|Home | About | Journals | Submit | Contact Us | Français|
Diarrhoea accounts for 1.8 million deaths in children in low- and middle-income countries (LMICs). One of the identified strategies to prevent diarrhoea is hand washing.
To assess the effects of hand washing promotion interventions on diarrhoeal episodes in children and adults.
We searched the Cochrane Infectious Diseases Group Specialized Register (27 May 2015); CENTRAL (published in the Cochrane Library 2015, Issue 5); MEDLINE (1966 to 27 May 2015); EMBASE (1974 to 27 May 2015); LILACS (1982 to 27 May 2015); PsycINFO (1967 to 27 May 2015); Science Citation Index and Social Science Citation Index (1981 to 27 May 2015); ERIC (1966 to 27 May 2015); SPECTR (2000 to 27 May 2015); Bibliomap (1990 to 27 May 2015); RoRe, The Grey Literature (2002 to 27 May 2015); World Health Organization (WHO) International Clinical Trial Registry Platform (ICTRP), metaRegister of Controlled Trials (mRCT), and reference lists of articles up to 27 May 2015. We also contacted researchers and organizations in the field.
Individually randomized controlled trials (RCTs) and cluster-RCTs that compared the effects of hand washing interventions on diarrhoea episodes in children and adults with no intervention.
Three review authors independently assessed trial eligibility, extracted data, and assessed risk of bias. We stratified the analyses for child day-care centres or schools, community, and hospital-based settings. Where appropriate, incidence rate ratios (IRR) were pooled using the generic inverse variance method and random-effects model with 95% confidence intervals (CIs). We used the GRADE approach to assess the quality of evidence.
We included 22 RCTs: 12 trials from child day-care centres or schools in mainly high-income countries (54,006 participants), nine community-based trials in LMICs (15,303 participants), and one hospital-based trial among people with acquired immune deficiency syndrome (AIDS) (148 participants).
Hand washing promotion (education activities, sometimes with provision of soap) at child day-care facilities or schools prevents around one-third of diarrhoea episodes in high income countries (rate ratio 0.70; 95% CI 0.58 to 0.85; nine trials, 4664 participants, high quality evidence), and may prevent a similar proportion in LMICs but only two trials from urban Egypt and Kenya have evaluated this (rate ratio 0.66, 95% CI 0.43 to 0.99; two trials, 45,380 participants, low quality evidence). Only three trials reported measures of behaviour change and the methods of data collection were susceptible to bias. In one trial from the USA hand washing behaviour was reported to improve; and in the trial from Kenya that provided free soap, hand washing did not increase, but soap use did (data not pooled; three trials, 1845 participants, low quality evidence).
Hand washing promotion among communities in LMICs probably prevents around one-quarter of diarrhoea episodes (rate ratio 0.72, 95% CI 0.62 to 0.83; eight trials, 14,726 participants, moderate quality evidence). However, six of these eight trials were from Asian settings, with only single trials from South America and sub-Saharan Africa. In six trials, soap was provided free alongside hand washing education, and the overall average effect size was larger than in the two trials which did not provide soap (soap provided: rate ratio 0.66, 95% CI 0.56 to 0.78; six trials, 11,422 participants; education only: rate ratio: 0.84, 95% CI 0.67 to 1.05; two trials, 3304 participants). There was increased hand washing at major prompts (before eating/cooking, after visiting the toilet or cleaning the baby's bottom), and increased compliance to hand hygiene procedure (behavioural outcome) in the intervention groups than the control in community trials (data not pooled: three trials, 3490 participants, high quality evidence).
Hand washing promotion for the one trial conducted in a hospital among high-risk population showed significant reduction in mean episodes of diarrhoea (1.68 fewer) in the intervention group (Mean difference 1.68, 95% CI 1.93 to 1.43; one trial, 148 participants, moderate quality evidence). There was increase in hand washing frequency, seven times per day in the intervention group versus three times in the control in this hospital trial (one trial, 148 participants, moderate quality evidence).
We found no trials evaluating or reporting the effects of hand washing promotions on diarrhoea-related deaths, all-cause-under five mortality, or costs.
Hand washing promotion probably reduces diarrhoea episodes in both child day-care centres in high-income countries and among communities living in LMICs by about 30%. However, less is known about how to help people maintain hand washing habits in the longer term.
This Cochrane Review summarises trials evaluating the effects of promoting hand washing on the incidence of diarrhoea among children and adults in day-care centres, schools, communities, or hospitals. After searching for relevant trials up to 27 May 2015, we included 22 randomized controlled trials conducted in both high-income countries (HICs) and low- and middle-income countries (LMICs). These trials enrolled 69,309 children and 148 adults.
How does hand washing prevent diarrhoea and how might hand washing be promoted
Diarrhoea causes many deaths in children below five years of age, mostly in LMICs. The organisms causing diarrhoea are transmitted from person to person through food and water contaminated with faeces, or through person-to-person contact. Hand washing after defecation, or after cleaning a baby's bottom, and before preparing and eating food, can therefore reduce the risk of diarrhoea. Hand washing can be promoted through group or individual training on hygiene education, germ-health awareness, use of posters, leaflets, comic books, songs, and drama.
What this review says
Hand washing promotion at child day-care facilities or schools in HICs probably prevents around 30% of diarrhoea episodes (high quality evidence), and may prevent a similar proportion in schools in LMICs (low quality evidence). Among communities in LMICs hand washing promotion prevents around 28% of diarrhoea episodes (moderate quality evidence). In the only hospital-based trial included in this review, hand washing promotion also had important reduction in the mean episodes of diarrhoea (moderate quality evidence). This is based on only a single trial with few participants and thus there is need for more trials to confirm this. Effects of hand washing promotion on related hand hygiene behaviour changes improved more in the intervention groups than in the control in all the settings (low to high quality evidence). None of the included trials assessed the effect of handwashing promotion on diarrhoeal-related deaths, all-cause under-five mortality, or the cost-effectiveness of hand washing promotions.
Hand washing promotion in HICs and LMICs settings may reduce incidence of diarrhoea by about 30%. However, less is known about how to help people maintain hand washing habits in the longer term.
Diarrhoea is a serious global public health problem, accounting for 1.8 million deaths annually especially among children under five years of age (Walker 2013). The yearly global diarrhoeal disease burden is estimated at 72.8 million disability adjusted life years (DALYs) lost through incapacitation and premature deaths, mainly in low- and middle-income countries (LMICs) (Murray 2012).
Diarrhoea contributes significantly to malnutrition in children through a combination of forced low-nutrient intake, reduced absorption, and increased nutrient excretion (WHO 2003). The malnutrition-infection complex is clearly reinforced during diarrhoea episodes, as poor nutritional status predisposes children to more severe and persistent diarrhoea, impaired growth and development, and higher case fatality rates (UNICEF/WHO 2009; Lee 2012).
Diarrhoeal disease pathogens are usually transmitted through the faecal-oral route (Curtis 2000). The pathways include ingestion of food and water contaminated by faecal matter, person-to-person contact, or direct contact with infected faeces (Eisenberg 2012). Some trials estimate that over 75% of all diarrhoea cases can be attributed to contaminated food and water (Curtis 2000; Maxwell 2012). Poor hygiene behaviours and improper handling practices of caregivers are associated with high levels of bacterial contamination of food and water (Iroegbu 2000; Mannan 2010; Pickering 2011).
Behaviours that encourage human contact with faecal matter include: improper disposal of faeces; children defecating on the floor; rags being used to cleanse the child after defecation; and lack of hand washing after defecation, handling faeces (including children's faeces) or cleansing the child's perineum and before handling food by caregivers and children (Pickering 2011). In particular, hand contact with ready-to-eat food (that is, food consumed without further washing, cooking, or processing/preparation by the consumer) represents a potentially important mechanism by which diarrhoea-causing pathogens contaminate food and water (UNICEF/WHO 2009). In addition, flies may serve as vectors of diarrhoea-causing pathogens to humans. Thus, consumption of food exposed to flies is associated with high risk of diarrhoea (Marino 2007).
Household economic status is significantly associated with diarrhoea prevalence (Woldemicael 2001), especially in low-income countries. Households may lack basic infrastructure for proper hygiene practices, such as facilities for proper disposal of excreta. In addition, even where available, these may not be adapted for children's use (Tumwine 2002; UNICEF/WHO 2009). This often leads to indiscriminate defecation in and around the premises, and to increased risk of excreta handling by mothers, caregivers, and children (Nielsen 2001). A trial in Eritrea found that the availability of a toilet facility in households was associated with a 27% reduction in the risk of diarrhoea (Woldemicael 2001). The same trial also found associations between the number of children living in the house and diarrhoea morbidity. In some cultures children's faeces are regarded as innocuous. For this reason adults may not wash their hands after handling children's faeces and may cleanse a child with their bare hands (Traore 1994; Curtis 2000). However, evidence suggests that children's faeces are equally as hazardous as adult faeces and may contain even higher concentrations of pathogens than those of adults due to the children's increased interactions with contaminated materials in their surroundings (Oketcho 2012).
Hygiene promotion interventions constitute one of a number of strategies identified by World Health Organization (WHO) for control of diarrhoea (UNICEF/WHO 2009). These constitute a range of activities aimed at encouraging individuals and communities to adopt safer practices within domestic and community settings to prevent hygiene-related diseases that lead to diarrhoea (WELL 1999; Ehiri 2001); hand washing is one such intervention. The practice of hand washing and the factors that influence hand washing behaviour among individuals in communities are complex (Whitby 2007); for example, washing hands with water only or with soap may be influenced by both knowledge of best practice and availability of water and soap (Curtis 2011). Also, hand washing may require infrastructural, cultural, and behavioural changes, which take time to develop, as well as substantial resources (for example, trained personnel, community organization, provision of water supply and soap) (Luby 2001a; UNICEF/WHO 2009). Consideration of the wide applicability and sustainability of hygiene interventions have recently come under critical review (Luby 2006 PAK; Ejemot-Nwadiaro 2008; Gould 2010; Curtis 2011; Huis 2012; Madhu 2012). For example, maintenance of the new hand washing behaviours that result from hand washing promotional interventions is vital in maximizing the associated potential health benefits. Apart from the challenges of sustaining new behaviour (hand washing) among the target communities, cost has been identified as a major factor that limits the sustainability of hand hygiene behaviour (Langford 2007 NPL; Hartinger 2010 PER). For example, to sustain the health benefits of newly acquired hand washing behaviours, it is also important that individuals and communities have access to resources that support hand washing, including water and soap. Thus, lack of access to hand washing resources may limit the potential impact of hand washing on health particularly for low-income households and communities.
Hand washing aims to decontaminate the hands and prevent cross transmission of diarrhoeal-causing pathogens (Ehiri 2001; Gurjeet 2013). Hand washing promotion employs direct approaches such as training and educating individuals or group of individuals about hygiene, diarrhoea transmission, the relationship between germs and health, demonstrating this relationship through leaflets, posters, drama, and songs (Whitby 2007; Curtis 2011). Washing hands with soap and water removes pathogens mechanically and may also chemically kill contaminating and colonizing flora, making hand washing more effective (Hugonnet 2000). Also washing hands with soap under running water or large quantities of water with vigorous rubbing was found to be more effective than several members of a household dipping their hands into the same bowl of water (often without soap) (Luby 2005), which is a common practice in many low-income countries, especially before household meals (Ehiri 2001). This may contribute to, rather than prevent, food contamination as pathogens present on contaminated hands of household members can be transferred to those who subsequently dip their hands in the same bowl of water (Prüss 2002).
Hand washing is a viable intervention in the control of diarrhoeal diseases. It is listed in the UNICEF/WHO 2009 seven-point plan for comprehensive control of diarrhoea. Hand washing requires infrastructural, cultural, and behavioural changes that take time and substantial resources to develop (Cave 1999; Yeager 1999; Luby 2001b). Given that resources spent on interventions to promote hand washing could be invested on other equally important public health programmes, it is important to ascertain that hand washing promotion is an efficient use of scarce health resources. In 2008, we published a review that assessed the broader question of the effectiveness of hand washing with soap in preventing diarrhoea as against other interventions such as provision of water, improvement of water quality (treatment of water), amongst randomized controlled trials (RCTs) (Ejemot-Nwadiaro 2008). A review by Curtis 2003, which examined the effectiveness of hand washing with soap in community-based trials, estimated that hand washing could reduce diarrhoea risk by up to 47%. Similarly, Fewtrell 2005 examined a range of water, sanitation, and hygiene interventions in LMICs, and estimated that hygiene interventions reduced diarrhoea incidence by 44%. However, both reviews included non-randomized trials. Curtis 2003 included cross-sectional trials which have inherent limitations with regard to establishment of causal relationships. Fewtrell 2005 presented evidence of publication bias in included trials. In this Cochrane Review, we assessed whether the estimate of effect observed only in RCTs is of similar magnitude to those seen in previous reviews and the applicability of hand washing interventions in reducing diarrhoeal diseases across wide population groups. We also included both institution-based and community-based trials in countries of any income level.
To assess the effects of hand washing promotion interventions on diarrhoeal episodes in children and adults.
Randomized controlled trials (RCTs), including cluster-RCTs.
Individuals (adults and children) in day-care centres or schools, patients in hospitals, communities, or households.
Activities that promoted hand washing after defecation or after disposal of children's faeces and before eating, preparing or handling foods; for example, small group discussions and larger meetings on hygiene education, germs-health awareness interventions, multimedia communication campaigns with posters, radio/TV campaigns, leaflets, comic books, songs, slide shows, use of T-shirts and badges, pictorial stories, dramas, and games. We included trials that focused exclusively on hand washing and those that had hand washing as part of a broader package of hygiene interventions if they undertook analyses of effects of hand washing on diarrhoea.
No hand washing promotion.
We defined diarrhoea as:
We attempted to identify all relevant trials regardless of language or publication status (published, unpublished, in press, and in progress).
We searched the following databases using the search terms and strategy described in Table 1: Cochrane Infectious Diseases Group Specialized Register (27 May 2015); Cochrane Central Register of Controlled Trials (CENTRAL), published in the Cochrane Library (2015, Issue 5); MEDLINE (1966 to 27 May 2015); EMBASE (1974 to 27 May 2015); and LILACS (1982 to 27 May 2015).
We also searched the following databases using diarrhoea, diarrhoea, and handwashing as search terms: PsycINFO (1967 to 27 May 2015); Science Citation Index and Social Sciences Citation Index (1981 to 27 May 2015); ERIC (Educational Resources Information Center; 1966 to 27 May 2015); SPECTR (The Campbell Collaboration's Social, Psychological, Educational, and Criminological Trials Register; 2000 to 27 May 2015); Bibliomap and TRoPHI (The Trials Register of Promoting Health Interventions) maintained by the Evidence for Policy and Practice Information and Co-ordinating Centre (www.eppi.ioe.ac.uk) (1990 to 27 May 2015); and The Grey Literature (www.nyam.org/library/grey.shtml; 2002 to 27 May 2015). We also searched the World Health Organization (WHO) International Clinical Trial Registry Platform (ICTRP) and the metaRegister of Controlled Trials (mRCT) for ongoing trials on 27 May 2015 using diarrhoea, diarrhoea, and handwashing as search terms. The PRISMA flow diagram is shown in Figure Figure11 below.
To obtain information on published, unpublished and ongoing trials, we contacted researchers in the field for unpublished and ongoing trials (October 2013).
We also examined reference lists of articles for relevant trials.
Three review authors (RIE, JC, and DA) independently screened titles and abstracts of relevant articles to assess their eligibility for inclusion in the review. We retrieved full-texts of articles that were deemed potentially relevant to the review for further assessment. Decision on inclusion was reached by consensus among all review authors. We scrutinized each trial report to ensure that we included multiple publications from the same trial only once. We listed the excluded trials and the reasons for their exclusion.
Three review authors (RIE, DA, and JC) independently extracted data on methods, types of participants, interventions, and outcomes from the selected trials using a standardized data extraction form. We resolved any disagreements by discussion and consensus among review authors. We requested unpublished data and additional information from published trials from relevant individuals, groups, and organizations.
We extracted the year of completion of the trial rather than the year of publication for identification of included trials. When such data were not reported we used the year of publication. In addition, we used a three-letter international code of the country were the trial was conducted. This was to give a clear time frame for the Cochrane Review (1977 to 2013). We extracted data on each trial site, including any measures of availability of water, soap, and literacy level of the communities. Where data were available, we extracted the socioeconomic status of trial participants since resources for effective hand washing (for example, running water and soap) may be more accessible to higher income households. We carefully summarized details of the intervention including: type of promotional activity, whether soap and water provision was part of the intervention, method of hand washing promoted (washing in a bowl or under running water), and procedure for hand washing.
We had intended to analyse episodes of diarrhoea as a dichotomous outcome, but the data reported by the trials did not permit this type of analysis. We analysed the outcome as count data, when either the incidence rate ratio and 95% confidence intervals (CIs), or the number of episodes of diarrhoea and the person-time at risk was reported; or as continuous data when the mean number of diarrhoea episodes and standard deviation (SD) were presented.
For individually RCTs, when continuous outcomes data were summarized as arithmetic means, we extracted the arithmetic means, SDs, and numbers of participants for the treatment and control groups. For count (rate) outcome data we extracted the number of episodes, the number of person-years at risk, and the number of participants for each intervention group, or we extracted a rate ratio and measure of variation (for example, CI) directly from the publication.
Cluster-RCTs require the use of different data extraction methods and analysis methods because trials with a cluster design require more complex analysis than trials that randomized individuals. Observations on participants in the same cluster tend to be correlated; therefore the intra-cluster variation must be accounted for during the analysis of the trial. If this correlation is ignored in the analysis and the same techniques are employed as for individually RCTs the resulting measure of effect remains a valid estimate, but the associated variance of the estimate will be underestimated leading to unduly narrow CIs. For meta-analysis this means that trials analysed without allowing for this design effect will receive too much weight.
For the cluster-RCTs, we extracted information on the number of clusters, average size of clusters, unit of randomization, whether the trials adjusted for clustering, and the statistical method used to analyse cluster trials. When a trial's analysis had adjusted for clustering, we extracted the point estimate and 95% CI. For count data we extracted the incidence rate ratio. If a trial had not adjusted for clustering, we extracted the same data as for the individually RCTs.
Two review authors (RIE and DA) independently assessed the risk of bias in included trials using the Cochrane 'Risk of bias' assessment tool (Higgins 2011). We assessed the risk of bias across the following domains: randomization sequence generation, allocation concealment, blinding, incomplete outcome data, selective reporting, and other biases. We classified our judgements as 'high', 'unclear' or 'low' risk of bias using criteria described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
In the blinding domain we acknowledged that double blinding is not possible in trials of hand washing interventions since there is no obvious placebo. However, outcome assessors could be blinded, and we assessed whether or not this had occurred. It is also difficult to assess losses to follow-up (incomplete outcome data) in open cluster-RCTs. Some adults and children may leave the trial, but others are born or enter the trial during the follow-up period; hence participant numbers are in constant flux. Inclusion of all randomized participants in the analysis is thus most clearly represented as the person-time at risk accrued as a percentage of maximum possible person-time at risk in each trial arm. Therefore, we reported on this measure and also on any loss to follow-up of both clusters and participants, and assessed this as low risk if at least 90%. We also assessed whether baseline characteristics were comparable across the intervention groups and assessed whether data was collected at similar time points for the intervention and control sites with a view to identifying selective reporting and other possible biases. The details are shown in Figure Figure22 and Figure Figure33.
We used the GRADE approach to assess the quality of evidence and interpret our findings. We imported data from Review Manager (RevMan) to GRADEpro 2014 to create a 'Summary of findings' table containing relevant information on the outcomes of interest. We then proceeded to downgrade the quality of evidence (if necessary) for each outcome across the following domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias for each trial that contributed to the outcome. We downgraded the evidence for each outcome by one level (for serious limitations), two levels (for very serious limitations), or left it at 'no limitations' when we found no reason to downgrade.
We included the pre-specified outcomes for the three independent settings in Summary of findings for the main comparison, Summary of findings 2, and Summary of findings 3.
We qualitatively compared included trials to ascertain the feasibility of pooling them together in a meta-analysis. Thus we identified three distinct settings which included: child day-care centres, community-based interventions, and hospital based trials; since the factors that affect hand washing practice may vary in these settings. We stratified the trials based on these settings for the meta-analysis and calculated risk ratios (RR) for dichotomous outcomes, mean difference (MD) for continuous outcomes measure on the same scale, and standardized mean difference (SMD) for continuous outcomes measured using different scales.
For all trials that did not adjust for clustering, we made approximate adjustments for clustering using estimates of the intra-cluster correlation coefficient (ICC) from other trials that did adjust for clustering and reported this statistic. We did this by multiplying the standard error for each trial by the square root of the design effect. We estimated the design effect as 1+(m-1)*ICC, where 'm' is the average cluster size and 'ICC' is the intra-cluster correlation coefficient (Higgins 2011).
We contacted authors of eligible trials for missing data or for additional information when the trials were less than 15 years old.
We checked for heterogeneity by visually inspecting the forest plots, applying the Chi² test with a P value of 0.10 indicating statistical significance, and also implementing the I² test statistic with a value of 50% used to denote moderate levels of heterogeneity. We used the random-effects model to pool data if we detected heterogeneity and it was still considered clinically meaningful to combine the trials. Due to the limited number of trials in each setting we were unable to explore potential sources of heterogeneity in depth. We explored and attempted to explain heterogeneity where possible using a pre-defined trial characteristic (provision of hand washing material (soap) as part of intervention, and type of promotional activity employed) and quality characteristics (whether outcome assessors were blind and whether trials had adjusted for clustering).
We assessed the possibility of publication bias by plotting a funnel plot if at least ten trials contributed to the treatment comparison. However, we did not undertake this due to an insufficient number of included trials.
We analysed the data using Review Manager (RevMan) and presented all results with 95% CIs. We stratified the analysis into three categories of settings – child day-care centres and school-based interventions (day-care centres or primary schools), community-based interventions, and intervention in people at high risk of diarrhoea (people with acquired immune deficiency syndrome (AIDS)). Also we stratified the analyses by the income status of the countries where the trials were conducted. Since the outcomes and methods of measuring behaviour changes were too variable to make meta-analysis meaningful, we tabulated the results.
We summarized continuous outcome data from individually RCTs using the MD value. Meta-analysis of individually RCTs was not undertaken due to the limited number of individually RCTs.
For count outcomes, we pooled incidence rate ratios (IRR) in Review Manager (RevMan) using the generic inverse variance method with the random-effects model. We used standard techniques for calculating standard errors from 95% CIs (Higgins 2008). When the outcomes and methods of measuring outcomes were too variable to make meta-analysis meaningful (for changes in hand washing behaviour) we tabulated the results. One trial performed child and site-level analyses (Haggerty 1988 COD); the 95% CIs were not provided for the site-level analysis. We therefore estimated the denominator from the number of children by trial arm by assuming that all those who had remained in the trial for at least nine weeks had a total of 12 weeks of follow-up. The numerator (average number of episodes per child) was provided at the cluster level. We classified this trial as cluster adjusted. One trial, Luby 2006 PAK, presented mean longitudinal prevalence of diarrhoea without presenting data on incidence of diarrhoea and hence we could not include it in the meta-analyses.
For trials that did not report on or were unclear on the method used to adjust for clustering, we either extracted information on the rate ratio and unadjusted 95% CI or, wherever possible, estimated the unadjusted rate ratios and 95% CIs from the total number of diarrhoea episodes and person-time at risk in each trial arm. Where data on person-time at risk were not directly provided by the trial authors, we estimated this as accurately as possible from the follow-up duration multiplied by the total number of children as the denominator for both intervention and control groups respectively. The measures of effect and CIs are presented in tables. One trial adjusted for clustering by comparing the mean incidence rate of intervention and non-intervention classrooms (Kotch 1989 USA), but only a cluster-adjusted 95% CIs for a different outcome (excess mean episodes) and not a rate ratio was presented. We took the cluster-adjusted estimate of the numerator (the mean incidence rate across the clusters) from the published data and estimated the person-time at risk crudely by multiplying the number of contacts every two weeks by the number of children and assuming this was equally distributed between the intervention and control groups. We classified this trial as not having adjusted for clustering
For all trials that did not adjust for clustering, we attempted to make an approximate adjustment using estimates of the ICC from one of the trials that did adjust for clustering and reported this statistic. Only two trials reported this statistic: one community-based trial, Luby 2003b PAK, and one trial in a child day-care centre, Roberts 1996 AUS. We assumed that these ICC estimates could be generalized to other community-based and child day-care centre or school-based trials respectively. We extracted the number of children and number of clusters from each unadjusted trial to estimate the average cluster size. We then followed standard methods (Higgins 2011) to estimate the design effect for each trial and multiplied the standard error for each trial by the square root of this design effect. This approximate adjustment increases the standard error (and hence width of CIs for the unadjusted trials) and appropriately reduces the weight given to such trial in the meta-analysis. We performed meta-analyses by pooling the estimates of the cluster adjusted and approximately adjusted trials together.
We planned to explore the possible causes of heterogeneity if we detected any using subgroup analysis. The subgroups used were: trial setting, provision of hand washing material (soap) as part of intervention, type of promotional activity employed), and quality characteristics (whether outcome assessors were blinded).
We undertook a sensitivity analysis to explore the robustness of our findings, including the trial size, duration of follow-up, differences in method of assessing the primary outcome, and differences in methodological quality (blinding of outcome assessors) of the included trials.
Our search yielded 47 additional potentially relevant trials, making a total of 84 when combined with the 37 search results of the original review (Ejemot-Nwadiaro 2008). In total, 22 trials met the inclusion criteria: 14 trials were included in the previous version of the review, Ejemot-Nwadiaro 2008, and we included eight new trials based on the updated search. We have described them in the 'Characteristics of included studies' section. One trial was in Danish (Ladegaard 1999 DEN), and the rest were written in English. Twelve trials were child day-care centres or school-based, nine trials were community-based, and one trial (Huang 2007 USA) was in a high-risk group. We have listed reasons for excluding 62 trials in the 'Characteristics of excluded studies' table.
All 12 trials in this group were randomized by cluster using primary schools (Bowen 2004 CHN; Talaat 2008 EGY; Pickering 2013 KEN), day-care centres (Black 1977 USA; Bartlett 1984 USA; Butz 1990 USA; Roberts 1996 AUS; Carabin 1997 CAN; Ladegaard 1999 DEN; Kotch 2003 USA; Zomer 2012 NED ), or classrooms in day-care centres (Kotch 1989 USA) as the unit of randomization. These trials were all conducted in high-income countries except for three trials conducted in LMICs Bowen 2004 CHN, (which was undertaken in Fujian province in China) Talaat 2008 EGY (which was conducted in Cairo, Egypt), and Pickering 2013 KEN (conducted in Nairobi, Kenya). The others trials were performed in Australia (Roberts 1996 AUS), Europe (Ladegaard 1999 DEN; Zomer 2012 NED), and North America (Black 1977 USA; Bartlett 1984 USA; Kotch 1989 USA; Butz 1990 USA; Carabin 1997 CAN; Kotch 2003 USA), where resources and materials for hand washing were relatively available and accessible.
All trials used multiple hygiene interventions, except Black 1977 USA, Bowen 2004 CHN, and Pickering 2013 KEN which used only a hand washing intervention. Though Pickering 2013 KEN was a three-arm trial that investigated hand sanitizer and hand washing with soap, we only considered the arm of hand washing with soap in this Cochrane Review, as such it is categorized as a hand washing only intervention. Kotch 2003 USA assessed the impact of provision of hand washing and diapering equipment on incidence and duration of infectious illness (including diarrhoea) in both children and staff. We have described the interventions in more detail in Table 2.
All but one of the included trials in child day-care centres or schools institution-based trials had intervention and control arms (monitoring only). Bowen 2004 CHN had three arms for the standard intervention, expanded intervention (which included the standard intervention and peer-monitoring of hand-washing), and control. It is important to note that the control group in most cases received quite frequent monitoring (estimating diarrhoea illness episodes on typically two-weekly basis). This monitoring may itself have influenced hand washing behaviour. The Carabin 1997 CAN trial attempted to tease out the effects of the intervention alone from 'monitoring'. The 'monitoring' effect in this trial was estimated as the difference in diarrhoea incidence rates within each arm over one year of the trial (September 1996 to November 1997). The crude effectiveness of intervention was estimated as the difference between the monitoring effect in the intervention group.
Twelve trials including 54,006 children met the inclusion criteria. Seven trials included children aged less than three years, one trial was in children under six years (Ladegaard 1999 DEN), and one trial was with children aged less than seven years (Butz 1990 USA). Bowen 2004 CHN involved children in the first grade at school in China; Talaat 2008 EGY included children in government elementary schools in Cairo, Egypt; and Pickering 2013 KEN involved children aged between five to 10 years in primary schools in Nairobi, Kenya. Hand washing behavioural changes and changes in knowledge, attitude, and belief on hygiene were assessed in the day-care providers (number not strictly reported) and children, while the primary outcome measures were assessed in the children.
The number of clusters ranged from four (Black 1977 USA) to 87 (Bowen 2004 CHN). Primary outcome measures were assessed across 278 day-care centres and 151 schools. Participants were exposed to mainly small and large group training sessions on hygiene education and germs-health theory, that employed multiple promotional techniques (for example, audio and video tapes, pamphlets, practical demonstrations, drama, posters, songs, games, or peer monitoring). Kotch 2003 USA employed the 'Keep-it-clean' module in training caregivers to standardise the interventions across the trial arms. The aim was to provide education about personal hygiene, diarrhoea transmission, treatment, and prevention, and the importance of techniques for hand washing. Intervention and control groups were generally comparable regarding important characteristics at baseline (Table 2).
All included trials measured the primary outcome, episodes of diarrhoea. Three trials reported proportion of people washing their hands and or changes in knowledge, attitude, and beliefs about hand washing (Kotch 1989 USA; Roberts 1996 AUS; Pickering 2013 KEN). No trials reported diarrhoea-related deaths, all-cause-under five mortality or cost-effectiveness data. However, Kotch 2003 USA reported that the cost of purchasing and installing one unit of the hand washing and diapering equipment was quite exorbitant at USD10,385 (USD7500 for the equipment and the rest for installation per classroom). Follow-up periods ranged from two to 12 months.
Five trials did not appear to have accounted for clustering in the analysis for any outcome measure (Black 1977 USA; Bartlett 1984 USA; Butz 1990 USA; Ladegaard 1999 DEN; Talaat 2008 EGY). Kotch 1989 USA adjusted for clustering by comparing the mean incidence rate of intervention and non-intervention classrooms, but only a cluster adjusted 95% CI for a difference outcome (excess mean episodes) and not a rate ratio was presented. Kotch 2003 USA reported controlling for clustering by estimating a random effect for the centres, but this does not seem to have been reflected in the results. In the other five cluster-adjusted trials, Bowen 2004 CHN presented only the school level analysis (mean illness and absence rates by school); Carabin 1997 CAN adjusted for clustering using a Bayesian hierarchical model, while Roberts 1996 AUS, Zomer 2012 NED and Pickering 2013 KEN estimated robust standard errors in a Poisson regression model.
We included nine cluster-RCTs that used entire communities (generally villages, squatter settlements, or neighbourhoods, except Han 1985 MMR, which used households) as units of randomization. These trials were conducted in LMICs in Africa (Haggerty 1988 COD), Asia (Han 1985 MMR; Stanton 1985 BGD; Luby 2003a PAK; Luby 2003b PAK; Luby 2006 PAK; Langford 2007 NPL; Nicholson 2008 IND), and South America (Hartinger 2010 PER).
Five trials evaluated hand washing only interventions (Han 1985 MMR; Luby 2003a PAK; Luby 2003b PAK; Langford 2007 NPL; Nicholson 2008 IND). Luby 2003a PAK had two hand washing arms, one with plain soap and one with antibacterial soap. These two arms had similar results and are combined in this Cochrane Review. Han 1985 MMR used plain soap. Luby 2003b PAK was a five-arm trial that investigated water quality interventions, hand washing, and a combination of the two; only the arm with antibacterial soap and hand washing education is considered in this review. Luby 2006 PAK conducted a follow-up trial to the Luby 2003b PAK trial, maintaining the initial randomization process to assess if learnt hygiene behaviours could be sustained over time without additional hygiene promotion intervention. Three other trials used multiple hygiene interventions that included hand washing with soap (the type of soap used is not described) (Stanton 1985 BGD; Haggerty 1988 COD; Hartinger 2010 PER). We have provided more detailed descriptions of the interventions in Table 3.
We included nine trials with 15,303 children. In the community-based trials, three trials were with very young children (< three years) (Haggerty 1988 COD; Langford 2007 NPL; Hartinger 2010 PER); two other trials were with children aged less than five years (Han 1985 MMR) or less than six years (Stanton 1985 BGD); and three involved older children up to 15 years of age (Luby 2003a PAK; Luby 2003b PAK; Luby 2006 PAK). Nicholson 2008 IND had four categories of participants: targeted children five years old, children less than five years old, children six to 15 years old, and adults in the families. The primary outcome measure (incidence of diarrhoea) was assessed in each of these categories with their corresponding control groups except for the adults reported as the 'whole family'. In this Cochrane Review we considered results from only the target group as the first three categories had similar effect size. Hand washing behavioural changes and changes in knowledge, attitude, and belief on hygiene were assessed in the mothers (number not strictly reported), while the primary outcome measures were assessed in the children.
The number of clusters varied from 18 (Haggerty 1988 COD) to 1923 (Stanton 1985 BGD). The participants were provided with hand washing materials and were involved in large-group hygiene education training, except for Luby 2006 PAK which was a follow-up trial. The intervention and control groups were socioeconomically comparable at baseline.
All included trials measured diarrhoea episodes except for Luby 2006 PAK, which measured mean longitudinal prevalence of diarrhoea; some trials also assessed different types of diarrhoea. Han 1985 MMR measured dysentery rates, and Luby 2003a PAK and Luby 2003b PAK also assessed the rate of persistent diarrhoea. None of the included trials reported diarrhoea-related deaths, all-cause-under five mortality, nor cost-effectiveness data. Langford 2007 NPL reported changes in hand washing from baseline to endline at hand washing junctures, Stanton 1985 BGD reported on changes in hand washing behaviour, while Nicholson 2008 IND reported it using soap wrapper collected as a measure of soap consumption as an indirect measure. Length of follow-up ranged from four to 12 months.
All trials adjusted for clustering in some way, except for Han 1985 MMR, Langford 2007 NPL, Nicholson 2008 IND, and Hartinger 2010 PER. Stanton 1985 BGD and Luby 2003a PAK adjusted for clustering by estimating rates at the group level; Luby 2003b PAK adjusted for clustering by calculating an ICC based on an analysis of variance level and design effect. Luby 2006 PAK though measured mean longitudinal prevalence of diarrhoea accounted for clustering using generalized estimating equations. Haggerty 1988 COD performed child and site level analyses; the 95% CIs were not provided for the site-level analysis. The numerator (average number of episodes per child) was provided at the cluster level.
We identified only one trial in a high-risk group (Huang 2007 USA). It individually randomized 148 adults with AIDS from one human immunodeficiency virus (HIV) clinic in the USA to receive intensive hand washing promotion delivered by specialist nurses (Huang 2007 USA). The intervention included hygiene education, hand washing demonstrations by nurses and participants, and weekly telephone calls to reinforce hand washing messages Table 4. The major outcomes reported were mean episodes of diarrhoea in each group and number of hand washing episodes per day. They reported the mean hand washing frequency per day at baseline and at the end of the intervention (Table 5).
We have listed the excluded trials and the reasons for exclusion in the 'Characteristics of excluded studies' section.
Five of the 12 trials used an adequate method to generate the allocation sequence (Roberts 1996 AUS; Carabin 1997 CAN; Bowen 2004 CHN; Talaat 2008 EGY; Zomer 2012 NED); the method was unclear in the others. The method used to conceal allocation was unclear in all trials. In cluster-RCTs, lack of concealment of allocation is not considered a major risk of bias since all clusters are usually randomized at the same time (Higgins 2011, Section 16.3.2).
Three trials reported blinding of the outcome assessors (Bartlett 1984 USA; Kotch 1989 USA; Roberts 1996 AUS); the rest were open trials. It was difficult to assess the number of randomized participants included in the analysis as this was reported at different levels (cluster, child, person time-at-risk). However, all trials were able to account for the number of randomized clusters included in the analysis.
Seven trials reported adequate comparability between the intervention and control groups with respect to diarrhoea incidence and sociodemographic characteristics (including mean total enrolment, percentage of drop outs, sex, age, and race composition of children enrolled, diapering, and toilet facilities) at baseline (Black 1977 USA; Bartlett 1984 USA; Butz 1990 USA; Ladegaard 1999 DEN; Bowen 2004 CHN; Talaat 2008 EGY; Pickering 2013 KEN). Investigators in Bowen 2004 CHN were forced to over- or under-sample certain regions to obtain more 'control' schools after the original control schools were sent intervention packs by mistake and thus excluded. This trial reported small differences in household sanitation and piped water at baseline, but no differences between schools in number of students, class size, or hygiene infrastructure. Comparability at baseline was unclear in the two other trials (Kotch 1989 USA; Roberts 1996 AUS), while it was considered inadequate in two trials; Kotch 2003 USA reported baseline differences in total number of children and boys in favour of the intervention which they believed may have influenced the outcome measure and Zomer 2012 NED acknowledged baseline imbalance in crude incidence diarrhoeal episodes per child-year of 3.0 for intervention versus 5.1 for the control but they applied statistical adjustments for this baseline characteristic. All trials reported collecting data at the same point in time for both the intervention and control groups.
Seven included trials reported adequate methods for generating allocation sequence (Stanton 1985 BGD; Luby 2003a PAK; Luby 2003b PAK; Luby 2006 PAK; Langford 2007 NPL; Nicholson 2008 IND; Hartinger 2010 PER). Only Luby 2003a PAK reported adequate allocation concealment; it was unclear in the other trials. Han 1985 MMR, Haggerty 1988 COD, Langford 2007 NPL and Hartinger 2010 PER reported blinding of outcome assessors, and the rest were open trials. Inclusion of all randomized participants in the analysis was unclear as it was reported at different levels of analysis (cluster, household, child) except for Nicholson 2008 IND, which reported 18% average attrition bias for all the subgroups in both arms.
Eight trials reported baseline similarity of diarrhoea morbidity and socioeconomic characteristics (including population/household size, socioeconomic status, hand washing and sanitary facilities, and sources of water supply) between the intervention and control groups (Han 1985 MMR; Stanton 1985 BGD; Luby 2003a PAK; Luby 2003b PAK; Luby 2006 PAK; Langford 2007 NPL; Nicholson 2008 IND; Hartinger 2010 PER). There were some differences at baseline in Haggerty 1988 COD (controls had diarrhoea episodes of longer duration than the intervention group). All the trials reported collecting data at the same period for intervention and control groups.
Huang 2007 USA did not clearly report the method of randomization or allocation concealment and did not use blinding. All 148 randomized participants were followed for the trial's one-year duration. Participants were similar at the start of the trial in terms of age, sex, ethnicity, hand washing episodes per day, CD4 count, HIV load, and prophylaxis for opportunistic infections. The results were presented as a continuous outcome only (mean and SD of number of diarrhoea episodes in each arm over the year). This should be viewed with caution as it is likely that the distribution of diarrhoea episodes may be highly skewed (the mean of 1.24 and SD of 0.9 episodes in the intervention arm imply a non-normal distribution of diarrhoea episodes). If so, the mean may not be the most appropriate measure of the 'average number' of episodes per participant. The trial reported collecting data at the same period for intervention and control groups.
See: Summary of findings for the main comparison Summary of findings table 1; Summary of findings 2 Summary of findings table 2; Summary of findings 3 Summary of findings table 3
We have presented the results as reported by each trial in Table 5 (behavioural change), Table 6, Table 7, Table 8 (incidence of diarrhoea), Table 9, and Table 10. For trials with cluster-adjusted results or where trials have been individually randomized, the data are summarized in forest plots. For trials where this was not possible, we have summarized the data in tables in the 'Data and analyses' section.
Overall, hand washing promotion reduced diarrhoea episodes by about a third (IRR 0.69, 95% CI 0.59 to 0.81; 11 trials, 50,044 children (Bowen 2004 CHN not included in analysis); Analysis 1.1). Most data were from high income countries (IRR 0.70, 95% CI 0.58 to 0.85; nine trials, 4664 participants; high quality evidence; Analysis 1.1), with only two trials from LMICs (IRR 0.66, 95% CI 0.43 to 0.99; two trials, 45,380 participants; low quality evidence; Analysis 1.1).
All trials showed a benefit from the intervention, except for Bowen 2004 CHN which showed no difference between each arm and for which it was not possible to calculate a rate ratio (the median episodes of diarrhoea were 0 per 100 student-weeks in the control group, standard intervention group, and expanded intervention) (Table 6). Roberts 1996 AUS showed greater risk reduction than other trials (IRR 0.50, 95% CI 0.36 to 0.69; one trial, 558 participants), possibly due to a more specific technique of hand washing used (an approximate "count to 10" to wash and "count to 10" to rinse).
All participants were monitored at least every two weeks to collect data on diarrhoea episodes. This monitoring itself may have helped to improve compliance with hand washing. Only Carabin 1997 CAN attempted to investigate this effect by assessing rates in both groups compared to the pre-intervention period. They found that monitoring alone appeared to reduce the incidence of diarrhoea (IRR 0.73, 95% CI 0.54 to 0.97; Table 6), and that the intervention effect did not appear to have any benefits over and above this monitoring effect when adjusted for age and gender (IRR 0.77, 95% CI 0.51 to 1.18; Table 6) or when adjusted for age, gender, season, and baseline incidence rate in each cluster (IRR 1.10, 95% CI 0.81 to 1.50; Table 6). However, monitoring was particularly frequent (daily) in this trial. In the Bowen 2004 CHN trial among first grade students in schools in China, monitoring may have been less intensive as in-class monitoring was carried out one day a week by teachers; reasons for absenteeism were noted when recorded. As the trial was school-based, no illness information was collected during weekends or school holidays. This design reduced the burden of data collection of teachers, but it may also have reduced the ability of the trial to detect differences in the incidence of diarrhoea between each trial arm.
Two trials, Black 1977 USA and Pickering 2013 KEN, focused only on hand washing intervention and there was no significant difference the effect estimate (IRR 0.69; 95% CI 0.43 to 1.09; two trials, 1045 participants). Nine trials (Bartlett 1984 USA; Kotch 1989 USA; Butz 1990 USA; Roberts 1996 AUS; Carabin 1997 CAN; Ladegaard 1999 DEN; Kotch 2003 USA; Talaat 2008 EGY; Zomer 2012 NED) involved multiple hygiene interventions (IRR 0.69; 95% CI 0.57 to 0.84; nine trials, 48,999 participants; Analysis 1.2). The implication of this aspect of hand hygiene interventions should be further investigated as we had few trials in each category to make a statement.
Three trials (Bartlett 1984 USA; Kotch 1989 USA; Roberts 1996 AUS) attempted blinding (of outcome assessors) and the benefit of hand washing seemed to be less, 26% reduction (IRR 0.74, 95% CI 0.56 to 0.98; three trials, 1303 participants), than in the other trials that did not blind outcome assessors (Black 1977 USA; Butz 1990 USA; Carabin 1997 CAN; Ladegaard 1999 DEN; Kotch 2003 USA; Talaat 2008 EGY; Zomer 2012 NED; Pickering 2013 KEN), 33% reduction (IRR 0.67, 95% CI 0.56 to 0.80; eight trials, 48,741 participants; Analysis 1.3).
Four trials reported measures of behavioural change (Kotch 1989 USA; Roberts 1996 AUS; Zomer 2012 NED; Pickering 2013 KEN). As described in Table 9, Kotch 1989 USA reported that hand washing behaviour based on 'event sampling scores' improved in the intervention classrooms compared with control classrooms. Roberts 1996 AUS reported that the intervention improved compliance with infection control procedures from 53% at baseline to = 80% at endline. This was associated with a lower illness incidence in children aged ≥ two years (RR 0.34, 95% CI 0.17 to 0.65), reflecting a two-third reduction in diarrhoeal episodes. Zomer 2012 NED reported significant increase in hand hygiene compliance for caregivers in intervention DCCs than in control but this did not seem to have effect on incidence of episodes of diarrhoea. Pickering 2013 KEN reported a statistically significant rate of hand washing with soap at intervention schools: 37% against 2% for the control for all toilet events (prevalence ratio 17.2; 95% CI 4.4 to 67.5), while the mean proportion (intervention: 0.70; control: 0.01) of students hand washing with soap before lunch events was equally significantly different between schools (prevalence ratio 143.0, 95% CI 38.9 to 525.6) (data not pooled; three trials, 1845 participants, low quality evidence; Table 9).
Overall, community based hand washing promotion reduced the incidence of diarrhoea by around a quarter (IRR 0.72, 95% CI 0.62 to 0.83; eight trials; 14,762 participants; high quality evidence; Analysis 2.1). Luby 2006 PAK reported mean longitudinal prevalence of diarrhoea for all children under observation with no apparent benefit of the intervention (Analysis 2.2). All the trials were conducted in LMICs; with six from Asia, one from South America, and one from Africa.
Three trials assessed the effect of intervention on the incidence rate of different categories of diarrhoea (Han 1985 MMR; Luby 2003a PAK; Luby 2003b PAK). Han 1985 MMR reported on dysentery, and Luby 2003a PAK and Luby 2003b PAK reported on persistent diarrhoea. None of the results were statistically significant (Table 6). Some trials reported the results by participant age (Han 1985 MMR; Stanton 1985 BGD; Luby 2003a PAK; Luby 2003b PAK; Nicholson 2008 IND), with no discernible trend of which age group intervention had greater diarrhoeal reductions (Table 6). Han 1985 MMR and Stanton 1985 BGD reported greater diarrhoeal reduction for children aged less than two years, while Luby 2003a PAK and Luby 2003b PAK reported greater reductions for older children. For Nicholson 2008 IND, the effect for the different age groups (five years old, less than five years, and six to 15 years) were similar.
Five trials (Han 1985 MMR; Luby 2003a PAK; Luby 2003b PAK; Langford 2007 NPL; Nicholson 2008 IND) promoted hand washing only while three trials promoted multiple hygiene interventions (Stanton 1985 BGD; Haggerty 1988 COD; Hartinger 2010 PER). The reduction in the risk of diarrhoea was greater in the trials that promoted hand washing only (IRR 0.63, 95% CI 0.52 to 0.78; 10,888 participants ) than in the trials that promoted multiple hygiene interventions (IRR 0.81, 95% CI 0.69 to 0.95; three trials, 3838 participants; Analysis 2.3). This aspect of hand hygiene interventions should be interpreted with caution as we had few trials in each category to make strong statement.
Four trials attempted blinding of outcome assessors and the benefit of hand washing appeared to be lower than in trials which did not blind outcome assessors (IRR 0.80, 95% CI 0.67 to 0.94; four trials, 3070 participants; versus IRR 0.63, 95% CI 0.48 to 0.83; four trials, 11,656 participants; Analysis 2.4).
Six trials provided soap free alongside hand hygiene promotional activities and the effect seemed to be larger in these trials than in those which did not provide soap (IRR 0.66, 95% CI 0.56 to 0.78; six trials, 11,422 participants; versus IRR 0.84, 95% CI 0.67 to 1.05; two trials, 3304 participants; Analysis 2.5).
With only a small number of trials, these differences may be due to chance or, even if real, it is difficult to discern which components (providing soap or focusing on hand washing only) may be most effective.
Stanton 1985 BGD adjusted for clustering and reported that the intervention group exhibited a greater increase in hygiene practices (IRR 1.48, 95% CI 1.01 to 2.21), though this increase is of borderline statistical significance (P = 0.056; Table 10). Langford 2007 NPL reports that at the end of the intervention, reported hand washing after cleaning the baby's bottom or before cooking, eating, or feeding the baby had increased in mothers from the intervention areas (McNemar's test, P < 0.01 for all four junctures), while hand washing practices remained unchanged in the control areas. Nicholson 2008 IND measured hand washing behaviour between trial groups indirectly by assessing soap consumption (soap wrapper collection) and reported median soap consumption per household per week of 235g for intervention households as against 45g for the controls (data not pooled; three trials, 3490 participants, high quality evidence; Table 10).
In Huang 2007 USA, the intensive hand washing intervention reduced the mean number of episodes of diarrhoea over the one-year period of trial (2.92 in control group; 1.24 in intervention group; a reduction of 1.68 episodes, 95% CI -1.93 to -1.43; 148 participants, moderate quality evidence Analysis 3.1).
At the beginning of the trial there was no difference in daily hand washing frequency between intervention and control groups (3.4 ± 1.1 in control group; 3.3 ± 0.98 in intervention group; Table 5), but at the end of the trial the intervention group reported hand washing seven times a day compared with four times daily in the control group (P < 0.05) (moderate quality evidence).
In the original review, Ejemot-Nwadiaro 2008, 14 trials met the inclusion criteria. We have included eight additional trials in this Cochrane Review update, giving a total of 22 included trials. One of the eight additional trials, Luby 2006 PAK, was a follow-up trial to Luby 2003b PAK. This trial involved no primary interventions. It assessed the sustainability of the Luby 2003b PAK hand hygiene interventions in preventing diarrhoea. The other trials had primary interventions.
Hand washing promotion at child day-care facilities or schools prevents around one-third of diarrhoea episodes in high income countries (high quality evidence). It may prevent a similar proportion in LMICs but only two trials from urban Egypt and Kenya have evaluated this (low quality evidence).
Hand washing promotion among communities in LMICs probably prevents around one-quarter of diarrhoea episodes (moderate quality evidence). However, six of these eight trials were from Asian settings, with only single trials from South America and sub-Saharan Africa. In six trials soap was provided free alongside education and behavioural change interventions. The overall effect size was larger than in the two trials that did not provide soap. The influence of this on the intervention effect estimate is not well understood.
The effect of hand washing promotion in a hospital-based setting among high-risk population had significant reduction in mean episodes of diarrhoea that favoured intervention group (moderate quality evidence). This is only from one trial.
The effect of the intervention on hand hygiene related behavioural outcome in all settings showed increase in proportion of hand washing or hand hygiene compliance at essential junctures (before eating/cooking and after visiting the toilet or cleaning the baby's bottom) favouring the intervention groups (unpooled data; reflecting a range of low to high quality evidence).
We found no trials evaluating or reporting the effects of hand washing interventions on diarrhoea-related deaths, all-cause-under five mortality, or costs.
We believe we identified all RCTs that met our inclusion criteria. We further categorised the included trials into three distinct settings in this Cochrane Review: child day-care centres or schools, community, and hospital. Although there were only a few trials included in each category, evidence favours hand washing intervention in preventing diarrhoea in all the settings. This suggests that the intervention exhibits population-wide health gains. However, most included trials in the institution subcategory were from childcare settings in high-income countries. Thus, we are not confident that this finding can be applied to schools in LMIC settings or alternative institutions. Also, only one hospital-based trial met the inclusion criteria, so evidence from this setting was limited.
We are unsure of the effect of this intervention in populations with participants above five years of age and adults, as 95% of the participants in which the primary outcome was measured were below five years of age. One trial, Talaat 2008 EGY, measured the primary outcome in participants with a mean age of eight years but did not stratify the results by age. Nicholson 2008 IND measured the primary outcome in participants of various ages (target children aged five years, children below five years of age, children aged between six to 15 years and adults) and stratified results by these independent subgroups and reported effect sizes, with no significant trend observed. Therefore the effect of the intervention may not be generalizable to all age groups.
All included trials were relatively small-sized and of short follow-up duration including intensive monitoring and they demonstrated significant reduction in the risk of diarrhoea after hand hygiene intervention. However, in one relatively large trial, Bowen 2004 CHN, and one with longer follow-up, Luby 2006 PAK, there were no apparent benefits as no significant differences between the incidence or longitudinal prevalence of diarrhoea was found. Therefore, we are unclear if the reductions in incidence of diarrhoea would be maintained if these trials had been larger and conducted over a longer time period.
The effect size was lower in child day-care centres or school-based trials that attempted blinding outcome assessors than in trials that did not (26% versus 33% reduction in the incidence of diarrhoea respectively). The same trend was observed for community-based trials, with 18% reduction for trials that attempted blinding of outcome assessors and 35% reduction for trials that did not attempt blinding. This suggests a possible introduction of bias in trials that did not attempt blinding. However, there were too few trials in each category to make strong conclusions.
We assessed the quality of evidence using the GRADE approach (GRADEpro 2014). In general, the evidence that hand washing reduces the incidence of diarrhoea in both child day-care centres in high-income countries and community settings in LMICs is considered high quality (Summary of findings for the main comparison; Summary of findings 2). Most trials were at high or unclear risk of detection or reporting bias due to no description of blinding of outcome assessors. However, this made negligible differences in our findings as restriction of the analysis to just the blinded trials found a slightly smaller but statistically significant effect size. In addition, the trials' results showed a lot of statistical heterogeneity. However, these inconsistencies did not affect the quality of evidence in these settings since all trials favoured the intervention though with varying effect size. We are therefore confident in the estimate of effect and further research is very unlikely to change our confidence in the estimate.
For the trials conducted in schools in LMICs, we considered the quality of evidence to be low due to indirectness as this limits our confidence in the effect estimate. The two trials, Talaat 2008 EGY and Pickering 2013 KEN, were conducted under experimentally controlled situations. Though they showed benefits in favour of the intervention groups, we are unsure if these benefits would be maintained if trials are replicated in a less controlled situations and in other settings.
Quality of evidence from unpooled data for the behavioural outcomes ranged from low to high in all the settings. These should be interpreted with caution as there were too few trials in each setting and method of assessment were too varied to make strong statements. The benefit of adopting an explicit behavioral change model is still unclear; this may influence the maintenance and sustainability of hand hygiene behaviour, as Whitby 2007 has opined that the strongest determinant of hand washing behaviour may be its habituation.The quality of evidence regarding the other outcomes (diarrhoea related deaths, all-cause-under five mortality, and cost-effectiveness) were not determined due to paucity of included trials providing data on which to make such judgements. Thus, further research is necessary to provide a basis for assessment of evidence to these factors critical to hand washing intervention in preventing diarrhoea.
We did not identify any potential biases in the review process.
The magnitude of intervention effect ( 30%) in both child day-care centres or schools and community settings we observed in this Cochrane Review did not differ significantly from that of the original review (Ejemot-Nwadiaro 2008). The effect size however remains lower in magnitude than previous reviews of hand washing interventions; 47% (Curtis 2003); and about 44% in Fewtrell 2004 and Fewtrell 2005 reviews. These differences may be attributed to choice of effect measure, mixed trial designs, and single setting. Curtis 2003 used odds ratios, known to inflate effects sizes for conditions such as diarrhoea with common event rates in the analyses. In this Cochrane Review we reported only rate ratios, which Guevara 2004 opines improves clinical interpretation of pooled effect estimates. Fewtrell 2005 presented evidence of publication bias, while Curtis 2003 included case-control and cross-sectional trials as well as prospective interventions. Both reviews considered only hand hygiene interventions conducted in LMICs. In this Cochrane Review we included only RCTs and mixed settings (child day-care centres or schools, community, and hospital based trials conducted in both developing and developed countries). However, they are all in agreement that hand hygiene interventions are effective in reducing diarrhoeal diseases.
Hand washing promotion leads to reduction in diarrhoea episodes in both child day-care centres in high-income countries and among communities living in LMICs by about 30%. The challenge is to find ways of encouraging people to maintain hand washing habits in the longer term.
The findings of this Cochrane Review show that further research to determine the efficacy of hand washing intervention in preventing diarrhoea will be unnecessary in child day-care centres in high-income countries and in communities in LMICs, although only one trial was conducted in Africa.
More trials conducted in child day-care centres or schools in LMICs are needed to enhance our ability to generalize the intervention effects. The need to conduct research that is of longer follow-up duration and uses a structured method of assessing the primary outcome is pertinent, since it has been observed that arbitrary use of methods may have significant effect on precision of estimates. Outcome assessors should be blinded so as to reduce the bias in estimates of effect size.
Evidence of hand washing on diarrhoea incidence in hospital based settings is still limited as we only found one trial that met the inclusion criteria. Therefore, further research in this area would be warranted.
We thank all trial authors that assisted us with information and clarifications regarding their trials. We are particularly grateful to Dr. S Luby of the Centers for Disease Control and Prevention (CDC) and Jonathan Kotch of University of North Carolina at Chapel Hill, USA. We thank Karin Schiöler and Jeppe Schroll for assisting with translation of the Danish trial. The first version of this review (Ejemot-Nwadiaro 2008) was technically completed during the Cochrane Review Finishing School attended by Regina Ejemot-Nwadiaro at Liverpool School of Tropical Medicine (LSTM) and organised by the Cochrane Infectious Diseases Group in June 2005. This document is an output from a project funded by the UK Department for International Development (DFID) for the benefit of LMICs. The views expressed are not necessarily those of DFID.
I have read the interesting Cochrane Review "Hand washing for preventing diarrhoea" conducted by you and your colleagues, published in The Cochrane Library 2009, issue 3. I would like to take the liberty to comment on the search strategies shown in Table 1:
From the attached search sets it appears that you may have missed 98 and 61 potentially relevant records in MEDLINE and EMBASE respectively. Of course, this does not mean that you have not identified all relevant and available trials but it still poses a risk which I suggest you address in your next update of the review. How I searched MEDLINE and EMBASE, via Ovid (other databases were not searched):
Set 1-11: Identical to the search shown in Table 1 (I assumed set 9 should be in upper case)
Set 12-16: I added handwashing$ as free text term and show how many records are missed (set 16: records published before 2008)
Set 17-22: Same as above, but added diarrhoea$ and diarrhea$ to the search (set 22: records published before 2008)
Also, it would be helpful to know how many records your retrieved in your initial searches, how many were excluded due to lack of relevance, methodological flaws etc., i.e. presented in a flowchart.
We agree with the contributor that there was an error in Table 1. We have corrected this. We do not believe that we have missed any relevant records, but as this review is due to be updated, we will investigate this further during the updating process. With regard to presenting the results in a flowchart, PRISMA diagrams were not expected in Cochrane Reviews at the time this review was initially produced. This will again be dealt with during the updating process.
Ole Frandsen Nørgaard of the Department of Computer Science, Faculty of Health Sciences, University of Copenhagen, Denmark identified slight anomalies in the search strategy used in preparing the original review (Ejemot-Nwadiaro 2008). We have incorporated his suggestions appropriately into this review update.
Regina Ejemot-Nwadiaro and Dachi Arikpo extracted and analysed data, and drafted the review. John Ehiri developed the protocol, drafted, and commented on the review. Julia Critchley extracted and analysed data, and edited the review. Martin Meremikwu helped finalize the data extraction form, drafted and commented on the review.
Regina Ejemot-Nwadiaro, John Ehiri, Dachi Arikpo, Martin Meremikwu and Julia Critchley declare that they have no conflicts of interest.
We have introduced the term 'promotion' into the title of this Cochrane Review update. We added methods for assessing blinding and changed our primary outcome measure in the protocol from the relative risk of at least one diarrhoea episode to the incidence rate ratio for diarrhoea episodes. We pooled rate ratios in our analyses rather than relative risks since all trials presented diarrhoea as episodes, and removed "or standard hygiene promotion" as a control because it is included in the "no hand washing promotion" control group. We added all-cause-under five mortality and cost-effectiveness as secondary outcome measures for this review update. We used GRADEpro 2014 to assess the quality of the evidence. In addition, we have included 'Summary of findings' tables in this update. Henry Ejere, a co-author on the protocol, did not participate in preparation of the original review nor this review update. Dachi Arikpo joined as a co-author in this review update.
*Hand Disinfection; Child Day Care Centers; Diarrhea [*prevention & control]; Randomized Controlled Trials as Topic; Schools
References to studies included in this review