WATER AND SANITATION PROGRAM: TECHNICAL PAPER 56933 Global Scaling Up Handwashing Project Scaling Up Handwashing Behavior: Findings from the Impact Evaluation Baseline Survey in Peru Sebastian Galiani and Alexandra Orsola-Vidal August 2010 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. Sebastian Galiani Global Scaling Up Handwashing is a WSP project focused Washington University in Saint Louis on learning how to apply innovative promotional approaches to behavior change to generate widespread and sustained Alexandra Orsola-Vidal improvements in handwashing with soap at scale among Water and Sanitation Program women of reproductive age (ages 15­49) and primary school- As an integral component of the Water and Sanitation aged children (ages 5­9). The project is being implemented Program's (WSP's) Global Scaling Up Handwashing by local and national governments with technical support Project, a cross­country impact evaluation (IE) study is from WSP. For more information, please visit www.wsp.org/ being conducted in Peru, Senegal, Tanzania, and Vietnam. scalinguphandwashing. This study is led by the World Bank's WSP IE Team. This Technical Paper is one in a series of knowledge products The project's Global IE Team oversees the IE design, designed to showcase project findings, assessments, and methodology, and country teams. It is led by Bertha Briceno lessons learned in the Global Scaling Up Handwashing (in its early stages the Global IE was led by Jack Molyneaux), Project. This paper is conceived as a work in progress to together with Alexandra Orsola-Vidal and Claire Chase. encourage the exchange of ideas about development issues. Professor Paul Gertler has provided guidance and advice For more information please email Alexandra Orsola-Vidal at throughout the project. Global IE experts also include wsp@worldbank.org or visit www.wsp.org. Sebastian Galiani, Jack Colford, Ben Arnold, Pavani Ram, Lia Fernald, Patricia Kariger, Paul Wassenich, Mark Sobsey, WSP is a multi-donor partnership created in 1978 and administered by and Christine Stauber. At the country level, the Peru IE the World Bank to support poor people in obtaining affordable, safe, and Team manages the in-country design, field activities, and sustainable access to water and sanitation services. WSP's donors include Australia, Austria, Canada, Denmark, Finland, France, the Bill & Melinda data analysis, and it is led by principal and co-principal Gates Foundation, Ireland, Luxembourg, Netherlands, Norway, Sweden, investigators Sebastian Galiani and Alexandra Orsola, with Switzerland, United Kingdom, United States, and the World Bank. operational assistance from Carlos Augusto Claux. Andres WSP reports are published to communicate the results of WSP's work Drenik has also provided significant research support during to the development community. Some sources cited may be informal the data analysis. documents that are not readily available. The findings, interpretations, and conclusions expressed herein are The Peru IE has also benefited from continuous support from entirely those of the author and should not be attributed to the World Eduardo Perez, the project's global task team leader, Rocío Bank or its affiliated organizations, or to members of the Board of Flórez Peschiera, the project's country task manager, and Executive Directors of the World Bank or the governments they the Global and Peru technical team comprised of Hnin Hnin represent. The World Bank does not guarantee the accuracy of the data included in this work. The maps were provided by the Map Design Unit Pyne, Jacqueline Devine, Nathaniel Paynter, Craig Kullmann, of the World Bank. The boundaries, colors, denominations, and other Catherine Amelink, Christianne Frischmuth, Doris Alfaro, information shown on any map in this work do not imply any judgment Carlos Augusto Claux, Jorge Aguela, and WSP support staff. on the part of the World Bank Group concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The technical body of the National Department for Health The material in this publication is copyrighted. Requests for Promotion (Ministry of Health) and the Environmental permission to reproduce portions of it should be sent to wsp@ Education Department (Ministry of Education) provided worldbank.org. WSP encourages the dissemination of its work and contributions to the initial impact evaluation concept design. will normally grant permission promptly. For more information, please visit www.wsp.org. The initial impact evaluation design was presented to the Ministry of Health and the Ministry of Education in Lima, © 2010 Water and Sanitation Program Peru, in April and May 2007. Executive Summary Background four countries to establish the causal effect of the interven- In response to the preventable threats posed by poor sanita- tion (treatment) on specific health and welfare outcomes. tion and hygiene, the Water and Sanitation Program (WSP) The IE includes several rounds of household and community launched two large-scale projects, Global Scaling Up Hand- surveys: pre-intervention (baseline), concurrent (longitudi- washing and Global Scaling Up Rural Sanitation,1 to im- nal), and post-intervention (endline). The surveys are de- prove the health and welfare outcomes for millions of poor signed to collect information on the characteristics of the people. Local and national governments are implementing eligible population and to track changes in desired these projects with technical support from WSP. outcomes. Global Scaling Up Handwashing aims to test whether This technical paper presents the findings of the WSP impact handwashing with soap behavior can be generated and sus- evaluation (IE) baseline survey in Peru and is one in a series tained among the poor and vulnerable using innovative of papers presenting IE findings from surveys conducted in promotional approaches. The primary objectives are to re- each project country. duce the risk of diarrhea in young children and increase household productivity by stimulating and sustaining the Peru Intervention behavior of handwashing with soap at critical times. Over- The handwashing project in Peru, implemented in 788 ran- all, the project aims to generate and sustain handwashing domly selected districts located in 104 provinces, comprises with soap practices among 5.4 million people living in a primary audience of mother/caregivers and children; the Peru, Senegal, Tanzania, and Vietnam, the four countries secondary targeted audience includes community-based where the project has been implemented to date. agents such as schoolteachers, health promoters, and local leaders. In Peru, the project objective is to reach women Handwashing with soap at critical times--such as after con- (ages 14­49) and children (ages 5­12) in order to stimulate tact with feces and before handling food--has been shown to and sustain handwashing behavior change in a total of substantially reduce the incidence of diarrhea. It reduces 1.3 million of those reached by project end. health risks even when families do not have access to basic sanitation and water supply. Despite this benefit, rates of The main components of the intervention include: handwashing with soap at critical times remain low through- · Mass media and promotional events at the provin- out the world. cial level that combine local radio and outreach ac- tivities in public spaces to promote behavior change In an effort to induce improved handwashing behavior, the among the primary target audience, and project intervention borrows from both commercial and · School and community social mobilization activities social marketing fields to bring about the desired outcomes. at the district level, including educational sessions and Behavior change communications campaigns and messages promotional events, to reinforce messages among the developed by the project have been designed and strategi- primary target audience, and promote capacity build- cally delivered across multiple, integrated channels, in mul- ing among the secondary target audience. tiple settings, to "surround" target audiences with handwashing promotion. Methodology and Design The IE study in Peru includes 120 of the 788 districts lo- One of the project's global objectives is to learn about and cated in 80 of the 104 provinces and covers a representative document the health and welfare impacts of the project inter- sample of the population targeted by the intervention. The vention. To measure the magnitude of these impacts, the IE is designed to separately assess the effects of the two project is implementing an impact evaluation (IE) using a main intervention components as explained above. In addi- randomized-controlled experimental design in each of the tion, it assesses the impact of the handwashing curricula implemented in primary schools. 1 For more information on Global Scaling Up Rural Sanitation, see www.wsp.org/ scalingupsanitation. www.wsp.org iii Findings from the Impact Evaluation Baseline Survey in Peru Executive Summary In Peru, the IE baseline survey was conducted from May Handwashing with soap behavior--Although almost all through August 2008, in a total of 3,526 households. Data caregivers report having washed their hands with soap at was collected on a range of indicators, including: household least once during the previous day, less than half confirmed characteristics, education, income, assets, water sources, having done so at times of fecal contact (46 percent of care- sanitation, handwashing facilities and behavior, child envi- givers associated handwashing with soap with toilet use and ronment, maternal depression, handwashing determinants, 42 percent with cleaning children's bottoms). Handwash- exposure to health interventions, relationship between fam- ing with soap was higher at times of cooking or food prepa- ily and school, diarrhea prevalence, acute lower respiratory ration (68 percent), but lower when feeding a child (34 infection (ALRI) and other health symptoms, child develop- percent). Handwashing with soap increased with income at ment, growth, anemia, and mortality. In addition, commu- every critical juncture. In nearly two-thirds of the house- nity questionnaires were conducted in all sample locations holds (64 percent) a handwashing station stocked with soap and structured observations of handwashing behavior, and water was observed within the household or the yard. water microbiology samples, and child fecal samples were The number of households with an observed handwashing collected in a subsample of 160 households. station with soap and water was higher in the jungle (72 percent) than along the coast (67 percent) or in the moun- Findings tains (62 percent). The higher the income, the closer the The main findings of the IE baseline survey in Peru are pre- handwashing station was to the toilet or kitchen facility. sented below. Over half of the caregivers (53 percent) appeared to have clean fingernails and about two-thirds had clean palms Household demographics (67 percent) or clean finger pads (68 percent). Size, age, education, income--Households averaged 5.3 members, with 1.4 children under age five. On average, Environmental contamination--Households with access to the household head was 37; around one-half of house- improved sanitation presented lower counts of bacteria in hold heads had attained secondary education and the hand rinses, drinking water, and on sentinel objects; house- majority (95 percent) were employed. The average holds with access to an improved water source showed higher monthly household income per capita was 165 Peruvian levels of water contamination. Water and caregivers' hand- nuevos soles (S/.). rinse samples from households with a handwashing station with soap and water had lower counts of bacteria, but counts Water, sanitation, and hygiene from child's hand-rinse samples and objects were higher in Access to water supply--Three-quarters of the households these households. When taking income levels into account, had access to improved sources of drinking water, but for there was a declining trend of Escherichia coli (E. coli) counts the poorest households, access to improved water sources with increased income. Households living along the coast decreased to 70 percent. Households living along the coast presented the highest E. coli counts in samples taken from the of Peru had higher access (86 percent) than those living in mother; households located in the jungle showed the highest the mountains (73 percent) or the jungle (62 percent). E. coli counts in objects and water. Access to sanitation--Half of the households had access to Child health improved sanitation. The highest percentage of access to Parasitical infestations--The most frequent parasites de- improved sanitation was observed among households lo- tected were Giardia and Blastocystis. On average, parasites cated along the coast (54 percent), while the lowest access were detected in 12 percent of the stool samples collected. was for households located in the jungle (33 percent). Ac- Prevalence of parasites was lower among households with cess for households located in the mountains was just below access to improved sanitation (7 percent) than those with the average (47 percent). Over 20 percent of all households unimproved sanitation (18 percent). Similarly, parasitical had no sanitation facilities of any type. prevalence was lower among households with access iv Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Executive Summary to improved water sources (8 percent) than those with with income level. A partial, plausible explanation could be unimproved water sources (25 percent). The lowest preva- that children in poor households were more likely to receive lence of parasites was found in households with a hand- iron supplements. washing station stocked with soap and water (3 percent) and highest in those without (29 percent). Parasitical preva- Nutrition and child development lence decreased as income increased; disaggregated by geo- Nutrition--The average child was breastfed for 12 months, graphic location, prevalence was twice as high in the although over 60 percent of caregivers gave their children mountains (22 percent) than in the jungle (11 percent) or infant formula during the first three days of life. Vitamin A the coast (9 percent). was given to 23 percent of the children and iron supple- ments to 22 percent. Diarrhea prevalence--Ten percent of children under the age of five had presented diarrhea symptoms in the previous 48 Growth measures--Arm circumference was found to be hours, 18 percent in the past seven days, and 20 percent in higher than the population mean, as were the body mass the past 14 days. Prevalence of diarrhea was higher in those index and the weight for length/height. By contrast, average households with unimproved sanitation (12 percent) and weight-for-age, length/height-for-age, and head circumfer- lower for those with improved sanitation (8 percent); how- ence were found to be lower than the population mean esti- ever, diarrhea prevalence was not lower in households with mated by the World Health Organization (WHO). On access to a handwashing station with soap and water nor in average, children coming from households without improved households with access to improved water sources, compared sanitation, improved water source, or soap and water at to those without access. Diarrhea prevalence appeared to be handwashing station tended to have a lower average z-score uncorrelated with income, but it varied noticeably by geo- for each anthropometric measure included in the analysis. All graphic location. For instance, diarrhea prevalence in the six measures increased with income. With respect to disag- jungle (13 percent) and the mountains (11 percent) was gregation by geographic area, all six measures indicated that twice as high than rates found along the coast (6 percent). children living along the coast were in a better situation than those living in the mountains and the jungle. Acute lower respiratory infection (ALRI) prevalence--On aver- age, 4 percent of children presented ALRI symptoms in the Child care environment--Three-quarters of the children (75 previous 48 hours, and 6 percent in the previous seven days. percent) appeared clean at the time of the interview but al- ALRI prevalence increased for those children living in house- most half of them had dirty fingernails (47 percent). The holds with unimproved sanitation and those with unim- overall cleanness of children (hands, clothes, fingernails, proved water sources. ALRI prevalence was higher for face) increased with income. The majority of the children children living in the mountains (6 percent) and lower for played both with toys (83 percent) and with adult house- those living along the coast (2 percent). As with diarrhea, hold members (84 percent). Each of these percentages in- similar percentages of households presented ALRI symptoms creased as income levels increased. in the previous week, irrespective of whether or not they had a handwashing station stocked with soap and water. Cognitive development--An index of child development was developed for specific skills for age, including com- Anemia--Three-quarters of the samples taken indicated the munication, social-personal, and gross motor skills. We presence of anemia. This proportion was lower for house- systematically observed a lower degree of development for holds with improved sanitation (70 percent) than those every type of skill in children from households without with unimproved sanitation (79 percent). Anemia presence improved sanitation, without improved water source, and was lower among households living in the jungle (70 per- without soap and water at the handwashing station. All cent) than those living along the coast (75 percent) or the the measures increased with income, but no clear-cut mountains (76 percent). An unexpected result was that the pattern was observed when disaggregated by geographic percentage of individuals suffering from anemia increased location. www.wsp.org v Findings from the Impact Evaluation Baseline Survey in Peru Abbreviations and Acronyms Abbreviations and Acronyms ALRI Acute Lower Respiratory Infection C Counterfactual or Control Group C-Schools Counterfactual or Control Group in Schools E. coli Escherichia coli ENAHO National Household Survey (Encuesta Nacional de Hogares) Hb Hemoglobin HH(s) Household(s) HW Handwashing IE Impact Evaluation T1 Mass Media Treatment or Treatment 1 T2 Social Mobilization Treatment or Treatment 2 T2-Schools Treatment 2 in Schools WHO World Health Organization WSP Water and Sanitation Program vi Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Contents Contents Executive Summary................................................................... iii Abbreviations and Acronyms ....................................................vi I. Overview .................................................................................... 1 1.1 Introduction ...................................................................... 1 1.2 Project Background ......................................................... 3 1.3 Project Components ........................................................ 4 1.4 Objectives of the Study.................................................... 4 II. Methodology .............................................................................. 6 2.1 Randomization ................................................................. 6 2.2 Study Design .................................................................... 7 2.3 Sampling Size and Strategy ............................................. 8 2.4 Variables for Data Analysis............................................. 10 2.5 Instruments for Data Collection ..................................... 11 III. Sample Representativeness ................................................... 14 3.1 Geographic Representativeness .................................... 14 3.2 Comparison Between WSP Baseline Study and Peru Population ...................................................... 15 IV. Findings ................................................................................... 20 4.1 General Household Characteristics ............................... 22 4.2 Water Source and Safe Water-Use Behavior ................. 30 4.3 Sanitation Facilities ........................................................ 32 4.4 Handwashing Behavior .................................................. 37 4.5 Mass-Media Consumption............................................. 42 4.6 Family-School Relationship ........................................... 45 4.7 Child Care Environment ................................................. 47 4.8 Child Development......................................................... 52 4.9 Diarrhea and Acute Lower Respiratory Infection Prevalence ..................................................................... 55 4.10 Anthropometric Measures and Anemia ....................... 58 4.11 Environmental Contamination and Parasitical Prevalence ................................................................... 63 V. Future Directions ..................................................................... 66 References ............................................................................... 67 Annexes 1: List of Districts Included in WSP Sample ............................. 68 2: Findings from Structured Observations of Handwashing Behavior ........................................................ 72 3: Test of Baseline Balance ...................................................... 74 www.wsp.org vii Findings from the Impact Evaluation Baseline Survey in Peru Contents Figures 1: Peru Impact Evaluation Sample Selection ........................... 9 2A: Distribution of Salaries Received in the Primary Occupation: Dependent Workers ....................................... 18 2B: Distribution of Salaries Received in the Primary Occupation: Independent Workers ..................................... 18 3: Distribution of Monthly Income per Capita......................... 19 4: Histograms of Child Development Measures' Z-Scores (Children <2) ....................................................................... 54 5: Histograms of Anthropometric Measures' Z-Scores (Children <2) ....................................................................... 59 6: Anthropometric Measures' Z-Scores by Sex and Months of Age (Children <2) .............................................. 61 Tables 1: Demographics .................................................................... 16 2: Educational Attainment ...................................................... 16 3: Occupation ......................................................................... 17 4: Percent Distribution of Water, Sanitation, and Hygiene Conditions by Geographic Area ......................................... 20 5: Correlations Between Water, Sanitation, Hygiene Conditions, and Income Group .............................................................. 23 6: Summary Statistics............................................................. 23 7: Percent Distribution of the Basic Socio-Demographic Characteristics .................................................................... 24 8: Percent Distribution of Individual's Education .................... 25 9: Actual Distribution of Students' Time ................................. 25 10: Percent Distribution of Household Assets and Non-Labor Income ............................................................. 26 11: Dwelling Characteristics .................................................... 27 12: Individual's Activity and Primary Work................................ 29 13: Households with Time Loss Due to Child Illness ............... 30 14: Access to Improved Water Sources ................................... 31 15: Type of Water Source ......................................................... 31 16: Safe Water-Use Behavior ................................................... 33 17: Access to Improved Sanitation .......................................... 34 18: Household Main Sanitation Facility Characteristics ........... 35 19: Improvement of Sanitation Facilities .................................. 35 20: Other Characteristics of Household Sanitary Condition .... 37 21: Household Cleanness......................................................... 37 22A: Self-Reported Handwashing Behavior with Soap by Income Quartile (Previous 24 Hours) ............................. 38 22B: Self-Reported Handwashing Behavior with Soap by Geographic Area (Previous 24 Hours) ........................... 38 viii Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Contents 23A: Observation of Handwashing Station with Soap and Water by Income Quartile ............................................ 39 23B: Observation of Handwashing Station with Soap and Water by Geographic Area .......................................... 39 24A: Observation of Handwashing Station Used After Going to Toilet..................................................................... 40 24B: Observation of Handwashing Station Used When Preparing Food or Feeding a Child .................................... 41 25A: Observations of Caregivers Hands by Income ................... 42 25B: Observations of Caregivers Hands by Geographic Area............................................................ 42 26A: Mass-Media Consumption by Observed Handwashing Station with Soap and Water ....................... 43 26B: Mass-Media Consumption by Geographic Area ................ 44 26C: Self-Reported Handwashing Behavior by Recall of Handwashing Campaign .................................................... 44 27A: Family-School Relationship by Access to Handwashing Station with Soap and Water .............................................. 46 27B: Family-School Relationship by Geographic Area ............... 46 28A: Soap Contribution to Schools by Observed Handwashing Station with Soap and Water ....................... 47 28B: Soap Contribution to Schools by Geographic Area ........... 47 29: Child Breastfeeding (Children <2) ....................................... 48 30: Infant/Young Child Feeding (Children <2) ........................... 49 31: Infant/Young Child Care Situation (Children <5) ................. 49 32A: Infant/Young Child Care Situation During Interview ........... 50 32B: Discipline Measures Towards Infant During Previous Month (Children <2) ............................................. 50 33: Infant/Young Child Learning Environment (Children <2)..... 51 34: Maternal Depression........................................................... 52 35A: Child Development Z-Scores by Sanitary Conditions (Children <2) ..................................................... 52 35B: Child Development Z-Scores by Income Quartile (Children <2) ....................................................................... 52 35C: Child Development Z-Scores by Geographic Area (Children <2) ....................................................................... 53 36A: Diarrhea Prevalence by Sanitary Conditions (Children <5) . 55 36B: Diarrhea Prevalence by Geographic Area (Children <5) ..... 55 37: Diarrhea Treatment by Income Quartile (Children <5) ........ 55 38A: ALRI Prevalence by Sanitary Conditions (Children <5) ...... 57 38B: ALRI Prevalence by Geographic Area (Children <5) ........... 57 39: ALRI Treatment by Income Quartile (Children <5) .............. 57 40A: Anthropometric Measures' Z-Scores by Sanitary Conditions (Children <2) ..................................................... 60 www.wsp.org ix Findings from the Impact Evaluation Baseline Survey in Peru Contents 40B: Anthropometric Measures' Z-Scores by Income Quartile (Children <2) .......................................................... 60 40C: Anthropometric Measures' Z-Scores by Geographic Area (Children <2) ............................................................... 60 41: Anemia Prevalence (Hb < 110 g/L) in Children < 2.............. 62 42A: Mean Escherichia coli Concentrations by Sanitary Conditions ....................................................... 63 42B: Mean Escherichia coli Concentrations by Income Quartile ............................................................. 63 42C: Mean Escherichia coli Concentrations by Geographic Area ........................................................... 63 43A: Parasites Prevalence in Stool Samples by Sanitary Conditions (Children <2) .................................. 64 43B: Parasites Prevalence in Stool Samples by Income Quartile (Children <2) .......................................................... 64 43C: Parasites Prevalence in Stool Samples by Geographic Area (Children <2) ...................................... 64 44A: List of Districts Selected to Receive Treatment 1 (Mass Media) ...................................................................... 68 44B: List of Districts Selected to Receive Treatment 2 (Community and School) .................................................... 69 44C: List of Districts Selected to Serve as Control .................... 70 45: Soap Use by Event Type as Measured by Structured Observation ........................................................................ 72 46: Bivariate Analysis of Factors Associated with Observation of Soap Use at Least Once During Fecal Contact ..................................................................... 73 Boxes 1: Health and Welfare Impacts ............................................... 10 2: Handwashing Behavior and Determinants ......................... 11 Map 1: Map of Peru with Descriptive Statistics by Administrative Department ......................................................................... 21 x Global Scaling Up Handwashing I. Overview 1.1 Introduction In December 2006, the Water and Sanitation Program potentially reaching more than 250 million people in more than (WSP) began implementation of two related large-scale sani- 20 countries by 2020. tation and hygiene projects with funding from the Bill & Melinda Gates Foundation. The interventions include the The handwashing project's global activities test innovative Global Scaling Up Handwashing Project and the Global approaches at scale and have four main objectives: Scaling Up Rural Sanitation Project. The goal of the hand- · Design and support the implementation of innova- washing project is to reduce the risk of diarrhea and therefore tive large-scale, sustainable handwashing programs increase household productivity by stimulating and sustain- in four diverse countries (Peru, Senegal, Tanzania, ing the behavior of handwashing with soap at critical times in and Vietnam). 5.4 million people in Peru, Senegal, Tanzania, and Vietnam. · Document and learn about the impact and sustainabil- Thus, on average, the project will improve the handwashing ity of innovative large-scale handwashing programs. behavior of over one million people per country. · Learn about the most effective and sustainable ap- proaches to triggering, scaling up, and sustaining Handwashing with soap at critical times (such as after con- handwashing with soap behaviors. tact with feces and before handling food) has been shown to · Promote and enable the adoption of effective hand- substantially reduce the incidence of diarrhea. It reduces washing programs in other countries and--through health risks even when families do not have access to basic the translation of results and lessons learned-- sanitation and water supply service. Despite this benefit, position handwashing as a global public health rates of handwashing with soap at critical times are very low priority into effective advocacy and applied knowl- throughout the world. edge and communication products. The project aims to test whether this handwashing behavior The handwashing project also aims to complement and im- can be improved among the poor and vulnerable using innova- prove upon existing hygiene behavior change and hand- tive promotional approaches. In addition, it will undertake a washing approaches, and to enhance them with novel structured learning and dissemination process to develop the approaches, including commercial marketing, to deliver evidence, practical knowledge, and tools needed to effectively handwashing with soap messages, along with broad and in- replicate and scale up future handwashing programs. clusive partnerships of government, private commercial marketing channels, and concerned consumer groups and WSP's vision of success is that at the end of project we will have NGOs. These innovative methods will be combined with demonstrated that handwashing with soap, at scale, is one of the proven community-level interpersonal communications most successful and cost-effective interventions to improve and and outreach activities, with a focus on sustainability. In protect the health of poor rural and urban families, especially addition, the project incorporates a rigorous impact evalua- children under age five. Moreover, we envision the effort to tion component to support thoughtful and analytical learn- develop the evidence, practical knowledge, and tools for effective ing, combined with effective knowledge dissemination and replication and scaling up of future handwashing programs, global advocacy strategies. www.wsp.org 1 Findings from the Impact Evaluation Baseline Survey in Peru Overview As reflected above, the process of learning, which is sup- Global Scaling Up Project Impact Evaluation ported in monitoring and evaluation components, is con- Rationale and Aims sidered critical to the project's success. As part of these The overall purpose of the IE is to provide decision makers efforts, the project will document the magnitude of health with a body of rigorous evidence on the effects of the hand- impacts and relevant project costs of the interventions. To washing and sanitation projects at scale on a set of relevant measure the magnitude of these impacts, the project is im- outcomes. It also aims to generate robust evidence on a plementing an impact evaluation (IE) using a randomized- cross-country basis, understanding how effects vary accord- controlled experimental design in each of the four countries ing to each country's programmatic and geographic con- to establish the causal effect of the intervention (treatment) texts, and generating knowledge of relevant impacts such as on specific health and welfare outcomes. The IE includes child cognitive development, anthropometric measures, several rounds of household and community surveys: pre- anemia, acute lower respiratory infection, and productivity intervention (baseline), concurrent (longitudinal), and post- of mother's time, among many others. intervention (endline). The surveys are designed to collect information on the characteristics of the eligible population The studies will provide a better understanding of at-scale and to track changes in desired outcomes. sanitation and hygiene interventions. The improved evidence will support development of large-scale policies This report is part of a series presenting the analysis of base- and programs, and will inform donors and policy makers line data collection surveys conducted in the implementation on the effectiveness and potential of the Global Scaling Up countries during 2008 and 2009. projects as massive interventions to meet global needs. 2 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Overview 1.2 Project Background In Peru, the handwashing project targets mothers/caregivers of children under Children under five represent the age five years old, and it is aimed at improving handwashing with soap practices. group most susceptible to diarrheal Children under five represent the age group most susceptible to diarrheal disease disease and acute respiratory infections, which are two major causes and acute respiratory infections, which are two major causes of childhood mor- of childhood morbidity and mortality in bidity and mortality in less developed countries. These infections, usually trans- less developed countries. ferred from dirty hands to food or water sources, or by direct contact with the mouth, can be prevented if mothers/caregivers wash their hands with soap at critical times (such as before feeding a child, cooking, eating, and after using a toilet or changing a child's diapers). In an effort to improve handwashing behavior, the intervention borrows from In an effort to improve handwashing both commercial and social marketing fields. This entails the design of com- behavior, the intervention borrows from munications campaigns and messages likely to bring about the desired behavior both commercial and social marketing fields. changes, and delivering them strategically so that the target audiences are "sur- rounded" by handwashing promotion. Some key elements of the intervention include: · Key behavioral concepts or triggers for each target audience · Persuasive arguments stating why and how a given concept or trigger will lead to behavior change, and · Communication ideas to convey the concepts through many integrated activities and communication channels. School initiative promotes handwashing with soap in Cajamarca www.wsp.org 3 Findings from the Impact Evaluation Baseline Survey in Peru Overview 1.3 Project Components The overall objective of the project is to improve the health of populations at risk of diarrhea and ALRI, especially in children under five years old, through a stra- tegic communications campaign aimed at increasing handwashing behavior with soap at critical times (before preparing food, feeding, or eating, and after going to toilet or changing diapers). In Peru, the handwashing project is In Peru, the handwashing project is implemented in 788 randomly selected dis- implemented in 788 randomly selected tricts in a total of 104 provinces. The intervention has the objective to stimulate districts in a total of 104 provinces. The and sustain handwashing behavior change in a total of 1.3 million women and intervention has the objective to reach children. The implementation comprises two different components: 5.9 million women and children. · Component 1--Mass Media and Promotional Events: Mass-communications campaign at the provincial level The communications strategy focuses on the availability and use of soap for handwashing and the need to wash hands with soap immediately before cooking or eating, and after going to the bathroom. It targets women ages 14 to 49, and children from 5 to 12 years of age. The main means of communication are local media (mainly radio) and unconven- tional media, such as market speakers. · Component 2--School & Community: Social mobilization at the district level This component comprises several activities to achieve an integral and sus- tainable change at the community level. It also targets women from 14 to 49 and children from 5 to 12 years of age, but it engages multiple actors in the community over a period of time; these actors participate and become agents of change. The specific activities include: i. Institutional development elements to ensure sustainability, includ- ing advocacy, partnership building, and capacity strengthening, ii. A communications campaign through local media and promotional events (street parades, local theaters, etc.) focused on the school and community, and iii. Training of community actors and agents of change (such as teach- ers, medical professionals, community leaders), and provision of educational handwashing sessions for mothers and children. 1.4 Objectives of the Study The objective of the IE is to assess the The objective of the IE is to assess the effects of the project on individual-level effects of the project on individual- handwashing behavior and practices of caregivers and children. By introducing level handwashing behavior and exogenous variation in handwashing promotion (through randomized exposure practices of caregivers and children. to the project), the IE also addresses important issues related to the effect of The IE also addresses important issues related to the effect of intended intended behavioral change on child health and development outcomes. In par- behavioral change on child health and ticular, it provides information on the extent to which improved handwashing development outcomes. behavior impacts infant health and welfare. 4 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Overview The IE aims to address the following primary research 3. Which promotion strategies are more cost-effective questions and associated hypotheses: in achieving desired outcomes? 1. What is the effect of handwashing promotion on handwashing behavior? The purpose of this report is to provide baseline informa- 2. What is the effect of handwashing promotion on tion for the selected indicators and outcomes of interest in- health and welfare? cluded in the survey. www.wsp.org 5 II. Methodology 2.1 Randomization To address the proposed research questions, a proper IE methodology that estab- lishes the causal linkages between the intervention and the outcomes of interest is needed. In order to estimate the causal In order to estimate the causal relationship between the project (treatment) and the relationship between the project outcomes of interest, the construction of an accurate counterfactual is required-- (treatment) and the outcomes of that is, one needs a comparison group that shows what would have happened to interest, the construction of an accurate the target group in the absence of the intervention. In the case of the project in- counterfactual is required--that is, one needs a comparison group that tervention, which is being implemented over a two-year period, it is possible that shows what would have happened to factors such as weather, macro-economic shocks, or other new and ongoing public the target group in the absence of the health, nutrition, sanitation, and hygiene campaigns, for example, could influence intervention. the same set of outcomes that are targeted by the project (e.g., diarrhea incidence in young children, health, and welfare). To account for factors external to the in- tervention, counterfactuals are created using comparison groups (control) that are equivalent to the treatment group on every dimension (observed and unobserved) except for the treatment, and thus account for time-varying factors that may affect the target population. Since a good counterfactual approximates what would have happened to treatments in the absence of the treatment, any differences in the aver- age outcome measurements of treatment and control groups following the imple- The randomization process, by which mentation can be understood as the causal effect of the intervention. a random selection of communities receives the treatment and the remaining serve as controls, generates The randomization process, by which a random selection of communities receives an appropriate counterfactual for the the treatment and the remaining serve as controls, generates an appropriate counter- purposes of the impact evaluation. factual for the purposes of the impact evaluation. Random assignment of treatment Survey team interviews caregivers 6 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Methodology to a sub-set of communities can ensure the treatment and comparison groups are equal, on average,1 and thus an appropriate counterfactual can be measured. A ran- domized experimental evaluation with a comparison group is valuable because it reduces the possibility that observed before-and-after changes in the intervention group are due to factors external to the intervention. If no control group is main- tained and a simple pre-to post-assessment is conducted of the project, one cannot attribute changes in outcomes to the intervention with certainty. The use of a random control group also helps to prevent other problems that affect our inference about the effects of the intervention. For example, communities chosen purposively as areas with a high likelihood of success for programs such as the project because of favorable local conditions (strong leadership, existing water and sanitation infrastructure, highly educated population, etc.) are likely to be different from areas that are considered less desirable for implementation. If a non-random control group is used, a comparison of treated and untreated areas would confuse the program im- pact with pre-existing differences, such as different hygiene habits, lower motivation, or other factors that are difficult to observe. This is known as selection bias. A random control group avoids these difficulties by ensuring that the communities that receive the program are no different on average than those that do not. 2.2 Study Design In order to assess the impact of each of the components of the project in the In order to assess the impact of each health of children younger than five years old, the evaluation study has two main of the components of the project in treatments, that is, one per component. These are the Mass Media Treatment at the health of children younger than five years old, the evaluation study has two the provincial level, also referred to as Treatment 1 (T1), and the Social Mobiliza- main arms, that is, one per component tion Treatment at the district level, also referred to as Treatment 2 (T2). As previ- or treatment. ously mentioned, in order to evaluate and identify the health impacts of each component, a counterfactual to T1 and T2 is needed, which we refer to as the Control (C). The three groups, T1, T2, and C include households with children under two years old at the time of the baseline. Additionally, the evaluation assesses the isolated impact of one subcomponent of Additionally, the evaluation T2: the promotion of handwashing behavior in primary schools, implemented in assesses the isolated impact of one a limited number of schools. This school effect can be estimated by comparing subcomponent of T2: the promotion of handwashing behavior in primary households with children who attend "treated" primary schools to its counterfac- schools, implemented in a limited tual, that is, households with children who attend similar primary schools, but number of schools. where handwashing promotion is not offered. Thus, to evaluate the impact of the school subcomponent, two additional groups are necessary: Treatment 2 in Schools (T2-Schools) and an extra counterfactual (C-Schools). This design allows us to investigate the impact of T1 and T2 (relative to con- trol districts), and also enables us to investigate the differential impact on 1 Technically, this is only true with infinite sample sizes, which is unaffordable and unnecessary. Instead, this study seeks to minimize the risk that the means of the treatment and comparison groups differ significantly. For details of mean comparison tests across treatment and control groups, please see Annex 3: Test of Baseline Balance. www.wsp.org 7 Findings from the Impact Evaluation Baseline Survey in Peru Methodology households that have children in treated schools from that on households that do not (T2 relative to T2-Schools). 2.3 Sampling Size and Strategy The primary objective of the project is to improve the health and welfare of young The sample size (total number of children. The sample size (total number of households) was chosen to capture a households) was chosen so as to minimum effect size of 20 percent on the key outcome indicator of diarrhea capture a minimum effect size of prevalence among children under two years old at the time of the baseline. The 20 percent on the key outcome selection of households with children in this age group was made under the as- indicator of diarrhea prevalence amongst children under two years old sumption that health outcome measurements for young children in this age range at the time of the baseline. are most sensitive to changes in hygiene in the environment. Data was collected for household members of all age ranges and the corresponding data analysis was conducted for older children and adults as well. Power calculations indicated that, in order to capture a 20 percent reduction in diarrhea incidence, around 600 households per treatment arm would need to be surveyed. Therefore, since the evaluation consists of three treatment groups and two control groups, the final sample incorporates approximately 3,000 households, each with children less than two years of age at the time the survey was conducted. An additional 500 households were added to the sample size in order to address potential attrition (loss of participants during the project); thus the minimal necessary sample size was 3,500 households (around 700 households per arm). To select the sample, the IE team used a three-stage sampling methodology: · Stage 1: Province Level From 195 total provinces in Peru, Pisco and Lima were excluded at the request of the implementation team.2 Of the remaining 193 provinces, 80 provinces were randomly chosen. Out of these 80 provinces, two groups of 40 provinces each were randomly formed: Group of Provinces 1 (GP1) and Group of Provinces 2 (GP2). · Stage 2: District Level Out of the first group of 40 provinces, GP1, 40 districts between 1,500 and 100,000 habitants were randomly chosen to receive T1. From the second group, GP2, 80 districts between 1,500 and 100,000 habitants were selected randomly; 40 of them were randomly assigned to receive T2, and the other 40 districts to serve as C to T1 and T2. · Stage 3: Household Level For each of the three sets of 40 districts (120 districts total) allocated to T1, T2, and C, 15-20 households with children under two years of age were selected at random in each district. Also, in each of the 40 districts 2 The province of Pisco was excluded because an earthquake had just hit the area. The province of Lima was excluded for being mainly urban and because most of its districts were too large for this type of intervention. 8 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Methodology allocated to T2, an additional set of 15­20 households with children under two and at least one sibling attending a treatment school was randomly chosen to assess the isolated effect of the school sub-component T2- Schools. Finally, in each of the 40 districts allocated to C, an additional set of 15­20 households with children under two and at least one sibling at- tending a no-treatment school was also randomly selected to serve as the counterfactual for T2-Schools (C-Schools). This sample selection process explained above is illustrated in Figure 1. The household survey was based on cluster sampling, and included a total of 120 districts chosen among 80 provinces (both choices made at random). The expectation was to conduct a total of 3,500 household questionnaires and 120 community questionnaires (one per district). By the end of the survey, data was By the end of the survey, data was collected from 3,576 households and 120 districts in 80 provinces. collected from 3,576 households and 120 districts in 80 provinces. In addition to the household survey, fecal samples from children under two years old, water samples taken from caregiver and child's hand rinses, drinking water, and a sentinel toy were collected with the purpose of assessing the health status of children and the level of fecal contamination in the household. These measures were taken from a subsample of 160 households. Structured observations of handwashing behavior were also collected in the same subset of 160 households. FIGURE 1: PERU IMPACT EVALUATION SAMPLE SELECTION 195 provinces (universe) 40 provinces 40 provinces 40 districts 40 clusters 40 clusters T2 T2-School (700 hhs) (700 hhs) Counterfactual Counterfactual 40 districts Counter- T1 factual C C-School (700 hhs) (700 hhs) (700 hhs) 40 clusters 40 clusters 40 clusters 40 districts www.wsp.org 9 Findings from the Impact Evaluation Baseline Survey in Peru Methodology 2.4 Variables for Data Analysis The IE aims to assess both the effect of the project on handwashing behavior and the In order to measure potential impacts effect on infant health and welfare. In order to measure potential impacts of the in- of the intervention the study collects tervention the study collects data on diarrhea, productivity, education, nutrition, data on diarrhea, productivity, child growth and development, iron deficiency, environmental contamination, para- education, nutrition, child growth and development, iron deficiency, site prevalence, and handwashing behavior and its determinants. environmental contamination, parasite prevalence, and handwashing behavior The above variables are collected through three different surveys: the baseline and its determinants. survey (collected before the intervention), a longitudinal survey (collected a total of 10 times before, during, and after the intervention), and a post-intervention survey (collected after the intervention has finalized). Box 1 and Box 2 summarize the variables measured and how measurements were performed. BOX 1: HEALTH AND WELFARE IMPACTS What Does the How Is It Being Measuring Evaluation Measure? Measured? Instrument Diarrhea prevalence Caregiver-reported Household symptoms collected in a questionnaire 14-day health calendar Productivity of mother's Time lost to own and Household time child's illness questionnaire Education benefits School enrollment and Household attendance questionnaire Child growth Anthropometric mea- In household collection sures:3 weight/height, arm of anthropometric and head circumferences measures Child development Caregiver reported Modified Ages & personal-social, Stages Questionnaire communication, and gross (ASQ)4 motor skills Anemia Hemoglobin concentration In household (<110g/L per international collection and analysis standards)5 of capillary blood using the HemoCue photometer Environmental Prevalence of E. coli In household contamination in: drinking water, hand collection of samples, rinses (caregiver & and microbiological children), sentinel toy analysis in lab Parasite prevalence Parasite prevalence in In household fecal samples collection of samples, and parasitological analysis in lab 3 Habicht 1974. 4 Bricker & Squires 1999. 5 Stoltzfus & Dreyfus 1999. 10 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Methodology BOX 2: HANDWASHING BEHAVIOR AND DETERMINANTS What Does the Evaluation How Is It Being Measured? Measuring Instrument Measure? Handwashing w/soap behavior Direct observation of handwashing sta- Household questionnaire tion stocked with soap and water Self-report handwashing with soap Household questionnaire behavior Observed handwashing with soap Structured observations behavior Determinants to handwashing Opportunity, ability, and motivation Household questionnaire with soap behavior 6 determinants observations of handwashing facilities and other dwelling characteristics, handwashing behavior, child discipline, maternal depression, handwashing deter- minants, exposure to health interventions, relation- ship between family and school, and mortality. · Health questionnaire: The health questionnaire was conducted in all 3,576 households in 120 districts to collect data on children's diarrhea prevalence, ALRI and other health symptoms, child develop- ment, child growth, and anemia. · Community questionnaire: The community ques- tionnaire was conducted in 120 districts to collect data on community/districts variables. · Structured observations: Structured observations were conducted in a subsample of 160 households Head circumference is measured to assess child health to collect data on direct observation of handwashing behavior. · Water samples: Water samples were collected in a 2.5 Instruments for Data Collection subsample of 160 households, to identify Escherichia The baseline survey was conducted May through August coli (E. coli) presence in hand rinses (mother and chil- 2008 and included the following instruments: dren), sentinel toy, and drinking water. · Household questionnaire: The household question- · Stool samples: Stool samples were collected in a naire was conducted in all 3,576 households in 120 subsample of 160 households to identify prevalence districts to collect data on household membership, of parasites in children's feces. education, labor, income, assets, dwelling charac- teristics, water sources, drinking water, sanitation, 6 The analysis for determinants to handwashing with soap behavior change is not included in this report. www.wsp.org 11 Findings from the Impact Evaluation Baseline Survey in Peru Methodology The post-intervention survey will be A total of ten longitudinal surveys will be conducted during the study. The post- conducted October through December intervention survey will be conducted October through December 2010 and will 2010 and will collect, at least, all the collect, at least, all the indicators collected during the baseline survey. indicators collected during the baseline survey. The survey instrument was drafted by the WSP global impact evaluation team, which is formed by experts from a variety of disciplines. The complete instru- ment, which included a set of household, community, and longitudinal question- naires, was translated into Spanish and pre-tested in a pilot survey including 60 households. Hemoglobin concentrations were measured in children under two years old at the household level using the HemoCue Hb 201 photometer, a portable device that allows for immediate and reliable quantitative results. Using sterile and disposable lancets (pricking needle), a drop of capillary blood was obtained from the child's second or third finger and collected in a cuvette, and then introduced into the HemoCue machine. Hemoglobin concentration appeared in the display screen of the device in about one minute, and results were trans- ferred to the questionnaire. Anthropometric measures were made according to standardized protocols using portable stadiometers, scales, and measuring tape.7 Water samples from a hand rinse, drinking water, and sentinel objects were ana- lyzed to determine presence of E. coli and other types of coliforms. The samples ® were collected within the household, inoculated using the Colilert reactive and transported to a lab. At the lab, samples were incubated at 35 degrees Celsius for 24 hours, and the results were read using an ultraviolet lamp. This procedure precluded sampling in areas where the cold chain could not be maintained. Fecal Health survey team carries equipment to measure health outcomes 7 Habicht 1974. 12 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Methodology samples were collected in the same subsample of house- supervisors, 30 health members, 45 interviewers, and 10 holds and transported to a central lab in Lima for parasito- observers. logical analysis. Field team supervisors were required to have previous After the questionnaires were administered, structured ob- fieldwork experience in conducting similar studies, a servations were conducted. During a five-hour period, the required level of superior technical education, and to researcher observed the handwashing behavior of the pri- show a satisfactory performance in all areas of training mary caregiver. Opportunities for handwashing for persons (anthropometry, biometrics, and especially question- other than the primary caregiver were also noted if the indi- naire training). Health specialists had to be standard- vidual came into the line of sight of the interviewer. During ized in order to collect anthropometric, anemia, and the five-hour period, the interviewer noted any opportunity Ages and Stages Questionnaire (ASQ) data. The Nutri- for handwashing and whether handwashing occurred dur- tional Research Institute (Instituto de Investigacion ing that time, as well as the details of the opportunity: the Nutricional), with support from the global IE team, type of critical event, the cleansing agent used (e.g., bar conducted the training for the collection of child- soap, liquid soap, mud), washing of both hands, and related data, and was in charge of the standardization in method of hand drying. Critical events of interest included the three measures (anthropometrics, anemia tests, and fecal contact (going to the toilet, defecating, or changing ASQ). Interviewers were required to complete the children's diapers), preparing food, eating, or feeding training satisfactorily and conduct at least three inter- children. views in under-the-average time. Finally, observers (for structured observations) had to complete the training Field team members administered the instruments. Each course successfully and conduct three four-hour obser- field survey team consisted of a team supervisor, two health vations, of which the trainers supervised at least one. members, and three interviewers. Those teams working in districts where structured observations of handwashing Specific training was designed for each member of the sur- behavior were collected included an extra person in charge vey team according to the specific skills required for the task of the observations. Thus, the field personnel for the col- to be performed in the field. lection of the baseline data included a total of 15 field www.wsp.org 13 III. Sample Representativeness 3.1 Geographic Representativeness The purpose of the IE design was to evaluate the causal effect of the intervention on a set of outcomes. As previously discussed, a randomized experimental design was used to ensure an accurate comparison between treatment and control groups. Thus, the evaluation design was intended to be representative of the population targeted by the intervention, rather than representative of the Peruvian population. The sample included in the IE study The sample included in the IE study is not representative of the Peruvian popula- is not representative of the Peruvian tion at the national level because the selection of provinces and districts was random population at the national level and not weighted by population, as would be necessary to be geographically repre- because the selection of provinces and sentative. Because populations differ across provinces and districts, the three-stage districts was random and not weighted by population. sampling design introduced a type of bias (with respect to geographical representa- tiveness) because selection probabilities varied across administrative units. In addition to the national scale, the sample is likewise not representative at the provincial, district, or household levels, due to the following reasons: · At the provincial level, Lima and Pisco were excluded from the overall sample of provinces, and out of the total 195 provinces in the country, only 80 provinces were selected (less than half of the total provinces). · Similarly, at the district level, only 120 districts were selected from over 1,800 districts in Peru (less than 10 percent of the total number of districts). More- over, the sample only included districts with populations between 1,500 and 100,000 inhabitants. An additional characteristic of the districts included in the IE sample is that they all had at least one primary school. Each of these factors suggests that selected districts need not be representative of all districts. Household members during survey interview 14 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Sample Representativeness · Lastly, at the household level, between 15 and 20 households were selected per district. Although the size of the district was taken into consideration in choosing the exact number of households, the population per district was not precisely weighted for representativeness. Rather, the criteria were: 10 households for districts under 2,250 residents; 15 households for dis- tricts between 2,250­6,000 residents; and 20 households for districts over 6,000 residents. Crucially, the IE sample only included households with at least one child less than two years old. These factors prevent the claims about the representativeness of sampled households. The IE sample was designed with the primary intention of producing internally valid estimates of program The IE sample was designed with the primary intention of producing internally impacts and would not be suitable valid estimates of program impacts and would not be suitable for computing coun- for computing country or district try or district level population statistics without appropriate corrections. For fur- level population statistics without ther details on the selected list of provinces and districts, please refer to Annex 1. appropriate corrections. 3.2 Comparison Between WSP Baseline Study and Peru Population In this subsection we compare some basic characteristics of the Peruvian popula- tion against characteristics of the individuals included in the IE subsample. The main reason behind this exercise was to confirm the external validity of the results presented throughout the document. We concentrated on four groups of vari- ables: demographics, educational attainment, occupation, and total household income per capita. We used the Peru 2007 National Household Survey/Encuesta Nacional de Hogares (ENAHO) data for the comparison (ENAHO 2007).8 Table 1 presents the demographics for both subsamples. The population The population included in the WSP included in the WSP impact evaluation baseline survey comprises a much impact evaluation baseline survey younger population than the general population. On average, the individuals comprises a much younger population than the general population. interviewed in the WSP survey were 18.4 years old, whereas the average age of total population was 28.3 years. The primary reason for this difference is that there were no childless households in the WSP sample. While the average num- ber of children under the age of five per household was 0.43 in Peru, this figure was 1.37 in the WSP sample. Regarding educational attainment, there appears to be no significant differences be- Regarding educational attainment, tween the individuals included in the WSP survey and total population (Table 2). there appears to be no significant differences between the individuals Although in this subsample there was a smaller proportion of individuals with no included in the WSP survey and total education, the proportion of those with trade, undergraduate, or graduate education population. was also smaller compared to the total Peruvian population older than 14 years old. In what follows, we focus on the occupational differences between both subsam- ples in order to assess the different possibilities of income generation. As we can 8 We excluded the Metropolitan Area of Lima from the ENAHO. The population considered in the ENAHO was selected following the restriction of age imposed by the WSP survey for each group of questions. Nominal income-related variables were adjusted by the inflation rate of 2008 obtained from the Instituto Nacional de Estadistica e Informatica (INEI). www.wsp.org 15 Findings from the Impact Evaluation Baseline Survey in Peru Sample Representativeness TABLE 1: DEMOGRAPHICS WSP Survey ENAHO Age (% Individuals): 0­4 26.0% 9.9% 5­9 13.0% 10.7% 10­14 9.7% 12.6% 15­19 7.6% 10.7% 20­24 8.6% 8.0% 25­29 9.8% 6.9% 30­34 8.6% 6.6% 35­39 6.0% 6.3% 40­44 3.3% 5.8% 45­49 2.1% 4.9% +50 5.5% 17.6% Average Age 18.64 28.27 Total Number of Children Under Five (% HHs): 0 0.0% 66.6% 1 66.7% 25.1% 2 29.2% 7.4% 3 3.9% 0.8% 4 0.1% 0.1% Average number of children under five (number of children) 1.37 0.43 see in Table 3, the percentage of individuals of the total population over 14 years old that had a job was almost 10% higher than that of the WSP subsample. Fur- thermore, there was a much higher proportion of individuals who "look after the home" in the WSP subsample (31.5%) compared to that of the total population (10%). This last result was probably driven by the presence of at least one child in the WSP survey, since a high proportion of women were the mothers of those young children and stayed at home in order to take care of them. TABLE 2: EDUCATIONAL ATTAINMENT WSP Survey ENAHO Level of Education Attained (% Individuals): No Education 3.41% 9.4% Kindergarten 4.8% 2.0% Primary 45.6% 42.2% Secondary 41.8% 32.3% Trade School 4.8% 7.1% University 2.9% 6.5% Higher 0.0% 0.6% 16 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Sample Representativeness TABLE 3: OCCUPATION WSP Survey ENAHO Last Week Activity (% Individuals): Working 57.2% 67.5% Not working, but has a job 1.4% 1.4% Looking for work 0.9% 2.3% Studying 5.9% 6.0% Looking after the home 31.5% 10.0% Rent earner 0.0% 1.2% Permanently unable to work 0.9% 2.3% Retired 0.1% ­ Not working and not looking for job 2.1% 9.3% Primary Employment Status (% Individuals): Self-employed 54.4% 37.8% Employee 36.6% 30.2% Employer or boss 0.4% 5.6% Worker with no remuneration 7.8% 23.5% Day laborer 0.9% ­ Other 0.1% 2.90% We also find important differences concerning primary employment status. In the The average salary of the dependent WSP subsample there was a much larger proportion of self-employed workers workers surveyed by the WSP survey than in the total population, 54.4% and 37.8% respectively. Also, the WSP popu- was 521.91 Peruvian nuevos soles (S/.), while the average salary of those lation had a smaller proportion of employers and workers with no remuneration, surveyed in the ENAHO was S/. 680.40. indicating a smaller household income per capita in the WSP sample. Finally, we present two measures of income: salaries received in the primary work and total household income per capita.9 Figure 2 presents the distribution of salaries divided into two groups: dependent and independent workers. The average salary of the dependent workers surveyed by the WSP survey was 521.91 Peruvian nuevos soles (S/.), while the average salary of those surveyed in the ENAHO was S/. 680.40. Moreover, the maximum salary earned by depen- dent workers in the WSP survey was almost S/. 1,000 less than the one earned by total dependent workers in Peru. The same difference applies to the subsample of independent workers included in the WSP survey, whose average income was S/. 332.70, while that of the total independent workers of Peru was S/. 381.70. 9 In the ENAHO we considered the gross salary for the dependent workers. For the independent workers, we included the payments received in kind, since the ENAHO does not divide the independent worker's income into monetary and inkind income (the WSP survey does not include income perceived in kind). The ENAHO measure of total HH income per capita includes: dependent workers' salary, independent workers' income, other labor income, domestic and foreign transfers, income received from the rent of household assets, and other extraordinary income. For these three income measures we used the imputed, deflated, and annualized variables provided by the ENAHO, which were inflation-adjusted and divided by 12 in order to have monthly values. www.wsp.org 17 Findings from the Impact Evaluation Baseline Survey in Peru Sample Representativeness FIGURE 2A: DISTRIBUTION OF SALARIES RECEIVED IN THE PRIMARY OCCUPATION: DEPENDENT WORKERS WSP Survey ENAHO .0025 .0025 .0020 .0020 .0015 .0015 Density Density .0010 .0010 .0005 .0005 0 0 0 500 1000 1500 2000 0 500 1000 1500 2000 Monthly Salary of Dependent Worker Monthly Salary of Dependent Worker kernel epanechnikov, bandwidth 45.0000 kernel epanechnikov, bandwidth 19.6230 FIGURE 2B: DISTRIBUTION OF SALARIES RECEIVED IN THE PRIMARY OCCUPATION: INDEPENDENT WORKERS WSP Survey ENAHO .005 .005 .004 .004 .003 .003 Density Density .002 .002 .001 .001 0 0 0 500 1000 1500 0 500 1000 1500 Monthly Salary of Independent Worker Monthly Salary of Independent Worker kernel epanechnikov, bandwidth 45.0000 kernel epanechnikov, bandwidth 10.7137 Even before considering other types of household income, one could predict that, on average, the total household income per capita was going to be much smaller in the household interviewed by the WSP survey. The main reasons for this are that households in our sample had on average a larger household size, as well as less labor income. Figure 3 presents the distribution of the total household income per capita. The average monthly income per capita among households included 18 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Sample Representativeness FIGURE 3: DISTRIBUTION OF MONTHLY INCOME PER CAPITA WSP Survey ENAHO .006 .006 .004 .004 Density Density .002 .002 0 0 0 200 400 600 800 1000 0 200 400 600 800 1000 Monthly Income per Capita Monthly Income per Capita kernel epanechnikov, bandwidth 45.0000 kernel epanechnikov, bandwidth 8.3893 in the WSP survey was S/. 165.30; on the contrary, Peru's individual's responses to be influenced by the individual's average monthly household income per capita was S/. 328.60. household income, possibly because for those households Therefore, since our subsample was on average poorer than with lower income, income level may have had a higher the average Peruvian households, we expected many of the marginal effect on the topics covered in this report. www.wsp.org 19 IV. Findings Throughout this report, we Throughout this report, we disaggregate all the findings by income and geo- disaggregate all the findings by graphic criteria, and for outcomes of interest (child development, diarrhea, etc.) income and geographic criteria, we also disaggregate the variables by sanitary conditions: access to water, sanita- and for outcomes of interest (child tion facilities, and a handwashing station. The importance of this group of vari- development, diarrhea, etc.) We also disaggregate the variables by sanitary ables is directly related to their effects over the probability of an individual getting conditions: access to water, sanitation sick due to unsanitary-environment related diseases. facilities, and a handwashing station. Table 4 presents summary statistics related to access to improved drinking-water source and improved sanitation facility,10 as well as access to an observed handwashing station with soap and water.11 On average, 47.8% of the surveyed households had access to improved sanitation. This figure rose to 54.4% for households located in a coastal area and declined to 32.5% for those located in the jungle. The number of households with access to improved water was higher; over 75.6% of the households had access to an improved water source. Again, this proportion was higher for those households located in a coastal area and lower for those in the jungle, 86.3% and 62% respectively. Finally, almost 65% of the households had a handwashing station with soap and water. Households in the jungle of Peru had the highest percentage of handwashing stations with soap and water. Map 112 presents a disaggregation of these variables by administrative department. The proportion of households having access to improved sanitation and improved water sources was clearly higher for the departments located near the Peruvian coast, as we have already mentioned. However, when using maps to show this information, we divide Peru into two large groups of departments with a very unequal percentage of households having improved sanitation and water source. TABLE 4: PERCENT DISTRIBUTION OF WATER, SANITATION, AND HYGIENE CONDITIONS BY GEOGRAPHIC AREA Geographic Area (*) As per JMP Definition Coast Jungle Mountain Total Access to Improved Sanitation Facility* (% HHs) 54.4% 32.5% 47.1% 47.8% Access to Improved Drinking-Water Source* (% HHs) 86.3% 62.0% 72.4% 75.6% Observed HW Station with Soap and Water (% HHs) 66.5% 72.3% 62.0% 64.4% 10 The "Access to Improved Sanitation Facility" and "Access to Improved Drinking-Water Source" variables were created following the definitions and recommendations made by the WHO/UNICEF Joint Monitoring Programme (JMP) for Water Supply and Sanitation (http://www.wssinfo.org/definitions/infrastructure.html). 11 The variable change "Observed HW Station with Soap and Water" responds to the number of households with an observed handwashing station stocked with soap AND water within the dwelling and/or yard premises. 12 The maps were computed without using sampling weights. 20 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings MAP 1: MAP OF PERU WITH DESCRIPTIVE STATISTICS BY ADMINISTRATIVE DEPARTMENT PERCENTAGE OF HOUSEHOLDS PERCENTAGE OF HOUSEHOLDS WITH A WITH IMPROVED WATER SOURCE HANDWASHING STATION WITH SOAP AND WATER COLOMBIA COLOMBIA ECUADOR ECUADOR BRAZIL BRAZIL PACIFIC PACIFIC OCEAN OCEAN LIMA 80­90 LIMA 70­80 BOLIVIA BOLIVIA 75­100 60­70 50­75 50­60 25­50 40­50 0­25 30­40 NO DATA NO DATA CHILE CHILE PERCENTAGE OF HOUSEHOLDS WITH IMPROVED SANITATION AVERAGE INCOME GROUP COLOMBIA COLOMBIA ECUADOR ECUADOR BRAZIL BRAZIL PACIFIC PACIFIC OCEAN OCEAN LIMA LIMA HIGH BOLIVIA BOLIVIA 60­80 UPPER-MIDDLE 40­60 MIDDLE 20­40 LOWER-MIDDLE 0­20 LOW NO DATA NO DATA CHILE CHILE www.wsp.org 21 Findings from the Impact Evaluation Baseline Survey in Peru Findings We then analyze the proportion of households having soap and water available at the handwashing station, disaggregated by department. We find that many de- partments having a very low proportion of households with improved sanitation and water source present a high proportion of households having soap and water at handwashing station. In order to provide an explanation for the results recently found, we show the distribution of the Peruvian departments according to some measure of house- hold wealth level. As previously mentioned throughout this document, the tabu- lation of the variables is disaggregated by total household income per capita quartiles, which is an important determinant of certain household characteristics (especially in this subsample, where there is a large proportion of poor individuals relative to the total Peruvian population). For this purpose, total household in- come was calculated considering the total monthly labor income provided by household members (salaries received in the first, second, and/or other jobs, in- come received from a pension plan, unemployment, and/or health insurance) and the total monthly household non-labor income (interest on investments, rents, scholarships, government transfers, donations, income received from household and/or agricultural production, etc.).13 Total household income per capita was calculated by dividing total household income by the total number of household members; the quartile classification was constructed by considering only one observation per household. The result of this classification is geographically displayed in Map 1 (Average Income Group). The relevant division of Peru according to the average income group by depart- ment seems to be a North-South classification. The correlation figures presented in Table 5 reinforces the weak relationship between these four variables. 4.1 General Household Characteristics An average household consisted of five Table 6 shows a brief summary of household basic socio-economic variables. An individuals, among whom there was average household consisted of five individuals, among whom there was more more than one child younger than five than one child younger than five years old. Household heads were 37 years old on years old. Household heads were average, half of them had some level of secondary education, and almost everyone 37 years old on average, half of them had some level of secondary education, was employed. Their average monthly income was S/. 482 (equivalent to and almost everyone was employed. US$17414), which varied highly across household heads (S/. 453). Other house- hold members were, on average, much younger (14.5 years old) and less educated (only 38.5% had some level of secondary education). More than a third of other household members were employed and their average monthly income was S/. 320 (equivalent to US$115). Finally, the average household income per capita was certainly low in comparison with the average Peruvian family (S/. 165, equivalent to US$59). 13 Interviewee responses related to income sources and income reception frequencies were standardized into a monthly frequency, considering months of 30 days. When specific information was not available, individual labor income was estimated by an earnings equation. These estimated incomes were not included when presenting labor income statistics. 14 The US-Nuevos Soles exchange rate was provided by the Central Bank of Peru, on March 15, 2010. 22 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 5: CORRELATIONS BETWEEN WATER, SANITATION, HYGIENE CONDITIONS, AND INCOME GROUP Observed HW Access to Improved Access to Improved Station with Soap Income Water Source Sanitation and Water Group Access to Improved Drinking-Water Source 1.000 Access to Improved Sanitation 0.248 1.000 Observed HW Station with Soap and Water 0.167 0.180 1.000 Income Group 0.068 0.254 0.132 1.000 TABLE 6: SUMMARY STATISTICS This figure was higher for poor households and lower for rich Standard households. There was a higher proportion of females (52.33%) Mean Deviation than males (47.67%) in this sample, but about 90.48% of the HH size 5.3 1.8 3,576 interviewed households had a male household head. Number of children under five years 1.4 0.6 HH Head: Table 8 presents the percent distribution of education for Age 36.9 11.6 individuals aged five years and older. A high proportion of HH head has secondary education them attended school, even in the case of poor households. (% HH heads) 50.4% -- Notwithstanding, 35% of the household heads had attained HH head is employed (% HH primary education only, while 50.4% of them had received heads) 95.2% -- secondary education. These figures were lower for the rest of Labor income (in S/.) 482.7 453.4 the household members and for poorer households. Other HH Members: Age 14.5 14.8 When asked about their weekly time distribution, currently Other HH member has secondary enrolled students answered that they spent most of time at education school (with no significant differences found between the (% other HH members) 38.5% -- sexes). The figures are summarized in Table 9. Only 2.2% of Other HH member is employed the males and 0.7% of the females had a paid job; 6.4% and (% other HH members) 37.0% -- 4.5% of the males and females, respectively, worked without Labor income (in S/.) 320.7 348.1 a salary. Regarding school and household related activities, HH monthly income per capita females tended to spend more time taking care of children (in S/.) 165.3 152.6 than males, and slightly more time doing school homework. Table 7 presents the distribution of basic household demo- The survey collected detailed information on the assets and graphic variables: age of the household members, household non-labor income that each household possesses, and on size, and total number of children under the age of five per the characteristics of the dwelling in which each household household. The mean and median age of the household mem- resides: type of dwelling; ownership situation; walls, floor, bers was 18.6 and 15 respectively. The higher concentration of and roof material; light source; cooking and heating fuel. individuals was among the younger ones. On average, poorer Table 10 presents a complete summary of household assets households were composed of younger members. The mean per income quartile. household size was 5.8 members for the poorest households and 5.1 for richest. While 74.1% of the poorest households Almost 20% of the households declared having income had five or more members, only 54.1% of the richest ones had sources other than labor, and this percentage was higher for the same number of members. Furthermore, the mean num- poorer households. The average non-labor income, consid- ber of children under the age of five per household was 1.4. ering only positive values, was S/. 126 per household. www.wsp.org 23 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 7: PERCENT DISTRIBUTION OF THE BASIC SOCIO-DEMOGRAPHIC CHARACTERISTICS Income Quartile 1st 2nd 3rd 4th Total Age: 0­4 7.2% 6.9% 6.2% 5.7% 26.0% 5­9 4.5% 3.6% 2.7% 2.2% 13.0% 10­14 3.6% 2.6% 1.6% 1.8% 9.7% 15­19 1.8% 1.8% 1.8% 2.2% 7.6% 20­24 1.3% 2.1% 2.1% 3.0% 8.6% 25­29 2.2% 2.5% 2.6% 2.6% 9.8% 30­34 2.4% 2.0% 2.2% 2.0% 8.6% 35­39 1.7% 1.5% 1.5% 1.4% 6.0% 40­44 1.0% 0.8% 0.6% 0.9% 3.3% 45­49 0.5% 0.6% 0.5% 0.5% 2.1% 50 1.3% 1.3% 1.4% 1.6% 5.5% Total 27.3% 25.5% 23.2% 23.9% 100.0% Age of HH head (average) 37.01 36.84 36.04 37.52 36.85 Age of other HH members (average) 13.35 14.01 14.99 16.12 14.55 HH head is male 39.6% 35.8% 38.5% 36.9% 37.7% (% HH heads) Other HH member is male 87.7% 93.8% 90.5% 90.0% 90.5% (% other HH heads) HH Size: 2 0.4% 0.7% 0.1% 0.4% 0.4% 3 10.2% 10.8% 17.9% 17.4% 14.1% 4 15.3% 23.3% 28.6% 28.1% 23.8% 5 23.6% 24.5% 24.6% 20.2% 23.2% 6 20.7% 16.4% 15.5% 17.5% 17.5% 7 12.5% 11.7% 5.5% 5.6% 8.8% 8 8.2% 5.6% 3.5% 6.3% 5.9% 9 5.4% 3.4% 2.8% 1.1% 3.2% 10 2.6% 2.2% 1.1% 2.3% 2.0% 11 0.9% 1.2% 0.1% 0.8% 0.8% 12 0.2% 0.2% 0.3% 0.4% 0.3% 13 0.0% 0.0% 0.1% 0.0% 0.0% 15 0.0% 0.3% 0.0% 0.0% 0.1% HH size (average) 5.8 5.5 4.9 5.1 5.3 Total Number of Children Under Five Years of Age: 1 56.3% 57.9% 71.1% 81.6% 66.7% 2 36.9% 38.1% 26.2% 15.7% 29.2% 3 6.4% 4.0% 2.7% 2.7% 3.9% 4 0.4% 0.0% 0.0% 0.0% 0.1% Number of children under five years of age (average) 1.5 1.5 1.3 1.2 1.4 24 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 8: PERCENT DISTRIBUTION OF INDIVIDUAL'S EDUCATION Income Quartile 1st 2nd 3rd 4th Total Number of HH heads that attended school (% HH heads) 93.8% 97.7% 98.7% 98.9% 97.3% Educational Attainment of HH Head (% HH Heads): Kindergarten 0.0% 0.1% 0.0% 0.0% 0.0% Primary 53.2% 35.4% 28.8% 23.5% 35.0% Secondary 43.6% 57.1% 48.3% 52.6% 50.4% Trade School 2.0% 5.6% 16.2% 11.5% 8.9% University 1.0% 1.9% 6.7% 12.1% 5.5% Higher 0.0% 0.0% 0.0% 0.4% 0.1% No Education 0.1% 0.0% 0.0% 0.0% 0.0% Other HH members ( 5 years old) attended school (% other HH heads) 93.0% 97.0% 97.3% 98.7% 96.4% Educational Attainment of Other HH Members (% HH Members Other Than HH Head): Kindergarten 8.3% 7.4% 4.2% 5.4% 6.4% Primary 63.8% 55.0% 42.3% 33.0% 49.0% Secondary 25.5% 33.0% 45.7% 52.0% 38.6% Trade School 1.2% 3.4% 4.9% 4.4% 3.4% University 0.4% 0.3% 2.6% 5.2% 2.0% No Education 1.0% 0.9% 0.5% 0.0% 0.6% TABLE 9: ACTUAL DISTRIBUTION OF STUDENTS' TIME Male Female Total Teenagers Spent Hours in (% HH Teenagers): School 94.5% 95.8% 95.2% Studying 96.6% 97.3% 96.9% Children care 65.7% 74.4% 70.0% Homework 64.3% 69.4% 66.9% Paid work 2.2% 0.7% 1.4% Unpaid work 6.4% 4.5% 5.5% As expected, poorer households (1st quartile) had on aver- age lower non-labor incomes than richer households (4th quartile), S/. 92 and S/. 197 respectively. This is not Dwelling characteristics are observed for each household surprising since a higher percentage of poorer households tended to work on and receive income from agricultural activities. This is reflected by the higher percentage of poor households possessing other plots of land, farm The figures show that the majority of the households, equipment, and a higher average number of animals per 78.3%, had a radio, cassette, or CD player. This percentage household (these animals are specifically "farm," not was higher for the richest households, 90.3%, but it was domestic, animals). also high for the poorest households, 72.2%. Owning www.wsp.org 25 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 10: PERCENT DISTRIBUTION OF HOUSEHOLD ASSETS AND NON-LABOR INCOME Income Quartile 1st 2nd 3rd 4th Total Average household non-labor income (in S/.) 92.1 112.4 139.0 197.4 125.9 HH Assets (% HHs): Radio, CD, cassette 72.2% 75.7% 75.2% 90.3% 78.3% TV 39.0% 63.3% 82.2% 85.5% 67.5% VCR 9.6% 18.4% 33.3% 55.1% 29.1% Computer 0.0% 1.2% 1.7% 9.1% 3.0% Bicycle 12.0% 20.1% 29.1% 29.7% 22.7% Motorbike 1.2% 3.2% 5.1% 5.0% 3.6% Car or Tractor 0.2% 0.4% 4.3% 8.6% 3.4% Refrigerator 3.3% 8.9% 13.1% 34.1% 14.8% Gas stove 19.2% 50.6% 69.4% 84.0% 55.7% Other type of stove 12.9% 12.7% 13.9% 17.6% 14.3% Blender 11.1% 17.6% 37.7% 62.8% 32.3% Toaster 0.0% 0.7% 1.4% 6.1% 2.0% Microwave 0.0% 0.0% 2.3% 6.2% 2.1% Washing machine 0.0% 1.8% 0.3% 3.7% 1.4% Water boiler 0.4% 1.1% 2.5% 10.9% 3.7% Other houses/properties 16.6% 8.8% 7.4% 6.9% 10.0% Machinery, equipment for family business 1.3% 2.9% 2.2% 4.0% 2.6% HH owns other piece of land (% HHs) 43.5% 29.7% 20.7% 15.9% 27.5% HH owns farm equipment (% HHs) 24% 16% 9% 11% 15% HH has animals (% HHs) 78% 69% 58% 54% 65% luxury items such as a TV or VCR will vary highly based on other walling materials like mud/bamboo/canvas, tin/zinc income status; for instance, 85.5% of the richest house- sheeting, and woven mats was rare, regardless of the income holds had a TV, while the percentage for poorest house- group. Tin/zinc sheeting was the most common roofing ma- holds was only 39%. On average, only 14.8% of households terial (55.1%), followed by brick (10.3%), woven mats had a refrigerator, and the figure was much lower for poor- (7.1%) and concrete (6.9%). In 53.1% of the dwellings the est households (3.3%). Regarding cooking stoves, 84% of floor was clay or dirt (this figure rose to 78.1% in the case of the richest households had a gas stove, while in the poorest the poorest households) and in 41.3% of the dwellings the households the percentage was only 19.2%. material used was concrete (polished or unpolished). The analysis of the household dwelling characteristics displayed The survey also included information regarding dwelling's by Table 11 shows that more than 95% of the households lived lighting source and type of fuel used for cooking and heat- in a detached, independent dwelling. The average number of ing the dwelling. In 75.5% of the surveyed households elec- rooms per dwelling was 2.97. Also, in 48.1% of the cases the tricity was the primary lighting source, with candles being owner of the dwelling (fully paid) was a household member, in the second alternative (14%), and kerosene the third (8%). 22.9% the dwelling was borrowed from a friend or family Forty-nine percent of the households used gas as the pri- member and in 8.2% of the cases the dwelling was rented. mary cooking fuel (13.3% of the poorest households), fol- lowed by wood (39.1% of the total number of households Concerning dwelling materials, 37.5% of the households and 71.6% of the poorest). Almost none of the households had walls made of un-backed brick/adobe, 22.3% and heated their dwelling (97.2%), and those that did used pri- 8.2% had brick and wood/logs walls, respectively. The use of marily a wood stove (2.6%). 26 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 11: DWELLING CHARACTERISTICS Income Quartile 1st 2nd 3rd 4th Total Type of Dwelling (% HHs): Detached house 95.4% 94.5% 95.1% 95.8% 95.2% Room in other dwelling 2.7% 3.4% 2.9% 2.9% 3.0% Other 1.9% 2.1% 2.0% 1.3% 1.9% Average number of dwelling's rooms 2.90 3.00 2.96 3.04 2.97 Dwelling Ownership (% HHs): HH member, still paying 4.1% 5.3% 4.5% 6.1% 5.0% HH member, fully paid 47.4% 48.2% 49.8% 46.9% 48.1% Rented 6.0% 6.7% 11.8% 8.4% 8.2% Family/Friend Loan 20.5% 20.7% 23.4% 27.2% 22.9% Other 22.1% 19.1% 10.6% 11.5% 15.8% Walling Materials (% HHs): Brick 6.0% 15.9% 23.1% 44.5% 22.3% Concrete 1.4% 5.9% 11.6% 6.4% 6.3% Unbaked brick, adobe 59.3% 39.9% 31.6% 19.0% 37.5% Wood, logs 9.3% 7.0% 7.2% 9.2% 8.2% Woven mats 2.0% 6.3% 6.0% 3.4% 4.4% Other 22.1% 24.9% 20.5% 17.5% 21.3% Roofing Materials (% HHs): Brick 0.5% 5.6% 9.2% 26.0% 10.3% Concrete 1.6% 6.4% 8.2% 11.4% 6.9% Wood, logs 0.6% 0.6% 0.3% 1.1% 0.7% Tin, zinc sheeting 61.7% 59.6% 55.0% 43.9% 55.1% Bamboo 2.2% 2.0% 4.5% 2.5% 2.8% Woven mats 4.2% 8.6% 9.1% 6.4% 7.1% Other 29.3% 17.2% 13.7% 8.8% 17.3% Flooring Materials (% HHs): Painted wood 0.69% 0.69% 0.40% 0.70% 0.62% Concrete 8.1% 18.7% 25.6% 39.1% 22.8% Clay, dirt floor 78.1% 56.3% 47.1% 30.6% 53.1% Unpolished concrete 9.2% 21.1% 22.7% 21.0% 18.5% Other 3.9% 3.2% 4.3% 8.6% 5.0% Dwelling Lighting Source (% HHs): No Lighting 0.4% 0.1% 0.2% 0.0% 0.2% Electricity 58.7% 72.9% 85.1% 85.5% 75.5% Kerosene 19.6% 8.4% 2.7% 1.0% 8.0% Candles 17.9% 16.9% 9.9% 11.1% 14.0% Other 3.5% 1.7% 2.1% 2.3% 2.4% Dwelling Cooking Fuel (% HHs): Gas 13.3% 40.8% 64.4% 79.2% 49.3% Wood 71.6% 46.1% 24.5% 13.7% 39.1% Peat/Manure 4.2% 2.2% 1.2% 1.4% 2.2% Other 11.0% 10.9% 9.9% 5.8% 9.4% (Continued ) www.wsp.org 27 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 11: (Continued) Income Quartile 1st 2nd 3rd 4th Total Dwelling Heating Fuel (% HHs): Do not heat dwelling 95.8% 96.8% 98.6% 97.6% 97.2% Wood Stove 3.8% 3.1% 1.4% 2.2% 2.6% Other 0.4% 0.1% 0.1% 0.2% 0.2% Table 12 presents information on the principal activity for any individuals over 15 years old. More than 95% of the household heads were employed in the week previous to the interview, but only 37% of the other household members older than 15 years were employed. For the poorest households, these figures were lower (93.6% and 24.1% for household heads and other HH members, respec- tively). The week before the interview, unemployed household heads were mainly looking after their homes and searching for work (33.7% and 20.9%, respec- tively). Regarding the other household members that were unemployed, they spent most of the week looking after their homes and studying (78% and 14.9%, respectively). The rest of the variables correspond to all employed individuals, household heads, and other household members. A very high proportion of those individuals that worked or helped the family generate income were self-employed (54.4%), espe- cially in the poorest households (66.4%). The rest of them were basically employ- ees (36.6%) or workers without remuneration (7.8%). The average monthly salary for the primary job was S/. 411. Those members em- ployed made the highest average salary, S/. 529, followed by those who were self- employed, S/. 333 (with the exception of those who responded "other" to type of employment). The weekly average number of hours worked was 45.6 hours a week; those employed by others or self-employed worked more hours than daily laborers and employers. On average, an individual had worked 9.5 months in the same job. On average, poorer households had worked 10.5 months in the same job, while richer households had worked in the same job for nine months. On average, 3.3% of the households Households were asked if they had lost work or school hours due to their children had lost work or school hours during getting sick, and results are summarized in Table 13. On average, 3.3% of the the previous 14 days to take care households had lost work or school hours during the previous 14 days to take care of their sick child. This percentage of their sick child. This percentage increased for households with unimproved increased for households with unimproved water sources, unimproved water sources, unimproved sanitation, or no handwashing station with soap and sanitation, or no handwashing station water. This percentage was relatively stable across different income levels. How- with soap and water. ever, when looking at these figures by geographic area we observed that a higher percentage of households living in the mountains (4.2%) had lost hours due to children's sickness than those living in the jungle (2.7%) or on the coast (1.5%). 28 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 12: INDIVIDUAL'S ACTIVITY AND PRIMARY WORK Income Quartile 1st 2nd 3rd 4th Total HH head is employed (% HH heads) 93.6% 95.3% 95.9% 95.8% 95.2% Other HH member is employed (% other HH members) 24.1% 30.9% 38.9% 50.5% 37.0% Last Week Activity--HH Head is Unemployed: Looking for work 13.6% 33.2% 25.0% 14.4% 20.9% Studying 0.0% 2.9% 3.2% 0.0% 1.4% Looking after the home 51.1% 19.9% 33.7% 22.2% 33.7% Not working and not looking for job 18.2% 7.7% 27.1% 8.5% 15.4% Other 17.1% 36.4% 11.1% 54.9% 28.7% Last Week Activity--Other HH Member is Unemployed: Looking for work 0.8% 1.1% 2.7% 0.4% 1.2% Studying 14.2% 15.4% 14.9% 15.3% 14.9% Looking after the home 78.5% 78.7% 77.9% 76.8% 78.0% Not working and not looking for job 5.0% 3.4% 3.2% 6.7% 4.5% Other 1.5% 1.5% 1.3% 0.8% 1.3% Primary Employment Status (% All Employed): Self-employed 66.4% 58.8% 50.0% 46.4% 54.4% Employee 14.1% 29.5% 43.9% 51.3% 36.6% Employer or boss 0.0% 0.6% 0.7% 0.0% 0.4% Worker without remuneration 17.6% 9.5% 5.1% 1.9% 7.8% Day laborer 1.8% 1.5% 0.3% 0.2% 0.9% Other 0.1% 0.1% 0.0% 0.1% 0.1% Monthly Salary: Self-employed 135 242 328 609 333 Employee 194 326 473 722 529 Employer or boss 120 264 335 447 305 Day laborer 183 264 266 343 235 Other -- 77 -- 2,000 1,004 Total 147 270 395 669 411 Hours Worked per Week: Self-employed 43.1 40.7 44.3 43.7 42.9 Employee 42.8 48.4 56.0 55.7 53.3 Employer or boss 36.0 55.3 35.5 31.0 43.8 Worker without remuneration 29.6 27.9 27.8 30.2 28.8 Day laborer 39.0 35.3 39.8 34.0 37.1 Other 21.0 38.8 -- 11.0 24.2 Total 40.6 41.8 48.5 49.6 45.6 (Continued ) www.wsp.org 29 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 12: (Continued) Income Quartile 1st 2nd 3rd 4th Total Months Worked in Past 12 Months: Self-employed 10.9 9.9 9.4 8.9 9.8 Employee 9.1 9.5 9.1 9.1 9.2 Employer or boss 12.0 11.7 8.5 11.2 10.0 Worker without remuneration 10.2 9.1 9.2 9.4 9.6 Day laborer 10.3 10.4 7.4 5.7 9.7 Other 12.0 12.0 -- 4.0 9.2 Total 10.5 9.8 9.2 9.0 9.5 TABLE 13: HOUSEHOLDS WITH TIME LOSS DUE TO CHILD ILLNESS % of HHs By Sanitary Conditions: Improved sanitation 2.80% Unimproved sanitation 3.70% Improved water source 3.10% Unimproved water source 3.70% HW station stocked w/soap & water 2.70% No HW station stocked w/soap & water 4.30% By Income Quartile: 1st 3.30% 2nd 3.60% 3rd 3.20% 4th 2.90% By Geographic Area: Coast 1.50% Jungle 2.70% Mountain 4.20% Overall 3.30% 4.2 Water Source and Safe Water-Use Behavior The survey also investigated household water source and the treatment that household members applied to drinking water. Questions related to water source are disaggregated by season (rainy versus dry season); however, as almost every Three-quarters of the households household had the same water source during the whole year, we present the re- (75.6%) had access to an improved sults only for the rainy season. Results are summarized in Tables 14 and 15. water source; this percentage was higher for the wealthiest percentiles. Three-quarters of the households (75.6%) had access to an improved water Households living along the coast of Peru had higher access to improved water source; this percentage was higher for the wealthiest percentiles. Households liv- sources (86.4%) than those living in the ing along the coast of Peru had higher access to improved water sources (86.4%) mountains (72.5%) or the jungle (62%). than those living in the mountains (72.5%) or the jungle (62%). 30 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 14: ACCESS TO IMPROVED WATER SOURCES % of HHs By Income Quartile: 1st 69.60% 2nd 76.70% 3rd 78.50% 4th 77.60% By Geographic Area: Coast 86.40% Jungle 62.00% Mountain 72.50% Overall 75.60% When taking a narrower look at the data, we found three the dwelling, and the rest (65.6%) were located elsewhere. In main sources of drinking water in the surveyed households: 70.5% of the households the water sources were covered, piped water from inside the dwelling (30.2%), piped water while in 25.3% the water sources were uncovered. The per- from public tap (17.2%), and piped water located in the yard centage of covered water sources in poorer households was (12.2%). On average, 19.7% of these water sources were lo- much lower than the average, 58.1%, and higher in richer cated in the household's own yard or plot, 14.6% within households, 75.8%. TABLE 15: TYPE OF WATER SOURCE Income Quartile 1st 2nd 3rd 4th Total HH Source of Water for Drinking Use (% HHs): Piped water, into dwelling 21.4% 26.2% 35.3% 37.9% 30.2% Piped water, into yard, plot 17.3% 10.5% 8.8% 12.0% 12.2% Piped water, public tap, standpipe 11.0% 20.6% 20.3% 17.2% 17.3% Tube well, bore hole 2.0% 2.7% 2.1% 1.2% 2.0% Dug well, protected 1.6% 0.8% 1.6% 2.4% 1.6% Dug well, unprotected 2.9% 2.1% 0.6% 0.7% 1.6% Spring water, protected 15.2% 14.8% 10.3% 6.8% 11.8% Spring water, unprotected 5.4% 1.4% 1.2% 0.5% 2.1% Tanker truck 0.2% 3.8% 9.7% 7.6% 5.3% Surface water 7.8% 3.8% 1.0% 2.9% 3.9% Other 15.3% 13.5% 9.0% 10.7% 12.1% Source Location (% HHs): In own dwelling 15.4% 17.2% 13.7% 11.6% 14.6% In own yard, plot 21.8% 17.1% 18.8% 21.4% 19.7% Elsewhere 62.8% 65.6% 67.6% 67.0% 65.6% Covered Source (% HHs): Covered 58.1% 71.8% 77.8% 75.8% 70.5% Open 39.5% 22.2% 19.7% 17.7% 25.3% Both covered and open 2.3% 6.0% 2.5% 6.5% 4.3% www.wsp.org 31 Findings from the Impact Evaluation Baseline Survey in Peru Findings Drinking water stored inside the house- hold's kitchen In the majority of the households In the majority of the households (80.3%) an adult female was in charge of col- (80.3%) an adult female was in charge lecting water from the source. The task was performed by an adult male only in of collecting water from the source. 16.6% of the households and by a child under 15 years old in 3.1% Among all the households that pay for the water (76.3% of the households), 49% of them received an unlimited amount. Table 16 summarizes water-use behavior. On average, 83.5% of the households stored water. Of those, 82.4% washed the storage container more than once a week, 14.4% washed it once per week, 3% rarely washed the storage container, and almost no one never washed the container. Almost 87% of the households that washed their storage container used soap, detergent, or bleach and 10.7% of them used only water. In comparison, in the poorest households a lower proportion of them used soap, detergent, or bleach and a higher proportion used only water. More than 85% of the households prepared the water before drinking it (79.3% in the case of the poorest households and 94.1% in the case of the richest ones); 88.2% did it every day during the week before the interview, 7.2% did it every other day, and 4.2% prepared the water only once or twice during the entire Less than half of the households had week. Boiling the water was the most common procedure for preparing the drink- access to improved sanitation facilities, ing water (96.8%).15 Also, in 5.6% of the poorest households, individuals let the and among the poorest households water stand and settle before drinking it. access to improved sanitation was as low as 31.6%. On the coast, access 4.3 Sanitation Facilities to improved sanitation increased to 54.4%, but the percentage of Since diarrheal disease is often the result of virus and bacteria propagation, keep- households with improved sanitation ing a clean and disinfected environment is crucial in its prevention, particularly facilities in the jungle is 32.6%. in handwashing and defecation stations. In this section we investigate the most 15 The interviewees were given the possibility to choose more than one procedure for preparing the drinking water. 32 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 16: SAFE WATER-USE BEHAVIOR Income Quartile 1st 2nd 3rd 4th Total Storage Container: Washing Frequency (% HHs): Never 0.2% 0.3% 0.1% 0.0% 0.1% Rarely 4.9% 2.6% 2.0% 2.6% 3.0% Once a week 20.3% 15.3% 11.1% 10.7% 14.4% More than once a week 74.6% 81.8% 86.9% 86.7% 82.4% How Water Container Is Washed (% HHs): Water only 18.5% 8.6% 4.2% 11.2% 10.7% Soap, detergent, bleach 77.9% 90.3% 95.0% 87.8% 87.6% Other 3.6% 1.2% 0.9% 1.0% 1.7% Water Treatment: Frequency (Past Seven Days, % HHs): Not in the past seven days 0.6% 0.3% 0.2% 0.3% 0.3% Every day 81.7% 89.7% 87.2% 93.7% 88.2% Every other day 9.7% 6.6% 8.5% 4.4% 7.2% Once or twice 8.1% 3.4% 4.2% 1.6% 4.2% Water Treatment (Past Seven Days, % HHs): Boiling treatment 95.0% 96.1% 97.7% 98.1% 96.8% Chlorine treatment 3.0% 5.1% 3.2% 2.4% 3.4% Let stand and settle 5.6% 2.1% 0.8% 0.6% 2.2% Other 0.5% 0.4% 0.4% 0.6% 0.5% common sanitation facilities available in the surveyed households. Table 17 shows that less than half of the households had access to improved sanitation facilities, and among the poorest households access to improved sanitation was as low as 31.6%. On the coast, access to improved sanitation increased to 54.4%, but the percentage of households with improved sanitation facilities in the jungle was 32.6%. When looking at the types of sanitation facilities (see Table 18), we observed the An average of 20.3% of the households most common type of toilet facility found in our sample was the flushed toilet had no sanitation facilities of any type, and the figure increased to 30.4% for piped to the sewer system (32.1%), followed by pit latrine without slab or open the poorest households. pit (27.3%). An average of 20.3% of the households had no sanitation facilities of any type, and the figure increased to 30.4% for the poorest households. Most of these facilities were located in the household yard (35.4%), inside the house- hold (33.4%), or in a nearby location less than a 10-minute walk away (22.3%) or in other locations more than a 10-minute walk away (7.3%). On average, more than 8% of the total toilet facilities were public and 28% of them were shared. Poorer households had similar percentages of public and shared toilet facili- ties than wealthier households (7.3% versus 8%, and 24.5% versus 26.5%). Regarding the safety of female household members when using the toilet facility during the night, only 74.7% of them declared being safe. This figure was lower for poorest households and higher for the richest ones (65.9% versus 82.8%, respectively). www.wsp.org 33 Findings from the Impact Evaluation Baseline Survey in Peru Findings Example of open pit latrine in household's backyard TABLE 17: ACCESS TO IMPROVED SANITATION % of HHs By Income Quartile: 1st 31.60% 2nd 42.70% 3rd 50.40% 4th 66.80% By Geographic Area: Coast 54.40% Jungle 32.60% Mountain 47.10% Overall 47.80% When asked about the satisfaction level with the sanitation facility, only 18.3% of the interviewees answered to be very satisfied, 41.7% to be somewhat satisfied, 19.2% less than satisfied and 22% completely dissatisfied. The level of satisfaction with the sani- tation facility improved as income increased. This is consistent with previous figures showing that poorer households had lower access to improved sanitation facilities. Table 19 summarizes household responses when asked about the main reasons for building or improving the toilet facility (only for those cases in which a household member actually built or improved their facility). On average, 41.5% of household heads put family's health consideration as the primary reason, followed by location The principal and most common and cleanness considerations (20.3%) and convenience (17.1%). When asked constraint mentioned by the households for building a private about the probability of installing a private toilet facility during the next 12 months, sanitation facility was the high cost 48.5% of the households declared a low probability and 14.5% of the households involved. declared a zero probability of doing so. The principal and most common constraint 34 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 18: HOUSEHOLD MAIN SANITATION FACILITY CHARACTERISTICS Income Quartile 1st 2nd 3rd 4th Total HH Main Sanitation Facility (% HHs): Flush, to piped sewer system 15.4% 27.6% 35.2% 50.4% 32.1% Flush, to other place 5.9% 6.2% 6.3% 7.1% 6.4% Ventilated improved pit latrine 8.7% 5.0% 4.3% 3.2% 5.3% Pit latrine with slab 3.1% 4.8% 5.3% 6.3% 4.9% Pit latrine without slab, open pit 30.4% 26.3% 32.2% 20.3% 27.3% Hanging toilet, latrine 1.3% 1.2% 1.2% 0.8% 1.1% No facilities 32.3% 26.2% 13.0% 9.6% 20.3% Other 3.0% 2.9% 2.6% 2.3% 2.7% Public toilet facilities (% HHs) 7.3% 8.7% 10.1% 8.0% 8.6% Location of Main Sanitation Facility (% HHs): Inside household 17.0% 33.8% 42.0% 41.1% 33.4% In own yard 39.0% 28.5% 32.6% 41.5% 35.4% Less than 10-min. walk 28.5% 27.2% 18.8% 14.6% 22.3% More than 10-min. walk 11.9% 8.8% 5.9% 2.3% 7.3% Other 0.2% 0.2% 0.0% 0.2% 0.2% No designated area 3.5% 1.6% 0.7% 0.3% 1.5% Sanitation facility is safe during night (% HHs) 65.9% 68.2% 82.1% 82.8% 74.7% Sanitation facility is shared with other households (% HHs) 26.5% 31.6% 29.1% 24.5% 27.9% Satisfaction with Sanitation Facility (% HHs): Very satisfied 17.0% 16.6% 20.5% 22.8% 19.2% Somewhat satisfied 35.6% 40.3% 47.8% 43.6% 41.8% Less than satisfied 20.5% 16.4% 12.9% 18.6% 17.1% Completely dissatisfied 26.9% 26.7% 18.9% 15.1% 21.9% TABLE 19: IMPROVEMENT OF SANITATION FACILITIES Income Quartile 1st 2nd 3rd 4th Total Principal Reason for Building or Improving Toilet (% HHs): No reason given 0.0% 1.9% 0.5% 0.0% 0.6% Convenience or location 24.7% 24.2% 14.5% 5.6% 17.1% More healthy for the family 46.8% 42.9% 32.6% 44.0% 41.5% Easier to keep clean 14.6% 11.2% 23.2% 32.1% 20.3% Privacy, dignity 4.1% 5.4% 7.0% 2.7% 4.8% Safety, security 4.2% 6.4% 9.4% 2.1% 5.5% Comfort 4.0% 3.7% 10.2% 12.0% 7.5% Other 1.7% 4.5% 2.7% 1.5% 2.6% (Continued ) www.wsp.org 35 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 19: (Continued) Income Quartile 1st 2nd 3rd 4th Total Probability of Future Toilet Installation (% HHs): High 7.9% 10.8% 9.1% 3.8% 8.3% Medium 19.2% 29.4% 26.4% 47.7% 28.8% Low 55.0% 51.3% 50.5% 29.9% 48.5% None 17.8% 8.6% 14.0% 18.6% 14.5% Principal Constraint for Installing Toilet (% HHs): No constraints 0.0% 0.5% 0.3% 0.0% 0.2% High cost 72.7% 70.8% 65.5% 40.4% 64.9% No one to build it 5.7% 14.2% 4.4% 5.4% 8.0% Materials not available 11.0% 5.7% 9.1% 8.8% 8.6% Water table, soil conditions 2.7% 0.1% 0.7% 2.6% 1.5% Savings, credit issues 3.8% 2.3% 0.5% 13.7% 4.5% Tenancy issues 1.6% 4.7% 10.9% 4.7% 4.8% Limited space 0.3% 0.5% 1.4% 18.3% 3.9% Other 2.3% 1.3% 7.3% 6.2% 3.6% mentioned by the households for building a private sanitation facility was the high cost involved (64.9%). Other constraints were the unavailability of materials (8.6%) and the unavailability of labor force (8%). The reasons expressed by the poorest households focused more heavily on cost considerations (72.7%); the rich- est households also focused on lack of savings and/or credit (13.7%). The most common practice for disposal Table 20 reports some final characteristics of household sanitary condition. In of child feces among the poorest 24.6% of the households, hardly any flies were observed near the sanitation facil- households was to throw the feces in ity, in 26.3% of them few flies were found, and in 23.8% of the households flies the bushes or in the ground. were always present and in abundance. Also, in 74.6% of the households there were no feces visible inside or around the household. The most common practice for disposal of child feces among the poorest households was to throw the feces in the bushes or in the ground; 23.8% of them threw the feces in the garbage and 19.8% disposed the feces in the toilet or latrine. Among the richest households, the most common practice was to dispose child feces in the garbage (65.5%); 26.4% of the households used the toilet or latrine for disposal, and 8.1% threw the feces directly to the ground or into a hole. Findings of the direct observation by the interviewers of the household cleanness are reported in Table 21. More than 63% of the households were considered to be clean; however, in 17.6% of the households food was found to be uncovered and in 47.2% of the observations garbage was observed in the kitchen or inside the house. 36 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 20: OTHER CHARACTERISTICS OF HOUSEHOLD SANITARY CONDITION Income Quartile 1st 2nd 3rd 4th Total Flies Near Sanitation Facility (% HHs): Always and many 30.2% 24.5% 21.6% 18.8% 23.8% Always and some 15.9% 15.4% 14.9% 16.1% 15.6% Sometimes and many 14.6% 10.2% 6.0% 8.1% 9.8% Sometimes and few 24.9% 27.0% 26.8% 26.7% 26.3% Rarely, hardly any 14.5% 22.9% 30.6% 30.3% 24.6% Visible Feces In/Around HH (% HHs): None 63.5% 72.0% 79.8% 83.2% 74.6% 1­5 feces 17.3% 17.7% 16.3% 10.8% 15.5% More than five feces 19.2% 10.4% 3.9% 6.0% 9.9% Disposal of Child Feces (% HHs): Bushes, ground 33.6% 24.8% 14.7% 6.8% 20.0% Pit, hole in the ground 11.4% 5.8% 4.2% 8.1% 7.4% Open sewer, drain 5.0% 5.7% 4.5% 0.9% 4.0% Toilet, latrine 19.8% 17.3% 21.7% 26.4% 21.3% Garbage 23.8% 40.1% 52.5% 65.5% 45.4% River 10.4% 5.7% 3.4% 2.5% 5.5% Basin, sink 9.0% 3.7% 4.4% 1.8% 4.7% Other 6.9% 11.9% 8.4% 3.1% 7.5% TABLE 21: HOUSEHOLD CLEANNESS Income Quartile 1st 2nd 3rd 4th Total HH is clean (% HHs) 48.5% 59.5% 66.3% 80.1% 63.4% HH has uncovered food (% HHs) 29.5% 19.9% 13.5% 7.2% 17.6% HH has garbage in kitchen or house (% HHs) 61.3% 47.9% 47.0% 32.4% 47.2% 4.4 Handwashing Behavior The handwashing project seeks to achieve health and welfare impacts by promot- ing handwashing with soap; therefore measuring handwashing behavior at critical junctures is crucial. The survey includes several modules aiming to measure hand- washing behavior. The questions include self-reported handwashing behavior with soap at critical moments, observations of handwashing station(s) stocked with soap and water (as well as its location), observations of mother's hands, and structured observations of handwashing behavior. Less than half of the caregivers reported handwashing with soap at times of fecal contact. . . . Regarding The interviewers asked caregivers to mention under what circumstances they used food handling, 68.3% of caregivers soap to wash their hands in the last 24 hours. Table 22A summarizes the answers associated having washed hands with disaggregated by critical juncture. Almost all caregivers (99.6%) confirmed having soap with cooking or preparing food washed their hands with soap at least once since yesterday, but handwashing with and 34.1% with feeding their children. www.wsp.org 37 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 22A: SELF-REPORTED HANDWASHING BEHAVIOR WITH SOAP BY INCOME QUARTILE (PREVIOUS 24 HOURS) Income Quartile 1st 2nd 3rd 4th Total Washed hands with soap at least once in previous 24 hr (% caregivers) 99.3% 99.6% 99.8% 99.6% 99.6% Washed Hands with Soap At Least Once in Previous 24 Hours During the Following Events (% Caregivers): Using the toilet (% caregivers) 33.8% 43.9% 49.7% 56.5% 46.0% Cleaning children's bottoms (% caregivers) 35.6% 40.6% 41.0% 49.7% 41.7% Cooking or preparing food (% caregivers) 68.2% 66.2% 67.0% 71.7% 68.3% Feeding children (% caregivers) 25.3% 35.4% 35.3% 40.7% 34.1% TABLE 22B: SELF-REPORTED HANDWASHING BEHAVIOR WITH SOAP BY GEOGRAPHIC AREA (PREVIOUS 24 HOURS) Geographic Area Coast Jungle Mountain Total Washed hands with soap at least once in previous 24 hours (% caregivers) 99.3% 99.6% 99.8% 99.6% Washed Hands with Soap At Least Once in Previous 24 Hours During the Following Events (% Caregivers): Using the toilet (% caregivers) 50.1% 36.2% 45.6% 46.0% Cleaning children's bottoms (% caregivers) 49.3% 20.1% 41.6% 41.7% Cooking or preparing food (% caregivers) 60.5% 59.7% 73.8% 68.3% Feeding children (% caregivers) 32.1% 27.5% 36.4% 34.1% soap at critical moments was much lower. Less than half of disaggregate these findings by income and geographical the caregivers reported handwashing with soap at times of area. The number of households with an observed hand- fecal contact (46% of caregivers associated handwashing with washing station with soap and water was much higher among use of toilet and 41.7% with cleaning children's bottoms). the wealthiest households (72.5%) than among the poorest Regarding food handling, 68.3% of caregivers associated having washed hands with soap with cooking or preparing food and 34.1% with feeding their children. Handwashing with soap increased with income at every juncture. Table 22B shows the same figures disaggregated by geographical area. Caregivers living in the jungle had the lowest rates of hand- washing with soap for all critical junctures. For instance, only 20.1% of caregivers in the jungle associated washing hands with soap with cleaning children's bottoms while that figure was 49.3% on the coast. Similarly, 36.2% of caregivers living in the jungle associated handwashing with soap with toilet use, compared to 50.1% on the coast. Despite the fact that practically all caregivers reported to wash hands with soap at least once since the previous day, only 64.4% of households had an observed handwashing Handwashing station stocked with water and soap station with both soap and water. Table 23A and Table 23B 38 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 23A: OBSERVATION OF HANDWASHING STATION WITH SOAP AND WATER BY INCOME QUARTILE Income Quartile 1st 2nd 3rd 4th Total Observed HW station with soap and water (% HHs) 55.4% 62.5% 67.3% 72.5% 64.4% Location of HW Station (% HHs): Inside toilet or kitchen facility 27.2% 33.7% 34.1% 37.6% 33.1% In yard, within three feet of toilet or kitchen facility 15.6% 16.8% 23.0% 20.6% 19.0% In yard, 3­10 feet from toilet or kitchen facility 20.5% 16.8% 21.7% 25.0% 21.0% In yard, more than 10 feet from toilet or kitchen facility 21.4% 21.9% 19.8% 13.2% 19.1% TABLE 23B: OBSERVATION OF HANDWASHING STATION WITH SOAP AND WATER BY GEOGRAPHIC AREA Geographic Area Coast Jungle Mountain Total Observed HW station with soap and water (% HHs) 66.5% 72.3% 62.0% 64.4% Location of HW Station (% HHs): Inside toilet or kitchen facility 50.0% 44.0% 22.7% 33.1% In yard, within three feet of toilet or kitchen facility 17.8% 18.3% 19.7% 19.0% In yard, 3­10 feet from toilet or kitchen facility 14.3% 20.0% 24.6% 21.0% In yard, more than 10 feet from toilet or kitchen facility 4.1% 29.3% 25.0% 19.1% (55.4%). The percentage was also higher in the jungle (72.3%), than on the coast Despite the fact that practically all (66.5%) or in the mountains (62%). The observed handwashing station was caregivers reported to wash hands with soap at least once since the previous located in the yard in almost 60% of the households, and inside the toilet or day, only 64.4% of households had an kitchen facility in 33% of households. The higher the income, the closer the observed handwashing station with handwashing station was to the toilet or kitchen facility. Thus, in 27.2% of the both soap and water. . . . In 16.5% of poorest households the handwashing station was inside the kitchen or toilet facil- the households no cleansing agent of ity compared to 37.6% in those households with the highest income. On the any type (no soap, mud or ash) was contrary, 21.4% of the poorest households had the handwashing station in the observed. yard more than 10 feet from either the kitchen or the toilet facility, while the percentage is only 13.2% in the richest households. Table 23B also shows that households living in the mountains had not only the lowest percentages of hand- washing stations with soap and water overall, but the location of the handwash- ing station also tended to be further from the kitchen or toilet facility. For instance, only 22.7% of the households in the mountains had the handwashing station inside the kitchen or toilet facility, compared to much higher percentages of households along the coast and in the jungle (50% and 44% respectively). If a different handwashing station was used to wash hands when preparing food or feeding a child than the one used after going to the toilet, both handwashing stations were observed and information regarding their characteristics was collected for all stations used. Thus, Table 24A summarizes characteristics of the handwashing station used after going to the toilet. There were two types of handwashing devices most commonly used, a basin or bucket (49.2%) and a tap or faucet (48.1%). In 86.8% of households, water was observed at the www.wsp.org 39 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 24A: OBSERVATION OF HANDWASHING STATION USED AFTER GOING TO TOILET Income Quartile 1st 2nd 3rd 4th Total Water is available at HW station (% HHs) 80.1% 84.6% 88.9% 92.7% 86.8% Location of HW Station (% HHs): Inside toilet facility 4.1% 13.9% 16.5% 23.8% 14.5% Inside cooking place 16.7% 13.7% 8.7% 4.7% 11.0% In yard, less than three feet away from toilet 18.0% 16.0% 24.3% 23.7% 20.5% Between 3 and 10 feet away from toilet 11.9% 12.5% 17.2% 22.5% 16.0% More than 10 feet away from toilet 33.2% 32.6% 26.1% 21.1% 28.2% No specific place 16.1% 11.3% 7.1% 4.4% 9.8% HW Device, Toilet (% HHs): Tap, faucet 43.9% 46.4% 45.2% 56.1% 48.1% Basin, bucket 52.7% 51.3% 53.2% 40.3% 49.2% Other 3.4% 2.3% 1.6% 3.5% 2.7% Soaps Available at HW Station (% HHs): Multipurpose bar soap 16.0% 15.0% 17.7% 11.2% 14.9% Beauty, toilet bar soap 17.6% 24.2% 32.0% 46.8% 30.6% Powder soap, detergent 41.4% 46.0% 44.2% 39.3% 42.7% No soap observed 34.7% 25.4% 21.7% 19.7% 25.1% Ash and Mud at HW Station (% HHs): Ash 1.3% 1.0% 0.2% 0.6% 0.8% Mud 27.9% 20.4% 15.2% 8.5% 17.7% Ash and Mud 6.6% 6.9% 2.9% 1.4% 4.4% No ash nor mud observed 64.2% 71.7% 81.7% 89.5% 77.2% No cleansing agents at HW station (no soap, nor ash, nor mud observed) (% HHs) 20.0% 15.8% 14.6% 15.9% 16.5% handwashing station; in 74.9%, there was at least one type located inside the kitchen or cooking area, 19.5% in an area of soap available. The most frequently observed types of located between three and 10 feet away from the kitchen, soaps were powder soap or detergent (42.7%), beauty or 16.9% in a yard less than three feet away from the kitchen, toilet bar soap (30.6%) and multipurpose bar soap (14.9%). and 9.4% in a place located more than 10 feet away from the Regarding the use of mud or ash, both ash and mud were kitchen. The observations of these facilities reveal that the found at the HW station in 4.4% of the households, and most common device was a container from which water was mud alone was found in 17.7% of the households. Finally, poured (62.7%) and a tap or faucet (36.1%). In 82.6% of in 16.5% of the households no cleansing agent of any type the households, water was observed at the handwashing sta- (no soap, mud, or ash) was observed. tion. Regarding the availability of soap, in 86.9% of the cases soap was observed; in those households in which soap was Table 24B presents the analysis of the same variables for available, powder or laundry soap and detergent were the those 44.2% households that used a different handwashing most observed type of soap (65.9%), followed by beauty or station to wash hands when preparing food or feeding a child toilet soap (13.2%) and multipurpose soap (7.8%). In than the one used to wash hands after going to the toilet (the 10.8% of the households mud was observed and in 5.3% of reported results correspond only to those handwashing sta- the households both ash and mud was found. Finally, in tions that are different than those reported in Table 24A). 16.5% of the households no cleansing agent of any type (no Results show that 45.3% of the handwashing stations were soap, mud, or ash) was observed. 40 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 24B: OBSERVATION OF HANDWASHING STATION USED WHEN PREPARING FOOD OR FEEDING A CHILD Income Quartile 1st 2nd 3rd 4th Total Water is available at HW station (% HHs) 83.5% 77.0% 80.3% 88.9% 82.6% Location of HW Device (% HHs): Inside toilet facility 2.0% 3.5% 0.3% 3.1% 2.2% Inside cooking area 48.7% 46.3% 44.2% 42.0% 45.3% In yard, less than three feet away from kitchen 13.3% 16.2% 21.6% 16.4% 16.9% Between 3 and 10 feet away from kitchen 21.5% 14.7% 22.2% 19.5% 19.5% More than 10 feet away from kitchen 8.3% 13.6% 5.8% 10.1% 9.4% No specific place 6.3% 5.8% 5.9% 9.0% 6.8% Type of HW Station (% HHs): Tap, faucet 15.2% 28.1% 41.5% 57.0% 36.1% Container from which water is poured 82.6% 71.0% 57.8% 41.8% 62.7% Other 2.2% 0.9% 0.8% 1.2% 1.2% Soaps Available at HW Station (% HHs): Multipurpose bar soap 11.9% 8.7% 7.4% 3.6% 7.8% Beauty, toilet soap 6.3% 6.4% 4.1% 33.8% 13.2% Powder or laundry soap, detergent 51.8% 63.2% 82.1% 66.1% 65.9% No soap observed 39.7% 27.4% 14.2% 10.8% 22.6% Ash and Mud at HW Station (% HHs): Ash 2.2% 0.5% 1.5% 0.0% 1.0% Mud 20.0% 14.2% 7.7% 2.8% 10.8% Ash and mud 6.9% 11.0% 3.7% 0.4% 5.3% No ash nor mud observed 70.9% 74.3% 87.1% 96.8% 82.8% No cleansing agents at HW station (no soap, nor ash, nor mud observed) (% HHs) 30.9% 16.9% 11.4% 9.5% 16.9% Tables 25A and 25B summarize the observations of mother hands. On average, On average, in 67.9% of the cases, in 67.9% of the cases, caregiver's palms appeared to be clean. This figure was caregiver's palms appeared to be clean. . . . When looking at the figures lower for the households with the lowest income (61.2%) and considerably by geographic location, the findings higher for those with the highest income (82.6%). Similarly, high-income house- show that those households living in holds appeared to have cleaner fingernails and finger pads (72.3% and 81.7% the jungle had cleaner hands in general respectively) than the poorest ones (44.6% and 61.7%). When looking at the (cleaner palms, fingernails, and finger figures by geographic location, the findings show that those households living in pads) than those living on the coast or the jungle had cleaner hands in general (cleaner palms, fingernails, and finger in the mountains. pads) than those living on the coast or in the mountains. The figures are consis- tent with those in Table 23B, which show households living in the jungle had the highest percentage of handwashing stations stocked with soap and water. Findings of structured observations of handwashing behavior are summarized in Annex 2. www.wsp.org 41 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 25A: OBSERVATIONS OF CAREGIVERS HANDS BY INCOME Income Quartile 1st 2nd 3rd 4th Total Caregiver's Fingernails Appear to Have . . . (% Caregivers): Visible dirt 27.3% 22.8% 30.0% 10.7% 22.7% Unclean appearance 28.1% 29.6% 19.9% 17.0% 23.7% Clean appearance 44.6% 47.6% 50.1% 72.3% 53.7% Caregiver's Palms Appear to Have . . . (% Caregivers): Visible dirt 18.1% 16.7% 19.2% 6.9% 15.2% Unclean appearance 20.7% 22.1% 14.1% 10.4% 16.8% Clean appearance 61.2% 61.1% 66.7% 82.6% 67.9% Caregiver's Finger Pads Appear to Have . . . (% Caregivers): Visible dirt 17.2% 17.7% 23.3% 6.7% 16.2% Unclean appearance 21.2% 22.3% 13.6% 11.6% 17.2% Clean appearance 61.7% 60.1% 63.1% 81.7% 66.7% TABLE 25B: OBSERVATIONS OF CAREGIVERS HANDS BY GEOGRAPHIC AREA Geographical Area Coast Jungle Mountain Total Caregiver's Fingernails Appear to Have . . . (% Caregivers): Visible dirt 24.2% 10.5% 24.1% 22.7% Unclean appearance 24.8% 26.1% 22.6% 23.7% Clean appearance 50.9% 63.3% 53.3% 53.7% Caregiver's Palms Appear to Have . . . (% Caregivers): Visible dirt 11.9% 6.1% 18.6% 15.2% Unclean appearance 16.6% 19.2% 16.5% 16.8% Clean appearance 71.5% 74.7% 64.9% 67.9% Caregiver's Finger Pads Appear to Have . . . (% Caregivers): Visible dirt 11.3% 6.2% 20.5% 16.2% Unclean appearance 18.3% 19.6% 16.1% 17.2% Clean appearance 70.4% 74.2% 63.4% 66.7% 4.5 Mass-Media Consumption On average, 16.9% of caregivers of A large part of this project's success depends on whether households are respon- children under two years old recalled a sive to the media environment, and whether they have any access to it. These handwashing campaign. findings are summarized in Table 26A. On average, 16.9% of caregivers of children under two years old recalled a handwashing campaign. Of those who recalled a handwashing campaign, 42.9% remembered a campaign message to, "Wash hands with water and soap," 44% recalled the slogan, "Washing hands prevents diarrhea," 39.5% remembered being told, "[You] must wash hands before eating or cooking," and 33.6% declared remembering a campaign whose theme was, "[You] must wash hands after using toilet." A higher percentage of households with soap and water at the handwashing station recalled the "Wash hands with water and soap" campaign than those without. However, this result does not pre- sent evidence for a causal relationship. The means of transmission that had the 42 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 26A: MASS-MEDIA CONSUMPTION BY OBSERVED HANDWASHING STATION WITH SOAP AND WATER Soap and Water at Handwashing Station Yes No Total Caregiver recalls any handwashing campaign (% caregivers) 16.9% 16.8% 16.9% Campaign Theme (% Caregivers): Wash hands with water and soap 47.5% 34.4% 42.9% Washing hands prevents diarrhea 45.4% 41.4% 44.0% Must wash hands before eating, cooking 36.4% 45.2% 39.5% Must wash hands after using toilet 31.0% 38.5% 33.6% Other 2.8% 5.9% 3.9% Means of Campaign Transmission (% Caregivers): School, teacher 10.5% 6.7% 9.1% Market 0.5% 0.1% 0.3% Radio 12.0% 16.3% 13.5% TV 1.1% 1.8% 1.4% Community organization 7.7% 15.0% 10.3% Health center, health agent 65.4% 65.7% 65.5% Other 13.7% 6.9% 11.3% Media Channel (% Caregivers): None 7.0% 10.1% 8.1% Radio 67.3% 68.3% 67.6% TV 57.6% 40.7% 51.6% Newspapers 8.4% 5.5% 7.4% Public address speakers 0.5% 0.6% 0.5% Other 0.4% 0.9% 0.6% largest reach were health centers and health agents (65.5%), followed by radio (13.5%), community organizations (10.3%), and schools (9.1%). As expected, the types of media consumed more frequently were radio (67.6%) and TV (51.6%). Table 26B disaggregates the same variables by geographical area. A higher percent- A higher percentage of caregivers living age of caregivers living in the mountains (20.5%) and in the jungle (16.9%) re- in the mountains (20.5%) and in the jungle (16.9%) recalled hearing, seeing, called hearing, seeing, or receiving handwashing campaigns than those living on or receiving handwashing campaigns the coast (9.8%). When analyzing the most common means of communication than those living on the coast (9.8%). we observe that caregivers living on the coast were more familiar with TV, while those living in the jungle and in the mountains relied more on radio communica- tion. Finally, 16% of caregivers living in the jungle were not familiar with any kind of media. As previously mentioned, no causal relationships can be inferred from these cross tabulations. Still, in order to search for any relevant correlation, it is interesting to compare the handwashing habits of caregivers who recalled any handwashing cam- paign to the habits of those who did not recall any campaign. Table 26C presents www.wsp.org 43 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 26B: MASS-MEDIA CONSUMPTION BY GEOGRAPHIC AREA Geographic Area Coast Jungle Mountain Total Caregiver recalls any handwashing campaign (% caregivers) 9.8% 16.9% 20.5% 16.9% Campaign Theme: Wash hands with water and soap 20.9% 48.5% 47.4% 42.9% Washing hands prevents diarrhea 52.1% 29.0% 44.2% 44.0% Must wash hands before eating, cooking 21.6% 29.8% 45.3% 39.5% Must wash hands after using toilet 19.2% 15.7% 39.8% 33.6% Other 5.3% 3.9% 3.5% 3.9% Means of Campaign Transmission: School, teacher 11.9% 4.0% 9.2% 9.1% Market 0.0% 0.0% 0.5% 0.3% Radio 3.1% 8.6% 16.8% 13.5% TV 0.0% 4.2% 1.3% 1.4% Community organization 24.5% 7.1% 7.3% 10.3% Health center, health agent 52.7% 81.3% 66.4% 65.5% Other 13.9% 0.2% 12.2% 11.3% Media Known: None 7.7% 15.8% 6.9% 8.1% Radio 53.0% 65.2% 75.6% 67.6% TV 72.7% 35.9% 43.6% 51.6% Newspapers 9.0% 2.3% 7.5% 7.4% Public address speakers 0.0% 2.5% 0.4% 0.5% Other 0.5% 1.5% 0.5% 0.6% TABLE 26C: SELF-REPORTED HANDWASHING BEHAVIOR BY RECALL OF HANDWASHING CAMPAIGN Recall of Any Handwashing Campaign Yes No Total Washed hands with soap at least once in previous 24 hrs (% caregivers) 99.3% 99.6% 99.8% Washed Hands with Soap At Least Once in Previous 24 Hours During the Following Events (% Caregivers): Bathing a child 30.4% 21.9% 23.3% Bathing oneself 23.2% 22.0% 22.2% Using toilet 49.5% 45.7% 46.0% Cleaning baby bottom 37.2% 43.0% 41.7% Cleaning latrine 1.4% 0.9% 1.0% Cleaning toilet 4.5% 5.3% 5.2% Returning home 10.1% 12.4% 12.0% Preparing food, cooking 77.4% 67.1% 68.3% Feeding children 39.7% 33.4% 34.4% (Continued ) 44 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 26C: (Continued) Recall of Any Handwashing Campaign Yes No Total Washing child's hands 8.4% 9.9% 9.7% Cleaning dishes 36.5% 33.6% 34.1% Doing laundry 38.1% 44.3% 43.3% Because they look dirty 10.4% 6.2% 6.9% the results. As almost all caregivers reported washing their hands with soap since yesterday, the results did not vary by income group. But when asked about every specific situation in which they washed their hands, a higher proportion of caregivers who recalled any handwashing campaign had washed their hands more frequently. However, this higher figure could be due to the fact that those households who recalled handwashing campaigns were aware of the social de- sirability of washing hands at particular critical times. 4.6 Family-School Relationship In this subsection we present information about the family-school relationship, since schools could be sources of sanitary-related diseases and of sanitary-related information and education. Table 27 shows that a very high proportion of caregivers participated in school activities (91.2%) and that 29.4% of them recalled some health and hygiene- related campaigns. A higher number of caregivers coming from households with a handwashing station with soap and water recalled a campaign promoted at the school. The most frequent campaign topics were personal hygiene (30.5%), oral hygiene (27.1%) and handwashing (20.4%). Also, a high percentage of caregivers (70.2%) admitted having contributed with the campaign by donating their time (32.7%), products (21.9%), or money (13.1%). When looking at the disaggregation by geographical area, we see that caregivers coming from the jungle or the mountains of Peru tended to collaborate more with the school in order to promote a better personal hygiene, not only by participating more in school activities but also by contributing more in school- organized health campaigns. Furthermore, caregivers had directly contributed to hygiene in the school environment as almost 75% of them sent soap to the school. Results are reported on Table 28A and 28B. As expected, this figure was slightly higher for those caregivers com- ing from households with soap and water at their School pupils in Lambayeque use handwashing dispenser distributed handwashing station. by the project www.wsp.org 45 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 27A: FAMILY-SCHOOL RELATIONSHIP BY ACCESS TO HANDWASHING STATION WITH SOAP AND WATER Handwashing Station with Soap and Water Yes No Total Caregiver participates in school activities (% caregivers) 92.1% 89.6% 91.2% Caregivers recalls any campaign on health and hygiene promoted by school (% caregivers) 30.8% 27.0% 29.4% Campaign Theme: Tuberculosis 3.4% 2.1% 3.0% Oral hygiene 30.5% 20.0% 27.1% Personal hygiene 23.9% 44.3% 30.5% Nutrition 7.1% 2.7% 5.7% Handwashing 20.5% 20.2% 20.4% Other 33.1% 23.9% 30.1% Ways of Contributing to the Campaign: Money 10.8% 18.0% 13.1% Products 23.2% 19.1% 21.9% Dissemination and calling people 2.4% 3.3% 2.7% Own time 29.7% 38.9% 32.7% Other 6.0% 0.5% 4.2% Did not contribute 32.1% 25.0% 29.8% TABLE 27B: FAMILY-SCHOOL RELATIONSHIP BY GEOGRAPHIC AREA Geographic Area Coast Jungle Mountain Total Caregiver participates in school activities (% caregivers) 84.2% 89.9% 95.5% 91.2% Caregiver recalls campaign on health and hygiene promoted by school (% caregivers) 29.9% 27.7% 29.5% 29.4% Campaign Theme: Tuberculosis 1.8% 0.0% 4.2% 3.0% Oral hygiene 22.2% 27.7% 29.9% 27.1% Personal hygiene 24.4% 38.4% 32.7% 30.5% Nutrition 7.4% 6.4% 4.6% 5.7% Handwashing 8.4% 16.8% 28.1% 20.4% Other 44.1% 24.5% 22.8% 30.1% Ways of Contributing to the Campaign: Money 9.6% 19.6% 14.1% 13.1% Products 32.5% 17.6% 16.4% 21.9% Dissemination and calling people 6.4% 0.0% 1.0% 2.7% Own time 19.9% 30.8% 40.6% 32.7% Other 2.7% 0.0% 5.9% 4.2% Did not contribute 36.4% 33.9% 25.3% 29.8% 46 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 28A: SOAP CONTRIBUTION TO SCHOOLS BY OBSERVED HANDWASHING STATION WITH SOAP AND WATER Soap and Water at Handwashing Station Yes No Total Caregiver Sent Soap to School (% Caregivers): Never 28.7% 19.3% 22.6% Sometimes 54.6% 60.2% 58.2% Many times 16.8% 20.5% 19.2% Reason For Not Sending Soap: Forgot 6.0% 5.5% 5.8% No money 9.4% 38.0% 22.2% Not important 5.3% 3.9% 4.7% None of their business 5.8% 5.5% 5.7% It would get lost or stolen 7.4% 4.8% 6.3% Other 57.1% 41.3% 50.1% TABLE 28B: SOAP CONTRIBUTION TO SCHOOLS BY GEOGRAPHIC AREA Geographic Area Coast Jungle Mountain Total Caregiver Sent Soap to School (% Caregivers): Never 11.7% 36.2% 26.4% 22.6% Sometimes 60.2% 47.6% 59.1% 58.2% Many times 28.1% 16.3% 14.5% 19.2% Reason For Not Sending Soap: Forgot 6.1% 7.9% 5.2% 5.8% No money 38.3% 12.9% 20.4% 22.2% Not important 9.4% 5.9% 3.2% 4.7% None of his business 17.8% 2.9% 3.2% 5.7% It would get lost or stolen 10.5% 1.4% 6.4% 6.3% Other 13.8% 66.3% 55.3% 50.1% When asked why some of them did not contribute soap, young children's health, nutritional status, and development only a small fraction mentioned money constraints (16.4%). (Black et al. 2008; Engle et al. 2007; Grantham-McGregor A higher proportion of caregivers coming from households et al. 2007; Victora et al. 2008; Walker et al. 2007). More- with soap and water declared having forgotten, or said it over, some of these factors have been found to be significant was not important or none of their business (17.7% in predictors of child outcomes beyond variation due to socio- total). The fact that they did have soap and water at their economic and education variables. To enable us to more handwashing station indicates some concern about their carefully tease out the potential effects of the interventions child's sanitation and cleanness. These figures contradict on child health, growth, and development, we gathered in- that view. formation on feeding practices, caregiving behavior, and caregiver well-being. 4.7 Child Care Environment It is largely recognized that characteristics of the caregiver Table 29 summarizes breastfeeding habits within the and the quality of care a child receives have huge impacts on interviewed households. The average breastfeeding time was www.wsp.org 47 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 29: CHILD BREASTFEEDING (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Average months breastfeeding 13.1 10.7 11.9 12.2 11.9 Still breastfeeding (% children) 88.0% 83.4% 76.3% 78.6% 81.6% Colostrum given during first three days (% children) 94.3% 88.8% 92.9% 91.9% 92.0% Liquid given during first three days, other than colostrum or breast milk (% children) 17.2% 21.6% 30.8% 31.8% 25.3% Liquid Other Than Breast Given During First Three Days (% Children): Instant formula 50.0% 63.4% 71.7% 58.3% 62.0% Milk (other than breast milk) 7.0% 15.6% 6.2% 17.9% 12.0% Plain water 4.9% 4.4% 16.5% 8.6% 9.4% Sugar, glucose water 1.8% 2.2% 0.1% 0.1% 0.9% Gripe water 32.1% 18.9% 9.7% 8.4% 15.2% Tea, infusions 10.5% 2.1% 4.0% 2.2% 4.2% Other 8.9% 5.0% 7.8% 7.2% 7.2% The average breastfeeding time was 12 months, and 92% of children received colostrum16 during the first three days 12 months, and 92% of children after childbirth. Although it is recommended that mothers feed only with breast received colostrum during the first milk during the first six months of life, about one-quarter of mothers also fed three days after childbirth. their babies liquids other than colostrum or breast milk during the first three days of life. These other liquids were mainly infant formula (62%), gripe water (15.2%), and milk (12%). The survey also included a section on child diet. Specifically, caregivers of infants under the age of two were asked about liquids and food given to their children in the day previous to the interview. Results are reported in Table 30. Breast milk was given to the majority of the children (77%), followed by plain water (47.6%), and other type of milk (33.1%). With respect to food, 73.3% of the children re- ceived solid or semi-solid food three times, on average. When asked about dietary supplements, 22.3% of caregivers declared giving iron pills or syrup to her child and 22.9% affirmed having given vitamin A. Children's overall cleanness (hands, The survey examined the care situation of the children under the age of five by clothes, fingernails, face) increases including questions related to cleanness and clothing, and about the attention and with income. care given by their caregiver. Table 31 shows that on average, during the week previous to the interview, every child under the age of two had been left almost one time in the charge of another child. Richer households tended to leave their chil- dren more times alone at home. The interviewer also observed the overall cleanness of children during the interview. Three-quarters of the children under the age of five had a clean aspect, 37.7% of them exhibited dirty hands, 46.5% displayed dirty 16 Colostrum is produced prior to mature breast milk during pregnancy and through the first 3­6 days of life. It contains not only necessary nutrients, but also properties that help protect the baby from viral and bacterial infections. 48 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 30: INFANT/YOUNG CHILD FEEDING (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Liquids Given Yesterday (% Children): Breast milk 81.9% 78.1% 73.9% 74.1% 77.0% Plain water 39.4% 43.9% 48.6% 58.7% 47.6% Infant formula 5.8% 4.2% 6.9% 7.4% 6.0% Fortified child food 7.9% 8.2% 7.6% 7.5% 7.8% Homemade gruel 23.9% 16.6% 14.2% 34.7% 22.3% Other milks 21.8% 32.1% 42.8% 36.1% 33.1% Fruit juice 12.3% 17.7% 23.8% 16.9% 17.6% Caffeine beverages 15.0% 15.0% 12.7% 18.1% 15.2% Other 17.8% 19.2% 18.0% 30.6% 21.3% % of children that were given solid or semi-solid food yesterday 67.1% 72.2% 75.5% 78.8% 73.3% Average number of times food was given yesterday 3.0 3.0 2.9 2.9 3.0 Food Given Yesterday (% Children): Grain-based food 85.4% 85.6% 89.6% 81.7% 85.6% Vitamin A food 77.2% 77.1% 79.8% 72.8% 76.7% Roots, potatoes 92.6% 90.5% 92.0% 84.6% 89.8% Fruits, vegetables 78.3% 81.4% 82.8% 92.4% 83.9% Meat red, white 82.7% 87.7% 93.6% 90.1% 88.7% Beans, peas, lentils 65.4% 65.3% 60.3% 51.0% 60.3% Oil, fats, butter 75.9% 79.0% 84.5% 87.4% 81.9% % of children that ever received vitamin A 25.6% 23.2% 24.0% 18.7% 22.9% % of children that were given iron pills or syrup 21.1% 25.7% 24.7% 17.6% 22.3% % of children that feed themselves 48.7% 52.0% 55.5% 55.5% 53.1% TABLE 31: INFANT/YOUNG CHILD CARE SITUATION (CHILDREN <5)17 Income Quartile 1st 2nd 3rd 4th Total Child was left at the charge of another child during past week (number of times)* 0.8 0.9 0.9 0.5 0.8 Child was left alone during previous week (number of times)* 0.2 0.6 0.3 0.7 0.4 Child appeared clean with no offensive odor (% children) 60.3% 71.7% 79.7% 89.6% 74.5% Child has dirty hands (% children) 46.2% 39.1% 35.9% 27.7% 37.7% Child has dirty finger nails (% children) 58.8% 48.2% 46.2% 30.0% 46.5% Child has pot-belly (% children) 19.7% 17.2% 7.6% 6.3% 13.1% Child has dirty face (% children) 38.1% 32.0% 30.5% 17.3% 30.0% Child wears clothes (% children) 42.3% 35.1% 27.5% 19.0% 31.7% Child wears shoes or has shoes available (% children) 83.1% 83.4% 83.0% 86.0% 83.8% 17 Note: The first two questions in Table 31 correspond only to children under two years old. www.wsp.org 49 Findings from the Impact Evaluation Baseline Survey in Peru Findings fingernails and 30% had a dirty face. In regards to clothing, 31.7% of the children were seen wearing clothes (of which 99.2% had dirty clothes) and 83.8% of them were wearing shoes (or shoes were available). Children's overall cleanness (hands, clothes, fingernails, face) increased with income. On average, caregivers devoted more Interviewers were asked to observe interaction between the caregivers and their than five hours per day taking care of children during the interview, and results are reported on Table 32A. More than their children. 90% of the caregivers kept the child in sight during the interview: 77.4% talked to the child, 51.4% played or interacted in order to promote his/her development and learning, 64.9% smiled to or laughed with the child, and 5.2% of the caregiv- ers spanked the child during the interview. Caregivers coming from wealthier households interacted more with their children during the interview. On average, caregivers devoted more than five hours per day taking care of their children. The survey also included a section of caregiver behavior towards child discipline (only for caregivers of children under the age of two). Findings are summarized in Table 32B and indicate that 56.6% of the caregivers explained to their children the reason why some behavior was inappropriate, 20.4% of caregivers shook their child during the last month, 48.3% of them shouted or yelled at the child, 26.6% spanked or slapped the child, and 6.7% used an insulting name. Although over one-fourth of the households reported having spanked or slapped their under TABLE 32A: INFANT/YOUNG CHILD CARE SITUATION DURING INTERVIEW Income Quartile 1st 2nd 3rd 4th Total Caregiver keeps child in sight (% caregivers) 88.0% 86.7% 91.1% 95.4% 90.0% Caregiver talks to child (% caregivers) 71.4% 78.9% 79.1% 81.5% 77.4% Caregiver promotes child's development/learning (% caregivers) 44.4% 46.3% 57.6% 59.7% 51.4% Caregiver smiles/laughs to child (% caregivers) 60.1% 61.7% 69.2% 70.0% 64.9% Caregiver spanks the child (% caregivers) 2.9% 8.3% 5.9% 3.8% 5.2% Average daily caring time 5.52 5.75 5.52 6.04 5.70 TABLE 32B: DISCIPLINE MEASURES TOWARDS INFANT DURING PREVIOUS MONTH (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Caregiver took away or forbade something (% caregivers) 18.3% 21.3% 33.3% 32.6% 26.3% Caregiver explained why the behavior was wrong (% caregivers) 43.3% 52.8% 60.7% 69.7% 56.6% Caregiver shook the child (% caregivers) 12.5% 22.4% 24.5% 22.4% 20.4% Caregiver shouted or yelled at the child (% caregivers) 36.8% 41.6% 60.1% 54.9% 48.3% Caregiver spanked, slapped the child (% caregivers) 17.0% 25.4% 34.2% 30.0% 26.6% Caregiver that hit the child on the bottom or elsewhere (% caregivers) 1.1% 3.7% 3.5% 1.0% 2.3% Caregiver that used an insulting name (% caregivers) 1.9% 6.4% 8.7% 9.8% 6.7% Caregiver thinks that physical punishment is necessary (%caregivers) 8.4% 5.7% 9.8% 2.2% 6.5% 50 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings Surveyors collected observations on child hygiene, care, and cleanness two-year-old child during the previous month, only 6.5% of the households agreed that physical punishment was necessary in order to raise and educate a child. Furthermore, there were specific questions related to household support for learn- While the majority of children played ing and development. These include the availability to play with objects, and the with an adult (83.9%) or were taken on frequency with which adults engaged children in various activities demonstrated an outing outside the home (91.8%) in the past three days, only about to promote language and cognitive development. Table 33 shows that 62.5% of one-quarter of caregivers read books the children under the age of two played with household objects and 82.6% of or told stories to the child in the past them played with toys. Only 4.9% of the children attend a nursery or child cen- three days. ter; this may be due to the fact that many centers only served children three to five years of age. While the majority of children played with an adult (83.9%) or were taken on an outing outside the home (91.8%) in the past three days, only about one-quarter of caregivers read books or told stories to the child in the past three days. The results reported in Table 33 reinforce previous findings that showed more time and effort dedication by the caregivers coming from households with higher incomes. TABLE 33: INFANT/YOUNG CHILD LEARNING ENVIRONMENT (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Child plays with household objects (% children) 57.9% 59.7% 65.3% 67.3% 62.5% Child plays with toys (% children) 73.5% 84.2% 85.9% 87.1% 82.6% Child attended early education programs (% children) 1.6% 9.1% 4.6% 4.4% 4.9% Adult reads books with child (% adults) 19.3% 21.9% 26.8% 33.9% 25.4% Adult tells stories to child (% adults) 19.9% 22.1% 23.7% 23.7% 22.3% Adult take child outside home (% adults) 86.9% 88.6% 94.6% 97.3% 91.8% Adult plays with child (% adults) 78.3% 77.6% 88.0% 92.3% 83.9% www.wsp.org 51 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 34: MATERNAL DEPRESSION Restless Could Not Felt Fearful Sleep Felt Lonely Felt Sad Enjoyed Life Get Going Never or rarely (% caregivers) 18.7% 13.2% 16.6% 4.6% 18.4% 34.6% Little of the time or occasionally (% caregivers) 25.5% 29.4% 22.9% 24.7% 39.4% 24.9% Sometimes or about half the time (% caregivers) 39.5% 34.4% 28.7% 37.7% 27.6% 23.5% Most or all of the time (% caregivers) 16.3% 23.0% 31.8% 33.0% 14.5% 17.0% Finally, this survey also considered maternal depression, as it the child had yet achieved various milestones (i.e., sitting, is an important determinant of the child's health environ- walking, saying some words, etc.). We measured three do- ment. Results show that 13.2% of the mothers felt depressed mains: communication skills, including pre-verbal bab- most or all of the time during the last seven days and 24.8% bling, as well as producing and understanding language; declared feeling depressed sometimes or about half the time. gross motor skills, including control of certain postures or Table 34 presents the most common symptoms of depres- coordination of movements requiring large muscle systems; sion for those mothers who answered being depressed and personal-social skills or behaviors related to engaging "Sometimes or about half the time," or "Most or all the with others, as well as to becoming independent. Scores on time." More than 70% of these mothers felt sad sometimes these types of outcomes have been useful for discriminating or most of the time, 60.5% felt lonely, 55.8% declared feel- between groups of children with different environmental ing fearful, and 57.4% experienced restless sleep. (poverty, etc.) and biological (stunting, etc.) profiles. The questions administered to each child were selected to mea- 4.8 Child Development sure a range of behaviors representing lower- to higher-than The survey included a section related to child development, average development per age range (based on U.S. estimates in which caregivers were asked a number of questions about of age-related behaviors, as international standards are not the child's reaction to specific stimuli (i.e., response to available). With this information, we computed a "degree mother's voice, reaction to seeing self in a mirror) or whether of child development" index per skill with higher scores TABLE 35A: CHILD DEVELOPMENT Z-SCORES BY SANITARY CONDITIONS (CHILDREN <2) Improved Sanitation Improved Water Source Soap and Water at HW Station Yes No Yes No Yes No Average communication skills-for-age z-score 0.12 0.00 0.09 0.06 0.07 0.04 Average gross motor skills-for-age z-score 0.21 0.07 0.13 0.16 0.14 0.08 Average personal-social skills-for-age z-score 0.17 0.02 0.12 0.02 0.13 0.00 TABLE 35B: CHILD DEVELOPMENT Z-SCORES BY INCOME QUARTILE (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Average communication skills-for-age z-score 0.06 0.06 0.09 0.13 0.06 Average gross motor skills-for-age z-score 0.16 0.05 0.16 0.27 0.06 Average personal-social skills-for-age z-score 0.03 0.03 0.10 0.23 0.09 52 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 35C: CHILD DEVELOPMENT Z-SCORES BY GEOGRAPHIC AREA (CHILDREN <2) Geographic Area Coast Jungle Mountain Total Average communication skills-for-age z-score 0.19 0.03 0.01 0.06 Average gross motor skills-for-age z-score 0.05 0.17 0.05 0.06 Average personal-social skills-for-age z-score 0.06 0.02 0.12 0.09 Several child development measures were collected during the survey representing a higher level of development in that domain. Table 35 presents the z-scores18 for these variables disaggregated by sanitary conditions, income, and geographic area. We systematically observed for every type of skill a lower degree of development in those We systematically observed for children from households without improved sanitation, without an improved water every type of skill a lower degree source, and without a handwashing station stocked with soap and water. Although we of development in those children from households without improved cannot infer any causal relationship between the variables in this bivariate analysis, the sanitation, without an improved water figures show a correlation between the sanitary conditions and the degree of child's de- source, and without a handwashing velopment. Furthermore, all of the measures increased with the income level, since pre- station stocked with soap and vious tables have showed that richer households can afford to provide healthier water. . . . When disaggregating the nourishment for younger children and to spend more time stimulating their develop- data by geographic area we did not ment. When disaggregating the data by geographic area we did not find a clear-cut find a clear-cut pattern, since in each area there is a skill for which the pattern, since in each area there is a skill for which the children coming from that area children coming from that area are are better than others. better than others. Figure 4 shows the histograms for the three variables' z-scores. All of them had a mean value equal to 0. The median values for the communication skills-for-age z-score, the gross motor skills-for-age z-score, and the personal-social skills-for- age z-score were ­0.06, 0.09, and 0.19, respectively. 18 A z-score, or standard score, indicates how many standard deviations an observation is below or above the average (mean). As the mean is normalized to zero, any negative z-scores would be below the mean, and any positive z-scores would be above the mean. www.wsp.org 53 Findings from the Impact Evaluation Baseline Survey in Peru Findings FIGURE 4: HISTOGRAMS OF CHILD DEVELOPMENT MEASURES' Z-SCORES (CHILDREN <2) Communication Skills-For-Age Z-Score 10 8 6 Percent 4 2 0 ­4 ­2 0 2 Gross Motor Skills-For-Age Z-Score 10 8 6 Percent 4 2 0 ­3 ­2 ­1 0 1 2 Social-Personal Skills-For-Age Z-Score 10 8 6 Percent 4 2 0 ­6 ­4 ­2 0 2 54 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings 4.9 Diarrhea and Acute Lower Respiratory which made them prone to contract any diseases related to Infection Prevalence sanitary and hygiene deficiencies. Tables 36 through 39 dis- Previous sections have shown that many of the interviewed play the analysis results of health-related questions for the households lacked access to improved water, improved group of children under the age of five. Specifically, we con- sanitation, and handwashing stations with soap and water, centrate on two diseases: diarrhea and ALRI. TABLE 36A: DIARRHEA PREVALENCE BY SANITARY CONDITIONS (CHILDREN <5) Improved Improved Water Soap and Water Sanitation Source at HW Station Yes No Yes No Yes No Child had diarrhea symptoms in previous 48 hours (% children) 7.6% 12.0% 10.0% 10.2% 10.2% 9.7% Child had diarrhea symptoms in previous week (% children) 16.6% 19.9% 18.5% 18.1% 18.5% 18.4% Child had diarrhea symptoms in past 14 days (% children) 17.7% 21.9% 19.4% 21.7% 19.8% 20.4% TABLE 36B: DIARRHEA PREVALENCE BY GEOGRAPHIC AREA (CHILDREN <5) Geographic Area Coast Jungle Mountain Total Child had diarrhea symptoms in previous 48 hours (% children) 6.04% 13.35% 11.39% 10.01% Child had diarrhea symptoms in previous week (% children) 14.08% 22.63% 19.87% 18.45% Child had diarrhea symptoms in past 14 days (% children) 14.94% 24.31% 21.81% 20.04% TABLE 37: DIARRHEA TREATMENT BY INCOME QUARTILE (CHILDREN <5) Income Quartile 1st 2nd 3rd 4th Total Child had diarrhea symptoms in previous 48 hours (% children) 10.2% 8.5% 12.3% 9.2% 10.0% Child had diarrhea symptoms in previous week 17.6% 16.6% 23.7% 16.1% 18.5% (% children) Child had diarrhea symptoms in past 14 days (% children) 19.1% 17.9% 24.9% 18.5% 20.0% Caregiver did seek public care provider (% caregivers) 94.9% 96.2% 85.0% 95.4% 93.2% Caregiver did not pay for the intestinal treatment (% caregivers) 75.9% 69.1% 36.4% 74.3% 60.0% Caregiver Did Seek Medical Advice (% Caregivers): Did not seek 45.2% 43.5% 66.3% 65.6% 55.1% Day visit to doctor 54.1% 54.5% 30.3% 31.7% 42.7% Other 0.7% 2.1% 3.4% 2.8% 2.2% Type of Treatment Given: No treatment 46.4% 25.2% 25.9% 56.4% 37.7% Pill or Syrup 45.3% 65.9% 65.9% 36.8% 54.2% Traditional remedies 2.9% 4.8% 1.2% 4.2% 3.1% Oral rehydration solution 1.6% 1.4% 2.5% 0.0% 1.5% Homemade sugar/salt water 0.1% 1.9% 3.6% 0.0% 1.5% Other 3.9% 1.1% 0.8% 0.2% 1.6% www.wsp.org 55 Findings from the Impact Evaluation Baseline Survey in Peru Findings The variable for diarrhea prevalence The variable for diarrhea prevalence was constructed on the basis of several symp- was constructed on the basis of toms reported by a child's caregiver and not on caregiver's self-diagnosis. Specifi- several symptoms reported by a child's cally, a child was declared to have diarrhea when he presented the following caregiver and not on caregiver's self- symptoms: three or more loose or watery stools per day, or one or more stools diagnosis. with blood and/or mucus (Baqui et al. 1991). Findings reveal that 10% of the children under the age of five presented diarrhea symptoms in the previous 48 hours, 18.4% presented symptoms in the past seven days and 20.4% in the past 14 days. For all the three recall periods, the prevalence of diarrhea was noticeably higher in those households with unimproved sanitation. Findings reveal that 10% of the children under the age of five presented Diarrhea prevalence was not lower in households with access to a handwashing diarrhea symptoms in the previous 48 station with soap and water (and an improved water source, to a smaller degree), hours, 18.4% presented symptoms in compared to those that did not have access. When disaggregating diarrhea preva- the past seven days and 20.4% in the lence by geographical region, we find that the situation was significantly worse for past 14 days. Diarrhea prevalence was households living in the jungle, where 24% of the children presented diarrhea not lower in households with access to symptoms in the past 14 days. For households living in the mountains this figure a handwashing station with soap and water (and an improved water source, reduced to 22% and for those living along the coast it further decreased to 15%. to a smaller degree), compared to Finally, we observed no strong relationship between income level and those that did not have access. diarrhea prevalence. On average, 55.1% of caregivers with children presenting diarrhea symptoms in the previous 48 hours did not seek medical advice, while 42.7% went to visit the doctor. In almost every case, assistance was provided by a public agent (93.2%) and a high proportion of caregivers did not pay for the treatment (60%). In 37.7% of the cases no treatment was received. Pill or syrup was given as treatment in 54.2% and traditional remedies in 3.1% of the cases. To analyze presence of parasites, stool samples are collected 56 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 38A: ALRI PREVALENCE BY SANITARY CONDITIONS (CHILDREN <5) Improved Improved Water Soap and Water Sanitation Source at HW Station Yes No Yes No Yes No Child had ALRI symptoms in previous 48 hours (% children) 2.6% 5.8% 3.8% 5.7% 4.5% 4.1% Child had ALRI symptoms in previous week (% children) 3.5% 7.3% 5.0% 7.1% 5.6% 5.5% TABLE 38B: ALRI PREVALENCE BY GEOGRAPHIC AREA (CHILDREN <5) Geographic Area Coast Jungle Mountain Total Child had ALRI symptoms in previous 48 hours (% children) 1.7% 3.3% 5.8% 4.3% Child had ALRI symptoms in previous week (% children) 2.4% 4.4% 7.4% 5.6% TABLE 39: ALRI TREATMENT BY INCOME QUARTILE (CHILDREN <5) Income Quartile 1st 2nd 3rd 4th Total Child had ALRI symptoms in previous 48 hours (% children) 6.0% 3.5% 5.2% 2.2% 4.3% Child had ALRI symptoms in previous week (% children) 8.0% 4.8% 6.0% 2.9% 5.6% Caregiver did seek public care provider (% caregivers) 85.8% 81.6% 53.7% 90.3% 79.0% Caregiver did not pay for the treatment (% caregivers) 90.5% 63.6% 28.7% 63.3% 65.4% Caregiver Did Seek Medical Advice (% Caregivers): Did not seek 52.6% 62.9% 71.8% 46.5% 59.7% Day visit to doctor 46.9% 36.5% 23.0% 53.6% 38.4% Other 0.6% 0.7% 5.2% 0.0% 1.9% Type of Treatment Given: No treatment 28.9% 30.0% 40.5% 10.3% 30.4% Pill or Syrup 62.5% 65.0% 58.4% 63.6% 62.0% Injection 1.3% 5.3% 4.4% 0.8% 3.0% Traditional remedies 7.0% 0.7% 0.0% 24.4% 5.5% Other 2.1% 4.3% 0.2% 0.6% 1.8% In order to construct the ALRI variable, we followed the methodology provided by the World Health Organization clinical case definition (WHO 2005). Specifically, a The prevalence of ALRI was lower child was identified as having ALRI when he/she presented the following symptoms: than diarrhea in our sample. ALRI constant cough or difficulty breathing, and raised respiratory rate (>60 breaths per prevalence increases to 7.3% among minute in children younger than 60 days old, >50 breaths per minute for children children living in households with aged 60­364 days, >40 per minute for children aged one to five years). unimproved sanitation and to 7.1% in the households with unimproved water sources. ALRI prevalence was higher The prevalence of ALRI was lower than diarrhea in our sample: only 4.3% of for children living in the mountains of children had ALRI symptoms in the previous 48 hours and the seven-day preva- Peru, where the effect of altitude over lence is 5.6%. ALRI prevalence increased to 7.3% among children living in house- respiratory difficulties seemed to be holds with unimproved sanitation and to 7.1% in the households with unimproved driving the results. www.wsp.org 57 Findings from the Impact Evaluation Baseline Survey in Peru Findings water sources. As expected, ALRI prevalence was higher for children living in the mountains of Peru, where the effect of altitude over respiratory difficulties seemed to be driving the results. As with diarrhea, similar percentages of households pre- sented ALRI symptoms in the previous week, despite whether or not they had a handwashing station stocked with soap and water. Of those that presented the ALRI symptoms in the previous 48 hours, 59.7% caregivers did not seek medical advice and 38.4% of them only made a day visit to the doctor. Seventy-nine per- cent of consulted care providers were public agents. Again, a very high percentage of caregivers did not pay for the treatment (65.4%). In 30.4% of the cases, children presenting ALRI symptoms received no treatment. The most frequent treatment was pills or syrup (62%), followed by traditional remedies (5.5%) and injections (3%). 4.10 Anthropometric Measures and Anemia The survey included anthropometric measures of children under the age of two: arm and head circumference, weight, and length/height. This information is im- portant in order to assess the average growth and development of the children. To analyze these variables, z-scores were computed using WHO's estimations of pop- On average, arm circumference was ulation mean and standard deviation for each of the aforementioned variables found to be higher than the population (WHO 2006, 2007). The histograms of the z-scores for each anthropometric mean, as well as the body mass index measure are presented in Figure 5. and the weight for length/height. On the contrary, the average weight, On average, arm circumference was found to be higher than the population length/height, and head circumference were found to be lower than the mean, as well as the body mass index and the weight for length/height. On the population mean estimated by the contrary, the average weight, length/height, and head circumference were found WHO. to be lower than the population mean estimated by the WHO. Table 40 presents the average z-scores for the six anthropometric measures disag- Children from households without gregated by sanitary condition, income level, and geographical area. Children improved sanitation, improved water coming from households without improved sanitation, improved water source, source, or a handwashing station or a handwashing station stocked with soap and water, tended to have a lower stocked with soap and water, tended to have a lower average z-score for each average z-score for each anthropometric measure included in the analysis. These anthropometric measure included in results confirm those found in the Child Development subsection. Physical de- the analysis. Physical development was velopment was positively correlated with household sanitary condition, although positively correlated with household no causal relationship can be inferred from this bivariate analysis. Again, all six sanitary condition, although no causal measures increased with income level, which could be driven by the fact that relationship can be inferred from this wealthier caregivers can and actually do provide their children with better nour- bivariate analysis. ishment during the first years of their lives. With respect to the disaggregation by geographical area, all six measures indicate that children living in coastal areas were in a better situation than those living in the mountains and the jungle. However, this does not preclude the fact that according to three out of six mea- sures, all children, independently of the geographical area considered, were un- derperforming compared to the mean value. 58 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings FIGURE 5: HISTOGRAMS OF ANTHROPOMETRIC MEASURES' Z-SCORES (CHILDREN <2) Arm circumference-for-age z-score Weight-for-age z-score 10 10 8 8 6 6 Percent Percent 4 4 2 2 0 0 0 4 2 0 2 4 6 4 2 0 2 4 Length/height-for-age z-score BMI-for-age z-score 10 10 8 8 6 6 Percent Percent 4 4 2 2 0 0 0 5 0 5 5 0 5 Weight-for-length/height z-score Head circumference-for-age z-score 10 10 8 8 6 6 Percent Percent 4 4 2 2 0 0 5 0 5 5 0 5 www.wsp.org 59 Findings from the Impact Evaluation Baseline Survey in Peru Findings TABLE 40A: ANTHROPOMETRIC MEASURES' Z-SCORES BY SANITARY CONDITIONS (CHILDREN <2) Improved Water Soap and Water Improved Sanitation Source at HW Station Yes No Yes No Yes No Average arm circumference-for-age z-score 0.69 0.35 0.57 0.33 0.60 0.35 Average weight-for-age z-score 0.10 0.44 0.23 0.43 0.19 0.45 Average length/height-for-age z-score 1.00 1.24 1.11 1.19 1.02 1.32 Average BMI-for-age z-score 0.67 0.36 0.54 0.42 0.56 0.40 Average weight-for-length/height z-score 0.59 0.34 0.48 0.40 0.53 0.32 Average head circumference-for-age z-score 0.06 0.30 0.12 0.40 0.17 0.21 TABLE 40B: ANTHROPOMETRIC MEASURES' Z-SCORES BY INCOME QUARTILE (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Average arm circumference-for-age z-score 0.20 0.41 0.55 0.85 0.51 Average weight-for-age z-score 0.65 0.40 0.22 0.16 0.28 Average length/height-for-age z-score 1.44 1.24 0.95 0.88 1.13 Average BMI-for-age z-score 0.28 0.45 0.48 0.83 0.51 Average weight-for-length/height z-score 0.25 0.41 0.42 0.77 0.46 Average head circumference-for-age z-score 0.40 0.31 0.04 0.00 0.19 TABLE 40C: ANTHROPOMETRIC MEASURES' Z-SCORES BY GEOGRAPHIC AREA (CHILDREN <2) Geographic Area Coast Jungle Mountain Total Average arm circumference-for-age z-score 0.72 0.18 0.46 0.51 Average weight-for-age z-score 0.06 0.55 0.35 0.28 Average length/height-for-age z-score 1.01 1.27 1.16 1.13 Average BMI-for-age z-score 0.72 0.28 0.44 0.51 Average weight-for-length/height z-score 0.64 0.22 0.41 0.46 Average head circumference-for-age z-score 0.09 0.47 0.19 0.19 Figure 6 presents the average z-score corresponding to each body mass index-to age z-score, the evolution of the aver- variable disaggregated by age and sex. Since this survey is a ages of the rest of the variables decreased with age, indicat- cross section of households, we cannot observe the evolu- ing two possible explanations. The first is that the gap tion over time of the anthropometrics variables for the between the sample mean and the population mean widens children under the age of two. Nevertheless, we can ana- during child's growth, in which case this evidence could be lyze the average z-score for the different groups of children interpreted as a worsening of child's physical development. arranged according to their age (in months), which gives The second explanation that can be derived is that the stan- us an approximation of the anthropometric measures' evo- dard deviation of each variable could be decreasing with age, lution over early child development. A very striking result which makes the situation more severe if the first explanation is that, with the exception of the evolution of the average is correct. 60 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings FIGURE 6: ANTHROPOMETRIC MEASURES' Z-SCORES BY SEX AND MONTHS OF AGE (CHILDREN <2) Arm circumference-for-age z-score Weight-for-age z-score 1.0 0.5 Male Female Male Female 0 0.5 0.5 0 1.0 0.5 1.5 0 5 10 15 20 25 0 5 10 15 20 25 Length/height-for-age z-score BMI-for-age z-score 0.5 1.0 Male Female Male Female 0.8 0.1 0.6 1.5 0.4 2.0 0.2 2.5 0 0 5 10 15 20 25 0 5 10 15 20 25 Weight-for-length/height z-score Head circumference-for-age z-score 1.0 0.2 Male Female Male Female 0 0.5 0.2 0.4 0 0.6 0.5 0.8 0 5 10 15 20 25 0 5 10 15 20 25 www.wsp.org 61 Findings from the Impact Evaluation Baseline Survey in Peru Findings Almost three-quarters of the samples Hemoglobin concentrations were obtained from children under the age of two in order taken indicate the presence of anemia. to estimate the percentage suffering from anemia, and results are reported in Table 41. This proportion is lower for households For households living in the mountains, the results were adjusted to account for differ- with improved sanitation but higher ences in altitude, since hemoglobin concentrations increase as an adaptive response to for households with improved water source. A surprising result is that the the lower partial pressure of oxygen and reduced oxygen saturation of blood (Nestel percentage of individuals suffering from 2002). Almost three-quarters of the samples taken indicated the presence of anemia. anemia increases with income level. This proportion was lower for households with improved sanitation, but higher for households with improved water source. The proportion was also higher among chil- dren living in the mountains. An unexpected result is that the percentage of individuals suffering from anemia increased with income level. A partial plausible explanation, con- sistent with the results shown in Table 30 could be that, on average, children in poor households were more likely to receive iron supplements, which could be a consequence of government and/or NGO programs targeting low-income families. TABLE 41: ANEMIA PREVALENCE (Hb < 110 g/L) IN CHILDREN < 2 % of HHs By Income Quartile: 1st 73.80% 2nd 72.10% 3rd 74.90% 4th 78.40% By Geographic Area: Coast 74.50% Jungle 69.50% Mountain 75.90% Overall 74.80% Hemoglobin concentrations are measured to test for anemia 62 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings 4.11 Environmental Contamination and Parasitical Prevalence To examine the presence of parasites and bacteria, the survey also collected stool The survey also collected stool and and environmental contamination samples on a subsample of 160 households. environmental contamination samples Baseline data on the presence of bacteria and parasites in the household may on a subsample of 160 households to examine the presence of parasites and allow us in the future to better understand the mechanism by which our treat- bacteria. ment operates, whether it is through the mother or the child. Also, data related to bacteria and parasites presence in household objects and water serves as a control for factors not related to our treatment that could also affect the output variable that we are interested in. In particular, the focus is set in the presence of bacteria such as E. coli, and parasites such as Giardia, Ascaris, and Blastocystis. Some kinds of E. coli can cause diarrhea, while others cause urinary tract in- Some kinds of E. coli can cause fections, respiratory illness and pneumonia, and other diseases. Still, other diarrhea, while others cause urinary tract infections, respiratory illness and kinds of E. coli are used as markers for water contamination. Table 42 presents pneumonia, and other diseases. . . . Consistent with previous findings, TABLE 42A: MEAN ESCHERICHIA COLI CONCENTRATIONS BY SANITARY households with access to improved CONDITIONS sanitation presented lower counts of Improved Water Soap and Water at the bacteria in each of the four samples Improved Sanitation Source HW Station taken, but households with access Yes No Yes No Yes No to an improved water source showed Log10 E. coli, PN/100ml: higher levels of water contamination. Mother 0.72 1.06 0.91 0.76 0.80 1.01 Child 0.55 0.56 0.61 0.36 0.60 0.47 Object 0.44 0.48 0.55 0.09 0.52 0.34 Water 0.42 0.79 0.63 0.45 0.55 0.66 TABLE 42B: MEAN ESCHERICHIA COLI CONCENTRATIONS BY INCOME QUARTILE Income Quartile 1st 2nd 3rd 4th Total Log10 E. coli, MPN/100ml: Mother 1.21 0.65 0.90 0.72 0.88 Child 0.48 0.67 0.63 0.35 0.56 Object 0.94 0.50 0.17 0.16 0.46 Water 1.16 0.25 0.55 0.36 0.59 TABLE 42C: MEAN ESCHERICHIA COLI CONCENTRATIONS BY GEOGRAPHIC AREA Geographic Area Coast Jungle Mountain Total Log10 E. coli, MPN/100ml: Mother 0.90 0.76 0.82 0.88 Child 0.57 0.56 0.50 0.56 Object 0.43 1.39 0.30 0.46 Water 0.60 1.26 0.34 0.59 www.wsp.org 63 Findings from the Impact Evaluation Baseline Survey in Peru Findings the logarithm of E. coli counts disaggregated by sanita- coming from the poorest households. Finally, households tion condition, income level and geographic area. Con- living in coastal areas presented the highest E. coli counts sistent with previous findings, households with access to in the samples taken from the mother, while in the jungle improved sanitation presented lower counts of the bacte- the highest E. coli counts were found in samples taken ria in each of the four samples taken, but households from objects and water. with access to an improved water source showed higher levels of water contamination. Samples collected from The parasitical analysis focused on three types of parasites: caregivers' hands and drinking water coming from house- Giardia, a parasite that colonizes and reproduces in the small holds with a handwashing station stocked with soap and intestine, causing giardiasis; Ascaris, a genus of parasitic water had lower counts of the bacteria, but the counts worms, which provokes an infection called ascariasis; and coming from the child and objects seemed to be higher. Blastocystis, which can produce the disease blastocystsis, for When taking into account income levels, there was a de- which the most frequently described symptoms are abdomi- clining trend of E. coli counts with income, though the nal pain, constipation, and diarrhea. Table 43A, 43B, and counts were also low for the sample taken from the child 43C summarize the results for these three parasites. TABLE 43A: PARASITES PREVALENCE IN STOOL SAMPLES BY SANITARY CONDITIONS (CHILDREN <2) Improved Water Soap and Water Improved Sanitation Source at HW Station Yes No Yes No Yes No Any parasites detected in stool samples (% HHs) 6.7% 17.5% 8.4% 25.0% 2.7% 29.2% Giardia detected in stool samples (% HHs) 1.5% 12.1% 4.3% 14.6% 1.3% 16.3% Ascaris detected in stool samples (% HHs) 0.2% 0.1% 0.2% 0.0% 0.2% 0.0% Blastocystis detected in stool samples (% HHs) 5.3% 8.1% 4.3% 15.5% 1.5% 16.4% TABLE 43B: PARASITES PREVALENCE IN STOOL SAMPLES BY INCOME QUARTILE (CHILDREN <2) Income Quartile 1st 2nd 3rd 4th Total Any parasites detected in stool samples (% HHs) 22.9% 3.3% 16.2% 1.5% 11.7% Giardia detected in stool samples (% HHs) 9.7% 0.3% 12.6% 1.3% 6.4% Ascaris detected in stool samples (% HHs) 0.4% 0.2% 0.0% 0.0% 0.2% Blastocystis detected in stool samples (% HHs) 13.8% 3.1% 7.2% 0.2% 6.6% TABLE 43C: PARASITES PREVALENCE IN STOOL SAMPLES BY GEOGRAPHIC AREA (CHILDREN <2) Geographic Area Coast Jungle Mountain Total Any parasites detected in stool samples (% HHs) 9.2% 11.4% 21.5% 11.7% Giardia detected in stool samples (% HHs) 6.5% 4.4% 6.8% 6.4% Ascaris detected in stool samples (% HHs) 0.0% 1.5% 0.2% 0.2% Blastocystis detected in stool samples (% HHs) 2.7% 7.0% 21.0% 6.6% 64 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Findings Caregiver's hands are tested for presence of Escherichia coli Parasites were detected in 12% of the stool samples, and the most frequent para- Parasites were detected in 12% of sites found were Giardia and Blastocystis (Ascaris affects only a minor percentage the stool samples, and the most of households). Prevalence of parasites was lower among households with access frequent parasites found were Giardia and Blastocystis (Ascaris affects only to improved sanitation (7%) than those with unimproved sanitation (18%). Sim- a minor percentage of households). ilarly, parasitical prevalence was lower among households with access to improved The lowest prevalence of parasites water sources (8%) than those with unimproved water (25%). The lowest preva- was found among households with lence of parasites was found among households with a handwashing station a handwashing station stocked with stocked with soap and water (3%) and the highest in those without such (29%). soap and water (3%) and the highest in The poorest households had the highest prevalence of parasites, although there those without such (29%). was a high and unexpected parasite presence in households located in the 3rd quartile of the income distribution. However, the prevalence of the different kinds of parasites was not homogeneous across income levels (poorest households display higher presence of Ascaris and Blastocystis, while those located in the 3rd quartile have a higher presence of Giardia). If the figures are disaggregated by geographical location, we observe the prevalence of parasites was twice as high in the mountains (22%) than in the jungle (11%) or the coast (9%). This is consis- tent with previous findings, as households in the mountains had the lowest access to improved water sources, improved sanitation, and a handwashing station with soap and water. www.wsp.org 65 V. Future Directions The data presented in the Findings section provides a snapshot of important human development indicators for a subsample of the Peruvian population. In addition, these data will be used in conjunction with endline data to achieve the primary goal of assessing the impacts of the handwashing project. As explained in the previous sections, the impact evaluation comprises a series of surveys, which include baseline, longitudinal, and post-intervention question- naires. At the time of this report's publication, the gathering of longitudinal data is ongoing. The collection of post-intervention data is expected to begin by the end of 2010. Data analysis and impact assessments Data analysis and impact assessments will be conducted during 2011, and a full will be conducted during 2011, and impact evaluation report will be published by the end of the year. a full impact evaluation report will be published by the end of the year. An enumerator conducts a household survey 66 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru References References Baqui, A. H., R. E. Black, M. Yunus, A. R. Hoque, Stoltzfus, R.J., M. L. Dreyfus. 1999. Guidelines for the use H. R. Chowdhury, R. B. Sack 1991. Methodological of iron supplements to prevent and treat iron deficiency issues in diarrhoeal diseases epidemiology: definition of anemia: a report of the International Nutritional Anemia diarrhoeal episodes. Int J Epidemiol. 20(4):1057­63. Consultative Group (INACG.) Washington, DC: The Black, R. E., L. H. Allen, Z. A. Bhutta, et al. 2008. Nutrition Foundation. Maternal and child undernutrition: global and re- Victora, C. G., L. Adair, C. Fall, P. C. Hallal, R. Martorell, gional exposures and health consequences. The Lancet. L. M. Richter. 2008. Maternal and child under- 371(9608):243­260. nutrition: consequences for adult health and human Bricker, D, and J. Squires. 1999. Ages and Stages Question- capital. The Lancet. 371(9609):340­357. naires: A Parent Completed, Child Monitoring System, Walker S. P., T. D. Wachs, J. Meeks Gardner, et al. 2nd Ed. Baltimore, MD: Paul Brookes. 2007. Child development: risk factors for adverse ENAHO. 2007. Encuesta Nacional de Hogares. Instituto outcomes in developing countries. The Lancet. 2007; Nacional de Estadistica e Informatica (INEI), Peru. 369(9556):145­157. Engle, P. L., M. M. Black, J. R. Behrman, et al. 2007. WHO/UNICEF Joint Monitoring Programme for Water Strategies to avoid the loss of developmental potential Supply and Sanitation Website (accessed June 2009). in more than 200 million children in the developing http://www.wssinfo.org/definitions/infrastructure.html world. The Lancet. 369(9557):229­242. World Health Organization. 2005. Pocket book of hospital Grantham-McGregor, S, Y. B. Cheung, S. Cueto, care for children: guidelines for the management of com- P. Glewwe, L. Richter, B. Strupp. 2007. Developmental mon illnesses with limited resources. WHO Press. potential in the first 5 years for children in developing ------. 2006. WHO child growth standards: length/ countries. Lancet. (9555):60­70. height-for-age, weight-for-age, weight-for-length, weight- Habicht, J.P. 1974. Estandarización de métodos epidemi- for-height and body mass index-for-age: methods and ológicos cuantitativos sobre el terreno [Standardization development. WHO Press. of quantitative epidemiological methods in the field]. ------. 2007. WHO child growth standards: head cir- Bol Oficina Sanit Panam. 76(5):375­384. cumference-for-age, arm circumference 2-for-age, triceps Nestel, P. and INACG Steering Committee. 2002. skinfold-for-age and subscapular skinfold-for-age: methods Adjusting hemoglobin values in program surveys. and development. WHO Press. (http://inacg.ilsi.org/file/Hemoglobin.pdf ) www.wsp.org 67 Findings from the Impact Evaluation Baseline Survey in Peru Annex 1: List of Districts Included in WSP Sample Annex 1: List of Districts Included in WSP Sample TABLE 44A: LIST OF DISTRICTS SELECTED TO RECEIVE TREATMENT 1 (MASS MEDIA) Treatment 1 Districts No. Region Province District Population 1 Amazonas Luya Santa Catalina 1,630 2 Amazonas Luya Santo Tomas 4,008 3 Ancash Bolognesi Cajacay 1,748 4 Ancash Bolognesi Huallanca 6,353 5 Ancash C. F. Fitzcarrald San Nicolas 3,762 6 Ancash Carhuaz Tinco 3,145 7 Ancash Huaylas Pamparomas 8,487 8 Ancash Sihuas Acobamba 1,773 9 Ancash Yungay Cascapara 1,872 10 Arequipa Arequipa San Juan de Siguas 1,633 11 Arequipa Arequipa Alto Selva Alegre 72,818 12 Arequipa Arequipa Cayma 75,908 13 Cajamarca San Miguel Bolivar 1,636 14 Cajamarca San Miguel Calquis 4,694 15 Cajamarca San Miguel San S. de Cochan 4,813 16 Cusco Acomayo Acomayo 5,062 17 Huanuco Ambo Colpas 2,872 18 Huanuco Ambo San Francisco 3,673 19 Huanuco Ambo Cayna 4,136 20 Huanuco Ambo Conchamarca 5,139 21 Huanuco Ambo Huacar 8,464 22 Huanuco Huanuco San F. de Cayran 5,056 23 Huanuco Lauricocha Jivia 1,928 24 Ica Chincha El Carmen 11,607 (Continued ) 68 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Annex 1: List of Districts Included in WSP Sample TABLE 44A: (Continued) Treatment 1 Districts No. Region Province District Population 25 Ica Chincha Grocio Prado 18,658 26 Junín Huancayo Huacrapuquio 1,589 27 Junín Huancayo Chupuro 2,494 28 Junín Jauja Parco 1,623 29 Junín Jauja Pancan 1,647 30 Junín Jauja Paca 1,658 31 Junín Jauja Pomacancha 2,244 32 Junín Jauja Marco 2,526 33 La Libertad S. de Chuco Santa Cruz de Chuca 3,478 34 La Libertad S. de Chuco Sitabamba 3,610 35 La Libertad S. de Chuco Santiago de Chuco 21,190 36 Madre de Dios Manu Huepetuhe 8,130 37 Moquegua Gral. Sanchez Cerro La Capilla 1,525 38 Moquegua Gral. Sanchez Cerro Ichuña 3,782 39 Pasco Oxapampa Palcazu 8,887 40 Tacna Tacna Pocollay 15,503 Total 340,761 TABLE 44B: LIST OF DISTRICTS SELECTED TO RECEIVE TREATMENT 2 (COMMUNITY AND SCHOOL) Treatment 2 Districts No. Region Province District Population 1 Amazonas Utcubamba Jamalca 8,137 2 Ancash A. Raymondi Chaccho 2,137 3 Ancash A. Raymondi Aczo 2,340 4 Apurímac Aymaraes Toraya 1,684 5 Arequipa Castilla Chachas 1,992 6 Arequipa Caylloma Huanca 1,919 7 Arequipa Caylloma Tisco 2,249 8 Arequipa Caylloma Caylloma 4,101 9 Ayacucho Huamanga S. de Pischa 1,643 10 Ayacucho Victor Fajardo Huancaraylla 1,796 11 Ayacucho Victor Fajardo Alcamenca 1,974 12 Cajamarca Jaen Chontali 10,344 (Continued ) www.wsp.org 69 Findings from the Impact Evaluation Baseline Survey in Peru Annex 1: List of Districts Included in WSP Sample TABLE 44B: (Continued) Treatment 2 Districts No. Region Province District Population 13 Cajamarca Jaen Santa Rosa 12,025 14 Huancavelica Huancavelica Moya 1,706 15 Huancavelica Huancavelica Nuevo Occoro 2,638 16 Huancavelica Huaytara Laramarca 1,845 17 Huancavelica Huaytara Huaytara 2,435 18 Huancavelica Huaytara Pilpichaca 5,410 19 Junín Chanchamayo Vitoc 2,301 20 Junín Chanchamayo San Ramon 24,663 21 Junín Chanchamayo Chanchamayo 25,565 22 Junín Chanchamayo Pichanaqui 40,625 23 La Libertad Pataz Ongon 1,574 24 La Libertad Pataz Pias 1,725 25 La Libertad Pataz S. de Challas 2,925 26 La Libertad Pataz Pataz 4,364 27 Lima Barranca Supe 21,693 28 Lima Cañete Asia 6,037 29 Lima Huaral Huaral 86,844 30 Loreto Requena Alto Tapiche 1,908 31 Piura Huancabamba Huarmaca 38,209 32 Piura Paita Colan 12,298 33 Piura Piura La Union 34,540 34 Piura Sechura Cristo Nos Valga 3,185 35 Puno Moho Moho 16,847 36 Puno Puno Chucuito 9,366 37 San Martin Huallaga El Eslabon 1,729 38 San Martin Huallaga Alto Saposoa 2,156 39 Tacna Jorge Basadre Locumba 1,692 40 Tacna Jorge Basadre Ite 1,763 Total 408,384 TABLE 44C: LIST OF DISTRICTS SELECTED TO SERVE AS CONTROL Control Districts No. Region Province District Population 1 Amazonas Chachapoyas Soloco 1,613 2 Amazonas Chachapoyas Chuquibamba 1,983 3 Amazonas Condorcanqui El Cenepa 11,236 4 Ancash A. Raymondi San Juan de Rontoy 1,605 5 Ancash A. Raymondi Chingas 2,071 (Continued ) 70 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Annex 1: List of Districts Included in WSP Sample TABLE 44C: (Continued) Control Districts No. Region Province District Population 6 Apurímac Aymaraes Lucre 2,391 7 Apurímac Aymaraes Tapairihua 2,770 8 Apurímac Aymaraes Chalhuanca 4,658 9 Apurímac Grau Curasco 1,742 10 Apurímac Grau Huayllati 1,915 11 Apurímac Grau Curpahuasi 2,540 12 Apurímac Grau Chuquibambilla 6,041 13 Arequipa Castilla Huancarqui 1,682 14 Arequipa Castilla Viraco 1,956 15 Arequipa Caylloma Lluta 1,859 16 Ayacucho Cangallo Chuschi 8,917 17 Ayacucho Huamanga San Jose de Ticllas 2,325 18 Ayacucho Huamanga Jesus Nazareno 15,248 19 Cajamarca San Ignacio Tabaconas 15,927 20 Cusco Chumbivilcas Chamaca 6,993 21 Cusco Chumbivilcas Llusco 7,325 22 Cusco Chumbivilcas Livitaca 11,403 23 Cusco Chumbivilcas Santo Tomas 24,614 24 Cusco Paucartambo Kosñipata 4,610 25 Huancavelica Huancavelica Huayllahuara 1,613 26 Huancavelica Huancavelica Huachocolpa 3,255 27 Huancavelica Huaytara San A. de Cusicancha 2,138 28 Huancavelica Huaytara Cordova 2,404 29 Junín Chanchamayo San Luis de Shuaro 7,193 30 Lambayeque Ferreñafe M. A. Mesones Muro 4,211 31 Lima Cañete Pacaran 1,588 32 Lima Cañete Calango 2,559 33 Lima Cañete San Antonio 3,460 34 Lima Cañete Mala 25,269 35 Lima Canta Santa Rosa de Quives 5,855 36 Loreto Alto Amazonas Balsapuerto 12,730 37 Loreto Requena Requena 26,969 38 Piura Huancabamba San M. de El Faique 9,430 39 Piura Paita Paita 69,401 40 Tumbes Tumbes La Cruz 8,092 Total 329,591 www.wsp.org 71 Findings from the Impact Evaluation Baseline Survey in Peru Annex 2: Findings from Structured Observations of Handwashing Behavior Annex 2: Findings from Structured Observations of Handwashing Behavior1 Structured five-hour observations were completed in 159 preparation events in 148 (93%), and water contact events households in Peru (see Table 45). These observations yielded in 64 (40%) households. Soap use was observed at least 2,234 events of interest during which the observer recorded once in 116 (73%) of households. the nature of the event, whether hands were washed, and whether hands were washed with soap. There were 341 fecal We analyzed self-report and rapid observation data to iden- contact events, 444 eating events, 273 feeding events, 368 tify factors associated with observation of soap use in the food preparation events, and 125 water contact events. Over- structured observation. Complete data were available for all, soap use was observed in 361 (16%) of the 2,234 events, this analysis for 115 households. The following factors were with soap use in 20% of fecal contact events, 25% of eating significantly associated with observation of soap use at a events, and just 2% of water contact events. fecal contact event: self-report of usually having soap and water at a handwashing place near the kitchen, observed At least one fecal contact event was observed in 139 (88%) presence of soap and water together at a handwashing sta- of 159 households. One or more eating events were ob- tion, and observed presence of soap and water together at a served in 141 (89%), feeding events in 132 (83%), food handwashing station in or near the toilet (Table 46). TABLE 45: SOAP USE BY EVENT TYPE AS MEASURED BY STRUCTURED OBSERVATION No. Events No. Events No. Households No. Households in Observed Accompanied by Observed with At Least Which Soap Use Was (N=2,234, %) Soap Use (%) One Event (N=159) Observed At Least Once All types 2234 361 159 116 (16%) (73%) Fecal contact2 341 68 139 58 (15%) (20%) (88%) (42%) Before eating 444 111 141 65 (20%) (25%) (89%) (46%) Before feeding a child 273 16 132 16 (12%) (6%) (83%) (12%) Before preparing 368 38 148 34 or serving food (16%) (10%) (93%) (23%) Water contact 12 3 64 3 (6%) (2%) (40%) (5%) 1 Analysis conducted by Pavani Ram. 2 Fecal contact includes defecation, toileting of any kind, and cleaning a child who has defecated. 72 Global Scaling Up Handwashing Findings from the Impact Evaluation Baseline Survey in Peru Annex 2: Findings from Structured Observations of Handwashing Behavior TABLE 46: BIVARIATE ANALYSIS OF FACTORS ASSOCIATED WITH OBSERVATION OF SOAP USE AT LEAST ONCE DURING FECAL CONTACT Factors Associated with Observation of Soap During Fecal Contact HH Observed to Use HH Observed NOT to Use Soap At Least Once During Soap At Least Once During 95% Fecal-Contact Event (%) Fecal-Contact Event (%) Odds Confidence Explanatory Variable (N=45 ) (N=70) P-Value Ratio Interval Self-report of usually having soap and water at a handwashing place near the kitchen 84% 66% 0.03 2.0 1.0­4.0 Observed soap and water together at a handwashing station 76% 56% 0.03 1.8 1.0­3.1 Observed soap and water together at a handwashing station specifically in or near the toilet 76% 53% 0.01 2.8 1.2­6.3 www.wsp.org 73 Findings from the Impact Evaluation Baseline Survey in Peru Annex 3: Test of Baseline Balance Annex 3: Test of Baseline Balance As mentioned in Section II: Methodology, a critical require- For the first comparison group--Treatment 1 vs. Control-- ment of the IE methodology is to create an appropriate the null hypothesis of mean equality at the 10% level counterfactual for the treatment group. This section pres- was rejected in 14.5% of the answers (40 out of 272 ents the mean comparison tests1 across treatment and con- answers); for the second comparison group--Treatment 2 vs. trol groups for an exhaustive group of variables included in Control--the null hypothesis of mean equality at the 10% the baseline survey. level was rejected in 11.4% of the answers (31 out of 280 answers); and for the last comparison group--Treatment 2/ Surveyed households possess many unobserved characteris- Schools vs. Control/Schools--the null hypothesis of mean tics not included in the database, and thus cannot be evalu- equality at the 10% level was rejected in 11.8% of the an- ated to see if they are balanced. However, if a sufficiently swers (33 out of 280 answers). large amount of observed variables are balanced across the different treatment groups, then there would be little reason Test of balance for the key variables included in the IE base- to believe that the unobserved variables are not balanced. line are presented in the following tables. The following tables present the mean comparison test across three different groups: · Comparison 1: Treatment 1 vs. Control · Comparison 2: Treatment 2 vs. Control · Comparison 3: Treatment 2/Schools vs. Control/ Schools 1 The standard errors used in those tests were clustered at the district level, allowing the possibility of intra-district correlation. 74 Global Scaling Up Handwashing INDIVIDUAL CHARACTERISTICS Treatment 2/ Treatment 1 Control 1 Treatment 2 Control/Schools P-value P-value Schools P-value www.wsp.org N Avg. N Avg. N Avg. N Avg. N Avg. Number of children under five 717 1.399 707 1.460 0.101 763 1.409 0.159 705 1.435 684 1.493 0.154 years of age (per HH) HH size 717 5.36 707 5.00 0.009 763 5.05 0.724 705 6.04 684 6.26 0.170 HH head is male 717 0.898 707 0.908 0.515 763 0.927 0.201 705 0.929 684 0.905 0.179 Age of HH head 713 37.964 706 35.392 0.006 762 35.757 0.673 702 38.470 683 38.269 0.752 HH heads attended school (over 710 0.952 706 0.955 0.836 763 0.969 0.203 704 0.970 682 0.959 0.262 all HH heads) Educational Attainment of HH Head Primary 674 0.445 672 0.382 0.218 735 0.410 0.558 677 0.411 649 0.473 0.210 Secondary 674 0.439 672 0.522 0.042 735 0.490 0.398 677 0.484 649 0.414 0.075 Trade school 674 0.052 672 0.052 0.993 735 0.061 0.538 677 0.040 649 0.049 0.494 University 674 0.064 672 0.043 0.185 735 0.039 0.740 677 0.065 649 0.062 0.848 Other HH member is male 3126 0.383 2831 0.371 0.266 3092 0.379 0.467 3555 0.416 3600 0.406 0.388 Age of other HH members 3123 15.004 2827 14.184 0.094 3088 14.776 0.216 3549 13.689 3594 13.609 0.819 Other HH members attended school (over all other HH 2106 0.936 1783 0.934 0.875 1993 0.963 0.012 2515 0.957 2557 0.958 0.891 members) Educational Attainment of Other HH Members Kindergarten 1941 0.062 1652 0.078 0.063 1911 0.061 0.057 2391 0.081 2434 0.096 0.136 Primary 1941 0.536 1652 0.518 0.637 1911 0.507 0.744 2391 0.622 2434 0.605 0.434 Secondary 1941 0.347 1652 0.350 0.920 1911 0.380 0.359 2391 0.269 2434 0.272 0.889 Trade school 1941 0.029 1652 0.039 0.260 1911 0.038 0.909 2391 0.018 2434 0.014 0.445 University 1941 0.025 1652 0.014 0.044 1911 0.015 0.878 2391 0.010 2434 0.014 0.492 Teenager Spent Time On School 819 0.928 710 0.962 0.372 695 0.965 0.814 1375 0.974 1393 0.933 0.229 Studying 815 0.958 710 0.973 0.229 695 0.967 0.559 1375 0.978 1393 0.953 0.258 Children care 815 0.710 711 0.729 0.541 695 0.683 0.189 1375 0.723 1395 0.720 0.941 Homework 815 0.752 711 0.713 0.266 695 0.722 0.818 1375 0.699 1395 0.691 0.832 Paid work 816 0.020 711 0.017 0.736 695 0.010 0.330 1375 0.007 1395 0.014 0.146 Unpaid work 815 0.189 711 0.091 0.012 695 0.056 0.207 1375 0.094 1395 0.080 0.648 75 76 INDIVIDUAL CHARACTERISTICS (Continued) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. HH head is employed 714 0.948 707 0.945 0.808 762 0.963 0.174 705 0.956 682 0.965 0.419 (over all HH heads) Last Week Activity, HH Head Looking for work 37 0.135 39 0.231 0.269 28 0.286 0.603 31 0.226 24 0.167 0.627 Looking after the home 37 0.324 39 0.462 0.203 28 0.357 0.355 31 0.484 24 0.542 0.680 Not working and not 37 0.216 39 0.128 0.349 28 0.250 0.215 31 0.161 24 0.208 0.659 looking for job Other 37 0.324 39 0.179 0.164 28 0.107 0.380 31 0.129 24 0.083 0.626 Other HH member is employed 1273 0.346 1055 0.344 0.941 1289 0.380 0.318 1119 0.350 1151 0.381 0.400 (over all other HH members) Last Week Activity, Other HH Members Looking for work 832 0.006 692 0.010 0.502 799 0.005 0.401 727 0.003 713 0.011 0.068 Studying 832 0.137 692 0.136 0.951 799 0.140 0.858 727 0.169 713 0.181 0.646 Looking after the home 832 0.802 692 0.803 0.943 799 0.801 0.932 727 0.779 713 0.750 0.356 Not working and not 832 0.038 692 0.040 0.900 799 0.038 0.868 727 0.030 713 0.046 0.410 looking for job Other 832 0.017 692 0.010 0.319 799 0.016 0.263 727 0.019 713 0.011 0.283 Primary Employment Status (over all employed individuals) Self-employed 1421 0.550 1290 0.557 0.825 1462 0.538 0.539 1294 0.590 1325 0.546 0.127 Employee 1421 0.265 1290 0.302 0.389 1462 0.325 0.507 1294 0.268 1325 0.303 0.330 Employer or boss 1421 0.001 1290 0.003 0.148 1462 0.005 0.425 1294 0.003 1325 0.004 0.807 Worker with no 1421 0.165 1290 0.123 0.224 1462 0.126 0.934 1294 0.134 1325 0.137 0.932 remuneration Day laborer 1421 0.018 1290 0.015 0.722 1462 0.005 0.234 1294 0.005 1325 0.008 0.630 Other 1421 0.001 1290 0.001 0.945 1462 0.000 0.309 1294 0.000 1325 0.002 0.145 Monthly salary 1148 334.22 1082 369.64 0.422 1245 391.42 0.617 1087 336.88 1101 351.01 0.634 Hours worked per 1404 42.551 1287 42.933 0.795 1448 41.823 0.395 1287 42.045 1322 40.735 0.260 week Global Scaling Up Handwashing Months worked in last 1405 9.731 1279 9.177 0.081 1446 9.577 0.177 1271 9.862 1307 9.431 0.174 12 months HOUSEHOLD ASSETS Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. www.wsp.org HH non-labor income 240 106.01 312 110.37 0.717 205 106.13 0.742 191 129.89 315 111.45 0.375 HH Assets Radio, CD, cassette 717 0.815 706 0.813 0.961 762 0.745 0.059 705 0.750 682 0.812 0.080 TV 717 0.513 707 0.532 0.780 762 0.552 0.720 705 0.609 682 0.595 0.832 VCR 717 0.197 707 0.248 0.189 762 0.241 0.867 705 0.254 682 0.224 0.452 Computer 717 0.015 707 0.020 0.509 762 0.012 0.276 705 0.016 682 0.025 0.239 Bicycle 716 0.196 707 0.184 0.773 762 0.213 0.459 705 0.240 682 0.217 0.579 Motorbike 717 0.032 707 0.025 0.568 762 0.039 0.200 705 0.035 682 0.032 0.786 Car or tractor 717 0.018 707 0.014 0.645 762 0.021 0.418 705 0.007 682 0.015 0.239 Refrigerator 717 0.075 707 0.107 0.240 762 0.079 0.250 705 0.118 682 0.087 0.227 Gas stove 717 0.329 707 0.383 0.451 762 0.419 0.604 705 0.445 682 0.346 0.178 Other type of stove 717 0.079 706 0.153 0.047 762 0.083 0.049 705 0.092 682 0.150 0.079 Blender 717 0.170 707 0.209 0.356 762 0.220 0.780 705 0.248 682 0.214 0.428 Toaster 717 0.013 707 0.004 0.146 762 0.007 0.518 705 0.010 682 0.007 0.663 Microwave 717 0.007 707 0.011 0.495 762 0.005 0.328 705 0.013 682 0.010 0.697 Washing machine 717 0.003 707 0.006 0.380 762 0.007 0.830 705 0.011 682 0.009 0.646 Water boiler 717 0.018 707 0.028 0.376 762 0.016 0.179 705 0.026 682 0.018 0.422 Other houses/properties 717 0.035 707 0.109 0.052 762 0.171 0.242 704 0.182 682 0.104 0.150 Machinery, equipment for family 717 0.031 707 0.023 0.507 761 0.021 0.859 703 0.020 682 0.015 0.578 business HH owns other piece of land 716 0.369 707 0.475 0.131 763 0.383 0.184 705 0.430 684 0.477 0.545 (over all HHs) HH owns farm equipment 716 0.260 707 0.201 0.252 763 0.215 0.779 705 0.214 684 0.200 0.794 (over all HHs) HH has animals 717 0.826 707 0.754 0.182 763 0.742 0.812 705 0.729 684 0.775 0.420 (over all HHs) Number of livestock owned per HH 717 2.787 707 2.337 0.162 763 1.992 0.192 705 2.009 684 2.401 0.203 (over all HHs) 77 78 DWELLING CHARACTERISTICS Treatment 2/ Treatment 1 Control 1 Treatment 2 Control/Schools P-value P-value Schools P-value N Avg. N Avg. N Avg. N Avg. N Avg. Dwelling Ownership (over all HHs) HH member, still paying 716 0.038 707 0.018 0.141 762 0.026 0.430 705 0.027 683 0.041 0.275 HH member, fully paid 716 0.536 707 0.454 0.072 762 0.459 0.908 705 0.482 683 0.515 0.550 Rented 716 0.084 707 0.130 0.131 762 0.118 0.701 705 0.150 683 0.111 0.264 Family/friend loan 716 0.226 707 0.198 0.362 762 0.277 0.006 705 0.197 683 0.145 0.072 Other 716 0.116 707 0.199 0.042 762 0.119 0.063 705 0.143 683 0.187 0.303 Type of Dwelling (over all HHs) Detached house 715 0.959 699 0.930 0.153 755 0.926 0.866 700 0.933 679 0.981 0.009 Room in other dwelling 715 0.027 699 0.027 0.958 755 0.050 0.164 700 0.034 679 0.009 0.055 Other 715 0.014 699 0.043 0.046 755 0.024 0.161 700 0.033 679 0.010 0.062 Dwelling Light Source (over all HHs) No lighting 714 0.000 703 0.001 0.312 761 0.011 0.025 699 0.016 683 0.004 0.064 Electricity 714 0.576 703 0.686 0.139 761 0.748 0.331 699 0.785 683 0.717 0.230 Kerosene 714 0.154 703 0.159 0.926 761 0.059 0.037 699 0.059 683 0.127 0.093 Candles 714 0.227 703 0.137 0.061 761 0.146 0.770 699 0.112 683 0.138 0.407 Other 714 0.043 703 0.017 0.216 761 0.037 0.261 699 0.029 683 0.013 0.192 Dwelling Cooking Fuel (over all HHs) Gas 714 0.237 703 0.296 0.416 761 0.293 0.967 699 0.313 683 0.233 0.281 Wood 714 0.718 703 0.587 0.120 761 0.618 0.694 699 0.568 683 0.672 0.198 Peat/manure 714 0.001 703 0.090 0.045 761 0.045 0.351 699 0.060 683 0.073 0.784 Other 714 0.043 703 0.027 0.478 761 0.045 0.522 699 0.059 683 0.022 0.148 Dwelling Heating Fuel (over all HHs) Do not heat dwelling 717 0.897 706 0.969 0.001 763 0.971 0.886 705 0.989 683 0.968 0.033 Wood stove 717 0.095 706 0.020 0.000 763 0.025 0.637 705 0.006 683 0.023 0.018 Other 717 0.008 706 0.011 0.708 763 0.004 0.327 705 0.006 683 0.009 0.576 Global Scaling Up Handwashing DWELLING CHARACTERISTICS (Continued) www.wsp.org Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. Walling Materials (over all HHs) Esteras 715 0.042 699 0.029 0.567 755 0.032 0.860 700 0.024 679 0.034 0.626 Brick 715 0.062 699 0.094 0.374 755 0.127 0.347 700 0.123 679 0.090 0.344 Concrete 715 0.014 699 0.054 0.013 755 0.057 0.911 700 0.070 679 0.019 0.026 Unbaked brick, adobe 715 0.664 699 0.584 0.393 755 0.576 0.928 700 0.543 679 0.580 0.683 Wood, logs 715 0.088 699 0.103 0.807 755 0.044 0.182 700 0.051 679 0.138 0.110 Other 715 0.130 699 0.136 0.909 755 0.164 0.564 700 0.189 679 0.138 0.369 Roofing Materials (over all HHs) Esteras 715 0.056 699 0.040 0.576 755 0.045 0.825 700 0.054 679 0.050 0.894 Brick 715 0.038 699 0.023 0.552 755 0.019 0.670 700 0.024 679 0.029 0.657 Concrete 715 0.017 699 0.052 0.074 755 0.038 0.549 700 0.053 679 0.024 0.069 Wood, logs 715 0.007 699 0.019 0.140 755 0.009 0.255 700 0.020 679 0.012 0.583 Tin, zinc sheeting 715 0.470 699 0.534 0.432 755 0.668 0.067 700 0.636 679 0.571 0.368 Bamboo 715 0.007 699 0.006 0.840 755 0.028 0.127 700 0.023 679 0.010 0.331 Other 715 0.406 699 0.328 0.342 755 0.193 0.039 700 0.190 679 0.303 0.060 Flooring Materials (over all HHs) Painted wood 713 0.004 699 0.009 0.495 753 0.009 0.937 700 0.017 679 0.013 0.708 Concrete 713 0.111 699 0.156 0.199 753 0.159 0.921 700 0.217 679 0.138 0.055 Clay, earthen floor 713 0.749 699 0.701 0.447 753 0.699 0.963 700 0.636 679 0.698 0.292 Unpolished concrete 713 0.093 699 0.076 0.563 753 0.098 0.366 700 0.091 679 0.082 0.709 Other 713 0.043 699 0.059 0.611 753 0.035 0.309 700 0.039 679 0.068 0.266 HH keeps food uncovered 651 0.257 685 0.225 0.404 704 0.207 0.616 658 0.240 663 0.270 0.487 (over all HHs) HH is clean 687 0.518 682 0.543 0.599 715 0.593 0.215 669 0.538 663 0.508 0.520 (over all HHs) HH has garbage in kitchen or 677 0.589 686 0.541 0.261 717 0.488 0.208 666 0.568 668 0.581 0.740 house (over all HHs) 79 80 TOILET FACILITIES Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. HH Main Toilet Facility (over all HHs) No facilities 717 0.240 707 0.221 0.709 763 0.239 0.722 704 0.203 683 0.218 0.755 Hanging toilet, latrine 717 0.010 707 0.017 0.561 763 0.001 0.178 704 0.003 683 0.018 0.397 Flush, to piped sewer system 717 0.132 707 0.226 0.082 763 0.229 0.955 704 0.339 683 0.217 0.052 Flush, to other place 717 0.047 707 0.095 0.042 763 0.101 0.853 704 0.101 683 0.088 0.699 Ventilated improved pit latrine 717 0.073 707 0.042 0.309 763 0.069 0.366 704 0.061 683 0.028 0.276 Pit latrine with slab 717 0.033 707 0.034 0.978 763 0.060 0.203 704 0.044 683 0.023 0.239 Pit latrine without slab 717 0.411 707 0.325 0.162 763 0.263 0.210 704 0.217 683 0.381 0.001 Other 717 0.053 707 0.040 0.553 763 0.037 0.888 704 0.031 683 0.028 0.765 Public toilet facilities 544 0.121 548 0.144 0.581 581 0.129 0.738 553 0.105 532 0.092 0.669 (over all HHs) Location of Main Toilet Facility Inside dwelling 717 0.156 707 0.202 0.364 763 0.224 0.669 705 0.278 684 0.211 0.200 In own yard 717 0.392 707 0.382 0.832 763 0.391 0.860 705 0.403 684 0.405 0.966 Less than 10-min. walk 717 0.340 707 0.291 0.254 763 0.248 0.262 705 0.213 684 0.269 0.199 More than 10-min. walk 717 0.071 707 0.088 0.432 763 0.110 0.388 705 0.077 684 0.076 0.979 No designated area 717 0.038 707 0.035 0.884 763 0.025 0.467 705 0.026 684 0.037 0.424 Other 717 0.003 707 0.001 0.567 763 0.003 0.683 705 0.004 684 0.003 0.752 Toilet facility is shared with other HHs 717 0.254 707 0.263 0.827 763 0.304 0.349 705 0.271 684 0.230 0.289 (over all HHs) Toilet facility is safe during night 715 0.738 707 0.745 0.849 763 0.773 0.424 704 0.781 683 0.761 0.601 (over all HHs) Disposal of Child Feces Bushes, ground 717 0.279 707 0.337 0.196 763 0.266 0.103 705 0.173 684 0.303 0.002 Pit, hole in the ground 717 0.100 707 0.092 0.716 763 0.087 0.799 705 0.096 684 0.076 0.364 Open sewer, drain 717 0.025 707 0.048 0.316 763 0.045 0.893 705 0.065 684 0.047 0.411 Toilet, latrine 717 0.209 707 0.163 0.128 763 0.215 0.126 705 0.237 684 0.209 0.486 Garbage 717 0.301 707 0.301 1.000 763 0.307 0.929 705 0.340 684 0.308 0.636 River 717 0.121 707 0.120 0.973 763 0.092 0.357 705 0.098 684 0.110 0.710 Basin, sink 717 0.114 707 0.098 0.622 763 0.060 0.164 705 0.062 684 0.104 0.129 Global Scaling Up Handwashing Other 717 0.064 707 0.071 0.732 763 0.041 0.064 705 0.045 684 0.064 0.358 WATER SOURCE Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value www.wsp.org N Avg. N Avg. N Avg. N Avg. N Avg. HH uses same sources all year 717 0.974 707 0.970 0.806 763 0.988 0.152 705 0.989 683 0.990 0.909 (over all HHs) HH Source of Drinking Water (over all HH) Tanker truck 717 0.026 707 0.000 0.062 763 0.004 0.180 705 0.007 683 0.015 0.579 Surface water 717 0.081 707 0.055 0.407 763 0.031 0.325 705 0.044 683 0.064 0.470 Piped water, into dwelling 717 0.233 707 0.223 0.870 763 0.249 0.667 705 0.250 683 0.228 0.724 Piped water, into yard, plot 717 0.201 707 0.160 0.370 763 0.232 0.159 705 0.247 683 0.170 0.173 Piped water, public tap, 717 0.095 707 0.051 0.280 763 0.043 0.749 705 0.045 683 0.053 0.784 standpipe Tube well, bore hole 717 0.013 707 0.024 0.447 763 0.010 0.366 705 0.013 683 0.004 0.262 Dug well, protected 717 0.026 707 0.017 0.672 763 0.008 0.316 705 0.010 683 0.006 0.576 Dug well, unprotected 717 0.004 707 0.040 0.132 763 0.007 0.163 705 0.009 683 0.026 0.271 Spring water, protected 717 0.121 707 0.262 0.062 763 0.257 0.957 705 0.230 683 0.261 0.731 Spring water, unprotected 717 0.077 707 0.041 0.176 763 0.031 0.623 705 0.023 683 0.038 0.398 Other 717 0.123 707 0.127 0.937 763 0.127 0.998 705 0.123 683 0.135 0.854 Source Location (over all HH) In own dwelling 406 0.074 436 0.128 0.346 396 0.278 0.123 355 0.282 411 0.148 0.173 In own yard, plot 406 0.355 436 0.431 0.361 396 0.313 0.210 355 0.287 411 0.372 0.335 Elsewhere 406 0.571 436 0.440 0.149 396 0.409 0.728 355 0.431 411 0.479 0.618 Covered Source (over all HH) Covered 404 0.597 433 0.607 0.913 393 0.687 0.462 355 0.645 407 0.582 0.569 Open 404 0.389 433 0.372 0.869 393 0.303 0.527 355 0.352 407 0.388 0.744 Both covered and open 404 0.015 433 0.021 0.659 393 0.010 0.394 355 0.003 407 0.029 0.186 HH Member Who Collects Water from Source Adult woman 405 0.847 436 0.846 0.990 396 0.886 0.372 355 0.865 411 0.820 0.285 Adult man 405 0.114 436 0.115 0.978 396 0.086 0.481 355 0.101 411 0.117 0.676 Girl (<15 years) 405 0.017 436 0.023 0.596 396 0.015 0.429 355 0.014 411 0.027 0.250 Boy (<15 years) 405 0.020 436 0.014 0.542 396 0.013 0.896 355 0.020 411 0.022 0.845 Other 405 0.002 436 0.002 0.958 396 0.000 0.322 355 0.000 411 0.015 0.075 81 82 WATER SOURCE (Continued) Treatment 2/ Treatment 1 Control 1 Treatment 2 Control/Schools P-value P-value Schools P-value N Avg. N Avg. N Avg. N Avg. N Avg. HH is satisfied with water quantity 715 0.734 704 0.724 0.806 763 0.738 0.767 702 0.708 681 0.686 0.642 (over all HHs) HH pays for water 716 0.603 706 0.564 0.564 761 0.662 0.140 705 0.674 683 0.698 0.723 (over all HHs) HH obtains fixed water quantity for the payment 428 0.339 389 0.524 0.017 495 0.442 0.359 465 0.447 466 0.470 0.801 (over all HHs) Water Treatment (Past 7 Days) Boiling treatment 603 0.954 649 0.948 0.816 689 0.972 0.314 635 0.980 618 0.963 0.338 Chlorine treatment 603 0.060 649 0.034 0.245 689 0.026 0.526 635 0.020 618 0.026 0.619 Let stand and settle 603 0.060 649 0.032 0.317 689 0.022 0.629 635 0.022 618 0.026 0.810 Other 603 0.015 649 0.015 0.969 689 0.000 0.121 635 0.002 618 0.010 0.242 HH has improved water source 717 0.690 707 0.737 0.533 763 0.801 0.360 705 0.804 683 0.722 0.269 HH has improved sanitation 717 0.278 707 0.386 0.075 763 0.461 0.195 704 0.544 683 0.335 0.001 HH has soap and water at HW 717 0.562 707 0.588 0.598 763 0.598 0.836 705 0.603 684 0.639 0.405 station Global Scaling Up Handwashing HANDWASHING FACILITIES Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value www.wsp.org N Avg. N Avg. N Avg. N Avg. N Avg. HH handwashing after using toilet 707 0.993 698 0.986 0.258 738 0.985 0.945 674 0.987 675 0.994 0.198 (over all HHs) Location of Handwashing Device Inside toilet facility 685 0.058 683 0.104 0.080 718 0.113 0.775 659 0.118 667 0.100 0.557 Inside cooking place 685 0.067 683 0.105 0.153 718 0.131 0.426 659 0.082 667 0.112 0.306 In yard less than 3 feet away 685 0.194 683 0.193 0.980 718 0.132 0.050 659 0.185 667 0.219 0.344 from toilet Between 10 feet and 3 feet 685 0.155 683 0.142 0.629 718 0.117 0.283 659 0.141 667 0.148 0.777 away from toilet More than 10 feet away from 685 0.352 683 0.335 0.743 718 0.373 0.451 659 0.340 667 0.315 0.631 toilet No specific place 685 0.174 683 0.120 0.109 718 0.134 0.638 659 0.134 667 0.105 0.308 Handwashing Device, Toilet Tap, faucet 563 0.584 600 0.643 0.381 620 0.655 0.862 570 0.695 597 0.631 0.367 Basin, bucket 563 0.352 600 0.335 0.806 620 0.319 0.813 570 0.295 597 0.337 0.548 Other 563 0.064 600 0.022 0.019 620 0.026 0.686 570 0.011 597 0.032 0.033 Water is available at handwashing station 561 0.850 599 0.871 0.531 619 0.889 0.601 570 0.854 594 0.877 0.510 (over all HHs) Soaps Available at Hand- washing Station Multipurpose bar soap 567 0.254 601 0.095 0.001 622 0.127 0.281 571 0.128 597 0.134 0.832 Beauty, toilet bar soap 567 0.224 601 0.245 0.645 622 0.241 0.941 571 0.254 597 0.214 0.361 Powder soap, detergent 567 0.347 601 0.486 0.003 622 0.471 0.729 571 0.478 597 0.524 0.256 No soap observed 567 0.275 601 0.306 0.483 622 0.273 0.422 571 0.282 597 0.295 0.725 Ash, Mud at Handwashing Station Ash 553 0.016 595 0.012 0.612 614 0.010 0.790 566 0.004 590 0.005 0.730 Mud 553 0.195 595 0.245 0.374 614 0.238 0.897 566 0.221 590 0.237 0.762 Ash and mud 553 0.045 595 0.034 0.542 614 0.024 0.511 566 0.037 590 0.049 0.524 Neither observed 553 0.743 595 0.709 0.617 614 0.728 0.766 566 0.739 590 0.708 0.615 83 84 HANDWASHING FACILITIES (Continued) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. HH handwashing before/after cooking, feeding a child 706 0.993 697 0.996 0.529 738 0.996 0.951 674 0.987 675 0.996 0.135 (over all HHs) Usual Handwashing Station Inside toilet facility 631 0.014 647 0.028 0.204 700 0.011 0.117 636 0.017 635 0.020 0.715 Inside cooking place 631 0.361 647 0.396 0.535 700 0.434 0.493 636 0.414 635 0.413 0.988 In yard less than 3 feet away 631 0.195 647 0.207 0.720 700 0.164 0.190 636 0.165 635 0.170 0.880 from kitchen Between 10 feet and 3 feet 631 0.189 647 0.162 0.388 700 0.154 0.757 636 0.173 635 0.191 0.567 away from kitchen More than 10 feet away from 631 0.127 647 0.124 0.913 700 0.139 0.602 636 0.170 635 0.131 0.242 kitchen No specific place 631 0.114 647 0.083 0.184 700 0.097 0.555 636 0.061 635 0.076 0.416 Handwashing Device Tap, faucet 245 0.208 243 0.263 0.389 283 0.314 0.407 288 0.299 237 0.283 0.807 Water (pouring) container 245 0.780 243 0.716 0.325 283 0.678 0.545 288 0.688 237 0.696 0.895 Other 245 0.012 243 0.021 0.494 283 0.007 0.229 288 0.014 237 0.021 0.637 Water is available at handwashing station 246 0.785 244 0.750 0.555 283 0.837 0.086 288 0.792 237 0.819 0.492 (over all HHs) Soaps Available at Handwashing Station Multipurpose bar soap 246 0.138 244 0.070 0.142 283 0.078 0.796 288 0.049 237 0.097 0.134 Beauty, toilet soap 246 0.081 244 0.107 0.478 283 0.110 0.933 288 0.101 237 0.114 0.703 Powder or laundry soap, 246 0.472 244 0.561 0.282 283 0.509 0.507 288 0.573 237 0.612 0.624 detergent No soap observed 246 0.341 244 0.352 0.874 283 0.385 0.648 288 0.365 237 0.287 0.292 Ash, Mud at Handwashing Station Ash 241 0.008 240 0.038 0.176 282 0.000 0.071 284 0.007 232 0.013 0.572 Mud 241 0.124 240 0.150 0.612 282 0.078 0.091 284 0.123 232 0.168 0.307 Ash and mud 241 0.087 240 0.058 0.439 282 0.025 0.140 284 0.053 232 0.069 0.611 Neither observed 241 0.780 240 0.754 0.766 282 0.897 0.037 284 0.817 232 0.750 0.308 Global Scaling Up Handwashing HANDWASHING BEHAVIOR Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. www.wsp.org Caregiver of child under the age of 2 washed hands with soap 720 0.999 712 0.997 0.647 765 0.996 0.751 707 0.994 693 0.999 0.169 since yesterday Last Moment of Hand Wash Since Yesterday Bathing a child 717 0.195 707 0.255 0.147 759 0.244 0.784 703 0.262 689 0.253 0.841 Washing child's hands 717 0.110 707 0.123 0.598 759 0.083 0.088 703 0.083 689 0.094 0.526 Washing dishes 717 0.459 707 0.410 0.212 759 0.348 0.109 703 0.331 689 0.437 0.006 Doing laundry 717 0.445 707 0.436 0.818 759 0.440 0.913 703 0.511 689 0.498 0.708 Looked dirty 717 0.066 707 0.102 0.087 759 0.045 0.002 703 0.053 689 0.115 0.007 Bathing oneself 717 0.153 707 0.204 0.180 759 0.219 0.696 703 0.259 689 0.224 0.367 Using toilet 717 0.389 707 0.426 0.362 759 0.397 0.484 703 0.385 689 0.356 0.427 Cleaning baby bottom 717 0.424 707 0.334 0.021 759 0.368 0.390 703 0.356 689 0.327 0.478 Cleaning latrine 717 0.010 707 0.021 0.152 759 0.008 0.087 703 0.024 689 0.015 0.219 Cleaning toilet 717 0.035 707 0.035 0.966 759 0.022 0.203 703 0.028 689 0.036 0.512 Returning home 717 0.130 707 0.123 0.768 759 0.119 0.809 703 0.134 689 0.129 0.824 Preparing food, cooking 717 0.763 707 0.717 0.129 759 0.675 0.219 703 0.643 689 0.704 0.094 Feeding children 717 0.351 707 0.369 0.611 759 0.278 0.007 703 0.282 689 0.335 0.168 Other 717 0.102 707 0.057 0.008 759 0.075 0.222 703 0.051 689 0.038 0.328 Best Way to Clean Hands Wipe on cloth 718 0.011 713 0.011 0.989 765 0.013 0.719 704 0.009 694 0.013 0.522 Wash with water alone 718 0.110 713 0.132 0.430 765 0.116 0.631 704 0.111 694 0.134 0.386 Wash with soap 718 0.864 713 0.847 0.591 765 0.854 0.854 704 0.865 694 0.840 0.421 Wash with ash, mud 718 0.003 713 0.000 0.157 765 0.000 704 0.000 694 0.003 0.149 Other 718 0.013 713 0.010 0.738 765 0.017 0.434 704 0.016 694 0.010 0.507 Caregiver's Fingernails Are Visibly dirty 719 0.303 714 0.284 0.687 767 0.210 0.093 703 0.225 694 0.281 0.225 Unclean in appearance 719 0.325 714 0.322 0.933 767 0.286 0.255 703 0.297 694 0.336 0.241 Clean 719 0.371 714 0.394 0.666 767 0.505 0.038 703 0.478 694 0.383 0.098 85 86 HANDWASHING BEHAVIOR (Continued) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. Caregiver's Palms Are Visibly dirty 719 0.220 714 0.225 0.894 767 0.154 0.095 703 0.137 694 0.225 0.036 Unclean in appearance 719 0.243 714 0.252 0.767 767 0.207 0.122 703 0.272 694 0.272 0.986 Clean 719 0.537 714 0.522 0.792 767 0.639 0.045 703 0.592 694 0.503 0.159 Caregiver's Finger Pads Are Visibly dirty 719 0.224 714 0.224 0.997 767 0.147 0.073 702 0.140 694 0.220 0.070 Unclean in appearance 719 0.249 714 0.265 0.597 767 0.210 0.068 702 0.255 694 0.277 0.532 Clean 719 0.527 714 0.511 0.775 767 0.643 0.026 702 0.605 694 0.503 0.104 MASS MEDIA Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. Caregiver recalls any 722 0.253 714 0.245 0.854 769 0.224 0.639 711 0.226 694 0.272 0.323 handwashing campaign FAMILY-SCHOOL RELATIONSHIP Treatment 2/ Treatment 1 Control 1 Treatment 2 Control/Schools P-value P-value Schools P-value N Avg. N Avg. N Avg. N Avg. N Avg. Caregiver Participation in School's Activities Parents association 389 0.2 88 296 0.176 0.025 301 0.256 0.150 702 0.249 688 0.196 0.280 Speeches, conferences 389 0.342 296 0.267 0.155 301 0.369 0.062 702 0.325 688 0.298 0.584 Kermesses 389 0.111 296 0.139 0.490 301 0.096 0.332 702 0.094 688 0.122 0.343 APAFA 389 0.584 296 0.720 0.025 301 0.611 0.083 702 0.598 688 0.670 0.145 Other 389 0.185 296 0.199 0.805 301 0.169 0.636 702 0.175 688 0.161 0.786 Does not participate 389 0.090 296 0.034 0.015 301 0.053 0.378 702 0.078 688 0.078 0.996 Caregiver recalls any campaign on health and hygiene 389 0.298 296 0.338 0.379 301 0.336 0.967 702 0.392 688 0.390 0.963 promoted by the school Global Scaling Up Handwashing CHILD DEVELOPMENT (% OF CHILDREN <2) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value www.wsp.org N Avg. N Avg. N Avg. N Avg. N Avg. Communication skills-for-age 557 0.051 581 0.060 0.910 601 -0.014 0.348 554 -0.057 533 -0.043 0.878 z-score Mobility skills-for-age z-score 556 -0.030 581 -0.011 0.818 601 0.001 0.880 554 -0.027 532 0.072 0.225 Social-personal skills-for-age 554 0.070 581 -0.009 0.333 599 -0.052 0.604 553 -0.080 533 0.079 0.054 z-score CHILD CARE SITUATION Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. Number of times child was left 735 0.707 732 0.657 0.692 786 0.641 0.877 729 1.060 717 1.024 0.799 at the charge of another child Number of times child was left 735 0.405 734 0.327 0.431 786 0.314 0.885 729 0.258 717 0.211 0.418 alone Child has clean aspect 955 0.625 982 0.667 0.329 1027 0.689 0.585 955 0.672 980 0.624 0.276 (over all children) Child has dirty hands 953 0.518 979 0.501 0.696 1020 0.429 0.109 951 0.449 976 0.513 0.212 (over all children) Child has dirty finger nails 948 0.584 973 0.597 0.803 1018 0.476 0.012 945 0.525 969 0.608 0.068 (over all children) Child has dirty face 955 0.452 981 0.414 0.405 1029 0.331 0.082 955 0.365 979 0.441 0.139 (over all children) Child wears clothes 955 0.439 976 0.411 0.510 1028 0.371 0.340 954 0.399 980 0.476 0.119 (over all children) Child wears dirty clothes 955 0.988 980 0.989 0.951 1030 0.993 0.295 956 0.985 981 0.989 0.524 (over all children) Child has pot-belly 944 0.144 979 0.153 0.783 1019 0.106 0.146 951 0.139 977 0.159 0.563 (over all children) Child wears shoes or has shoes available 956 0.844 986 0.853 0.698 1032 0.834 0.471 962 0.823 982 0.869 0.067 (over all children) 87 88 CHILD CARE SITUATION (Continued) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. Child plays with household objects 734 0.580 733 0.673 0.015 784 0.626 0.157 729 0.647 718 0.636 0.739 (over all children) Child plays with toys 734 0.800 733 0.809 0.701 784 0.800 0.710 729 0.813 718 0.788 0.352 (over all children) Number of children's books or 736 0.255 736 0.292 0.679 788 0.208 0.322 730 0.225 718 0.230 0.940 pictures Child attended early education 734 0.030 732 0.040 0.489 783 0.041 0.944 726 0.032 718 0.033 0.914 programs Adult reads books with child 733 0.225 731 0.274 0.113 784 0.241 0.262 729 0.263 717 0.225 0.164 Adult tells stories to child 731 0.197 732 0.265 0.025 784 0.227 0.170 729 0.254 717 0.247 0.824 Adult takes the child outside the 734 0.913 733 0.943 0.163 784 0.926 0.459 729 0.918 718 0.911 0.769 house Adult plays with child 734 0.869 733 0.868 0.949 784 0.857 0.659 729 0.842 718 0.831 0.660 Average daily caring time 988 4.924 1028 5.038 0.767 1062 5.704 0.141 1001 5.393 1015 4.858 0.249 ACUTE LOWER RESPIRATORY INFECTION AND DIARRHEA SYMPTOMS PREVALENCE (% OF CHILDREN <5) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. ALRI in previous 48 hrs 1003 0.103 1031 0.029 0.001 1074 0.049 0.252 1012 0.053 1017 0.031 0.306 ALRI in previous week 1003 0.139 1031 0.040 0.001 1074 0.065 0.274 1012 0.073 1017 0.043 0.272 Diarrhea in previous 48 hrs 1003 0.082 1031 0.098 0.368 1074 0.101 0.888 1012 0.084 1017 0.077 0.639 Diarrhea in previous week 1003 0.153 1031 0.167 0.586 1074 0.162 0.844 1012 0.139 1017 0.140 0.990 Household lost working hours 3832 0.063 3534 0.014 0.001 3852 0.019 0.507 4232 0.010 4236 0.020 0.335 due to child illness (over all HHs) Global Scaling Up Handwashing www.wsp.org ANTHROPOMETRIC MEASURES AND ANEMIA (CHILDREN <2) Treatment 1 Control 1 Treatment 2 Treatment 2/Schools Control/Schools P-value P-value P-value N Avg. N Avg. N Avg. N Avg. N Avg. BMI-for-age z-score 719 0.463 709 0.361 0.263 769 0.405 0.599 709 0.471 692 0.455 0.844 Head circumference-for-age 714 -0.332 707 -0.238 0.290 769 -0.310 0.454 714 -0.261 690 -0.281 0.825 z-score Length/height-for-age z-score 722 -1.380 711 -1.314 0.559 765 -1.336 0.841 712 -1.398 690 -1.396 0.979 Arm circumference-for-age 631 0.397 635 0.300 0.364 684 0.350 0.663 634 0.314 609 0.271 0.695 z-score Weight-for-length/height z-score 718 0.374 707 0.287 0.336 768 0.342 0.519 706 0.357 690 0.392 0.674 Weight-for-age z-score 724 -0.502 713 -0.502 1.000 774 -0.515 0.900 716 -0.524 694 -0.487 0.723 Anemia (Hb <110 g/L) 652 0.701 605 0.711 0.777 632 0.731 0.503 565 0.701 596 0.711 0.744 MICROBIOLOGY AND PARASITOLOGY Treatment 2 Control P-value N Avg. N Avg. Log10 E. coli, MPN/100ml, Child 74 0.463 86 0.584 0.472 Log10 E. coli, MPN/100ml, Mother 74 0.829 86 0.702 0.469 Log10 E. coli, MPN/100ml, Object 72 0.621 82 0.595 0.902 Log10 E. coli, MPN/100ml, Water 74 0.814 85 0.432 0.055 Stool sample, Ascaris detected 74 0.000 86 0.023 0.134 Stool sample, Blastocystis detected 74 0.108 86 0.105 0.949 Stool sample, Giardia detected 74 0.041 86 0.105 0.115 Any parasite detected 74 0.122 86 0.186 0.311 89