Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania

Transcription

1 Water and Sanitation Program: technical Paper Public Disclosure Authorized Public Disclosure Authorized Global Scaling Up Rural Sanitation Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Bertha Briceño and Ahmad Yusuf September 2012 Public Disclosure Authorized Public Disclosure Authorized 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.

2 Bertha Briceño and Ahmad Yusuf, Water and Sanitation Program Today, 2.5 billion people live without access to improved sanitation. Of these, 75 percent live in rural communities. To address this challenge, WSP is working with governments and local private sectors to build capacity and strengthen performance monitoring, policy, financing, and other components needed to develop and institutionalize large scale, sustainable rural sanitation programs. With a focus on building a rigorous evidence base to support replication, WSP combines Community-Led Total Sanitation, behavior change communication, and sanitation marketing to generate sanitation demand and strengthen the supply of sanitation products and services, leading to improved health for people in rural areas. For more information, please visit This technical paper is one in a series of knowledge products designed to showcase project findings, assessments, and lessons learned through WSP s Scaling Up Rural Sanitation initiatives. This paper is conceived as a work in progress to encourage the exchange of ideas about development issues. For more information please Bertha Briceño and Ahmad Yusuf at wsp@worldbank.org or visit WSP is a multi-donor partnership created in 1978 and administered by the World Bank to support poor people in obtaining affordable, safe, and sustainable access to water and sanitation services. WSP s donors include Australia, Austria, Canada, Denmark, Finland, France, the Bill & Melinda Gates Foundation, Ireland, Luxembourg, Netherlands, Norway, Sweden, Switzerland, United Kingdom, United States, and the World Bank. WSP reports are published to communicate the results of WSP s work to the development community. Some sources cited might be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are entirely those of the author and should not be attributed to the World Bank or its affiliated organizations, or to members of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The map was produced by the Map Design Unit of the World Bank. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank Group concerning the legal status of any territory or the endorsement or acceptance of such boundaries Water and Sanitation Program

3 Global Scaling Up Rural Sanitation Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Bertha Briceño and Ahmad Yusuf September 2012

4 Acknowledgments The Tanzania Impact Evaluation Team oversees the in-country design and field activities; investigators include Sebastian Martinez, Bertha Briceño, Aidan Coville, and (in its early stages) Alicia Salvatore. The survey analyzed in this report was intended to serve as a baseline to the impact evaluation of the program but unanticipated problems with reliability of data resulted in the cancelation of field work in five out of the 10 districts originally planned and the impossibility of using the data to validate the randomized design, as it was originally intended. The project s global impact evaluation task team is led by Bertha Briceño (in its early stages, it was led by Jack Molyneaux), together with Alexandra Orsola-Vidal and Claire Chase. Professor Paul Gertler has provided guidance and advice throughout the study. Advisors also include Sebastian Galiani, John M. Colford, Benjamin Arnold, and Pavani Ram. The authors are grateful to Jason A. Cardosi, Patrick Mwakilama, Kaposo Mwambuli, and Yolande Coombes, the country task team for project implementation in Tanzania; to Wambui Gichuri, former Regional Team Leader for the WSP-Africa region; and to Eduardo Perez, overall task team leader. Generous financial support was provided by the Bill and Melinda Gates Foundation. ii Scaling Up Rural Sanitation

5 Executive Summary Since 2007, the Water and Sanitation Program (WSP) has provided technical assistance to local and national governments implementing large rural sanitation and handwashing promotion programs in various countries. 1 In Tanzania, handwashing with soap and sanitation programs were phased into 10 rural districts in the second half of This report presents summary descriptive statistics for key demographic, socioeconomic, hygiene, health, and child development variables based on a survey of approximately 1,500 households in the Masasi, Musoma rural, Rufiji, Iringa rural, and Mpwapwa districts. It offers a glimpse at the general status of sanitation and hygiene practices in some of the program s target areas before the beginning of implementation activities. The household survey, conducted by Muhimbili University of Health and Allied Sciences (MUHAS), was carried at the end of 2008 and early in The original goal of the baseline survey was to validate the design of the impact evaluation of the program. However, unexpected problems with data reliability and the enumerators training resulted in data collection being cancelled in five of the 10 districts that were part of the evaluation. The statistics presented have been positively validated against existing household surveys, including the 2010 Tanzania Demographic and Health Survey (DHS). Summary of Findings Characteristics of Targeted Households In the targeted areas in rural Tanzania, the typical household is headed by a male and comprises five members. Most houses are single detached dwellings with mud or brick walls and clay floors. Households typically use kerosene for lighting and wood for cooking, and about half of households own a few animals and a bicycle. The mean age of household members studied was 18.2, and 80% of household members had some schooling. The average self-reported monthly income by the household head was 114,769 TZS (approximately US$73.00). 2 Although self-reported income is not always a reliable measure and must be interpreted with caution, the income of the household head can serve as a proxy for total household income. Water Sources Water is typically collected by women (94.2% of households in our study). On average, only about half of households (49.7%) had access to an improved water source, which generally was not in their dwellings or on their plots. Drinking water typically came from three main sources: piped water from public tap (24.2%), an unprotected dug well (27.1%), and surface water (19.2%). On average, 31.5% of these water sources were located in the household s own yard or plot, 13.4% within the dwelling, and the rest (55.1%) were located elsewhere, suggesting that most households are investing time in collecting water. The majority of households stored drinking water at home, but only 24.4% reportedly treated the water before drinking. Of these, less than half (43%) mentioned boiling water before drinking as a treatment. This suggests that improving water safety practices in these districts could help prevent waterborne diseases. Handwashing Behavior and Sanitation Handwashing behavior is known to be difficult to assess. In this study, we relied on two sources: self-reported handwashing at critical times and, as a proxy measure, spot-check observations of whether the household had a designated place for handwashing with both soap and water. An additional measure assessed the cleanliness of the caretaker s hands through direct observation again to serve as a proxy indicator of handwashing with soap behavior. Although most caregivers (more than 78%) reported having used soap to wash hands at least once in the previous 24 hours, when asked about the circumstances, only 33.7% 1 For more information on Global Scaling Up Rural Sanitation and Handwashing, visit 2 Approximate conversion rate as of June 2012 was 1,500 Tanzanian shillings (TZS) to 1 US dollar (USD). iii

6 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Executive Summary mentioned handwashing at one or more critical juncture and only 21.2% mentioned handwashing after using the toilet. In terms of sanitation, rural Tanzania mostly relies on basic latrines that do not separate humans from excreta and are described as fixed-point open defecation. The most common type of toilet facility reported was a pit latrine, present in almost 89% of the studied households. Only 12.7% of households reported having a pit latrine with slab. Open defecation was reported in 9.2% of the households. Child Care and Nutrition Because caregiver characteristics and child care quality can dramatically affect the health, nutritional status, and development of young children, the study collected information on feeding practices, caregiving behavior, and caregiver well-being. In the studied households, most children under the age of two played with household objects (85.9%) and, in more than half of the cases, an adult sang songs with the child. However, in the three days prior to the survey, less than 15% of caregivers reported reading (13.7%) or telling stories (14.5%) to the child. In terms of diet, almost 90% of children had been given plain water since the previous day and breast milk consumption, as expected, was very high (89%) for children under two. Most children (72.1%) consumed homemade gruel and one-fifth of the children were given beverages such as tea or coffee (20.3%). With respect to the previous day s food, 70.3% of the children had received solid or semi-solid food an average of 2.41 times. Conclusion The survey revealed that there was limited baseline knowledge of the critical handwashing times among the target households prior to the program, indicating room to improve handwashing behavior. Likewise, the survey indicated limited access to improved water sources, a scarcity of pit latrines with slabs, and a non-negligible percentage of open defecation practice in the studied households. Underlying challenges also included unsafe facilities for small children and poor practices related to disposal of child feces. The data presented in this technical report provides a snapshot of the conditions of the target population prior to the start of the sanitation and handwashing programs. An impact evaluation of the programs, which will rely exclusively on post-intervention data, will be carried out during 2012; a full report will be published in The study hopes to enable a close examination of the links between poor sanitation, handwashing behavior, and health, and provide evidence for future projects in rural Tanzania. iv Scaling Up Rural Sanitation

7 Contents Executive Summary... iii Acronyms... vii I. Introduction... 1 II. Characteristics of Targeted Households... 3 III. Water Sources... 7 IV. Handwashing Behavior... 9 V. Sanitation VI. Child Care Environment VII. Child Health and Anthropometric Measures VIII. Conclusion Figures 1: Map of Tanzania Showing 10 Project Districts : Distribution of Sampled Population by Age Group and Gender (Population Pyramid)... 4 Tables 1: Summary Statistics : Percent Distribution of the Basic Socio-Demographic Characteristics : Percent Distribution of Households Assets (% HHs) : Dwelling Characteristics and Materials by Observation : Dwelling Energy Source : Water Source : Drinking Water : Self-Reported Handwashing Behavior with Soap by District : Household Main Sanitation Facility Characteristics (% HHs) : Household Main Sanitation Facility Sharing and Safety (% HHs) v

8 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Contents 11: Other Characteristics of Households Sanitary Condition (% HHs) : Child Breastfeeding (Children < 2) : Infant/Young Child Feeding (Children < 2) : Infant/Young Child Learning Environment (Children < 2) : Anemia Prevalence (% of Children < 2) : Nutritional Status of Children < : Estimates of Z-Scores (in SD) from Selected Datasets vi Scaling Up Rural Sanitation

9 Abbreviations and Acronyms ALRI(s) ARI(s) CLTS DHS HH MUHAS SD(s) TZS WHO WSP Acute Lower Respiratory Infection(s) Acute Respiratory Infection(s) Community-Led Total Sanitation Demographic and Health Survey Household Muhimbili University of Health and Allied Sciences Standard Deviation(s) Tanzanian Shilling World Health Organization Water and Sanitation Program vii

10

11 I. Introduction Since 2007, the Water and Sanitation Program (WSP) has provided technical assistance to local and national governments implementing large rural sanitation and handwashing promotion programs in various countries, under the umbrella of two related projects, Global Scaling Up Handwashing and Global Scaling Up Rural Sanitation. 3 The goal is to contribute to addressing the poor hygiene and sanitation conditions of large rural populations in the developing world. In Tanzania, the handwashing with soap and rural sanitation programs were phased into 10 rural districts in the second half of Handwashing with soap at critical times (such as after contact with feces and before handling food) has been shown to substantially reduce the risk of diarrhea and acute respiratory infections (ARIs). The handwashing project aims to test whether handwashing with soap behavior can be generated and sustained among the poor and vulnerable using innovative promotional approaches. In a similar fashion, scaling-up rural sanitation focuses on learning how to combine the approaches of community-led total sanitation (CLTS), behavior change communications, and social marketing of sanitation to generate sanitation demand and strengthen the supply of sanitation products and services at scale, leading to improved health for people in rural areas. RWANDA BURUNDI DEM. REP. OF CONGO Lake Tanganyika UGANDA Lake KARAGWE Victoria MUSOMA SUMBAWANGA paper presents findings from the survey carried to assess the pre-program conditions of sanitation and handwashing in five out of the 10 districts where the program is being implemented in Tanzania. The household survey was commissioned to the MUHAS. It was carried out between the end of 2008 and early 2009 in approximately 1,500 households in the districts of Masasi, Musoma rural, Rufiji, Iringa rural, and Mpwapwa. The survey intended to serve as a baseline to the implementation of the program, but unanticipated problems with reliability of data and enumerators training resulted in the cancelation of data collection in five out of the 10 districts originally planned and the impossibility of using the data in longitudinal analysis as baseline to validate the evaluation randomized design, as it was originally intended. FIGURE 1: MAP OF TANZANIA SHOWING 10 PROJECT DISTRICTS IGUNGA KITETO DODOMA MPWAPWA IRINGA KONDOA KENYA Dar es Salaam RUFIJI LEGEND/KEY PROJECT DISTRICTS* NATIONAL CAPITAL INTERNATIONAL BOUNDARIES *Based on District boundaries of KILOMETERS One of the global project s objectives is to learn about and document the long-term health and welfare impacts of the country program s interventions. This technical ZAMBIA MALAWI Lake Malawi MASASI MOZAMBIQUE INDIAN OCEAN IBRD AUGUST For more information on Global Scaling Up Rural Sanitation and Handwashing, visit 1

12 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Introduction In spite of difficulties experienced during data collection, this technical note presents summary descriptive statistics for key demographic, socioeconomic, hygiene, health, and child development variables that were considered the simplest in survey application and compare well with information obtained from other surveys. The figures presented here have been positively validated against existing information from available household surveys, in particular the 2010 DHS. The general aspects presented provide guidance as to the general status of sanitation and hygiene practices in the rural areas surveyed and some of the districts where the programs took place, prior to implementation activities. Findings are occasionally cross-tabulated by administrative location (districts and corresponding regions) or sanitary conditions; however, no confident relation should be assumed between these variables based on a simple empirical association. 2 Scaling Up Rural Sanitation

13 II. Characteristics of Targeted Households Table 1 shows a brief summary of household basic socioeconomic characteristics. The main eligibility criteria used in the survey was that the household had at least one child under two years of age, in accordance to the target population of the program. We found that the average household comprised 4.9 individuals; among them there were, on average, 2.4 male individuals. Consistent with the household eligibility criteria of having at least one child under two years of age, on average there were 1.4 children younger than five years old per household; 68% of households interviewed had exactly one child under five. The mean age of the household members was 18.2 and 80% of household members have had some schooling. Approximately 85% of the 1,586 interviewed households have a male household head. The average age of the household head is 37.6 years old; approximately 85% have had some level of school attendance and 90.2% reported working TABLE 1: SUMMARY STATISTICS Mean Household (HH) size 4.9 Number of children under five yrs per HH 1.4 at the time of survey, eligible HH only Number of males per HH 2.4 HH Head: Male (%) 84.8% Age 37.6 Ever attended school (%) 84.8% Worked outside household in last % months, age > 14 (%) Monthly Labor Income (in TZSs) 114,769 All HH Members: Age 18.2 Ever attended school (%) 80.0% Worked outside household in last % months, age > 14 (%) outside the home. Finally, self-reported monthly income by the household head was 114,769 TZS. Although selfreported income is not always a reliable measure and has to be interpreted with caution, income of the head of household could be thought as a proxy for total household income. With an average household size of 4.9 members, equivalent per capita reported monthly income would be approximately 23,422 TZS. As a reference, reported mean per capita household monthly income in the 2007 Tanzania Mainland Household Budget Survey for rural areas was 28,418 TZS. The following tables provide further analysis of the sociodemographic and socio-economic characteristics of the household. Table 2 presents the distribution of basic household demographic variables: age of the household members TABLE 2: PERCENT DISTRIBUTION OF THE BASIC SOCIO- DEMOGRAPHIC CHARACTERISTICS Age (% Individuals): Total 100 Percentage of HHs with following # of children under 5: 1 child children or more children 4.1 Total

14 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Characteristics of Targeted Households The targeted population is young; on average, more than 50% of the household members are under 14 years old. This is consistent with figures from rural mainland Tanzania. and the total number of children under the age of five per household. The target population was very young; on average, more than 50% of the household members were under 14 years old. Figure 2 presents the sample population distribution by age group and gender, reflecting compliance with the enrollment criteria. Table 3 shows assets of households and other ownerships. In terms of assets, more than half of the households (54.9%) had a radio or a cassette or CD player. Items associated with more well-off households such as TVs or VCRs were scarce: 2.1% and 1.3% of the households reported having such items. On average, 39.5% of the households had one or more bicycles and 71.6% had mattresses. We also observed that the large majority of the households owned mosquito nets and bed frames, 79.2% and 80.1% respectively. 4 Most households possessed some type of farm tools (89.4%) and a little more than half of the households owned animals (52%). The typical household in the program areas of rural Tanzania is headed by a male and comprises five members living in a single detached dwelling with clay floor and mud or brick walls. The analysis of the household dwelling type and construction materials is summarized in Table 4. Most households (82.5%) allowed enumerators to enter the household to observe dwelling characteristics. Among the observable households, we found that the most common type of dwelling was a detached, independent FIGURE 2: DISTRIBUTION OF SAMPLED POPULATION BY AGE GROUP AND GENDER (POPULATION PYRAMID) Percentage of population in the sample by age groups Males (%) Females (%) 4 Tanzania has had a successful mosquito net program for the past 10 years; nets are free to all pregnant mothers and children under five years of age, so, it is likely that these household nets were not purchased. 4 Scaling Up Rural Sanitation

15 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Characteristics of Targeted Households TABLE 3: PERCENT DISTRIBUTION OF HOUSEHOLD ASSETS (% HHs) Bed frame 80.1 Mosquito net 79.2 Mattress 71.6 Radio, CD, cassette 54.9 Stove (gas or other) 40.4 Bicycle 39.5 Cell phone 21.6 Clothes Iron (charcoal) 10.3 Sewing machine 4.2 Other house/other buildings 2.2 TV 2.1 Motorbike 1.4 VCR, DVD 1.3 Machinery, equipment 0.9 Clothes Iron (electric) 0.8 Blender 0.8 Automobile or truck 0.3 HH owns farm equipment 89.4 HH owns other piece of land besides home 46.8 HH has animals 52.0 Number of different livestock owned by HH (average) 0.96 The average household uses kerosene for lighting, wood for cooking, and in about half of the cases owns a few animals and a bicycle. TABLE 4: DWELLING CHARACTERISTICS AND MATERIALS BY OBSERVATION HH allows enumerator to observe housing characteristics (% HHs) Type of Dwelling (% observable HHs): 82.5 Detached house 43.9 Buildings 24.8 Apartment 6.3 Multi-family house 5.5 Room in a larger dwelling 0.9 Improvised housing unit 0.7 NA/Missing * 17.9 TOTAL 100 Walling Materials (% observable HHs): Unbaked brick, adobe 34.5 Mud 30.7 Brick 24.9 Other (including tin sheeting, bamboo) 5.5 Wood, logs 2.8 Concrete 1.5 TOTAL 100 Roofing Materials (% observable HHs): Tin, zinc sheeting 47.9 Other (including canvas, brick, concrete, adobe) 39.5 Bamboo 5.7 Mud 5.2 Wood, logs 1.8 TOTAL 100 Flooring Materials (% observable HHs): Clay, earthen floor 78.8 Concrete 15.1 Other (incl. tile, linoleum, parquet) 6.0 TOTAL

16 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Characteristics of Targeted Households dwelling (43.9%). The most common construction material of walls was unbaked brick (34.5%); followed by mud and baked brick (30.7% and 24.9%, respectively). The use of other walling materials such as concrete or wood was rare. Tin or zinc sheeting was the most common type of roofing material (47.9%). In 78.8% of the observable dwellings, the floor was clay, or earthen floor, and only 15.1% of observable households had concrete floors. The survey also included information regarding the type of energy source used for lighting and cooking in the dwelling (see Table 5). For dwelling lighting source, the large majority of the households (86.1%) used kerosene, with wood being the second alternative (7.2%). Most of the households (91%) used wood as the primary cooking fuel, followed by charcoal (5.9%). TABLE 5: DWELLING ENERGY SOURCE Dwelling Lighting Source (% HHs): Kerosene 86.1 Wood 7.2 Electricity 2.3 Other (incl. peat, candles, battery, solar, charcoal) 2.1 No Lighting 1.6 Gas 0.6 Total Dwelling Cooking Fuel (% HHs): Wood 91.0 Charcoal 5.9 Kerosene 1.8 Other (incl. coal, peat) 1.1 No Fuel for Cooking 0.2 Total Scaling Up Rural Sanitation

17 III. Water Sources The pre-program survey also investigated the household water source and the treatment that household members apply to drinking water. Questions related to water source were disaggregated by season (rainy versus dry season); however, because almost every household had the same water source throughout the year, we present results for the rainy season only. Results are summarized in Tables 6 and 7. TABLE 6: WATER SOURCE JMP defined improved water source (% HHs) 49.7 Source of Water for Drinking Use (% HHs): Dug well unprotected 27.1 protected 7.7 Piped water into public tap, standpipe 24.2 yard, plot, dwelling 1.3 Surface water 19.2 Tube well, worewhole 10.1 Rainwater 5.1 Spring water unprotected 3.3 protected 1.4 Other sources, water Vendor 0.8 TOTAL 100 Source location (% HHs): Elsewhere 55.1 In own yard, plot 31.5 In own dwelling 13.4 TOTAL 100 Covered source (% HHs): Open 64.7 Covered 30.6 Other answer 4.8 TOTAL 100 Who mainly collects water in HH (% HHs): Adult woman 94.2 Adult man 3.2 Child < 15 years 1.6 Other members 1.1 TOTAL 100 TABLE 7: DRINKING WATER Stores drinking water in home (% HHs) 86.0 Frequency of washing drinking water storage (% HHs that indicated storing drinking water at home): More than once per week 59.8 Once per week 19.2 Rarely 18.8 Never wash 2.3 TOTAL Cleansing agent used to wash drinking water storage (% HHs that indicated washing water storage): Soap, detergent, bleach 53.4 Water only 33.5 Other 10.0 Mud 2.5 Ash 0.7 TOTAL Prepares water before drinking (% HHs): No 68.4 Yes 24.4 Sometimes 7.2 Method of preparing water (% HHs that indicated preparing water before drinking)*: Boil 43.2 Stand and settle water before drinking 32.3 Strain through a cloth 22.9 Filter 7.3 Chlorine 5.7 Iodine 0.8 *For HH which answered both Yes and Sometimes. Multiple responses are permitted. 7

18 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Water Sources On average, only about half of households (49.7%) have access to an improved water source, which is generally not in their dwellings or on their plots. In rural Tanzania, collecting water is a woman s task (94.2% of households). As Table 6 shows, almost half of the households (49.7%) reported having access to an improved water source. When taking a narrower look at the data, we find three main sources of drinking water in the surveyed households: piped water from public tap (24.2%), unprotected dug well (27.1%), and surface water (19.2%). On average, 31.5% of these water sources were located in the household s own yard or plot, 13.4% within the dwelling, and the rest (55.1%) are located elsewhere, suggesting most of households could be investing some significant amount of time collecting water. Regarding features of the water source, 30.6% of the households used a water source that was covered, while 64.7% reported using uncovered water sources. In the vast majority of the households (94.2%), an adult female was in charge of collecting water from the source. The task was performed by an adult male in only 3.2% of the households and by a child under 15 years old in 1.6% of the households. Finally, we explored practices related to drinking water consumption (Table 7). The majority of households stored drinking water at home, but only 24.4% reported treating the water before drinking it. Of these, less than half (43%) mentioned boiling the water as one of the treatments. This suggests that water safety practices could be considerably improved in these districts in order to prevent waterborne diseases. 8 Scaling Up Rural Sanitation

19 IV. Handwashing Behavior Handwashing with soap at critical times (such as after contact with feces and before handling food) has been shown to substantially reduce the risk of diarrhea. Children under five years of age represent the age group most susceptible to diarrheal disease and acute lower respiratory infections (ALRIs), two major causes of childhood morbidity and mortality in less developed countries. These infections, usually transferred from dirty hands to food or water sources or to the mouth directly through hand-to-mouth contact, can be prevented if mothers/caregivers wash their hands with soap at critical times such as before feeding a child, cooking or eating, and after using a toilet or having contact with a child s feces. In this analysis, these are considered as critical junctures. Handwashing behavior is difficult to assess. In this study, we relied on two sources: self-reported handwashing at critical times and, as a proxy measure, spot-check observations of whether the household had a designated place for handwashing with both soap and water available. An additional measure assessed the cleanliness of the caretaker s hands through direct observation, which is another proxy indicator of handwashing with soap behavior. In the rural Tanzanian areas surveyed, a majority of the caregivers more than 78% reported having used soap to wash hands at least once in the previous 24 hours. However, when asked under what circumstances they last used soap to wash hands, only 33.7% mentioned at least one of the critical times (see Table 8). Most caregivers reported using soap to wash their hands at least once in the previous 24 hours, but when asked the circumstance in which they last washed, only 33.7% mentioned a critical time and only 21.2% mentioned after using the toilet. TABLE 8: CAREGIVERS SELF-REPORTED HANDWASHING BEHAVIOR WITH SOAP BY DISTRICT (% PREVIOUS 24 HOURS) DISTRICT (REGION) Iringa (Iringa) Masasi (Mtwara) Mpwapwa (Dodoma) Musoma (Mara) Rufiji (Pwani) TOTAL Washed hands with soap at least once during the following events During at least one critical juncture Using the toilet Cooking or preparing food Feeding children Cleaning children s bottom Washed hands with soap at least once in previous 24 hours

20 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Handwashing Behavior Only 21.2% mentioned using the toilet among the occasions in which they had used soap to wash hands since the previous day, and only 9.1% mentioned cleaning children s bottom. In relation to the other critical times, 12.5% of caregivers reported that they had washed hands with soap at times of cooking or food preparation, and 9.7% of caregivers mentioned when feeding a child. On average, self-reported handwashing with soap was relatively higher among wealthier than poorer households for most critical junctures. Table 8 also shows disaggregation by districts and regions. Among all districts, there were slight differences, showing Mara with the lowest rates of handwashing and Dodoma with the highest. However, when inspecting wealth and regional breakdowns there were no clear patterns; population in these rural districts seemed fairly homogenous. 10 Scaling Up Rural Sanitation

21 V. Sanitation Although sanitation coverage in Tanzania is relatively high since a latrinisation program was introduced by the government in the early 1970s, 5 most of rural Tanzania still relies on basic latrines, which do not separate humans from excreta and are described as fixed-point open defecation. Over the years, many of those basic latrines lacked adequate maintenance or fell into disrepair. In this section, we investigate the most common sanitation facilities available in the households surveyed, and some of their features. Table 9 presents the different types of sanitation facilities reported. Unfortunately, there were ex-post reports of problems with the interpretation that enumerators gave to the type of facilities, in particular the flush to pit latrine classification, which apparently was understood as any latrine with water present by the side. However, the presence of water likely was used as a method of anal cleansing, rather than for flushing. Taking this caveat into consideration, we could no longer use the original classification and thus grouped pit latrines of any type. Pit latrines thus reached 89% predominance from adding the 35.6% of households that reported having a pit latrine without slab or open pit, the 39.6% that The most common type of toilet facility was a pit latrine, present in almost 89% of households. Only 12.7% of households reported having a pit latrine with slab. Open defecation was reported in 9.2% of households. TABLE 9: HOUSEHOLD MAIN SANITATION FACILITY CHARACTERISTICS (% HHs) HH Main Sanitation Facility: Pit latrine 88.9 with flush (water present) 39.6 without slab, open pit 35.6 with slab 12.7 ventilated improved 1.1 Open defecation 9.2 Flush other (piped sewer system, septic tank or elsewhere) 0.9 Composting toilet 0.4 Other (e.g. bucket, hanging toilet) 0.4 Total 100 Location of Main Sanitation Facility: In household yard 81.1 Less than 10 min walk from the household 10.9 Other 5.4 Household 1.5 Inside household 1.0 TOTAL World Bank,

22 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Sanitation Most facilities were located in the household s yard (81.1%). There was no apparent gender bias in the use of facilities; men and women reportedly shared them in 96.7% of households interviewed. reported having a latrine with what can be understood as flush-water present by the side, the 1.1% that reported having a ventilated improved pit latrine, and the mere 12.7% of households that reported having a pit latrine with slab. These numbers seem consistent with the 2010 DHS survey for rural mainland, which found that 71.4% of households used a pit latrine without slab or open pit, while 6.3% of households used a pit latrine with slab. Open defecation was reported in 9.2% of the households, while for rural mainland, the DHS reported 17.8% of open defecation (no facility/bush/field). Regarding location, most of the sanitation facilities were located in the household yard (81.1%) and another 10.9% were located within a 10-minute walk from the dwelling. We present findings about household main sanitation facility safety and sharing in Table 10. A large majority of the sanitation facilities were not shared with other households in the neighborhood (86.2%) and were reported as private (89.9%). Of those reportedly sharing (13.8%), in most cases they shared with two or three TABLE 10: HOUSEHOLD MAIN SANITATION FACILITY SHARING AND SAFETY (% HHs) Sanitation facility in HH is not exclusive for men or women 96.7 private 89.9 not shared with other HHs in the neighborhood 86.2 shared with other HHs: other HH (% of shared HHs) other HHs (% of shared HHs) other HHs (% of shared HHs) other HHs (% of shared HHs) 30.4 Average number of people sharing the toilet 5.94 Women in HH feel that sanitation facility is safe to use during day time 87.8 is safe to use during night 82.0 provides enough privacy during defecation 75.5 HH most knowledgeable member reports children < 5 not feeling safe using sanitation facility 51.3 With the main sanitation facility, the HH is somewhat satisfied 29.6 very satisfied 26.7 less than satisfied 23.5 completely dissatisfied Scaling Up Rural Sanitation

23 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Sanitation other households. In addition, there was no apparent gender bias in the use of facilities; men and women reportedly shared them in 96.7% of households interviewed. The DHS for rural mainland found 80.2% of households not sharing facilities and 12.6% sharing with another household. Notably, 51.3% of respondents (most knowledgeable household member) reported not feeling that facilities were safe for children under 5 years old. In addition, we explored feelings of safety and privacy for female household members, and they generally responded positively: 82% felt it was safe to use the facility at night and more than 87% felt safe using it during the day, while more than 75% reportedly felt they had privacy during defecation. Notably, 51.3% of respondents (the most knowledgeable household member) reported that they did not feel the facilities were safe for children under five years of age. Finally, when asked about satisfaction with the sanitary facility, 19.6% of households declared complete dissatisfaction. Table 11 reports other aspects of household sanitary conditions and practices. In 36.2% of households, flies were reported as always present near the sanitation TABLE 11: OTHER CHARACTERISTICS OF HOUSEHOLDS SANITARY CONDITION (% HHs) Flies at/near place for defecation always 36.2 sometimes 51.7 rarely 11.9 Visible Feces in/around HH: None to 5 feces to 10 feces 4.1 More than 10 feces 2.9 TOTAL Disposal of Child Feces: Single response 96.2 Toilet, latrine (% of single response) 79.4 Open disposal incl. bushes, ground, drain (% of single response) 19.5 Garbage (% of single response) 1.0 Other (% of single response) 0.1 TOTAL Multiple reponses* 3.8 *Multiple entries are allowed for this question. 13

24 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Sanitation facility. Flies can act as vectors of transmission, and some studies have associated them with certain diarrhea-causing pathogens. When handling children s feces, most households relied on disposal to the latrine or sanitation facility (almost 80%); however, a considerable proportion reported that they were disposed of in the open (19.5%). Finally, we asked about the presence of visible feces, as they are the most common source of diarrhea pathogens in the environment. In most cases in this survey, no visible animal or human feces were observed inside or around the household (72.8%). However, in 20.2% of the households, one to five pieces of feces were visible. Unsafe disposal of children s feces has increasingly gained attention as a factor associated with diarrhea risk. In this survey, most households reported a consistent practice surrounding child feces disposal (over 96% chose a unique method of disposal). Among these, most households relied on disposal in the latrine or sanitation facility (almost 80%), but a considerable proportion reported that child feces were disposed of in the open (19.5%). 14 Scaling Up Rural Sanitation

25 VI. Child Care Environment It is largely recognized that characteristics of the caregiver and the quality of care a child receives have huge impacts on young children s health, nutritional status, and development beyond variation due to socio-economic and education variables. We collected information on feeding practices, caregiving behavior, and caregiver well-being in order to more carefully tease out the potential effects of the behavior change interventions on child health and development. Table 12 summarizes breastfeeding habits within the interviewed households. The average breastfeeding duration was 15 months, and 93% of the children received colostrum 6 during the first three days after childbirth. Only 15.7% of the mothers fed their babies liquids other than colostrum or breast milk during the first three days of life. Of those children who were fed other liquid, more than two-thirds were provided with plain water. TABLE 12: CHILD BREASTFEEDING (CHILDREN < 2) Child ever breastfed at the time of survey (% children < 2) 98.4 Still breastfeeding (% children ever breastfed) 87.9 Average months of breastfeeding (for children ever breastfed) 15.0 During the first three days after birth,. colostrum was fed (% children < 2) 93.0 liquid other than colostrum or breast milk was fed (% children < 2) 15.7 Single Response 87.0 Plain water 60.6 Sugar-salt solution 10.8 Milk other than breast 7.5 Other liquids 7.4 Sugar, glucose water 5.0 Tea or gripe water 4.6 Instant formula 4.1 Multiple Responses Colostrum is produced prior to mature breast milk during pregnancy and through the first 3 6 days of life. It contains not only necessary nutrients but also properties that help protect the baby from viral and bacterial infections. 15

26 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Child Care Environment The survey also included a section on children s diet, summarized in Table 13. 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. Almost 90% of the children were given plain water since the previous day; breast milk consumption (89%) was very high as expected for children under the age of two. Homemade gruel was also consumed by a majority of children (72.1%). Interestingly, one-fifth of the children were given tea or coffee (20.3%). With respect to TABLE 13: INFANT/YOUNG CHILD FEEDING (CHILDREN < 2) Liquids given the day before survey taken (% children): Plain water 89.5 Breast milk 89.0 Homemade porridge or gruel 72.1 Caffeine beverages 20.3 Fortified child food 8.1 Fruit juice 6.1 Other milks or instant formula 3.2 % of children that were given solid or semi-solid food the day before survey taken 70.3 Food given (% children who had solid or semisolid food the day before): Cereal, rice, other grain-based food 94.7 Other fruits, vegetables 89.9 Beans, peas, lentils 86.2 Fruits or vegetables rich in vitamin A (carrots, yams, mango, papaya, green leaf vegetables) 82.4 Meat red, white 80.3 Potatoes, yucca, other roots 73.9 Oil, fats, butter 48.9 Average number of times solid or semi-solid food was given the day before survey taken meal meals meals meals 8.4 Feed themselves with utensils or hands (% children who had solid or semi-solid food the day before) 68.6 Received vitamin A supplement in the last 6 months (% children < 2) 56.4 Given iron pills or syrup in the last seven days before survey taken (% children < 2) Scaling Up Rural Sanitation

27 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Child Care Environment food given the day before, 70.3% of the children received solid or semi-solid food 2.41 times on average. When asked about dietary supplements, 25.1% of caregivers declared giving iron pills or syrup to their children in the seven days prior to the survey and 56.4% mentioned having provided vitamin A supplement to their children at least once in the previous six months. A large number of caregivers (68.6%) also mentioned that children who had been given solid or semi-solid food the day before had also fed themselves. Most children under the age of two are breast-fed (89%) and play with household objects (85.9%). In more than 50% of cases, adults sing songs with their children. Caregivers reported rarely reading books (13.7%) or telling stories to the child (14.5%) in the three days prior to the survey. Furthermore, there were specific questions related to household support for learning and development, including the frequency with which adults engaged children in various activities recognized as promoting language and cognitive development. Table 14 shows that 85.9% of the children under the age of two played with household objects, and in more than 50% of cases, adults had sung songs with their children. Only 2.4% of the children had attended an early education program. Although most respondents reportedly took their children outside the home in the three days prior to the survey (71.5%), a low number of caregivers reported reading books (13.7%) or telling stories to their children (14.5%) in the same period. TABLE 14: INFANT/YOUNG CHILDREN LEARNING ENVIRONMENT (CHILDREN < 2) Child plays with household objects (% children) 85.9 Adult take child outside home (% caregivers associated with children < 2) 71.5 Adult sings songs with child (% caregivers associated with children < 2) 50.1 Adult spend time naming, counting, drawing with child (% caregivers associated with children < 2) 23.8 Adult tells stories to child (% caregivers associated with children < 2) 14.5 Adult reads books with child (% caregivers associated with children < 2) 13.7 Child attended early education programs (% children)

28 VII. Child Health and Anthropometric Measures The survey also intended to provide information on anemia prevalence, children s diarrhea prevalence, ALRI, and other health symptoms; child development; and child growth in order to generate evidence on the association with sanitation and hygiene interventions. However, validation of the symptom-related data and nutritional indexes against information from other surveys suggested measurement problems during data collection. In light of this, results are not reported in detail. Presence of anemia was measured by hemoglobin concentration in children between six months and two years of age (<110g/L per international standards), and findings are summarized in Table 15. The majority (79.3%) of samples taken indicated the presence of anemia. Estimates from the 2010 DHS reported 58.1% prevalence of anemia in rural mainland and between 68.7% and 81.1% for children between six months and two years of age, which seemed consistent with the levels reported in our sample. There were no major differences for households with improved water source (78.4%) or unimproved water source (80.5%). Regarding different levels by districts, the percentage of anemia prevalence is highest among children living in Pwani (88.9%) and lowest among children living in Dodoma (71.6%). TABLE 15: ANEMIA PREVALENCE (% OF CHILDREN < 2) (Hb < 110 g/l) District (Region): Rufiji (Pwani) 88.9 Masasi (Mtwara) 83.7 Iringa (Iringa) 80.3 Musoma (Mara) 72.8 Mpwapwa (Dodoma) 71.6 Sanitary Conditions: Improved sanitation 82.1 Unimproved sanitation 76.0 Improved water source 78.4 Unimproved water source 80.5 Place for handwashing w/ soap and water 81.3 No place for handwashing w/ soap and water 79.2 Overall 79.3 Diarrhea was defined as the reported presence of three or more loose or watery stools over a 24-hour period or one or more stools with blood and/or mucus present, using symptom data reported from caregivers. Acute lower respiratory infection was defined using the clinical case definition of the World Health Organization (WHO 2005), which diagnoses a child as having an ALRI when he/she presents constant cough or difficulty breathing and raised respiratory rate symptoms, as reported by the caregiver. Whereas estimates from the 2010 DHS reported prevalence of diarrhea among children under five at 14.6% in mainland, 13.6% in mainland rural areas, and approximately 20% for children under 24 months old, the findings in relation to caregiver-reported diarrhea for our sample of children under two was below 2.0% (between 19 and 30 cases reported out of a total of 2,164 children under two years old). Similarly, caregiver-reported two-day and seven-day ALRI prevalence was positive only in eight and nine cases in the sample, respectively (approximately 1,500 households), while the 2010 DHS survey reported 4.7% prevalence of ARI for children under 24 months old. 7 The inconsistency suggests serious under-reporting in the sample and doubts about reliability of data symptom collected. In light of this, results are not presented in detail here. The survey also intended to include baseline anthropometric measures of children under the age of two, such as weight, height, and arm and head circumference. This information is important in order to assess the average growth and development of the children. To analyze the child growth findings, anthropometric Z-scores were assigned by comparing children in the sample to the WHO reference population median and standard deviation (SD) for each of the aforementioned variables. 8 The reference population is designed to be internationally applicable regardless of ethnicity, socioeconomic status, or feeding practices. The Z-score for each measure indicates the number of SD units from the median 7 In the DHS, the prevalence of symptoms of ARI was estimated by asking mothers whether their children under age 5 had been ill with a cough accompanied by short, rapid breathing in the two weeks preceding the survey. (DHS, 2010) 8 World Health Organization, 2005 and 2006.

29 Scaling Up Handwashing and Rural Sanitation: Findings from a Baseline Survey in Tanzania Child Health and Anthropometric Measures of the reference population. The WHO guidelines for child growth use a Z-score cutoff of less than 2 SDs below the median of the reference population for low growth levels (stunted, wasted, and underweight) and less than 3 SDs from the median, indicating severity of the condition (severely stunted, wasted, and underweight). In addition, children with a weightfor-height index of more than 2 SDs above the median reference population are considered obese. Table 16 presents information on children s nutritional status based on the indexes. Unfortunately, these suggest low quality of measurements during data collection as well. Validation of these figures against comparable datasets posed doubts about their reliability (Table 17). In our sample, an average of 58.3% of children were stunted and 35.1% were severely stunted (height-for-age Z-score); 2.2% were wasted or malnourished and 0.4% were severely wasted (weight-forheight Z-score); and 10.6% were underweight and 2.7% were severely underweight (weight-for-age Z-score). For example, the mean height-for-age Z-score was 3.56, a relatively large number indicating generalized stunting; this figure was higher than the rural mean Z-score of 1.8 reported in the 2010 DHS and other comparable datasets. When trimming down the height-for-age to levels feasible by WHO standards (between 6 to +6), near 20% of the sample was lost. After trimming, the mean height-for-age Z-score became 2.38, which could be plausible. However, high variability of height-for-age Z-scores (not shown here) relative to other countries also suggested poor quality measurements. Finally, the finding of 9.1% of children reportedly overweight was unlikely in comparison with the 0.6% reported in the DHS for rural mainland. Similarly, the weight-for-height mean Z-score was positive, which was inconsistent with the other indexes and similar datasets. In light of these apparent inconsistencies and the problems experienced during data collection, we abstain from reporting more detailed figures here. TABLE 16: NUTRITIONAL STATUS OF CHILDREN < 2 Height-for-age Weight-for-height Weight-for-age Background characteristics DHS 2010, rural mainland * Percentage below 3 SD Percentage below 2 SD Mean Z-score (SD) Percentage below 3 SD Percentage below 2 SD Percentage Mean Z over +2 SD score (SD) Percentage below 3 SD Percentage below 2 SD Percentage Mean Z over +2 SD score (SD) WSP, 5 districts *Source: 2010 Tanzania Demographic and Health Survey (DHS) TABLE 17: ESTIMATES OF Z-SCORES (IN SD) FROM SELECTED DATASETS Height-for-age Weight-for-height Weight-for-age Dataset mean Z-score (SD) mean Z-score (SD) mean Z-score (SD) WSP Tanzania, five districts Tanzania DHS, 2010 (rural mainland) WSP Senegal WSP Himachal Pradesh, India WSP Madhya Pradesh, India Rural Water Project Kenya 1.85 na na Tamil Nadu, India 1.98 na na WSP Peru WSP East Java, Indonesia