes	Location: low-income squatter settlements, Karachi, Pakistan Duration: April 2002 to April 2003

---

Luby 2006

---

Methods	Cluster-randomized trial Allocation sequence: computer generated Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 9 months Cluster-adjustment method: adjusted
Participants	Number: 18 clusters, with 544 households (262 intervention, 282 control) Inclusion criteria: households with at least 1 child < 5 years; provided informed consent Exclusion criteria: not stated Age range: < 15 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group training using slide shows, pamphlets, and video tapes Control: 2. No receipt of products expected to change risk of diarrhoea but provided them with regular supply of children's books, note books, etc
Outcomes	1. Primary diarrhoea rates 2. Persistent diarrhoea rates
Notes	Location: multi-ethnic squatter settlements in Central Karachi, Pakistan Duration: April 2003 to December 2003

---

Roberts 2000

---

Methods	Cluster-randomized trial Allocation sequence: computer generated Allocation concealment: unclear Blinding: assessors Inclusion of participants in the analysis: unclear Length of follow up: 9 months Cluster-adjustment method: adjusted
Participants	Number: 23 day-care centres, with 558 children Inclusion criteria: day-care centres licensed in the Australian Capital Territory; children < 3 years as at January 1996; attendance for at least 3 days per week; have no underlying chronic illness that predisposes to infection Exclusion criteria: not stated Age: < 3 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Large group training 2. Booklets/newsletters 3. Songs about hand washing for children Control: 4. No intervention
Outcomes	1. Diarrhoeal rate 2. Knowledge, attitude, and practice of hand washing
Notes	Location: Australian Capital Territory, Australia Duration: March to November 1996

---

Stanton 1987

---

Methods	Cluster-randomized trial Allocation sequence: table of random numbers Allocation concealment: unclear Blinding: none Inclusion of participants in the analysis: unclear Length of follow up: 6 months Cluster-adjustment method: adjusted
Participants	Number: 1923 families (937 intervention, 986 control) Inclusion criteria: not stated Exclusion criteria: not stated Age: < 6 years
Interventions	Intervention (see  Table 2 for detailed description): 1. Small group discussion (only women or children) 2. Larger demonstrations (mixed audience) 3. Posters, games, pictorial stories, and 'flexiflans' for illustrations Control: 4. No intervention
Outcomes	1. Diarrhoeal rates 2. Change in knowledge, attitude, and practice of water sanitation behaviours
Notes	Location: Urban Dhaka, Bangladesh Duration: October 1984 to May 1985

---

 ^aSee  Table 2 for a detailed description of the interventions.

 

Characteristics of excluded studies [ordered by study ID]

---

Study	Reason for exclusion
Ahmed 1993	Observational study examining risk factors for diarrhoeal infections
Alam 1989	Main intervention was provision of water supply through hand pumps
Barros 1999	Observational study examining risk factors for diarrhoeal infections
Clemens 1987	Observational study examining risk factors for diarrhoeal infections
Curtis 2001	No concurrent control
Doebbeling 1992	Outcome measure (incidence of nosocomial infection) not specific to diarrhoea episodes but to incidence of gastrointestinal infections in general
Dyer 2000	Intervention was instant hand sanitizer
Guinan 2002	Observational study
Hammond 2000	Intervention did not involve hand washing
Khan 1982	Case-control study
Larson 2003	No relevant outcome measures (measured colony-forming units of bacteria)
Larson 2004	Outcome measure not specific to incidence of diarrhoea
Lee 1991	Controlled before-and-after study
Luby 2001	Observational study
Luby 2004b	Nonrandomized study
Master 1997	Outcome measure not specific on diarrhoeal episodes
Morton 2004	Outcome measure not specific on diarrhoeal episodes
Peterson 1998	Observational study examining risk factors for diarrhoeal infections
Pinfold 1996	No comparable baseline information on diarrhoeal episodes provided
Shahid 1996	No comparable baseline information provided
Sircar 1987	No comparable baseline information on diarrhoea episodes provided
White 2003	Outcome measure not specific to diarrhoeal morbidity
Wilson 1991	Controlled before-and-after study

---

 

Comparison 1. Institutional-based trials: hand washing promotion vs no intervention

---

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of diarrhoea: cluster-adjusted rate ratios	2		Incidence rate ratio (Random, 95% CI)	0.61 [0.40, 0.92]
2 Incidence of diarrhoea: not cluster-adjusted rate ratios			Other data	No numeric data

---

 

Analysis 1.2 Comparison 1 Institutional-based trials: hand washing promotion vs no intervention, Outcome 2 Incidence of diarrhoea: not cluster-adjusted rate ratios.
Incidence of diarrhoea: not cluster-adjusted rate ratios

---

Study	Incidence rate ratio	9% confidence interval
Bartlett 1988	0.89	0.71 to 1.11
Black 1981	0.52	0.34 to 0.78
Butz 1990	0.72	0.54 to 0.94
Kotch 1994	0.84	0.73 to 0.98
Ladegaard 1999	0.67	0.43 to 1.05

---

 

Comparison 2. Community-based trials: hand washing promotion vs no intervention

---

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

---

 1 Incidence of diarrhoea: cluster-adjusted rate ratios

4

Incidence rate ratio (Random, 95% CI)

0.68 [0.52, 0.90]

---

 2 Incidence of diarrhoea: not cluster-adjusted rate ratios

Other data

No numeric data

---

 3 Incidence of diarrhoea: stratified by soap provision and type of intervention

4

Incidence rate ratio (Random, 95% CI)

Subtotals only

---

3.1 Soap provided and focus on hand washing

2

Incidence rate ratio (Random, 95% CI)

0.49 [0.39, 0.62]

---

3.2 No soap provided and multiple hygiene interventions

2

Incidence rate ratio (Random, 95% CI)

0.84 [0.67, 1.05]

---

 4 Episodes

1

148

Mean Difference (IV, Random, 95% CI)

-1.68 [-1.93, -1.43]

---

 

Analysis 2.2 Comparison 2 Community-based trials: hand washing promotion vs no intervention, Outcome 2 Incidence of diarrhoea: not cluster-adjusted rate ratios.
Incidence of diarrhoea: not cluster-adjusted rate ratios

---

Study	Incidence rate ratio	95% confidence interval
Han 1989	0.70	0.54 to 0.92

---

 

Table 1. Detailed search strategies

---

Search set	CIDG SR^a	CENTRAL	MEDLINE^b	EMBASE^b	LILACS^b
1	handwashing	handwashing	hand wash*	hand wash$	handwashing
2	diarrhea	hand washing	hand disinfec*	hand disinfec*	diarrhea
3	diarrhoeal diseases	hand cleansing	hand clean*	hand clean$	1 and 2
4	—	hand hygiene	hand hygiene	hand hygiene	—
5	—	1 or 2 or 3 or 4	hand sterility	hand sterility	—
6	—	diarrhea	HANDWASHING	HANDWASHING	—
7	—	5 and 6	1 or 2 or 3 or 4 or 5 or 6	1 or 2 or 3 or 4 or 5 or 6	—
8	—	—	diarrhea	diarrhea	—
9	—	—	diarrhoea	diarrhoea	—
10	—	—	8 or 9	8 or 9	—
11	—	—	7 and 10	7 and 10	—

---

 ^aCochrane Infectious Diseases Group Specialized Register.
^bSearch terms used in combination with the search strategy for retrieving trials developed by The Cochrane Collaboration (Higgins 2006); upper case: MeSH or EMTREE heading; lower case: free text term.

 

Table 2. Intervention details

---

Trial	Promotional activity	Classification^a	Message content	Hand washing method	Hand washing style^b	Material provision	Water availability
Institution-based
Bartlett 1988	1. Large group meetings (directors and caregivers) 2. Provision of posters and handouts depicting the procedures taught	1. Hygiene education 2. Participatory learning*	Staff and child hand washing, diapering, food handling, and environmental cleaning	Unclear	Not specified	Not specified	Adequate
Black 1981	Large group education	Hygiene education	Staff and child hand washing before handling food and after defecation	Water with bar soap and paper towels	Unclear	By the day-care centres' management	Adequate
Bowen 2007	1. Large group training 2. Posters, videotape, wall charts, games 3. Soap 4. Take home packs 5. Peer trainers and peer-monitoring	1. Hygiene education 2. Behaviour modification	Hand washing before eating and after toileting	Water with soap	Under running water	Supplies of soap to schools in "Expanded Intervention"; 1 bar of soap to homes in both expanded and standard intervention	Adequate (criteria for taking part in trial)
Butz 1990	Large group training (in-home instruction to day-care providers)	1. Hygiene education 2. Provision of soap/hand rinse material	1. Modes of transmission of pathogens in the home 2. Indications of hand washing 3. Use of vinyl gloves and disposable diaper changing pad 4. Use of an alcohol-based hand rinse (if unable to wash hand with water plus soap)	Water with soap	Not specified	All supplies provided by researchers	Adequate
Carabin 1999	1. Large group hygiene training (educators) 2. Handouts	Hygiene education	1. Wash hands before lunch and after using the toilets 2. Clean toys with bleach 3. Use of reminder cues for hand washing 4. Clean the sand box with bleach 5. Open windows at least 30 min every day	Unclear	Not specified	Unclear	Adequate
Kotch 1994	1. Large group training 2. Curriculum for caregivers	Hygiene education	1. Hand washing of children and staff 2. Disinfection of diapering areas and toilet 3. Physical separation of diapering areas from food preparation and serving areas 4. Hygienic diaper disposal	Water with soap plus disposable towel	Under running water	Unclear	Adequate
Ladegaard 1999	Small group practical demonstration	1. Hygiene education 2. Participatory learning^c	1. Hand washing after stool contact 2. Information on disease spread and when to wash hands to prevent diarrhoea	Water with soap	Under running water	Unclear	Adequate
Roberts 2000	1. Large group training 2. Booklets/newsletters 3. Songs about hand washing for children	1. Hygiene education 2. Behaviour modification	1. Hand washing before eating and after toileting or changing a diaper (staff and child) 2. Wash toys daily in dishwashers	Water with soap	Under running water	Unclear	Adequate
Community-based
Haggerty 1994	Large group training	Hygiene education	1. Hand washing before meal preparation and eating 2. Hand washing after defecation (wash both hand and buttocks for children) 3. Proper disposal of children's faeces 4. Disposal of animal faeces from yard	Unclear	Not specified	Unclear	Unknown
Han 1989	Small group education (households)	1. Hygiene education 2. Provision of hand washing material	Hand washing: 1. After defecation 2. Before preparing or eating food	Water with bar soap	Not specified	Plain bar soap provided by researcher	Unknown
Luby 2004a	1. Large group training using slide shows, pamphlets, and video tapes; education at weekly field visits 2. Education at weekly field visits	Hygiene education	Hand washing: 1. Before preparing food 2. Before eating food	Water with plain or antibacterial soap	Water from a pitcher (though not clearly stated)	Soap provided by researchers	Unknown
Luby 2006	1. Large group training using slide shows, pamphlets, and video tapes 2. Education at twice-weekly visits	1. Hygiene education 2. Provision of hand washing material	Hand washing: 1. After stool contact/ defecation 2. Before food preparation/handling/eating 3. Before feeding infants	Water with antibacterial soap	Not specified	Soap provided by researchers	Unknown
Stanton 1987	1. Small group discussion (only women or children) 2. Larger demonstrations (mixed audience) 3. Posters, games, pictorial stories, and 'flexiflans' for illustrations	Hygiene education	1. Hand washing before food preparation 2. Defecation away from the house and in a proper site 3. Suitable disposal of waste and faeces	Unclear	Not specified	Unclear	Inadequate
High-risk group (AIDS patients)
Huang 2007	Demonstration by nurses and patients	Hygiene education	1. Hand washing after toileting, before food preparation/handling, eating 2. After cleaning infants who had defecated 3. Before and after sex	Water with soap	Under running water	Unclear (probably not relevant in this population)	Adequate

---

 ^aMessage classification.
^bWhether done under running water; in a bowl by an individual or by several people.
^cParticipatory learning involves a process that helps engage learners in an active role of inquiry in which they share experiences and reflect critically on practice in a context that many group members find stimulating and relatively safe (Martin 1997).

 

Table 3. Methodological quality assessment

---

Trial	Sequence generation	Allocation concealment	Blinding	Inclusion^a	Comparability^b	Time of collection^c
Institution-based
Bartlett 1988	Unclear	Unclear	Assessor	Unclear	Adequate	Adequate
Black 1981	Unclear	Unclear	None	Unclear	Adequate	Adequate
Bowen 2007	Adequate	Unclear	None	Adequate	Adequate	Adequate
Butz 1990	Unclear	Unclear	None	Unclear	Adequate	Adequate
Carabin 1999	Adequate	Unclear	None	Unclear	Unclear	Adequate
Kotch 1994	Unclear	Unclear	Assessor	Unclear	Unclear	Adequate
Ladegaard 1999	Unclear	Unclear	None	Unclear	Adequate	Adequate
Roberts 2000	Adequate	Unclear	Assessors	Unclear	Unclear	Adequate
Community-based
Haggerty 1994	Unclear	Unclear	Assessor	Unclear	Inadequate	Adequate
Han 1989	Unclear	Unclear	None	Unclear	Adequate	Adequate
Luby 2004a	Adequate	Adequate	None	Unclear	Adequate	Adequate
Luby 2006	Adequate	Unclear	None	Unclear	Adequate	Adequate
Stanton 1987	Adequate	Unclear	None	Unclear	Adequate	Adequate
High-risk group (AIDS patients)
Huang 2007	Unclear	Unclear	None	Adequate	Adequate	Adequate

---

 ^aInclusion of randomized participants in the analysis was reported at different levels of analysis (cluster, child, person-at-risk levels).
^bComparability between intervention and control groups with respect to baseline characteristics (see methods).
^cData collected at similar time periods for intervention and control sites.

 

Table 4. Incidence of diarrhoea

---

Trial	Cluster adjusted?	Outcome and result	Method of assessment	Sample size
Institution-based
Bartlett 1988	No	Diarrhoea rate per child-year of observation Intervention: 0.71 (95% CI 0.65 to 0.77) Control: 0.81 (95% CI 0.75 to 0.87)	1. Active day-care centre-based surveillance (weekly visits plus daily telephone calls to identify diarrhoeal illness 2. Family-based surveys (questionnaire every 2 weeks)	26 day-care centres with 374 children (196 intervention, 178 control) aged 0 to 3 years
Black 1981	No	Diarrhoea incidence/100/child-weeks of observation Intervention: 4.2/100/child-week Control: 8.1/100/child-week	Daily record of attendance plus diarrhoea occurrence for each child by day-care personnel	4 day-care centres (2 intervention, 2 control) with 116 children < 3 years
Bowen 2007	Yes	Median episodes of diarrhoea per 100 student weeks Expanded intervention: 0 per 100 student-weeks Standard intervention: 0 per 100 student-weeks Control: 0 per 100 student-weeks	Teachers trained using standardized case definitions to identify 10 symptoms or signs of illness and record these among students in class, 1 day per week; if parent's reported infection as cause of absence, teachers recorded name of syndrome and asked parent if child suffered any of 10 individual symptoms; verified verbally that reports of diarrhoea met case definition	3962 children within 87 primary schools
Butz 1990	No	Proportion of diarrhoea days per month Diarrhoea episodes/child-days Intervention: 93/10,159 Control: 133/10,424	Daily symptom record for each child by care providers	24 family day-care homes with 108 children (58 intervention, 50 control) aged 1 month to 7 years
Carabin 1999	Yes	Diarrhoea incidence: episodes/100 child-days at risk Incidence rate ratio (95% Bayesian credible interval) 1.10 (0.81 to 1.50), adjusted for age and gender Intervention alone: 0.77 (0.51 to 1.18) Monitoring alone: 0.73 (0.54 to 0.97)	Daily record of diarrhoea episodes on calendar by educators	52 day-care centres with 1729 children aged 18 months to 3 years
Kotch 1994	Yes	Diarrhoea rates: incidence density (episodes/child-year) Intervention (< 2 years): 4.54 Intervention (> 2 years): 2.85 Control (< 2 years): 5.12 Control (> 2 years): 2.79 All: risk ratio 1.19, 95% CI -0.48 to 1.96	Telephone interview methodology (biweekly calls to families) 5 week interval visits to day-care centres	24 day-care centres with 389 children < 3 years
Ladegaard 1999	No	Diarrhoea episodes/child-month Intervention: 33/848 Control: 61/1052 (34% reduction from 3.25 days per child in favour of children 3 years or more)	Information on absenteeism recorded on a form by child-care provider	8 day-care centres with 475 children (212 intervention, 263 control) aged 6 years and below
Roberts 2000	Yes	Diarrhoeal rates: episodes/child-year Intervention: 1.9 episodes/child-year Control: 2.7 episodes/child-year All: risk ratio 0.50 (95% CI 0.36 to 0.68) < 2 years: risk ratio 0.90 (95% CI, 0.67 to 1.19) > 2 years: risk ratio 0.48 (95% CI 029 to 0.78) (Adjusted for clustering by centre, confounding variables (age, sex, weight at birth, breastfeeding status, child care history, and home factors), and interactions between age and intervention status, and between having a sibling who attends child care and intervention status)	1. Biweekly telephone interviews (parents reports of symptoms) 2. Observation for compliance of recommended practices every 6 weeks	23 day-care centres (11 intervention, 12 control) with 558 children under 3 years
Community-based
Haggerty 1994	Yes	Diarrhoea rates (mean episodes of diarrhoea ) Intervention site: 0.071 Control site: risk ratio 0.075 (risk ratio 0.94, 95% CI 0.85 to 1.05; P = 0.3)	1. Observation recording form 2. Diarrhoeal morbidity form	18 sites (9 intervention, 9 control) with 1954 children aged 3 months to 35 months
Han 1989	No	Incidence density ratio 1. Diarrhoea < 2 years: 0.69 (95% CI 0.48 to 1.10) > 2 years: 0.67 (95% CI 0.45 to 0.98) All: 0.70 (95% CI 0.54 to 0.92) 2. Dysentery < 2 years: 0.59 (95% CI 0.22 to 1.55) > 2 years: 1.21 (95% CI 0.52 to 2.80) All: 0.93 (95% CI 0.39 to 2.23)	Daily surveillance (24 h recall) for diarrhoea and dysentery	350 households (162 intervention, 188 control) with 494 children (236 intervention; 258 control) under 5 years
Luby 2004a	Yes	Incidence density of diarrhoea (number of new episodes of diarrhoea divided by the at-risk person-weeks of observation) Mean incidence 1. Primary diarrhoea Intervention: Antibacterial soap: 2.02 Plain soap: 1.91 Control: 4.06 2. Persistent diarrhoea Intervention: Antibacterial soap: 0.14 Plain soap: 0.12 Control: 0.17	Weekly observational visits to households	36 neighbourhoods (25 intervention, 11 control) with 4691 children (3163 intervention, 1528 control) aged < 15 years
Luby 2006	Yes	Diarrhoea episodes/100 child-weeks: for diarrhoea and persistent diarrhoea 1. Risk ratio: 0.57 (95% CI 0.35 to 0.86) 2. Diarrhoea, mean incidence: 3.71 3. Persistent diarrhoea, mean incidence: 0.09 -52% (-100% to 100%)	Weekly observational visits to households	18 clusters (544 households; 262 intervention; 282 control) with children < 15 years
Stanton 1987	Yes	Rate of diarrhoea per 100 person-weeks Incidence density ratio 0.75 (95% CI 0.66 to 0.84; P < 0.0001) < 2 years: 0.54 (95% CI 0.43 to 0.66) > 2 years: 0.68 (95% CI 0.54 to 0.85)	1. Biweekly histories of diarrhoea for children of all households 2. Single prolonged on-site visit to each sentinel family for hand washing-related behaviour observation	1923 families (937 intervention, 986 control) with children aged < 6 years
High-risk group (AIDS patients)
Huang 2007	Not applicable	Mean episodes of diarrhoea over study period (1 year) Intervention group: 1.24 (+/- 0.9) Control group: 2.92 (+/- 0.6)	Daily hand washing diary to record number of hand washing episodes per day and diarrhoea diary to record stool frequency and characteristics; weekly telephone calls from study nurse to ascertain episodes of these outcomes	75 in hand washing group, 73 controls

---

 CI: confidence interval.

 

Table 5. Behavioural change

---

Trial	Cluster adjusted?	KAP^a changes	Detail	Intervention		Control		P
				n	N	n	N
Institution-based
Kotch 1994	Yes	Hand washing behaviour, based on event sampling scores^b	After changing a diaper	—	0.75	0.37	—	—
			After contact with child's mucus, saliva, vomit, etc	—	0.66	0.21	—	—
Roberts 2000	Yes	Compliance for hand washing by children in 11 intervention centres by a score^c; measured as risk ratio of diarrhoeal episodes (relative to control centres) with 95% confidence intervals	1 (4 centres)	—	0.52 (0.37 to 0.75)	—	—	—
			2 (4 centres)	—	0.53 (0.37 to 0.76)	—	—	—
			3 (3 centres)	—	0.43 (0.27 to 0.70)	—	—	—
			Children ≥ 2 years	—	0.34 (0.17 to 0.65)	—	—	—
Community-based
Stanton 1987	Yes	Comparison of hygienic practices after intervention (risk ratio 1.48, confidence interval 1.01 to 2.21)	—	39	79 (39/79 = 49%)	25	75 (25/75 = 33%)	—
High-risk group (AIDS patients)
Huang 2007	Not applicable	Frequency of hand washing per day	At baseline and at the end of study	—	3.3 (+/- 0.98) 7 times daily	—	3.4 (+/- 1.1) 4 times daily	< 0.05

---

 ^aKAP: knowledge, attitude, and practice.
^bEvent sampling scores (0 = none; 0.5 = partially correct; 1.0 = as recommended in the training).
^cCompliance score: 1 = lowest compliance rate (53% to 69%); 2 = moderate compliance rate (70% to 79%); 3 = high compliance rate (≥ 80%).