WATER AND SANITATION PROGRAM: TECHNICAL PAPER 90744 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior: An Analysis of Data from Multiple Impact Evaluations in the Global Scaling Up Handwashing Project Pavani K. Ram, Michelle W. Sahli, Benjamin Arnold, John M. Colford, Claire Chase, Bertha Briceño, Alexandra Orsola-Vidal, and Paul Gertler August 2014 The Water and Sanitation Program is a multi-donor partnership administered by the World Bank to support poor people in obtaining affordable, safe, and sustainable access to water and sanitation services. By Pavani K. Ram, Michelle W. Sahli, Benjamin Arnold, John M. Colford, Claire Chase, Bertha Briceño, Alexandra Orsola-Vidal, and Paul Gertler We express deep gratitude to the participants of the Impact Evaluation endline surveys, and extend sincere thanks to the country principal investigators and the research teams in Peru, Senegal, and Vietnam for operationalizing the data collection efforts. Many thanks to Amy Pickering and Craig Kullmann for their thoughtful reviews of earlier drafts of this document. Global Scaling Up Handwashing is a project by the Water and Sanitation Program (WSP) focused on applying innovative behavior change approaches to improve handwashing with soap behavior among women of reproductive age (ages 15–49) and primary school-age children (ages 5–9). It was implemented by local and national governments with technical support from WSP in four countries: Peru, Senegal, Tanzania, and Vietnam. For more information, please visit www.wsp.org/ scalinguphandwashing. WSP is a multidonor partnership created in 1978 and administered by the World Bank to support poor people in obtaining affordable, safe, and sustainable access to water and sanitation services. WSP’s donors include Australia, Austria, Denmark, Finland, France, the Bill & Melinda Gates Foundation, Luxembourg, Netherlands, Norway, Sweden, Switzerland, United Kingdom, United States, and the World Bank. WSP reports are published to communicate the results of WSP’s work to the development community. Some sources cited may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are entirely those of the author and should not be attributed to the World Bank or its affiliated organizations, or to members of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to worldbankwater@worldbank.org. WSP encourages the dissemination of its work and will normally grant permission promptly. For more information, please visit www.wsp.org. © 2014 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior: An Analysis of Data from Multiple Impact Evaluations in the Global Scaling Up Handwashing Project Pavani K. Ram, Michelle W. Sahli, Benjamin Arnold, John M. Colford, Claire Chase, Bertha Briceño, Alexandra Orsola-Vidal, and Paul Gertler August 2014 Executive Summary There is increasing interest in improving handwashing in During structured observations of primary caregivers, low- and middle-income countries. However, there is a pau- 1,467 critical events were observed in Peru, 444 in Senegal, city of information on the measurement of handwashing and 1,421 in Vietnam. At these critical events, hands were behavior from many low- and middle-income countries, washed with soap in 14 percent of events in Peru, 13 per- because most prior studies have been carried out in South cent in Senegal, and 10 percent in Vietnam. Handwashing Asia. There is an unmet need to estimate handwashing be- with soap was observed during 34 percent of fecal contact havior using practical measures that yield valid indicators events in Peru, 25 percent in Senegal, and 24 percent in of handwashing behavior across cultural and geographic Vietnam. In all three countries, caregivers were 1.6 to 3.6 contexts. The validity of rapid handwashing measures was times more likely to be observed washing hands with soap if evaluated by comparing them to handwashing behavior they lived in homes with observed soap at the handwashing measured during five-hour structured observations. place used after defecation, compared to caregivers living in homes without soap and water at that place. Similarly, care- Handwashing was measured in the Impact Evaluation givers were 2 to 2.4 times more likely to be observed wash- of the Global Scaling Up Handwashing project, carried ing hands with soap if soap and water were observed at the out by the Water and Sanitation Program (WSP) in Peru, place where hands are washed before food preparation. Dis- Senegal, and Vietnam. Global Scaling Up Handwashing tance of the handwashing location from either the latrine or tested the effects of at-scale implementation of handwash- the food preparation place was not associated with observed ing promotion on various outcomes, including behavior handwashing with soap. Self-reported handwashing was not and health, in four countries—Peru, Senegal, Tanzania, and associated with observed handwashing behavior in multiple Vietnam—using cluster-randomized controlled trial designs. countries. In Peru and Vietnam, adjustment for wealth did not alter the associations between the rapid handwashing For each country, data on self-reported and observed hand- measures and observed handwashing with soap. In Senegal, washing measures from primary caregivers of young chil- none of the rapid handwashing measures were significantly dren was analyzed during endline surveys among the control associated with observed handwashing with soap in models arms of the evaluations conducted in Peru, Senegal, and including wealth. Vietnam. Structured observations were carried out among a subset of households participating in endline surveys in each country. The relationship between each of the rapid This multicountry analysis of the validity of rapid hand- handwashing measures was evaluated, and handwashing be- washing measures confirms the utility of observing hand- havior was observed, using regression models for all events washing materials at the places where people wash hands, at observed, as well as events restricted to fecal contact. Logistic the times most necessary for washing them (after fecal con- regression was used to model the relationship between the tact and before food preparation). The findings described rapid handwashing measure and the probability that hands here also reinforce the global imperative of improving were washed during the observed event, accounting for the handwashing behavior for prevention of the leading causes repeated nature of structured observation data. of death in young children. iv Global Scaling Up Handwashing Contents I. Introduction ............................................................................... 1 II. Methods ..................................................................................... 3 III. Data Analysis ............................................................................. 5 IV. Results ....................................................................................... 6 V. Discussion ............................................................................... 14 References ............................................................................... 17 Tables 1: Measures of Handwashing Behavior Assessed in Global Scaling Up Handwashing, by Method and Level of Data Collection ................................................................ 3 2: Timeline and Sample Size of Control Populations in Endline Surveys of the Impact Evaluation of Global Scaling Up Handwashing, by Country ................................. 6 3: Descriptive Analysis of Rapid Handwashing Measures, All Countries......................................................................... 7 4: Prevalence-Adjusted Kappa Scores Reflecting Moderate or Greater Agreement between Rapid Handwashing Meaures in Endline Surveys in Peru, Senegal, and Vietnam, 2009–2011 ........................................................... 8 5: Frequency of Observation of Critical Events for Handwashing, and Handwashing Behavior, during Endline Structured Observations, by Country, 2011 ......... 10 6: Associations between Rapid Handwashing Measures and Observed Handwashing Behavior, Endline Surveys and Structured Observations in Peru, Senegal, and Vietnam, 2009–2011 .......................................................... 11 Annex 1: Supplemental Tables.......................................................... 19 www.wsp.org v I. Introduction Because of the compelling evidence that handwashing re- for use in large survey populations. In addition, data collec- duces diarrhea and respiratory illness, two of the leading tion for the measure would not require multiple visits, or causes of child mortality globally (Aiello et al. 2008; Ejemot skill or training in the data collector beyond what is typical et al. 2008; Luby et al. 2005; Luby et al. 2004), this practice for most community-level studies of health or hygiene. is increasingly being promoted in low- and middle-income countries. Public health practitioners promoting hand- Rapid handwashing measures include: washing seek to evaluate program impacts on individual • self-reported handwashing behavior, behavior. Researchers examine strategies to improve hand- • observation of handwashing materials in the home, washing and investigate its role in improving health. Prac- • handwashing behavior demonstrated upon request, titioners and researchers alike need to measure individuals’ and handwashing, a behavior often perceived to be challenging • hand cleanliness on visual inspection. to measure because of the prevalent social desirability of washing hands (Ram 2013). Several rapid handwashing measures have been evaluated against health outcomes, given that improved health is the Structured observation is often considered the best way to ultimate goal of any handwashing promotion program. In measure handwashing behavior. During structured obser- three studies from Bangladesh and one from Nepal, care- vation, an observer studies the target individual(s), such giver-reported handwashing behavior has been associated as the primary caregiver of a young child or all household with neonatal mortality (Rhee et al. 2008), child diarrhea members, from within the home/courtyard. The observer mortality (Unicomb et al. 2010), child diarrhea morbidity records opportunities for handwashing (e.g., potential fecal (Luby et al. 2011a), and child pneumonia (Silk et al. 2010). contact at times such as toileting or cleaning a child who Observation of water at a handwashing place has twice been has defecated), whether or not the target individual washes shown to be associated with fewer episodes of respiratory hands, how hands are washed and dried, and other details illness, with both studies set in Bangladesh (Manun’Ebo et al. of interest. The duration of structured observations is sev- 1997; Szklo and Nieto 2007). Another study, also carried eral hours (Ibid.), often five, to allow the observer to wit- out in Bangladesh, found that soap use by mothers during ness a number of opportunities for handwashing. demonstration was significantly associated with lower prev- alence of diarrhea among their children, compared to the Although structured observation is objective and yields children of mothers who did not use soap during demon- detailed information on target individuals’ handwashing stration (Luby et al. 2011a). In the same study population, behavior, it is resource intensive. A five-hour duration im- observation of visibly clean fingerpads on a child’s hands plies that a fieldworker can only complete one structured was associated with reduced diarrhea prevalence (Ibid.). observation in a day, making personnel costs prohibitive. Training for structured observation is more intensive than Rapid handwashing measures have also been evaluated for other approaches to measuring handwashing behavior. against observed handwashing behavior, as measured by Practitioners and researchers who cannot carry out struc- structured observations. For example, several studies have tured observations because of budgetary or logistical con- found that study populations tend to overreport their straints seek valid measures of handwashing behavior that handwashing behavior severalfold, when compared to are more rapid. structured observation (Manun’Ebo et al. 1997; Stanton et al. 1987; Biran et al. 2008; ICDDR,B 2008; Danquah An important characteristic of a rapid handwashing mea- 2010). The presence of water at the handwashing place sure is the ease of data collection—for example, through an used after defecation has been associated with observed interview or rapid observation of a household environment. handwashing with soap during structured observation Such a measure would be logistically and financially feasible (Luby et al. 2009). In India, observed soap use during a www.wsp.org 1 Validity of Rapid Measures of Handwashing Behavior Introduction handwashing demonstration was associated with observed in sub-Saharan Africa and Latin America. Analysis of data soap use after fecal contact during structured observation from multiple countries based on a common set of study (Biran et al. 2008). methods and instruments would address concerns about the comparability of studies from different contexts. There- Most community-level assessments of the validity of hand- fore, this study sought to validate rapid handwashing washing behavior measures performed in low- and middle- measures against observed handwashing behavior through income settings were conducted in a handful of countries, structured observation in three countries—Peru, Senegal, mostly in South Asia. This geographical focus limits the and Vietnam. The opportunity to analyze data on similarly generalizability of the existing evidence on rapid hand- measured handwashing measures from three countries was washing measures. Validation studies using data from other provided by the Impact Evaluation of Global Scaling Up sites would inform practitioners and researchers needing Handwashing, which was initiated in 2006 by the Water to use rapid handwashing measures elsewhere, for example and Sanitation Program (WSP). 2 Global Scaling Up Handwashing II. Methods WSP has led an intensive randomized controlled design to Participant selection, adherence to human subject research evaluate the impacts of at-scale handwashing promotion on guidelines, and general data collection methods are de- health, growth, household productivity, and handwashing scribed in detail in the individual country reports from behavior. The handwashing promotion interventions de- the Impact Evaluation (Chase and Do 2012; Galiani et al. ployed in the Global Scaling Up countries were based on 2012). Only endline data from control populations were the FOAM framework (Focus, Opportunity, Ability, and used for these analyses. Motivation), which has been described in detail elsewhere (Coombes and Devine 2010). As part of the Impact Evalu- Table 1 describes the various measures of handwashing be- ation of Global Scaling Up Handwashing, handwashing be- havior, the data collection method, and the levels at which havior was measured. A common study methodology was the data relevant to handwashing behavior were collected. developed to measure handwashing and largely similar data In brief, handwashing was measured using self-reports of collection instruments were deployed across the Impact handwashing at critical times, rapid observations of hand- Evaluation countries. This analysis included data from the washing materials in the home, respondent hand cleanli- Impact Evaluation of Global Scaling Up Handwashing on ness, and, in a subset of households, structured observations handwashing behavior measured in control populations in of handwashing behaviors at critical times. endline surveys in Peru, Senegal, and Vietnam. TABLE 1: MEASURES OF HANDWASHING BEHAVIOR ASSESSED IN GLOBAL SCALING UP HANDWASHING, BY METHOD AND LEVEL OF DATA COLLECTION Method of Data Indicator Collection Level of Data Collection Measures of observed handwashing behavior (basis of comparison for validation of rapid handwashing measures) Handwashing with soap at any type of event Structured observation Individual—caregivers Handwashing with soap after fecal contact Structured observation Individual—caregivers Handwashing with soap before food preparation Structured observation Individual—caregivers Handwashing with soap before feeding a child Structured observation Individual—caregivers Handwashing with soap before eating Structured observation Individual—caregivers Rapid handwashing measures Presence of soap anywhere in the home Rapid observation Household Presence of soap and water at the handwashing place used after Rapid observation Household defecation Presence of soap and water at the handwashing place used before Rapid observation Household food preparation Distance between toileting place and handwashing place Rapid observation Household Distance between food preparation place and handwashing place Rapid observation Household (continued) www.wsp.org 3 Validity of Rapid Measures of Handwashing Behavior Methods TABLE 1: (Continued) Method of Data Indicator Collection Level of Data Collection Time taken to show soap upon request Rapid observation Household Cleanliness index of caregiver hands (index based on observation Rapid observation Individual—caregiver of nails, palms, and fingerpads, dichotomized with score Ͻ7 considered “not clean” and score Ն7 considered “clean”) Handwashing with soap after fecal contact during previous day Self-report Individual—caregiver Handwashing with soap before food preparation during previous Self-report Individual—caregiver day Handwashing with soap before feeding a child during previous day Self-report Individual—caregiver Handwashing with soap before eating during previous day Self-report Individual—caregiver During a household visit for the endline survey, the sur- included in the endline cross-sectional surveys. A survey team vey team asked the household head or appropriate designee member carried out a five-hour observation to record details of to describe household-level characteristics, including de- handwashing practices. The observer recorded opportunities mographic details of household members, access to water for handwashing, hereafter referred to as events, and whether supply and sanitation, and handwashing facilities. The in- and how hands were washed and dried at those times. terviewer asked whether the household members typically wash their hands after defecation or before food handling In Vietnam, only the caregiver of the youngest child under and observed the fixed location where hands were report- 5 years old was recruited per household, whereas in Senegal, edly washed and materials at that location. An interviewer multiple caregivers were recruited if present and consent- sat privately with the caregiver of a child under the age of ing. In Peru, a small number of households were found to 5 and asked about the caregiver’s handwashing behavior, have multiple caregivers. and inspected the caregiver’s hands for cleanliness (palm, fingerpads, and fingernails). Structured observations were carried out in a randomly se- lected subset of households in each country because it was infeasible to carry them out in the entire set of households 4 Global Scaling Up Handwashing III. Data Analysis Detailed definitions of the handwashing measures are pro- errors were calculated using robust error variances in the vided in Annex 1: Supplemental Tables (Table S1). log-binomial regression models. The events observed dur- ing structured observation were divided into four catego- All data analysis was conducted at the country level. Data ries of critical events: fecal contact, food preparation, child from the three countries included in this analysis were not feeding, and eating. Fecal contact included defecation or aggregated. toilet use of the caregiver, as well as contact with child feces. Agreement between the various rapid handwashing mea- In Vietnam, only the primary caregiver was interviewed sures described in Table 1 was evaluated using kappa scores. in each household. In Peru and Senegal, when multiple For example, agreement between self-report of handwash- caregivers were present in a participating household, each ing with soap after defecation with observation of soap at caregiver was interviewed. For the purposes of this analy- the handwashing place near the toilet was assessed. Kappa sis, it was not possible to ascertain which of the caregiv- is considered a measure of “true agreement,” in that it de- ers was observed during any single event in the structured scribes agreement taking into account the agreement that observation. Thus, the datasets of rapid handwashing mea- would be expected to occur by chance. Because kappa scores sures and structured observations were restricted. In Peru, can be underestimated when the prevalence of one or more structured observation events occurring in households with of the conditions under study is high or low, prevalence- multiple caregivers were excluded from the analyses. In adjusted kappa scores were also calculated using previously Senegal, given the potential for a large loss in the sample described methods (Byrt et al. 1993). A number of authors if households with two or three caregivers were removed, have recommended different cut-offs for interpretation of those events occurring in households with more than three kappa scores, although many are overlapping (Szklo and caregivers were excluded and a random number generator Nieto 2007). This study used Altman’s cutoffs: poor agree- in SAS version 9.2 (a commonly used statistical program) ment (k Ͻ 0.2), fair (k ϭ 0.2 to Ͻ 0.4), moderate (k ϭ 0.4 was used to randomly choose one caregiver’s responses in to Ͻ0.6), good (k ϭ 0.6 to 0.8), and very good (k Ն 0.8 households with two or three caregivers listed. For the principal study objective, to validate rapid hand- The relationship between each of the rapid handwashing washing measures against observed handwashing behavior, measures and observed handwashing behavior was evalu- the analysis was restricted to those households with data ated using regression models for all events observed, as well from a structured observation. Data on handwashing be- as events restricted to fecal contact. Unadjusted risk ratios havior among primary caregivers was examined, and equiv- were estimated and adjusted risk ratios were calculated alence between households with structured observation using log-binomial regression models, including wealth as a data and those without was assessed. A dataset was then covariate, given the frequent description of associations be- constructed in which each event observed during the struc- tween wealth and measures of handwashing behavior (Luby tured observation was included as a record. For example, if and Halder 2008; Ram et al. 2010); within each country 10 events were observed in household Y, the dataset con- dataset, a wealth index was created using principal com- tained 10 records associated with household Y. Multilevel ponent analysis of ownership of assets such as radio and log-binomial regression was used to model the relationship television. Exploratory analyses evaluating water scarcity, between the rapid handwashing measure (independent distance to water source, and location of toilet in household variable) and the probability that hands were washed dur- or yard indicated that these variables did not act as potential ing the observed event (dependent variable). Because of the confounders (data not shown). repeated nature of structured observation data, standard www.wsp.org 5 IV. Results Table 2 describes the timelines of endline data collection among the three countries. Only one set of measures was and sample sizes for control arms, by country. In all, end- found to have moderate or greater agreement in all three line data was collected from 1,368 Peru households, 757 countries: Senegal households, and 1,105 Vietnam households. • Soap and water observed at the handwashing place used postdefecation, and soap and water observed at Handwashing with soap after fecal contact was reported by the handwashing place used before preparing food about two thirds of caregivers in Peru and Vietnam, and by 0 Peru (0.83) 45 percent of caregivers in Senegal (see Table 3). Soap was 0 Senegal (0.56) present in 83 percent of households in Peru, 90 percent in 0 Vietnam (0.84) Senegal, and 98 percent in Vietnam. Soap and water to- gether were observed at the place used to wash hands after The following sets of measures were found to have moder- defecation in 67 percent of households in Peru, 27 percent ate or greater agreement in both Peru and Vietnam, but not in Senegal, and 82 percent in Vietnam. Soap and water in Senegal: were observed at the place used to wash hands before food • Self-reported handwashing after fecal contact, and preparation in 67 percent of households in Peru, 19 percent soap observed anywhere in the home in Senegal, and 79 percent in Vietnam. Hands were rated 0 Peru (0.46) as clean for a majority of caregivers in all three countries. 0 Vietnam (0.50) • Soap observed anywhere in the home, and soap Agreement between various rapid handwashing measures and water observed at the handwashing place used behavior in each country sample was evaluated (Annex 1, postdefecation Tables S2–S4). Table 4 describes the sets of handwashing 0 Peru (0.68) measures for which there was moderate or greater agreement 0 Vietnam (0.81) TABLE 2: TIMELINE AND SAMPLE SIZE OF CONTROL POPULATIONS IN ENDLINE SURVEYS OF THE IMPACT EVALUATION OF GLOBAL SCALING UP HANDWASHING, BY COUNTRY Endline Peru Senegal Vietnam Dates 2/2011 to 6/2011 3/2011 to 7/2011 10/2010 to 1/2011 Number of households 1,368 757 1,105 Number of caregivers 1,379 1,411 1,064 Number of households with 286 88 200 structured observation data 6 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior Results TABLE 3: DESCRIPTIVE ANALYSIS OF RAPID HANDWASHING MEASURES, ALL COUNTRIES Peru Senegal Vietnam Handwashing Measures N n (%) N n (%) N n (%) Self-reported Handwashing with soap 1,409 930 (66) 1,338 600 (45) 1,064 723 (68) after fecal contact during the previous day Handwashing with soap 1,409 950 (67) 1,338 258 (19) 1,064 333 (31) before food preparation during the previous day Handwashing with soap 1,409 284 (20) 1,338 43 (3) 1,064 388 (36) before feeding a child during the previous day Handwashing with soap 1,409 573 (41) 1,338 311 (23) 1,064 170 (16) before eating during the previous day Rapid observation Presence of soap anywhere in 1,368 1,119 (82) 757 678 (90) 1,064 1,045 the home (Households) (Households) (Households) (98) Presence of soap and water at 1,362 912 (67) 753 213 (28) 1,063 947 (89) the handwashing place used (Households) (Households) (Households) postdefecation Presence of soap and water at 1,367 915 (67) 753 149 (20) 1,064 946 (89) the handwashing place used (Households) (Households) (Households) before food preparation High cleanliness index of caregivers hands* Hands rated clean based on 1,395 919 (66) 1,280 1,026 (80) 1,064 687 (65) observation of nails, palms, and fingerpads * Index based on observation of nails, palms, and fingerpads, dichotomized with score <7 considered “not clean” and score ≥7 considered “clean” www.wsp.org 7 Validity of Rapid Measures of Handwashing Behavior Results TABLE 4: PREVALENCE-ADJUSTED KAPPA SCORES REFLECTING MODERATE OR GREATER AGREEMENT BETWEEN RAPID HANDWASHING MEASURES IN ENDLINE SURVEYS IN PERU, SENEGAL, AND VIETNAM, 2009–2011 Handwashing before Handwashing before Handwashing before Cleanliness Index of Handwashing after Anywhere in Home Caregiver Hands * Food Preparation Soap and Water Soap and Water Observed at the Feeding a Child Soap Observed Postdefecation Place Used to Fecal Contact Handwashing Prepare Food Observed at Place Used Reported Reported Reported Reported Eating All Countries Reported handwashing — Peru after fecal contact (0.46) Vietnam (0.50) Reported handwashing — — Senegal Senegal before food preparation (0.61) (0.41) Reported handwashing — — — Senegal Senegal before feeding a child (0.55) (0.61) Reported handwashing — — — — before eating Soap observed — — — — — Peru Peru Senegal anywhere in home (0.68) (0.68) (0.52) Vietnam Vietnam (0.81) (0.79) Soap and water — — — — — — Peru observed at (0.83) handwashing place Senegal used postdefecation (0.56) Vietnam (0.84) Soap and water — — — — — — — observed at the place used before preparing food Cleanliness index of — — — — — — — — caregiver hands* *Index based on observation of nails, palms and fingerpads, dichotomized with score <7 considered “not clean” and score ≥7 considered “clean” 8 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior Results • Soap observed anywhere in the home, and soap and 444 observed in Senegal, and 1,421 observed in Vietnam water observed at the handwashing place used before (Table  5). Overall, primary caregivers washed hands with preparing food soap at a minority of all events observed in Peru (14 per- 0 Peru (0.68) cent), Senegal (13 percent), and Vietnam (10 percent). 0 Vietnam (0.79) Events of the following types were designated as critical events, because of their potential relevance to pathogen For the analyses of validity of rapid handwashing measures, transmission to or from hands: fecal contact, food prepara- data from four households, each with two caregivers, was tion, eating, or feeding a child. For the critical events of in- removed in Peru. In Senegal, where households had two or terest, handwashing with soap was observed in the minority more caregivers, 11 households were removed, each with (Table 6). For example, handwashing of any kind (with or more than three caregivers listed. without soap) ranged from 61 percent to 74 percent after fecal contact events, but handwashing with soap was ob- Structured observations were completed in 286 (21 per- served at only 24 to 34 percent of such events. Soap use cent), 88 (12 percent), and 200 (18 percent) households was less frequently observed before food preparation (7 to in Peru, Senegal, and Vietnam, respectively. Differences in 8 percent), eating (6 to 14 percent), and feeding events (4 household- and caregiver-level characteristics among house- to 9 percent). holds with and without structured observation data were evaluated (Table S5). In Peru, households with structured Table 6 details the associations between rapid handwash- observations had a toilet located in the household or yard ing measures and observed handwashing behavior in Peru, and reported frequent scarcity of water at the source more Senegal, and Vietnam. In all three countries, caregivers often than households without structured observation; who lived in homes with observed soap and water at the however, the differences were relatively minor. There were handwashing place used after defecation were more likely no statistically significant differences between caregivers to be observed washing hands with soap than caregivers liv- with and without structured observations in Peru. In Sen- ing in homes without soap and water at that place: Peru egal, there were statistically significant differences between (RR ϭ 1.59, 95% CI ϭ 1.11–2.28), Senegal (RR ϭ 2.63, households with and without structured observation data. 95% CI ϭ 1.36–5.10), and Vietnam (RR ϭ 3.61, 95% Households with structured observation data had lower CI ϭ 1.53–8.50). Similarly, caregivers were more likely wealth scores and were less likely to have improved sani- to be observed washing hands with soap if soap and water tation than their counterparts. In addition, differences in were observed at the place where hands are washed before caregiver hand cleanliness score, self-reported handwashing food preparation: Peru (RR ϭ 2.02, 95% CI ϭ 1.40–2.92), with soap after fecal contact, and self-reported handwash- Senegal (RR ϭ 2.44, 95% CI ϭ 1.11–5.34), and Vietnam ing before eating suggested poorer handwashing behavior (RR  ϭ 2.20, 95% CI ϭ 1.14–4.24). Also, observation of in the structured observation group than in the group with- soap at the place used after defecation, irrespective of the out structured observations in Senegal. In Vietnam, there presence of water, was associated with observed handwash- were statistically significant differences between the two ing with soap in Senegal and Vietnam. Observation of soap groups in possession of a refrigerator, type of fuel used for at the place used to wash hands before food preparation was cooking, and highest education level attained, indicating a associated with observed handwashing with soap in Peru somewhat higher socioeconomic status among households and Vietnam, but not in Senegal. In Senegal, soap retrieval with structured observation than those without structured within 60 seconds was found to be strongly associated with observation. observed handwashing behavior (RR ϭ 8.21, 95% CI ϭ 1.68–40.08). However, in Vietnam, the inverse association Data on handwashing behavior among primary caregivers was found, such that primary caregivers living in households was available for 278 households in Peru, 77 households in which soap retrieval occurred within 60 seconds were in Senegal, and 199 households in Vietnam. Among pri- less likely to be observed washing hands with soap. In Viet- mary caregivers, there were 1,467 events observed in Peru, nam alone, self-reported handwashing after fecal contact or www.wsp.org 9 Validity of Rapid Measures of Handwashing Behavior Results TABLE 5: FREQUENCY OF OBSERVATION OF CRITICAL EVENTS FOR HANDWASHING, AND HANDWASHING BEHAVIOR, DURING ENDLINE STRUCTURED OBSERVATIONS, BY COUNTRY, 2011 Observation Peru* Senegal# Vietnam Number of structured observations completed 278 77 199 Number of events observed among all household members 2,911 1,742 2,416 Number of critical events among all household members 1,627 966 1,507 Number of events observed among primary caregivers 1,467 444 1,421 Number of events among primary caregivers when hands were washed with or 725 (50) 189 (43) 961 (68) without soap (% of events) Number of events among primary caregivers when hands were washed with soap 207 (14) 52 (13) 136 (10) (% of events) Number of critical events observed among primary caregivers 725 266 858 Number of critical events among primary caregivers when hands were washed 431(59) 112 (42) 318 (37) with or without soap (% of events) Number of critical events among primary caregivers when hands were washed 106 (15) 29 (11) 103 (12) with soap (% of events) Number of fecal contact events among primary caregivers 168 47 289 Number of fecal contact events among primary caregivers after which hands were 103 (61) 35 (74) 182 (63) washed with or without soap (% of fecal contact events) Number of fecal contact events among primary caregivers after which hands were 57 (34) 12 (25) 69 (24) washed with soap (% of fecal contact events) Number of food preparation events among primary caregivers 361 70 181 Number of food preparation events among primary caregivers before which hands 258 (71) 25 (36) 60 (33) were washed with or without soap (% of food preparation events) Number of food preparation events among primary caregivers before which hands 27 (7) 5 (7) 14 (8) were washed with soap (% of food preparation events) Number of eating events among primary caregivers 92 97 125 Number of eating events among primary caregivers before which hands were 40 (43) 37 (38) 31 (25) washed with or without soap (% of eating events) Number of eating events among primary caregivers before which hands were 13 (14) 10 (10) 8 (6) washed with soap (% of eating events) Number of feeding events among primary caregivers 104 52 263 Number of feeding events among primary caregivers before which hands were 30 (29) 15 (29) 45 (17) washed with or without soap (% of feeding events) Number of feeding events among primary caregivers before which hands were 9 (9) 2 (4) 12 (5) washed with soap (% of feeding events) *Excludes structured observations in households with more than one primary caregiver #Excludes structured observations in households with more than three primary caregivers 10 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior Results before feeding a child, and observed hand cleanliness, were clean to be observed washing hands with soap after fecal each associated with observed handwashing with soap at contact in Peru (RR ϭ 2.20, 95% CI ϭ 1.67–4.16) and any observed event. Distance of the handwashing location Vietnam (RR ϭ 2.63, 95% CI ϭ 1.40–4.95). The only from either the latrine or the food preparation place was not other measure significantly associated with observed hand- associated with observed handwashing with soap. washing with soap after fecal contact was self-reported handwashing with soap after defecation, a finding only de- In Peru and Vietnam, adjustment for wealth did not alter tected in Vietnam (RR ϭ 3.33, 95% CI ϭ 1.45–7.67). the associations between the rapid handwashing measures and observed handwashing with soap. In Senegal, none of The extremely low levels of soap use for handwashing at the rapid handwashing measures were significantly associ- food preparation, eating, and feeding events during struc- ated with observed handwashing with soap in models in- tured observation made it impossible to estimate the asso- cluding wealth. ciation between rapid handwashing measures and observed handwashing with soap at these critical times. Caregivers whose hands were not observed to be clean were less likely than caregivers whose hands were noted to be TABLE 6: ASSOCIATIONS BETWEEN RAPID HANDWASHING MEASURES AND OBSERVED HANDWASHING BEHAVIOR, ENDLINE SURVEYS AND STRUCTURED OBSERVATIONS IN PERU, SENEGAL, AND VIETNAM, 2009–2011 Associations between rapid handwashing measures behavior and observed handwashing Peru N ϭ 1,446§ events Senegal N ϭ 435§ events Vietnam N ϭ 1,410§ events Unadjusted Adjusted* Unadjusted Adjusted* Unadjusted Adjusted* Measure of Handwashing Relative Risk Relative Risk Relative Risk Relative Risk Relative Risk Relative Risk Behavior (95% CI**) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Structured observation of handwashing behavior at any event compared to: Self-report of handwashing with soap in the last 24 hours: after fecal contact 0.93 0.92 1.83 1.24 2.94 2.93 (0.67–1.29) (0.66–1.29) (0.84–4.00) (0.65–2.40) (1.53–5.64)*** (1.53–5.64) before preparing food 1.31 1.34 1.35 1.22 0.97 0.97 (0.92–1.87) (0.94–1.92) (0.59–3.10) (0.63–2.34) (0.60–1.58) (0.59–1.58) before feeding a child 0.78 0.74 undefined# undefined# 2.22 2.22 (0.50–1.20) (0.47–1.16) (1.45–3.41) (1.45–3.41) before eating 1.37 1.39 1.06 1.76 1.24 1.24 (0.98–1.90) (0.99–1.93) (0.35–3.28) (0.64–4.83) (0.72–2.13) (0.72–2.14) Rapid observation of: soap in the home 1.15 1.14 4.23 2.05 undefined# undefined# (0.80–1.65) (0.79–1.63) (0.58–31.16) (0.31–13.74) (continued) www.wsp.org 11 Validity of Rapid Measures of Handwashing Behavior Results TABLE 6: (Continued) Peru N ϭ 1,446§ events Senegal N ϭ 435§ events Vietnam N ϭ 1,410§ events Unadjusted Adjusted* Unadjusted Adjusted* Unadjusted Adjusted* Measure of Handwashing Relative Risk Relative Risk Relative Risk Relative Risk Relative Risk Relative Risk Behavior (95% CI**) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Handwashing place used after defecation soap observed 1.27 1.23 3.02 1.54 3.29 3.32 (0.87–1.87) (0.86–1.86) (1.54–5.93) (0.69–3.41) (1.42–7.61) (1.44–7.69) water observed 1.81 1.79 2.37 1.13 undefined# undefined# (1.06–3.07) (1.05–3.05) (1.08–5.20) (0.59–2.19) soap and water 1.59 1.58 2.63 1.19 3.61 3.64 observed (1.11–2.28) (1.10–2.26) (1.36–5.10) (0.63–2.26) (1.53–8.50) (1.55–8.58) handwashing station 1.00 1.01 2.31 1.19 1.23 1.26 Յ3 meters from (0.72–1.38) (0.72–1.40) (1.01–5.30) (0.52–2.72) (0.79–1.93) (0.78–2.04) latrine Handwashing place used before food-related event soap observed 1.68 1.67 2.47 1.40 (0.78 2.03 2.03 (1.12–2.53) (1.11–2.22) (1.17 – 5.20) – 2.52) (1.07 – 3.84) (1.07 – 3.86) water observed 2.16 2.14 1.20 1.20 2.47 2.52 (1.01–4.61) (0.99–4.36) (0.56–2.56) (0.68–2.12) (0.73–8.33) (0.73–8.68) soap and water 2.02 2.01 2.44 1.49 2.20 2.20 observed (1.40–2.92) (1.38–2.91) (1.11–5.34) (0.84–2.64) (1.14–4.24) (1.14–4.26) handwashing station 0.81 0.82 2.29 2.02 0.95 0.95 Յ3 meters from food (0.57–1.16) (0.57–1.18) (0.93–5.61) (0.86–4.74) (0.62–1.47) (0.62–1.48) preparation place Soap retrieved in Յ60 0.60 0.50 8.21 6.84 0.20 0.16 seconds (0.29–1.24) (0.23–1.08) (1.68–40.08) (1.16–40.26) (0.06–0.63) (0.03–0.83) Hand cleanliness 1.09 1.08 4.11 1.76 2.75 2.75 index Ն7 (0.77–1.54) (0.77–1.53) (0.93–18.15) (0.37–8.37) (1.66 – 4.55) (1.66–4.55) Structured observation of handwashing behavior after fecal contact event compared to: Self-report of hand washing with soap in the last 24 hours after defecation 1.03 1.02 3.11 3.33 3.32 (0.65–1.63) (0.64–1.61) (0.99–9.63) (1.45–7.67) (1.44–7.65) Rapid observation of: soap in the home 0.92 0.93 1.02 undefined# undefined# (0.52–1.63) (0.52–1.66) (0.17–6.16) 12 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior Results TABLE 6: (Continued) Peru N ϭ 1,446§ events Senegal N ϭ 435§ events Vietnam N ϭ 1,410§ events Unadjusted Adjusted* Unadjusted Adjusted* Unadjusted Adjusted* Measure of Handwashing Relative Risk Relative Risk Relative Risk Relative Risk Relative Risk Relative Risk Behavior (95% CI**) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Handwashing place used after defecation soap observed 0.89 0.91 2.45 5.60 5.53 (0.52–1.78) (0.53–1.55) (0.83–7.24) (0.85–37.04) (0.84–36.49) water observed 1.05 1.07 1.60 undefined# undefined# (0.53–2.07) (0.54–2.12) (0.48–.29) soap and water 1.03 1.05 2.08 6.44 6.37 observed (0.63–1.70) (0.64–1.73) (0.74–5.90) (0.96–43.21) (0.95–42.67) handwashing station 1.26 1.25 2.50 1.45 1.43 Յ3 meters from (0.76–2.04) (0.76–2.05) (0.65–9.63) (0.90–2.34) (0.88–2.34) latrine Hand cleanliness 2.20 2.21 2.63 2.68 index Ն7 (1.67–4.16) (1.17–4.17) (1.40–4.95) (1.43–5.00) * Adjusted for wealth index score ** CI: confidence intervals *** Associations shown in bold significant at p < 0.05 # Undefined due to zero observations in some cells § Total events observed less those in which information on handwashing was missing (21 events in Peru, 9 events in Senegal, and 11 events in Vietnam) www.wsp.org 13 V. Discussion There is interest globally in improving handwashing be- soap, were much less common at critical times other than havior (http://www.globalhandwashingday.org). However, fecal contact (Table 5), making it difficult to estimate as- there is a paucity of information on handwashing behavior sociations between rapid handwashing measures and ob- from many low- and middle-income countries, with most served handwashing with soap at these critical times. prior studies carried out in the South Asian subcontinent. Hence, there is a substantial unmet need to estimate hand- Handwashing with Soap Is Infrequent in All Three washing behavior globally using a methodology that is valid Countries for measuring handwashing behavior across cultural and Low rates of handwashing with soap were observed. Only geographic contexts. This multicountry analysis sought one third of fecal contact events in Peru and one quarter in to describe handwashing behavior in the absence of hand- both Senegal and Vietnam were followed by handwashing washing promotion, and to validate rapid handwashing with soap. Soap use was substantially lower at other critical measures against observed handwashing behavior through times, when pathogens can be transmitted from hands to structured observation in three countries—Peru, Senegal, food, or one’s own mouth, or the mouth, mucosa, or hands and Vietnam. Improving handwashing behavior remains a of a child. These low frequencies of handwashing prevented priority for each of these three countries, where handwash- identification of which rapid handwashing measures serve ing with soap is practiced during only a minority of critical as reasonable proxies of handwashing at critical times other times when pathogens can be transmitted to or from hands. than fecal contact. More importantly, these data reinforce The agreement between the various measures was evalu- the substantial opportunity and imperative to improve ated, and the extent to which each rapid handwashing mea- handwashing with soap to decrease child diarrhea and re- sure was associated with observed handwashing with soap, spiratory infections (Curtis et al. 2009). as measured by structured observation of behavior, was assessed. The findings reported here reinforce the impor- Water and Soap Together at Locations Designated tance of using objective measures of handwashing, rather for Handwashing Associated with Observed than simply asking respondents to describe their own be- Handwashing havior. Observation of materials at designated handwashing The findings reported here confirm the work of Luby and locations yielded valid and internally consistent measures of colleagues, who found in Bangladesh that having soap at the handwashing with soap overall. Observed hand cleanliness place to wash hands after toileting, and having water at that is promising as a proxy measure for handwashing at the spe- place, were independently associated with observed hand- cific critical time of fecal contact. Multicountry evidence of washing with soap after fecal contact (Luby et al. 2009). association with observed behavior provides a strong basis The analysis of baseline data from the Impact Evaluation for the use of rapidly observed measures as proxies of hand- of Global Scaling Up Handwashing in Peru similarly found washing behavior when structured observation is infeasible. that having soap and water together at one or more des- Therefore, the findings of associations between the presence ignated handwashing places was associated with observed of soap and water at designated handwashing locations, and soap use following at least one fecal contact event (Ram observed handwashing behavior, in multiple countries af- et al. 2014). firm the recent inclusion of observations of handwashing locations and materials into the standard modules of both The majority of households in each country had soap present the Demographic and Health Surveys (DHS) and Multiple somewhere in the home. Soap is a highly valued commodity Indicator Cluster Surveys (MICS). in many low-income settings and its use may be limited in order to keep household expenditures low. Also, soap present No measures that were associated with observed handwash- in a home may be used for a number of purposes, includ- ing at specific critical times, other than fecal contact, were ing bathing, dishwashing, and laundry. In contrast to soap identified. Handwashing overall, and handwashing with presence anywhere in the home, soap kept at a designated 14 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior Discussion handwashing place suggests a prioritization of the product at children’s hands. Only hand cleanliness was associated for handwashing. Certainly, the same location may be used with handwashing with soap after fecal contact in both Peru for washing hands as for washing dishes or clothes. But the and Vietnam, suggesting that it is worthwhile to continue convenience of washing hands may be substantially increased to explore the use of this measure in some contexts. It is by having all the materials needed to carry out the behavior important to better understand why hand cleanliness obser- at the location where hands need to be washed (Kamm et al. vation is incongruous with observed handwashing behavior 2011). Individuals leaving a latrine may be more likely to in Senegal. It is possible that differentiating gradations of wash hands if the soap and water are consistently available cleanliness by visual inspection may be more difficult with at a location in or near the latrine; if they have to fetch soap the darker skin complexions often found in sub-Saharan from the home either before going into the latrine or after Africa, compared to relatively lighter skin complexions coming out of it, they may not remember, or may not feel more commonly found elsewhere. they have sufficient time to carry out the behavior. Self-Reported Measures Not Consistently Associated Curtis describes habit as a learned, automated behavior that with Observed Handwashing is reinforced by cues (Curtis et al. 2011). The presence of Self-reported measures have been shown to overestimate soap and water at a location commonly used for handwash- observed handwashing behavior in numerous countries ing may provide a visual cue, an immediate prompt to the (Manun’Ebo et al. 1997; Stanton et al. 1987; Biran et al. behavior at critical times (e.g., a handwashing station vis- 2008; ICDDR,B 2008; Danquah 2010; Byrt et al. 1993; ible near the latrine for use after defecation). Consistently Sim and Wright 2005; Curtis et al. 2009) so the finding of a maintaining the materials needed for handwashing at the lack of consistent association between self-reported and ob- same location, and thus providing visual cues tied to the site served handwashing across countries is not novel. The analy- of hand contamination, may foster a handwashing habit. sis presented here underscores previously expressed concerns about using self-report as the sole approach to measuring The finding of the lack of an association between proximity handwashing behavior. Alternatives to measuring event- of the handwashing location and observed handwashing be- specific handwashing behavior by self-report include struc- havior is notable. It warrants further inquiry into the extent tured observation, video observation, or sensor-based tech- to which a proximal location is required for handwashing to nologies; all of these can be intrusive, time-consuming, occur habitually at critical times. Perhaps habit can be formed personnel-intensive, and costly (Ram 2013). For settings in as long as the location is fixed, and the cues and convenience which such intensive resources are not available, it is impor- are preserved, even if the location is not immediately inside tant to validate similarly the use of other questionnaire-based or next to the place where hands must be washed. Not only approaches to measuring event-specific handwashing behav- does this finding imply that investigators need not invest ior, including Likert-scale questions (e.g., do you always, time in measuring distance between the site where the critical sometimes, or never wash your hands after defecation), as event occurs (e.g., latrine, cooking area) and the site where well as indices indicating a handwashing habit (Aunger et al. handwashing takes place, but it also suggests that individuals 2010; Stevenson et al. 2009; Verplanken and Orbell 2003). may wash their hands consistently as long as they maintain the necessary materials in a fixed location that is somewhat Wealth and Handwashing Behavior more distant, if they cannot set up a handwashing station im- Prior studies have demonstrated the important relationship mediately next to a latrine or food preparation area. between wealth and soap availability in the home, as well as observed handwashing behavior (Luby and Halder 2008; Observed Hand Cleanliness Ram et al. 2010). Compared to poor households, wealthier In Bangladesh, cleanliness of children’s hands has been households may be able to purchase soap more regularly, shown to be associated with reduced diarrhea risk (Luby et al. may be able to prioritize the use of soap for handwashing 2011a), whereas mothers’ hand cleanliness has not. This as opposed to other purposes, or may be more aware of the study evaluated mother’s hand cleanliness but did not look health benefits or the social desirability of handwashing. www.wsp.org 15 Validity of Rapid Measures of Handwashing Behavior Discussion Whereas in Senegal, wealth was an important confounder Second, structured observations were carried out in only a for the relationship between observed handwashing materi- relatively small subset of households in each country. The als and observed handwashing behavior, the associations in subsets may not have been representative of the larger pop- models including wealth were stable in Peru and Vietnam. ulations of the Impact Evaluation countries, as noted by This lack of consistency in the effects of wealth suggests the differences between households taking part in structured need to better understand the potentially variable role of observation and those not taking part (Table 6). wealth in influencing access to soap in different cultural or geographical contexts. Third, small numbers of certain critical events, such as eat- ing and feeding a child, were observed, making it impossi- Challenge to Validating Rapid Handwashing Measures ble to determine which rapid handwashing measures might Using Observed Handwashing be associated with handwashing at the specific times when Given emerging data on the association between observed pathogens may be transmitted to or from hands. To ad- handwashing behavior and improved health outcomes, ob- dress the small numbers of specific critical events observed, served handwashing behavior serves as a reasonable basis of events consisting of potential contact with one’s own feces evaluating rapid handwashing measures (Luby et al. 2011b). (defecation) were combined with events of potential con- However, there are important concerns about reactivity, the tact with a child’s feces (cleaning a child who has defecated). extent to which individuals behave in their usual way dur- This approach of group critical times related from a micro- ing a structured observation, when an outsider is present bial transmission perspective may have ignored behavioral in their home or compound (Ram et al. 2010; Cousens realities. For example, latrine use events were combined et al. 1996). The overarching goal of most handwashing with cleaning the anus of a child who had defecated into a programs is neither cosmetic nor social, but rather to re- single “fecal contact.” Handwashing behaviors may differ duce the burden of preventable infections, particularly in based on the individual’s perception of disgust associated children. It is important, therefore, that rapid handwashing with the feces; in many cultures, there is substantial disgust measures be further validated using health outcomes when associated with touching one’s own feces but somewhat less feasible (Luby et al. 2011a). with touching children’s feces. Ideally, a larger number of specific critical events would have allowed the investigation Limitations of the extent to which rapid handwashing measures were as- Apart from reactivity to structured observation, which is sociated with observed handwashing at each type of event, described above, there are several limitations to the data and acknowledging the unique drivers of behavior at each type analyses presented in this article. First, although data col- of event when pathogens may be transmitted to and from lection instruments and training guidance documents were hands. designed centrally, with only minor adaptations, training of survey teams was carried out by different investigators Conclusions within each country, potentially leading to differences in This multicountry analysis has shown that observation of survey administration and data collection that could have handwashing materials at the places where people wash affected the estimates of agreement and validity. Still, the hands, at the times most necessary for washing (after fecal common survey tools used in each country allowed for contact and before food preparation), is a valid measure of greater comparison across countries than is typically pos- handwashing with soap in multiple cultural and geographic sible when comparing disparate studies, often carried out contexts. There continues to be an overarching need for de- by entirely different research teams. Moreover, to be widely veloping valid measures of handwashing behavior that can applicable, handwashing measures must withstand applica- be collected in an efficient and inexpensive fashion. The tion by investigators of various skill levels. To that end, the structured observation data indicating low rates of soap use validity of observations of soap and water at a handwashing for handwashing at times of pathogen transmission reinforce place as appropriate proxies for observed handwashing be- the global imperative to improve handwashing behavior for havior is strengthened. prevention of the leading causes of death in young children. 16 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior References References Aiello, A. E., R. M. Coulborn, V. Perez, and E. L. Larson. Danquah, L. 2010. “Measuring Hand Washing Behaviour: 2008. “Effect of Hand Hygiene on Infectious Disease Methodological and Validity Issues.” Paper presented Risk in the Community Setting: A Meta-Analysis.” at South Asia Hygiene Practitioners’ Workshop. Am J Public Health 98 (8): 1372–1381. Dhaka, Bangladesh. Aunger, R., W. P. Schmidt, A. Ranpura, et al. 2010. Ejemot, R. I., J. E. Ehiri, M. M. Meremikwu, and “Three Kinds of Psychological Determinants for J. A. Critchley. 2008. “Hand Washing for Preventing Hand-washing Behaviour in Kenya.” Soc Sci Med 70 Diarrhoea.” Cochrane Database Syst Rev 1, Art. No.: (3): 383–391. CD004265. DOI: 10.1002/14651858.CD004265. Bentley, M., M. Boot, J. Gittelsohn, and R. Stallings. pub2. 1994. The Use of Structured Observations in the Study of Galiani, S., P. Gertler, and A. Orsola-Vidal. 2012. “Pro- Health Behaviour. The Hague, The Netherlands: IRC moting Handwashing Behavior in Peru: the Effect of International Water and Sanitation Centre. Large-Scale Mass Media and Community Level Inter- Biran, A., T. Rabie, W. Schmidt, S. Juvekar, S. Hirve, and ventions.” World Bank Policy Research Working Paper V. Curtis. 2008. “Comparing the Performance of In- Series No. 6257. Washington, DC: World Bank. dicators of Hand-washing Practices in Rural Indian ICDDR,B. 2008. “Handwashing Behavior in Rural Ban- Households.” Trop Med Int Health 13 (2):278–285. gladesh.” Health and Science Bull 6 (3). http://www. Byrt, T., J. Bishop, and J. B. Carlin. 1993. “Bias, Preva- icddrb.org/what-we-do/publications/cat_view/52- lence and Kappa,” J Clin Epidemiol 46 (5): 423–429. publications/10042-icddrb-periodicals/10048-health- and-science-bulletin-bangla-and-english/10271-vol-6- Chase, C., and Q.-T. Do. 2012. “Handwashing Behavior no-3-english-2008/11786-handwashing-behavior-in- Change at Scale: Evidence from a Randomized Evalu- -rural-bangladesh. ation in Vietnam.” World Bank Policy Research Work- ing Paper Series No. 6207. Washington, DC: World Kamm, K. B., D. R. Feikin, G. Bigogo, et al. 2011. “As- Bank. sociations with Handwashing in the Home and Re- spiratory and Diarrheal Illness in Children under Five Coombes, Y., and J. Devine. 2010. “Introducing FOAM: Years Old in Rural Western Kenya.” Paper presented A Framework to Analyze Handwashing Behaviors to at American Society of Tropical Medicine and Hy- Design Effective Handwashing Programs.” Washing- giene. Philadelphia, Pennsylvania. ton, DC: Water and Sanitation Program, World Bank. Luby, S. P., M. Agboatwalla, D. R. Feikin, et al. 2005. “Ef- Cousens, S., B. Kanki, S. Toure, I. Diallo, and V. Curtis. fect of Handwashing on Child Health: A Randomised 1996. “Reactivity and Repeatability of Hygiene Be- Controlled Trial.” Lancet 366 (9481): 225–233. haviour: Structured Observations from Burkina Faso.” Soc Sci Med 43 (9): 1299–1308. Luby, S. P., M. Agboatwalla, R. M. Hoekstra, M. H. Rah- bar, W. Billhimer, and B. H. Keswick. 2004. “Delayed Curtis, V., S. Cousens, T. Mertens, E. Trarore, B. Kanki, Effectiveness of Home-Based Interventions in Reduc- and I. Diallo. 1993. “Structured Observations of Hy- ing Childhood Diarrhea, Karachi, Pakistan.” Am J giene Behaviors in Burkina Faso: Validity, Variability, Trop Med Hyg 71 (4): 420–427. and Utility,” Bull World Health Organ 71 (1): 23–32. Luby, S. P., and A. K. Halder. 2008. “Associations among Curtis, V., W. Schmidt, S. Luby, R. Florez, O. Toure, and Handwashing Indicators, Wealth, and Symptoms of A. Biran. 2011. “Hygiene: New Hopes, New Hori- Childhood Respiratory Illness in Urban Bangladesh.” zons.” Lancet Infect Dis 11 (4): 312–321. Trop Med Int Health 13 (6): 835–844. Curtis, V. A., L. O. Danquah, and R. V. Aunger. 2009. Luby, S. P., A. K. Halder, T. M. Huda, L. Unicomb, “Planned, Motivated and Habitual Hygiene Behav- and R. B. Johnston. 2011a. “Using Child Health iour: An Eleven Country Review.” Health Educ Res 24 (4): 655–673. www.wsp.org 17 Validity of Rapid Measures of Handwashing Behavior References Outcomes to Identify Effective Measures of Hand- Rhee, V., L. C. Mullany, S. K. Khatry, et al. 2008. “Mater- washing.” Am J Trop Med Hyg 85 (5): 882–892. nal and Birth Attendant Hand Washing and Neonatal Luby, S. P., A. K. Halder, T. Huda, L. Unicomb, and R. B. Mortality in Southern Nepal.” Arch Pediatr Adolesc Johnston. 2011b. “The Effect of Handwashing at Rec- Med 162 (7): 603–608. ommended Times with Water Alone and with Soap Silk, B. J., S. Doshi, D. Dutt, et al. 2010. “Hand Hygiene on Child Diarrhea in Rural Bangladesh: An Observa- and Radiographically-Confirmed Pneumonia among tional Study.” PLoS Med 8 (6): e1001052. Young Children Living in Urban Dhaka, Bangladesh: Luby, S. P., A. K. Halder, C. Tronchet, S. Akhter, Preliminary Results from a Case-Control Study.” A. Bhuiya, and R. B. Johnston. 2009. “Household Paper presented at American Society of Tropical Medi- Characteristics Associated with Handwashing with cine and Hygiene. Atlanta, Georgia. Soap in Rural Bangladesh.” Am J Trop Med Hyg 81 (5): Sim, J., and C. C. Wright. 2005. “The Kappa Statistic in 882–887. Reliability Studies: Use, Interpretation, and Sample Manun’Ebo, M., S. Cousens, P. Haggerty, M. Kalengaie, Size Requirements.” Phys Ther 85 (3): 257–268. A. Ashworth, and B. Kirkwood. 1997. “Measuring Stanton, B. F., J. D. Clemens, K. M. Aziz, and M. Rah- Hygiene Practices: A Comparison of Questionnaires man. 1987. “Twenty-Four-Hour Recall, Knowl- with Direct Observations in Rural Zaire.” Trop Med edge-Attitude-Practice Questionnaires, and Direct Int Health 2 (11): 1015–1021. Observations of Sanitary Practices: A Comparative Ram, P. K., A. K. Halder, S. P. Granger, et al. 2010. “Is Study.” Bull World Health Organ 65 (2): 217–222. Structured Observation a Valid Technique to Measure Stevenson, R. J., T. I. Case, D. Hodgson, R. Porzig- Handwashing Behavior? Use of Acceleration Sensors Drummond, J. Barouei, and M. J. Oaten. 2009. “A Embedded in Soap to Assess Reactivity to Structured Scale for Measuring Hygiene Behavior: Development, Observation.” Am J Trop Med Hyg 83 (5): 1070–1076. Reliability and Validity.” Am J Infect Control 37 (7): Ram, P. 2013. “Practical Guidance for Measuring Hand- 557–564. washing Behavior: 2013 Update.” Working Paper. Szklo, M., and F. J. Nieto. 2007. Epidemiology: Beyond Washington, DC: Water and Sanitation Program, the Basics, 2nd ed. Sudbury, Massachusetts: Jones & World Bank. http://www.wsp.org/sites/wsp.org/files/ Bartlett Publishers. publications/WSP-Practical-Guidance-Measuring- Unicomb, L. E., T. M. N. Huda, R. B. Johnston, et al. Handwashing-Behavior-2013-Update.pdf. 2010. “The Relationship between Diarrheal Death Ram, P., B. Briceño, C. Chase, et al. 2014. “Analysis of and Measures of Water, Sanitation, and Hygiene Handwashing Behaviors Measured in Baseline Impact among Children < 5 Years of Age in Bangladesh.” Evaluation Surveys: Findings from Peru, Senegal, and Paper presented at American Society of Tropical Medi- Vietnam.” Technical Paper. Washington, DC: Water cine and Hygiene. Atlanta, Georgia. and Sanitation Program, World Bank. Verplanken, B., and S. Orbell. 2003. “Reflections on Past Behavior: A Self-Report Index of Habit Strength.” J Appl Soc Psychol 33 (6): 1313–1330. 18 Global Scaling Up Handwashing Validity of Rapid Measures of Handwashing Behavior Annex 1: Supplemental Tables Annex 1: Supplemental Tables TABLE S1: DEFINITIONS OF HANDWASHING MEASURES Presence of Soap Anywhere in the Home • The indicator is defined as presence of at least one type of soap observed by the enumerator anywhere in the home. • Following observations of handwashing places, enumerators asked the household respondent to show soap typically used for washing hands, irrespective of where it was located in the home. All households that allowed observation of soap anywhere in the home are included in this analysis. Presence of Soap and Water at a Fixed Handwashing Place Used Post-Defecation • The indicator is defined as the presence of at least one type of soap observed by the enumerator at the handwashing place reportedly used after defecation. • Questions to describe the location of the handwashing place and materials observed at that place followed an introductory question regarding whether hands are usually washed after defecation. Households in which the respondents indicated washing hands after defecation, and where respondents allowed observation of the location of the handwashing place and the presence of soap and water at that place, were eligible for analysis. • Enumerators recorded whether the handwashing place was inside the toilet or cooking place, or elsewhere in the yard. In Peru and Senegal, if the handwashing place was located elsewhere in the yard, the distance from the toilet was recorded (< 3 feet from the toilet, 3 to 10 feet from the toilet, more than 10 feet from the toilet). In Vietnam, if the handwashing place was located elsewhere in the yard, enumerators recorded distance in meters, but the precoded categories approximated the ones used in Peru and Senegal. Enumerators recorded the type of soap present at the handwashing place. In Peru and Senegal, the types of soap observed were beauty bar soap, multipurpose bar soap, and powder/detergent soap. In Vietnam, the types of soap observed were liquid soap, multipurpose bar soap, and powder/detergent soap. For the indicator, a household was considered as having soap at the handwashing place if at least one type of soap, irrespective of type, was present at the handwashing place. The presence of water was recorded at the handwashing place, irrespective of the type of device located therein. Presence of Soap and Water at a Fixed Handwashing Place Used before Food Preparation • The indicator is defined as the presence of at least one type of soap observed by the enumerator at the handwashing place reportedly used before food preparation. • Questions to describe the location of the handwashing place and materials observed at that place followed an introductory question regarding whether hands are usually washed before food preparation. Observations of soap and water were carried out if the handwashing place used before food preparation differed from the handwashing place used after defecation. • Households identified as eligible for this analysis were, first, those that showed a handwashing place used before food preparation that was distinct from the handwashing place used after defecation, and for which observation of the location, soap, and water were all completed. Also included in the analysis were those households in which the handwashing place used after defecation was located in the kitchen and was the same place used to wash hands before food preparation. Households in which the respondent indicated not usually washing hands before food preparation, or that had no specific place for washing hands before food preparation, were also added. (continued) www.wsp.org 19 Validity of Rapid Measures of Handwashing Behavior Annex 1: Supplemental Tables TABLE S1: (Continued) Presence of Soap and Water at a Fixed Handwashing Place Used before Food Preparation • Questions to describe the location of the handwashing place and materials observed at that place followed an introductory question regarding whether hands are usually washed before food preparation. All households in which the respondents indicated washing hands before food preparation were eligible for analysis, and where respondents allowed observation of the location of the handwashing place and the presence of soap and water at that place. • Enumerators recorded whether the handwashing place was inside the toilet or cooking place, or elsewhere in the yard. The distance of the handwashing place from the cooking place was recorded similarly to the distance from the toilet, as described above for the postdefecation handwashing place. Soap and water observations were also recorded similarly. Cleanliness Index of Caregiver Hands (Index Based on Observation of Nails, Palms, and Fingerpads) • This is a nine-point index based on the enumerator’s observation of the cleanliness of the nails, palms, and fingerpads of individual caregivers. Each aspect of the hand was rated on a three-point scale, ranging from one point for visible dirt to three points for clean appearance. These points were totaled to compute the hand cleanliness index. • Caregivers who did not allow observation of one or more aspects of the hand were not included in the analysis of the cleanliness index. The distribution of the index scores among caregivers in the control arm is reported below. Caregivers were also dichotomized according to cleanliness, with an index of 7 or higher denoting “clean hands,” and an index score less than 7 denoting “unclean hands.” Self-Reported Handwashing with Soap at Critical Times during Previous Day • The indicator is defined as self-reported handwashing with soap at one of the four critical times during the previous day. • Individual caregivers were asked whether they had washed hands with soap at least once during the previous day (since the same time the day before the enumerator’s visit). If they reported washing hands, they were asked in an unprompted fashion about the context of when hands were washed with soap and all other times that hands were washed with soap during the previous day. Although information about a number of critical times was captured, those of principal interest are as follows: after fecal contact, before food preparation or cooking food, before eating, or before feeding a child. • Caregivers described as being alone at the time of the interview were included in the analysis, because the presence of others may have influenced caregivers’ responses to handwashing questions. Observed Handwashing with Soap at Critical Times • Structured observation data can be analyzed and reported in numerous ways. Here, handwashing behavior was analyzed at the caregiver level, rather than at the household level. Observed handwashing behavior is reported below for each event, overall and by type of critical event observed. The events of interest were the same as those for self-reported handwashing behavior: after fecal contact, before food preparation, before eating, and before feeding a child. In particular, behavior as measured among primary caregivers is reported. 20 Global Scaling Up Handwashing TABLE S2: AGREEMENT BETWEEN RAPID HANDWASHING MEASURES BEHAVIOR IN ENDLINE SURVEYS, PERU, 2011 Soap and Water www.wsp.org Reported Observed at Reported Reported Handwashing Soap Handwashing Soap and Water Cleanliness Handwashing Handwashing before Reported Observed Place Observed at the Index of after Fecal before Food Feeding a Handwashing Anywhere in Used Post Place Used to Caregiver Peru Contact Preparation Child before Eating Home Defecation Prepare Food Hands * Reported Kappa — 0.04 0.04 0.03 0.03 –0.01 — 0.03 handwashing after (95%CI) (–0.01 – 0.10) (0.01 – 0.07) (–0.01 – 0.08) (–0.02 – 0.08) (–0.06 – 0.05) (–0.03 – 0.08) fecal contact Prev-adj 0.15 0.06 0.04 0.46 0.10 0.12 Kappa n ϭ 1,373 n ϭ 1,373 n ϭ 1,373 n ϭ 1,364 n ϭ 1,358 n ϭ 1,349 Reported Kappa — 0.03 –0.03 0–.05 — –0.04 0.11 handwashing before (95%CI) (0.00 – 0.06) (–0.08 – 0.02) (–0.09 – 0.0) (–0.09 – 0.02) (0.06 – 0.16) food preparation Prev-adj 0.04 –0.02 0.20 0.14 0.21 Kappa n n ϭ 1,373 n ϭ 1,373 n ϭ 1,364 n ϭ 1,363 n ϭ 1,349 Reported Kappa — –0.11 0.00 — –0.01 0.00 handwashing before (95%CI) (–0.15 – –0.06) (–0.02 – 0.02) (–0.04 – 0.03) (–0.03 – 0.04) feeding a child Prev-adj 0.05 0.00 0.01 0.12 Kappa n n ϭ 1,373 n ϭ 1,364 n ϭ 1,363 n ϭ 1,349 Reported Kappa — 0.01 — –0.05 –0.09 handwashing before (95%CI) (–0.02 – 0.05) (–0.10 – –0.01) (–0.14 eating Prev-adj 0.06 –0.04 – –0.05) Kappa n n ϭ 1,364 n ϭ 1,363 –0.09 n ϭ 1,349 Soap observed Kappa — 0.58 0.58 0.05 anywhere in home (95%CI) (0.53 – 0.63) (0.53 – 0.63) (0.00 – 0.10) Prev-adj 0.68 0.68 0.26 Kappa n n ϭ 1,373 n ϭ 1,378 n ϭ 1,351 (continued) 21 22 TABLE S2: (Continued) Soap and water Kappa — — — — 0.81 0.14 observed at (95%CI) (0.78 – 0.85) (0.09 – 0.20) handwashing place Prev-adj 0.83 0.24 used Kappa n nϭ1,372 n ϭ 1,345 postdefecation Soap and water Kappa — — 0.15 observed at the (95%CI) (0.09 – 0.20) place used to Prev-adj 0.24 prepare food Kappa n n ϭ 1,349 Cleanliness index of Kappa — caregiver hands* (95%CI) Prev-adj Kappa n *Index based on observation of nails, palms, and fingerpads, dichotomized with score <7 considered “not clean” and score ≥7 considered “clean” Global Scaling Up Handwashing TABLE S3: AGREEMENT BETWEEN RAPID HANDWASHING MEASURES BEHAVIOR IN ENDLINE SURVEYS, SENEGAL, 2011 Soap and Reported Soap and Water Water www.wsp.org Reported Reported Handwashing Soap Observed at Observed at Cleanliness Handwashing Handwashing before Reported Observed Handwashing the Place Index of after Fecal before Food Feeding a Handwashing Anywhere in Place Used Used to Caregiver Senegal Contact Preparation Child before Eating Home Postdefecation Prepare Food Hands * Reported Kappa — 0.10 0.05 0.08 0.01 0.08 — 0.09 handwashing after (95%CI) (0.05 – 0.14) (0.03 – 0.08) (0.03 – 0.13) (–0.02 – 0.04) (0.03 – 0.13) (0.05 – 0.13) fecal contact Prev-adj 0.21 0.27 0.17 0.11 0.15 0.15 Kappa n ϭ 1,338 n ϭ 1,338 n ϭ 1,338 n ϭ 1,338 n ϭ 1,329 n ϭ 1,280 Reported Kappa — 0.05 0.00 0.02 — 0.05 0.03 handwashing before (95%CI) (0.01 – 0.10) (–0.05 (0.00 – 0.03) (–0.01 – 0.10) (0.01 – 0.05) food preparation Prev-adj 0.61 – 0.06) 0.02 0.41 0.03 Kappa n n ϭ 1,338 0.33 n ϭ 1,338 n ϭ 1,331 n ϭ 1,280 n ϭ 1,338 Reported Kappa — 0.05 0.00 — 0.04 0.01 handwashing before (95%CI) (0.01 – 0.9) (–0.01 – 0.01) (0.00 – 0.09) (0.00 – 0.02) feeding a child Prev-adj 0.55 0.00 0.61 0.02 Kappa n n ϭ 1,338 n ϭ 1,338 n ϭ 1,331 n ϭ 1,280 Reported Kappa — 0.01 — –0.04 0.01 handwashing before (95%CI) (–0.01 – 0.03) (–0.09 – 0.01) (–0.02 – 0.04) eating Prev-adj 0.02 0.31 0.01 Kappa n n ϭ 1,338 n ϭ 1,331 n ϭ 1,280 Soap observed Kappa — 0.09 0.05 0.08 anywhere in home (95%CI) (0.07 – 0.10) (0.04 – 0.06) (0.02 – 0.14) Prev-adj 0.10 0.05 0.52 Kappa n n ϭ 1,402 n ϭ 1,404 n ϭ 1,280 Soap and water Kappa — — — — 0.39 0.06 observed (95%CI) (0.33 – 0.44) (0.03 – 0.09) at handwashing Prev-adj 0.56 0.06 place used Kappa n n ϭ 1,395 n ϭ 1,271 postdefecation (continued) 23 24 TABLE S3: (Continued) Soap and water Kappa — 0.04 observed at the (95%CI) (0.02 – 0.06) place used to Prev-adj 0.04 prepare food Kappa n n ϭ 1,273 Cleanliness index of Kappa — caregiver hands* (95%CI) Prev-adj Kappan *Index based on observation of nails, palms, and fingerpads, dichotomized with score <7 considered “not clean” and score ≥7 considered “clean” Global Scaling Up Handwashing TABLE S4: AGREEMENT BETWEEN RAPID HANDWASHING MEASURES BEHAVIOR IN ENDLINE SURVEYS, VIETNAM, 2011 Soap and Reported Soap and Water Water www.wsp.org Reported Reported Handwashing Soap Observed at Observed at Cleanliness Handwashing Handwashing before Reported Observed Handwashing the Place Index of after Fecal before Food Feeding a Handwashing Anywhere in Place Used Used to Caregiver Vietnam Contact Preparation Child before Eating Home Postdefecation Prepare Food Hands * Reported Kappa — 0.03 0.11 0.05 0.03 0.06 — 0.11 handwashing after (95%CI) (0.00 – 0.07) (0.06 – 0.16) (0.02 – 0.08) (0.00 – 0.06) (0.01 – 0.11) (0.04 – 0.17) fecal contact Prev-adj 0.03 0.12 0.07 0.50 0.36 0.20 Kappa n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 n ϭ 1,063 n ϭ 1,064 Reported Kappa — 0.09 –0.02 0.00 — 0.03 0.05 handwashing before (95%CI) (0.03 – 0.15) (–0.07 – 0.04) (–0.01 – 0.01) (0.00 – 0.05) (0.01 – 0.10) food preparation Prev-adj 0.19 0.07 0.06 0.06 0.06 Kappa n n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 Reported Kappa — 0.00 0.00 — 0.04 0.10 handwashing before (95%CI) (–0.06 – 0.05) (–0.01 – 0.01) (0.02 – 0.07) (0.05 – 0.15) feeding a child Prev-adj 0.06 0.09 0.09 0.27 Kappa n n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 n ϭ 1,064 Reported Kappa — 0.00 — –0.01 0.00 handwashing before (95%CI) (0.00 – 0.01) (–0.02 – 0.01) (–0.03 – 0.04) eating Prev-adj 0.01 -0.01 0.03 Kappa n n ϭ 1,064 n ϭ 1,064 nϭ1,064 Soap observed Kappa — 0.23 0.15 0.02 anywhere in home (95%CI) (0.14 – 0.32) (0.07 – 0.23) (–0.01 – 0.04) Prev-adj 0.81 0.79 0.37 Kappa n n ϭ 1,063 n ϭ 1,064 n ϭ 1,064 Soap and water Kappa — — — — 0.58 0.08 observed (95%CI) (0.50 – 0.66) (0.03 – 0.13) at handwashing Prev-adj 0.84 0.33 place used Kappa n n ϭ 1,063 n ϭ 1,063 postdefecation (continued) 25 26 TABLE S4: (Continued) Soap and water Kappa — — 0.07 observed at the (95%CI) (0.02 – 0.12) place used to Prev-adj 0.32 prepare food Kappa n n ϭ 1,064 Cleanliness index of Kappa — caregiver hands* (95%CI) Prev-adj Kappa n *Index based on observation of nails, palms, and fingerpads, dichotomized with score <7 considered “not clean” and score ≥7 considered “clean” Global Scaling Up Handwashing TABLE S5: CHARACTERISTICS OF HOUSEHOLDS AND CAREGIVERS WITH AND WITHOUT ENDLINE STRUCTURED OBSERVATION (SO) DATA, AMONG CONTROLS, PERU, SENEGAL, AND VIETNAM, 2011 Peru Senegal Vietnam www.wsp.org Households Households Households Households Households Households with SO without SO with SO without SO with SO without SO n ϭ 286 n ϭ 1,086 n ϭ 88 n ϭ 669 n ϭ 200 n ϭ 864 Characteristic Number (%) Number (%) p-value# Number (%) Number (%) p-value# Number (%) Number (%) p-value# Wealth Mean wealth 0.013 0.11 0.13 –0.50 0.49 <0.001 4.24 4.17 0.24 index (sd) (1.01) (0.31) (2.5) (2.5) (1) (1) Water- and sanitation-related characteristics JMP defined 146 572 0.75 51 483 <0.01 127 605 0.07 improved (52) (53) (58) (72) (64) (70) sanitation Toilet is located in 218 771 <0.01 57 496 0.11 170 763 0.28 the household or (79) (71) (67) (75) (85) (88) yard JMP defined 252 1,045 0.10 64 533 0.09 197 837 0.21 improved water (95) (92) (74) (81) (99) (97) source Reports treating 227 877 0.78 15 104 0.64 171 764 0.25 drinking water (87) (88) (20) (18) (86) (88) every day in the past seven days (any method) Handwashing behavior measures Presence of soap 228 891 0.64 75 603 0.16 199 846 0.23## anywhere in the (81) (82) (85) (90) (99) (98) home Presence of soap 192 720 0.58 18 195 0.08 181 766 0.35 and water at the (68) (67) (20) (29) (91) (89) handwashing place used postdefecation (continued) 27 28 TABLE S5: (Continued) Households Households Households Households Households Households with SO without SO with SO without SO with SO without SO n ϭ 286 n ϭ 1,086 n ϭ 88 n ϭ 669 n ϭ 200 n ϭ 864 Characteristic Number (%) Number (%) p-value# Number (%) Number (%) p-value# Number (%) Number (%) p-value# Handwashing behavior measures Presence of soap 187 728 0.87 13 136 0.21 179 767 0.77 and water at the (67) (67) (15) (20) (90) (89) handwashing place used before food preparation Ability to bring soap 31 112 0.34 42 262 0.25 13 56 0.77 § in Յ one minute (89) (65) (86) (79) (93) (90) Caregivers with Caregivers Caregivers with Caregivers Caregivers with Caregivers SO without SO SO without SO without SO n ϭ 300 n ϭ 1,115 n ϭ 171 n ϭ 1,240 n ϭ 200 n ϭ 864 Number (%) Number (%) p-value# Number (%) Number (%) p-value# Number (%) Number (%) p-value# Female 299 1,115 1.00+ 169 1,232 0.48+ 186 808 0.79 (99.5) (99.5) (99) (99.7) (93) (94) Mean age in 30.16 29.31 0.01 31.00 30.9 0.86 35.30 35.63 0.75 years (sd) (7.5) (8.0) (9.0) (8.7) (13.1) (12.5) Ever attended 281 1,045 0.93 36 362 0.03 192 837 0.87 school (94) (94) (21) (29) (98) (98) Education level 0.60 <0.01 attained* Kindergarten or 128 447 Data not 42 189 primary school (47) (43) available (23) (23) Lower secondary$ 125$ 497$ Data not 88 456 (46) (48) available (46) (54) Upper secondary — — Data not 31 132 available (16) (16) Trade school 21 87 Data not 29 60 (college), university, (8) (8) available (15) (7) or higher Global Scaling Up Handwashing TABLE S5: (Continued) Handwashing behavior measures www.wsp.org Presence of soap 192 720 0.58 18 195 0.08 182 765 0.34 and water at the (68) (67) (20) (29) (91) (89) handwashing place used postdefecation Presence of 150 530 0.18 48 397 0.37 88 354 0.43 handwashing (55) (51) (56) (62) (44) (41) station Յ3 meters from latrine Presence of soap 187 728 0.87 13 136 0.21 179 767 0.77 and water at the (67) (67) (15) (20) (90) (89) handwashing place used before food preparation Presence of 217 726 0.02 51 372 0.74 134 491 <0.01 handwashing (78) (71) (58) (56) (67) (57) station Յ3 meters from food preparation place All three aspects 137 509 0.31 94 758 0.01 93 389 0.72 of hands rated (46) (46) (60) (68) (47) (45) clean Self-reported 190 740 0.41 55 545 <0.01 142 581 0.31 handwashing with (64) (67) (34) (46) (71) (68) soap after fecal contact during previous day Self-reported 203 747 0.70 35 223 0.40 54 279 0.15 handwashing with (68) (67) (22) (19) (27) (32) soap before food preparation during previous day (continued) 29 30 TABLE S5: (Continued) Handwashing behavior measures Self-reported 53 231 0.26 2 41 0.13 70 319 0.61 handwashing (18) (21) (1) (3) (35) (37) with soap before feeding a child during previous day Self-reported 112 461 0.25 23 288 <0.01 29 141 0.53 handwashing with (38) (41) (14) (24) (15) (16) soap before eating during previous day * Chi square comparing distribution of education level among those with structured observation to those without # p-values were calculated using a chi-square test for categorical data comparisons and t-tests for continuous data comparisons ## Fisher’s exact test to test for differences between groups used due to small cell size § For those households in which soap was not readily observable $ In Peru, this category refers to any secondary school Global Scaling Up Handwashing