Portal de Periódicos da CAPES

Effectiveness and sustainability of water, sanitation, and hygiene interventions in combating diarrhoea

2009; Taylor & Francis; Volume: 1; Issue: 3 Linguagem: Inglês

10.1080/19439340903141175

1943-9342

Hugh Waddington, Birte Snilstveit,

Global Maternal and Child Health

Abstract This paper presents a synthetic review of impact evaluations examining effectiveness of water, sanitation, and hygiene interventions in reducing diarrhoea among children. The evaluations were conducted in 35 low- and middle-income countries during the past three decades. The paper challenges the existing consensus that water treatment at point-of-use and hygiene interventions are necessarily the most effective and sustainable interventions for promoting reduction of diarrhoea. The analysis suggests that sanitation ‘hardware’ interventions are highly effective in reducing diarrhoea morbidity. Moreover, while there is a wealth of trials documenting the effectiveness of water treatment interventions, studies conducted over longer periods tend to show smaller effectiveness and evidence suggests compliance rates and therefore impact may fall markedly over time. Keywords: water supplywater qualitysanitationhygienechild healthdiarrhoeaimpact evaluationsynthetic review Acknowledgements Thanks to Howard White, Lorna Fewtrell, and Ron Bose for comments and to participants at the Perspectives on Impact Evaluation conference in Cairo, April 2009 and Campbell Collaboration Colloquium in Oslo, May 2009. The authors are responsible for all remaining errors. Thanks also to Steve Luby, Marion Jenkins, Alix Peterson Zwane and Mark Sobsey who shared unpublished or forthcoming papers. The views contained in this paper are those of the authors and do not necessarily reflect the views of the International Initiative for Impact Evaluation (3ie). Notes 1. Methods to reduce faecal contamination of the environment such as fly spraying are not included in this review. See, for example, Chavasse et al. (1999 Chavasse, D. C. 1999. Impact of fly control on childhood diarrhoea in Pakistan: community-randomised trial. Lancet, 353(9146): 22–25. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]). 2. The figure is highly simplified; ‘exogenous’ factors affecting disease risk at individual level include household size, age, nutritional and health status, and personal immunity. 3. Wright et al.’s (2004 Wright, J., Gundry, S. and Conroy, R. 2004. Household drinking water in developing countries: a systematic review of microbiological contamination between source and point-of-use. Tropical medicine and international health, 9(1): 106–117. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) systematic review finds substantial evidence for (re-) contamination between source and point-of-use. 4. Only two evaluations included in this review collected information on timesavings. Pattanayak et al. (2007 Pattanayak, S. K. 2007. Promoting latrine use: midline findings from a randomized evaluation of a community mobilization campaign in Bhadrak, Orissa, North Carolina: Research Triangle Institute. Working Paper, No. 07_02 [Google Scholar]) estimate a reduction of 17 minutes per family member per day in walking for defecation associated with improved sanitation in rural India. In rural Nigeria, Blum et al. (1990 Blum, D. 1990. The Imo state (Nigeria) Drinking Water Supply and Sanitation Project, 1. Description of the project, evaluation methods, and impact on intervening variables. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84(2): 309–315. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) estimate reduced time from six hours to 45 minutes per household per day during the dry season associated with hand-pump installation, mainly benefiting adolescent girls and young women (although an increase of zero to 12 minutes in the wet season due to less reliance on the sole use of rainwater). In addition, Wang et al. (1989 Wang, Z. S. 1989. Reduction of enteric infectious disease in rural China by providing deep-well tap water. Bulletin of the World Health Organization, 67(2): 171–180. [PubMed], [Web of Science ®] , [Google Scholar]) estimate timesavings of 20 minutes per household per day from a village water supply improvement in China. 5. Estimates of cost-effectiveness in terms of dollars (US$) per disability-adjusted life year (DALY) averted from improved hygiene and sanitation suggest that hygiene promotion is the most efficient, at US$3/DALY averted, followed by sanitation promotion, at US$11/DALY, and finally sanitation construction, at up to US$270/DALY (Cairncross and Valdmanis 2006 Cairncross, S. and Valdmanis, V. 2006. “Water supply, sanitation and hygiene promotion. In: D. Jamison et al”. In Disease control priorities in developing countries, 2nd, 771–792. New York: Oxford University Press. [Google Scholar]). Estimates of cost-effectiveness of improved water suggests the impact of community connection in terms of disability-adjusted life years, estimated at US$94/DALY, is less than half that for household connection, but substantially above comparable estimates of point-of-use water treatment – for example, Clasen et al. (2007a Clasen, T. 2007a. Cost-effectiveness of water quality interventions for preventing diarrhoeal disease in developing countries. Journal of water and health, 5(4): 599–608. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) estimate US$53/DALY averted from chlorination. 6. Owing to concerns of external validity, we also exclude studies based on disease reporting to health facilities, including deWilde et al. (2008 DeWilde, C. K. 2008. An integrated method for evaluating community-based safe water programmes and an application in rural Mexico. Health policy and planning, 23(6): 452–464. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) and Wang et al. (1989 Wang, Z. S. 1989. Reduction of enteric infectious disease in rural China by providing deep-well tap water. Bulletin of the World Health Organization, 67(2): 171–180. [PubMed], [Web of Science ®] , [Google Scholar]) and those based on case-control design (for example, Clemens and Stanton 1987 Clemens, J. D. and Stanton, B. F. 1987. An educational intervention for altering water-sanitation behaviors to reduce childhood diarrhea in urban Bangladesh I. Application of the case-control method for development of an intervention. American journal of epidemiology, 125(2): 284–292. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Daniels et al. 1990 Daniels, D. L. 1990. A case-control study of the impact of improved sanitation on diarrhoea morbidity in Lesotho. Bulletin of the World Health Organization, 68(4): 455–463. [PubMed], [Web of Science ®] , [Google Scholar]). Those seeking formal health care are unlikely to be a random sample of the population. 7. The searches in the databases included in Fewtrell et al. (2005 Fewtrell, L. 2005. Water, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis. Lancet infectious diseases, 5: 42–52. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) – Pubmed, Embase, LILACs and the Cochrane Library – were limited to papers published since 2003, while the Web of Science databases (including Science Citation Index Expanded, Social Sciences Citation Index, Conference Proceedings Citation Index – Science), JOLIS, IDEAS, BLDS and Google Scholar were searched back to 1998. As the Google searches generated a large number of results, which were ordered by relevance, we limited our reviews to the first 1000 results. The study report provides the list of organisations contacted (Waddington et al. 2009b Waddington, H. . Water, sanitation and hygiene interventions to combat childhood diarrhoea in developing countries. Study Report. Synthetic Review 1 SR001. International Initiative for Impact Evaluation. June2009, New Delhi. [Google Scholar]). 8. In the calculation of ES precision, while most studies appear to have adjusted for clustering at community level, not all studies adjusted for clustering at household level, where multiple observations were from the same household, or at individual level, where multiple observations were taken from the sample individual over time. 9. Effect sizes are calculated as relative risk ratios, average rate ratios, longitudinal prevalence ratios, or odds ratios. Risk measures the probability of being ill during the measurement period, odds the conditional probability of illness divided by the probability of not being ill over the measurement period, rate, or incidence density measures the average risk over the measurement period measured in average number of ‘episodes’, and longitudinal prevalence is more closely associated with duration of illness, usually measured as the proportion of days of illness during the measurement period. Both odds and longitudinal prevalence ratios will tend to be further away from the value of one (the point of no effect) than risk or rate ratios, this difference being larger the greater the disease incidence in the sample. It was not possible to convert these ratios into a common metric. However, we note that where the risk of disease is low, as in the majority of studies included in this review, the measures produce similar results (Kleinbaum et al. 1982 Kleinbaum, D., Kupper, L. and Morgenstern, H. 1982. Epidemiologic research: principles and quantitative methods, Belmont, CA: Lifetime Learning Publications. [Google Scholar]). We also examined whether ratio type accounts for differences in effects observed across interventions. 10. Longitudinal prevalence of diarrhoea is preferred on theoretical grounds and empirically is more strongly associated with child mortality and weight gain than incidence (Morris et al. 1996 Morris, S. S. 1996. Is prevalence of diarrhea a better predictor of subsequent mortality and weight gain than diarrhea incidence?. American journal of epidemiology, 144(6): 582–8. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]). Different interventions may affect measures of incidence and prevalence differently – for example, Gross et al. (1989 Gross, R. 1989. The impact of improvement of water supply and sanitation facilities on diarrhea and intestinal parasites: a Brazilian experience with children in two low-income urban communities. Revue saude publica, 23(3): 214–220. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) note that hygienic practices such as removal of faeces from the yard may have greater impact on spell duration than incidence. 11. For some studies, the effect sizes were adjusted using multivariate regression with control variables, often due to concerns of confounding. Some of the control variables used, for example, carers’ education and observed hygiene practices, may be instruments for compliance, which would tend to inflate ITT impact estimates towards the TOT effect. We examine whether adjustment systematically affects impact estimates. 12. An additional source of bias which we are unable to examine here is censoring of observations due to mortality: where WSH interventions save lives, which are more prone to disease than average, a positive impact on mortality could yield perverse outcomes in terms of morbidity. 13. A recall period of two weeks is usually the limit of what is considered as a reasonable period in reporting disease morbidity before significant bias sets in, as used for example in household surveys such as the DHS. There are studies arguing that a recall period greater than 48 hours is unreliable, particularly when providing data on family members other than oneself (Boerma et al. 1991 Boerma, J. 1991. Accuracy and completeness of mothers’ recall of diarrhea occurrence in pre-school children in demographic and health surveys. International journal of epidemiology, 20: 1073–1080. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]). Arguably, the risk of bias would be less when caregivers are asked about young children. 14. Where possible, we collected or synthesised estimates for children. However, we were obliged to include estimates for all ages for six evaluations that did not provide separate effect sizes for children (Khan 1982 Khan, M. U. 1982. Interruption of shigellosis by handwashing. Transactions of the Royal Society of Tropical Medicine and Hygiene, 76(2): 164–168. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Xiao et al. 1997 Xiao, S., Lin, C. and Chen, K. 1997. Evaluation of effectiveness of comprehensive control for diarrhea diseases in rural areas of East Fujian and analysis of its cost–benefit. Chinese journal of preventive medicine, 31(1): 40–41. [Google Scholar], Walker et al. 1999 Walker, I. 1999. Ex-post evaluation of the Honduran social investment fund (FHIS 2), Tegucigalpa, Honduras: ESA Consultores. [Google Scholar], Iijima et al. 2001 Iijima, Y. 2001. Prevention of bacterial diarrhea by pasteurization of drinking water in Kenya. Microbiology and immunology, 45(6): 413–416. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Quick et al. 2002 Quick, R. E. 2002. Diarrhea prevention through household-level water disinfection and safe storage in Zambia. American journal of tropical medicine and hygiene, 66(5): 584–589. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Doocy and Burnham 2006 Doocy, S. and Burnham, G. 2006. Point-of-use water treatment and diarrhoea reduction in the emergency context: an effectiveness trial in Liberia. Tropical medicine & international health, 11(10): 1542[Crossref], [PubMed], [Web of Science ®] , [Google Scholar]). Inclusion of these estimates did not alter meta-analysis results. 15. Two multiple arm trials report separately the impact of water treatment or water treatment plus safe storage against a single control group (Reller et al. 2003 Reller, M. E. 2003. A randomized controlled trial of household-based flocculant-disinfectant drinking water treatment for diarrhea prevention in rural Guatemala. American journal of tropical medicine and hygiene, 69(4): 411–419. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Luby et al. 2006 Luby, S. P. 2006. Combining drinking water treatment and hand washing for diarrhoea prevention, a cluster randomised controlled trial. Tropical medicine & international health, 11(4): 479–489. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]). Two multiple arm sanitation studies report separate impact of sewer connection or latrine provision versus a single control group (Walker et al. 1999 Walker, I. 1999. Ex-post evaluation of the Honduran social investment fund (FHIS 2), Tegucigalpa, Honduras: ESA Consultores. [Google Scholar], Pradhan and Rawlings 2002 Pradhan, M. and Rawlings, L. B. 2002. The impact and targeting of social infrastructure investments: lessons from the Nicaraguan Social Fund. World Bank economic review, 16(2): 275–295. [Crossref], [Web of Science ®] , [Google Scholar]), while one reports separately improved drainage or improved drainage and sewer connection versus a single control group (Moraes et al. 2003 Moraes, L. R. 2003. Impact of drainage and sewerage on diarrhoea in poor urban areas in Salvador, Brazil. Transactions of the Royal Society of Tropical Medicine and Hygiene, 97(2): 153–158. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]). For each study, we calculate a weighted mean effect size for pooled meta-analysis, but conduct additional sub-group analysis of separate arms. 16. The correlation between estimates was calculated as the sample-weighted mean of the correlation of treatment groups and the correlation of the control groups. The correlation between control arms was assumed 1 where the same control group is used as comparator and 0 otherwise. The correlation between treatment arms was assumed to be 0 when combining results from different treatment groups and 1 when combining results from the same treatment groups over time. When combining results across different individuals within the same treatment group the correlation was assumed 0.5, which estimates variance at the mid-point between the two extreme cases of treating comparisons as independent (with correlation coefficient equal to 0) and most likely underestimating the variance, or treating them as perfectly correlated (correlation coefficient of 1) and most likely overestimating the variance. See Borenstein et al. (2009 Borenstein, M. 2009. Introduction to meta-analysis, Chichester: John Wiley. [Crossref] , [Google Scholar], Chapters 24 and 25) for more on this. Meta-analysis results were not sensitive to these assumptions. 17. The choice of random effects model was made given likely heterogeneity arising from contextual factors, including location of the study, baseline environmental risk and diarrhoea incidence, and underlying behavioural factors, all of which would likely invalidate the assumption that the ‘true’ effect of the intervention is fixed across studies. 18. The list of studies that examined the relationship between WSH and diarrhoea morbidity but did not meet the inclusion criteria for rigorous impact evaluation is provided in the study report (Waddington et al. 2009b Waddington, H. . Water, sanitation and hygiene interventions to combat childhood diarrhoea in developing countries. Study Report. Synthetic Review 1 SR001. International Initiative for Impact Evaluation. June2009, New Delhi. [Google Scholar]). 19. Universidad Rafael Landivar (1995 Universidad Rafael Landivar. 1995. Contra la morbilidad infantil: filtros artesanales y educacion. Estudios sociales IV epoca, : 1–66. [Preventing infant morbidity: artisanal filters and education.] [Google Scholar]) examined water treatment, hygiene education, and a combination of the two, using factorial (‘a, b, a+b’) design, finding some evidence for complementarity. Treatment arms involving hygiene education were excluded from our analysis because inclusion in the hygiene education treatment was based on self-selection. Kremer et al. (2008 Kremer, M., et al., 2008. Trickle down: diffusion of chlorine for drinking water treatment in Kenya. Available from http://www.stanford.edu/group/SITE/SITE_2008/segment_1/papers/de_giorgi_SITE_Trickle_Down.pdf (Accessed: 6 November 2008). [Google Scholar]) also examine water treatment at source and at POU using factorial design, finding no additional impact when the interventions were combined; this result was not included in our analysis of multiple interventions, defined as those combining water, sanitation, and/or hygiene. 20. Lou et al. (1990 Lou, E. A. 1990. Effectiveness evaluation and cost-effectiveness estimate for diarrhoea control by environment improvement in rural area. Zhonghua liu xing bing xue za zhi, 11(3): 170–174. [PubMed] , [Google Scholar]) was excluded from meta-analysis because we were unable to calculate the precision of reported estimates. 21. A regression coefficient less than one indicates the variable is associated with a bigger effect on diarrhoea reduction in the treatment group relative to the control group while a coefficient of greater than one indicates less effectiveness in reducing diarrhoea in the control group. 22. These results understate the placebo effect documented by Schmidt and Cairncross (2009 Schmidt, W. P. and Cairncross, S. 2009. Household water treatment in poor populations: is there enough evidence for scaling up now?. Environmental science & technology, 43(4): 986–992. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]); we were unable to obtain full text copies of two randomised placebo-controlled trials reported therein of POU water quality interventions that demonstrated zero impact on diarrhoea morbidity (Austin 1993 Austin, C. J. 1993. Investigation of in-house water chlorination and its effectiveness for rural areas of the Gambia, New Orleans: Dissertation, Tulane University School of Public Health and Tropical Medicine. [Google Scholar], Blanton et al. 2008). 23. Aziz et al. (1990a Aziz, K. M. 1990a. Reduction in diarrhoeal diseases in children in rural Bangladesh by environmental and behavioural modifications. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84(3): 433–438. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Ahmed et al. (1993 Ahmed, N. U. 1993. A longitudinal study of the impact of behavioural change intervention on cleanliness, diarrhoeal morbidity and growth of children in rural Bangladesh. Social science and medicine, 37(2): 159–171. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Bateman et al. (1995 Bateman, O. M. Prevention of diarrhea through improving hygiene behaviors: the Sanitation and Family Education (SAFE) Pilot Project experience, Special publication: ICDDR, B, No. 42. Dhaka: International Centre for Diarrhoeal Disease Research 1995 [Google Scholar]), Conroy et al. (1999 Conroy, R. M. 1999. Solar disinfection of water reduces diarrhoeal disease: an update. Archives of disease in childhood, 81(4): 337–338. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Chiller et al. (2006 Chiller, T. M. 2006. Reducing diarrhoea in Guatemalan children: randomized controlled trial of flocculant-disinfectant for drinking water. Bulletin of the World Health Organization, 84: 28–35. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Garrett et al. (2008 Garrett, V. 2008. Diarrhoea prevention in a high-risk rural Kenyan population through point-of-use chlorination, safe water storage, sanitation, and rainwater harvesting. Epidemiology and infection, 136(11): 1463–1471. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Haggerty et al. (1994 Haggerty, P. A. 1994. Community based hygiene education to reduce diarrhoeal disease in rural Zaire: impact of the intervention on diarrhoeal morbidity. International journal of epidemiology, 23(5): 1050–1059. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Hoque et al. (1996 Hoque, B. A. 1996. Sustainability of a water, sanitation and hygiene education project in rural Bangladesh: a 5-year follow-up. Bulletin of the World Health Organization, 74(4): 431–437. [PubMed], [Web of Science ®] , [Google Scholar]), Huttley et al. (1990 Huttly, S. R. 1990. The Imo state (Nigeria) Drinking Water Supply and Sanitation Project, 2. Impact on dracunculiasis, diarrhoea and nutritional status. Transactions of the Royal Society of Tropical Medicine and Hygiene, 84(2): 316–321. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Pattanayak et al. (2007 Pattanayak, S. K. 2007. Promoting latrine use: midline findings from a randomized evaluation of a community mobilization campaign in Bhadrak, Orissa, North Carolina: Research Triangle Institute. Working Paper, No. 07_02 [Google Scholar]), Pinfold and Horan (1996 Pinfold, J. V. and Horan, N. J. 1996. Measuring the effect of a hygiene behaviour intervention by indicators of behaviour and diarrhoeal disease. Transactions of the Royal Society of Tropical Medicine and Hygiene, 90(4): 366–371. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Pradhan and Rawlings (2002 Pradhan, M. and Rawlings, L. B. 2002. The impact and targeting of social infrastructure investments: lessons from the Nicaraguan Social Fund. World Bank economic review, 16(2): 275–295. [Crossref], [Web of Science ®] , [Google Scholar]), Stanton et al. (1988 Stanton, B. F., Clemens, J. D. and Khair, T. 1988. Educational intervention for altering water-sanitation behaviour to reduce childhood diarrhea in urban Bangladesh: impact on nutritional status. American journal of clinical nutrition, 48(5): 1166–1172. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]), Torun (1982 Torun, B. 1982. “Environmental and educational interventions against diarrhoea in Guatemala. In”. In Diarrhea and malnutrition: interactions, mechanisms and interventions, Edited by: Chen, L. C. and Scrimshaw, N. S. New York: Plenum Press. [Google Scholar]). 24. Of the studies which collected data across wet and dry seasons, a number found bigger impact of WSH interventions on diarrhoea morbidity during rainy season (for example, Ahmed et al. 1993 Ahmed, N. U. 1993. A longitudinal study of the impact of behavioural change intervention on cleanliness, diarrhoeal morbidity and growth of children in rural Bangladesh. Social science and medicine, 37(2): 159–171. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Luby et al. 2006), while others found bigger impacts during dry season (for example, Jensen et al. 2003 Clasen, T. F. 2005. Household-based ceramic water filters for the prevention of diarrhea: a randomized, controlled trial of a pilot program in Colombia. American journal of tropical medicine and hygiene, 73(4): 790–795. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Stauber et al. 2009 Stauber, C. E. 2009. A randomized controlled trial of the concrete biosand filter and its impact on diarrheal disease in Bonao, Dominican Republic. American journal of tropical medicine and hygiene, 80(2): 286–293. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar], Tiwari et al. 2009). 25. Where intervention date was not reported we used length of data collection period (Jalan and Ravallion 2003 Jalan, J. and Ravallion, M. 2003. Does piped water reduce diarrhea for children in rural India?. Journal of econometrics, 112(1): 153–173. [Crossref], [Web of Science ®] , [Google Scholar], Khanna 2008 Khanna, G. The impact on child health from access to water and sanitation and other socioeconomic factors. HEI Working Paper No. 02/2008. 2008. Geneva: Graduate Institute of International and Development Studies [Google Scholar], Bose 2009 Bose, R. 2009. The impact of water supply and sanitation interventions on child health: evidence from DHS surveys, Mimeo, New Delhi: International Initiative for Impact Evaluation. [Google Scholar]). 26. The study report presents the full analysis of compliance (Waddington et al. 2009b Waddington, H. . Water, sanitation and hygiene interventions to combat childhood diarrhoea in developing countries. Study Report. Synthetic Review 1 SR001. International Initiative for Impact Evaluation. June2009, New Delhi. [Google Scholar]). 27. Although not examining sustainability of diarrhoea morbidity reduction, Cairncross and Shordt (2004 Cairncross, S. and Shordt, K. 2004. It does last! Some findings from a multi-country study of hygiene sustainability. Waterlines, 22(3): 4–7. [Crossref] , [Google Scholar]) conclude from a multi-country study that collected data between one and nine years after intervention that improved hygiene behaviours were sustained. 28. A number of water quality and hygiene studies refer to diffusion theory in explaining their results. Quick et al. (2002 Quick, R. E. 2002. Diarrhea prevention through household-level water disinfection and safe storage in Zambia. American journal of tropical medicine and hygiene, 66(5): 584–589. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) argue of a trial of POU water chlorination and safe storage in Zambia that the technology satisfied four out of five characteristics of innovations that are particularly important in explaining their adoption: (1) it had a relative advantage over the alternative – boiling, which was time consuming and expensive – and was made readily available during the trial; (2) it was compatible with the perceived needs of the target population; (3) it was simple to understand and use and therefore of low complexity; and (4) it had a high degree of trialability. Only in (5) observability of results was the technology argued to be lacking. In a trial of drinking water disinfectant and hand-washing with soap in a squatter settlement in Karachi, Luby et al. (2004 Luby, S. P. 2004. Delayed effectiveness of home-based interventions in reducing childhood diarrhea, Karachi, Pakistan. American journal of tropical medicine and hygiene, 71(4): 420–427. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) argue that households with refrigerators observed benefits earlier than those without refrigerators because they ‘behaved like early adopters, that is, their early reduction in diarrhea incidence suggests they quickly adopted the necessary behavior change to benefit from the interventions’ (p. 425). 29. See Stockman et al. (2007 Stockman, L. 2007. Point-of-use water treatment and use among mothers in Malawi. Emerging infectious diseases, 13(7): 1077–1080. [Crossref], [PubMed], [Web of Science ®] , [Google Scholar]) for a comparison of sustainability in adoption rates for water disinfectant in Malawi and Zambia.