Summary of findings for the main comparison. Summary of findings table 1.

Education and hygiene promotion intervention compared with no intervention for preventing diarrhoea in low‐ and middle‐income countries
Patient or population: adults and children Settings: LMICs Intervention: education and hygiene promotion intervention that includes promotion of safe child faeces disposal among other promoted behaviours Comparison: no intervention
Outcomes	Anticipated absolute effects (95% CI)		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)**	Comments
	Risk with no intervention	Risk with intervention
Diarrhoea prevalence Cluster RCTs	3 episodes per person per year	2.79 episodes per person per year (2.52 to 3.12)	RR 0.93 (0.84 to 1.04)	12,040 (2 studies)	⊕⊕⊝⊝ Lowa,b,c,d	The intervention may make little or no difference to diarrhoea prevalence.
Diarrhoea incidence Cluster RCTs	3 episodes per person per year	2.13 episodes per person per year (1.77 to 2.58)	Rate ratio 0.71 (0.59 to 0.86)	2549 (2 studies)	⊕⊕⊝⊝ Lowa,d,e,f	The intervention may reduce diarrhoea incidence.
Diarrhoea prevalence Controlled cohort studies: Sanitation Hygiene Education and Water Supply in Bangladesh (SHEWA‐B) intervention	3 episodes per person per year	2.73 episodes per person per year (1.92 to 3.84)	RR 0.91 (0.64 to 1.28)	˜2000 (2 studies)	⊕⊝⊝⊝ Very lowa,g,h,i	We are uncertain whether or not the intervention reduces diarrhoea prevalence.
Diarrhoea prevalence Controlled cross‐sectional studies: Health Extension Package intervention (Ethiopia)	3 episodes per person per year	0.78 episodes per person per year (0.48 to 1.26)j	OR 0.26 (0.16 to 0.42)	1660 (2 studies)	⊕⊝⊝⊝ Very lowa,b,d,k	We are uncertain whether or not the intervention reduces diarrhoea prevalence.
*The assumed risk for diarrhoea is taken from Walker 2012 and represented an estimated mean for the incidence of diarrhoea in LMICs. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). **RCTs begin as high‐certainty evidence and observational studies as low‐certainty evidence (Guyatt 2008) CI: confidence interval; LMICs: low‐ and middle‐income countries; OR: odds ratio; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: we are very uncertain about the estimate.

^aDowngraded one level for serious risk of bias: the outcome was self‐reported diarrhoea, and was susceptible to bias as all studies were unblinded.
 ^bNo serious inconsistency.
 ^cDowngraded one level for indirectness: only two studies in low‐income countries. Both conducted in rural settings, one in Rwanda and one in Democratic Republic of Congo. Diarrhoea was only measured in children aged < 3 years in Haggerty 1994 DRC.
 ^dNo serious imprecision.
 ^eNo serious inconsistency: there was considerable statistical heterogeneity (I² = 82%); however, there was consistency in the direction of the effect. Possible reasons for heterogeneity included the location of the studies; Stanton 1987 BGD was conducted in urban Bangladesh and Hashi 2017 ETH in rural Ethiopia. Furthermore, the studies used different definitions of diarrhoea and different age groups (aged less than six years for Stanton 1987 BGD and less than five years for Hashi 2017 ETH).
 ^fDowngraded one level for indirectness: only two studies, one in an urban Asian setting (Bangladesh) and one in an African rural setting (Ethiopia).
 ^gDowngraded one level for inconsistency: substantial statistical heterogeneity (I² = 55%).
 ^hDowngraded one level for indirectness: only two studies, both conducted in Bangladesh and evaluating the same intervention that was specifically tailored to Bangladesh.
 ⁱDowngraded one level for imprecision: small sample size and large CIs which included important effects in both directions.
 ^jCalculated using the OR as an approximation for RR.
 ^kDowngraded one level for indirectness: only two studies, both conducted in rural Ethiopia and evaluating an intervention specifically designed for Ethiopia.