UNDERSTANDING HAND WASHING BEHAVIOR RESULTS OF FORMATIVE RESEARCH ON HAND WASHING IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA SUMMARY OF FINDINGS

INTERNATIONAL RESCUE COMMITTEE / BUREAU OF POPULATION, REFUGEES, AND MIGRATION

DECEMBER 2011

EXECUTIVE SUMMARY

OBJECTIVES OF THE STUDY

Diarrhea is one of the leading killers, globally and in refugee camps, and strong evidence indicates that hand washing with soap is one of the most effective interventions available to reduce the incidence of diarrhea. Yet little is known about why people in refugee camps do or do not wash their hands, and what approaches will increase hand washing.

Diarrheal diseases kill close to two million children each year; tens of thousands of these deaths occur in refugee camps. Hand washing with soap has been shown to reduce the risk of diarrhea by up to about a half1. There is less documentation, however, on how governments and aid agencies can increase hand washing with soap.

To address this gap, the International Rescue Committee (IRC), in collaboration with the London School of Hygiene and Tropical Medicine (LSHTM), conducted formative research in three refugee camps in Thailand, Kenya and Ethiopia, with funding from the Bureau for Population, Migration, and Refugees (PRM). The study aimed to understand hand washing behavior in long term refugee camp populations, and to explore how formative research could be used to improve hygiene promotion programs.

The IRC and LSHTM team conducted formative research in three refugee camps in Thailand, Kenya and Ethiopia, with two objectives: 1. To understand the motivations, barriers and enabling factors for hand washing in long term refugee camp populations. 2. To produce a case study of how formative research on hand washing behavior could be used to develop communication plans to increase hand washing in long term refugee settings.

The research team used five principal data collection methods: structured observation, household surveys, semi-structured interviews, focus group discussions, and trials of improved behaviors. A major finding of the study was the lack of correlation between reported knowledge and actual behavior; and the value of both quantitative and qualitative data in informing hygiene promotion efforts. The following recommendations are based on the study’s findings: 1. Soap distribution in refugee camps should meet Sphere Minimum Standards, including additional soap designated for laundry;

SETTING The study took place from January to May 2011 in three long-term camps. Demographic and other background information on the camps is summarized in Table 1. TABLE 1 - INFORMATION ON THE THREE REFUGEE CAMPS IN THE STUDY Site

Ban Mai Nai Soi, Thailand

Kakuma, Kenya

Shimelba, Ethiopia

Year established

1989

1992

2001

3. Hand washing promotion programs in refugee camps should experiment with and document alternative approaches;

Countries of origin of refugee populations

Burma

South Sudan, Ethiopia, DR Eritrea Congo and Somalia

4. Formative research in camps for refugee populations should investigate the various information sources in a camp;

Ethnic composition

Karenni

Sudanese, Somali

Tigrinya, Kunama

Total population

15,620

77,358

9,442

Relevant services provided by the IRC

Water, Sanitation, Hygiene Promotion

Sanitation, Hygiene Promotion

Water, Sanitation, Hygiene Promotion

2. Hygiene programs in refugee camp programs should review the equipment commonly used to facilitate hand washing in the household;

5. When reliable data on actual hand washing rates are needed, researchers should use structured observation rather than other methods, such as self-report; 6. Behavior trials and properly prepared focus group discussions should be included in all formative research efforts in refugee camps. The report also includes findings specific for each of the three refugee camps.

METHODS The research team used five principal data collection methods: Quantitative methods: …
Structured observation …
Household questionnaire-based surveys Qualitative methods: …
Semi-structured interviews …
Focus group discussions …
Trials of improved behaviors Key Information about these method is summarized in Table 2.

2

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

3

TABLE 2 - RESEARCH METHODS METHOD

Information Sought

Observed hand washing practices linked to critical times for hand washing Structured observation

Household survey

1. 2. 3. 4. 5.

After defecation Before eating Before preparing food Before feeding a child After handling a child’s feces

Hand washing/hygiene knowledge, self-reported hygiene practices, access to and placement of soap, water storage containers and latrines, household water quality, and engagement with communication channels and social activities

Hygiene practices, access to soap, access to health services and health care information Semi-structured interviews

Hygiene promotion activities and communication channels

Sample Size and Sampling Method

Participants/ Respondents

40 Households

Households randomly selected using a random walk procedure

Behavior trials

4

All members of selected households

Sample sizes (calculated to provide 5% precision or better for all variables, and to permit testing of relationships between variables) Ban Mai Nai Soi

Female caretakers (mothers) of < 5 yrs children where present otherwise, in order of preference: male caretaker, female > 15 yrs, male > 15 yrs

Kakuma

382

Shimelba

360

Selected members of refugee population

New mothers Purposive sampling, 11 interviews per site, chosen from different geographical areas and ethnic groups within the camp

Mothers and fathers of children < 5 yrs

Research project (IRC & LSHTM) staff using translators

2 groups of new mothers per site

Disposal practices for children’s feces

Purposive sampling, 6-8 women per group

2 groups of mothers of children < 5 yrs per site

Hygiene promotion activities being implemented and motivation levels of staff towards different activities

Purposive sampling, 6-8 women per group

2 groups of hygiene promoters per site

Purposive sampling,

Mothers of children < 5 yrs

8 Households per site

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

For behavior trials, members of participating households attempted to wash their hands at all critical times and reported back about the experience, which provided an insight into the barriers and enabling factors involved. At the start of the trial, researchers gave household heads a bar of soap and discussed with them what they should do to ensure hand washing at critical times by all family members, such as where to place soap and water and how to supervise children to ensure proper hand washing with soap. After two days researchers interviewed household heads again to assess how the trial was going and to observe actions taken by the household to facilitate hand washing. The trial ended after 5 days, and the researchers interviewed the household heads a final time to explore their overall experiences.

RECOMMENDATIONS AND SUPPORTING FINDINGS IMPACT OF SOAP AND WATER AVAILABILITY The barriers, enabling factors and motivations for hand washing in long-term refugee populations were similar to populations in non-camp rural settings in developing countries. Barriers included prioritization of soap use for laundry over hand washing; enabling factors included having a convenient hand washing station; motivations included disgust, properly nurturing children, and the comfort of having clean hands.

Key informants Purposive sampling, 6-8 women per group

Structured observation was used to measure the prevalence of hand washing at critical times in the three refugee camps. Enumerators arrived at participating households in the early morning, a time when more events of interest were likely to occur. Positioning themselves at a suitable vantage point within the compound (and moving as necessary), they recorded entries in their data collection sheets every time a critical time for hand washing (an ‘event’) occurred for a household member. Each entry included the gender/age category of the person concerned, the event, whether they washed hands and, if so, whether they used soap. For focus group discussions, groups of mothers were shown pictorial cards representing potential motivations for hand washing (disgust, physical comfort, religious purity, properly nurturing their children, attractiveness, social conformity and social respect) and were asked which motivations were relevant and important in their families and community.

347

Motivations for hand washing

Perceived barriers and benefits associated with regular hand washing with soap

Selected members of refugee population

Systematic random sampling (probability proportional to size)

General information about life in the camps

Focus group discussions

Enumerators

The team developed tools for each of these methods prior to deployment to the first camp, Shimelba, where they field-tested and finalized the tools. The tools included the household survey questionnaire, structured observation recording sheets, semi-structured interview checklists, focus group discussion guides, behavior trial guide and consent forms for the various data collection methods. They later field tested and adapted the tools for each of the two remaining camps, but avoided changes that would prevent comparison and aggregation of the results across the three camps.

Research project (IRC & LSHTM) staff using translators

Research project (IRC & LSHTM) staff using translators

Water availability appears to be higher in refugee camps than in many rural communities in developing countries. The household survey found water availability in the three refugee camps to be 12.5 liters per person per day (l/p/d) in Ban Mai Nai Soi, 16.0 l/p/d in Kakuma, and 19.5 l/p/d in Shimelba. The figure for Ban Mai Nai Soi is likely an underestimate as refugees were frequently observed using water at the water points for bathing and laundry but this water was not included in the household consumption calculations. These figures are below the UNHCR target of 20 l/p/d, but the quantities of water available appeared to be sufficient for drinking, cooking and personal hygiene in all three camps. Several elements, including the absence of queues for water or complaints about water availability, suggest that water availability is not a constraint in the three camps studied.

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

5

Soap availability is higher than in many non-camp rural settings, but is lower in Kakuma and Shimelba, because of irregular distribution, and use of soap for other purposes. Almost all households in Ban Mai Nai Soi cited free distributions as their primary source of soap, compared to around 60% of households in Shimelba and less that 20% in Kakuma. Figure 1 shows the sources of soap for households in each camp. Soap availability in the household was higher in camps where free distribution was the main source of soap. Regardless of the primary source of soap, roughly two-thirds of the households in the three camps cited laundry as their priority for soap use, with most of the remainder prioritizing bathing. Mothers liked to separate soap for different purposes, notably separating laundry soap from bathing soap. Somali mothers designated separate bars of soap for hand washing after using the toilet, and for hand washing at other times. Recommendation 1: Soap distribution in refugee camps should meet Sphere Minimum Standards, including additional soap designated for laundry. Hand washing rates are low despite the availability of soap and water. The proportion of defecation events followed by hand washing with soap was 20% across all camps, close to the 17% reported from studies in 12 different countries where no intensive hygiene promotion had taken place². Rates in non-camp settings ranged from 3% to 42%. The study did not find a significant difference in the prevalence of hand washing with soap despite the fact that 96% of households had soap in the Ban Mai Nai Soi, compared with 84% in Shimelba and only 35% in Kakuma. Hence, whilst a necessary enabler, soap availability alone has little impact on hand washing rates. FIGURE 1 – SOAP AVAILABILITY AND SOURCE OF HOUSEHOLD SOAP IN THREE REFUGEE CAMPS

Households which had soap and water at designated hand washing sites encountered other limitations which discouraged regular hand washing. Mothers in all three camps spoke of the importance of soap and water at a prominent place in the compound as a visual cue for hand washing and for making hand washing convenient. They expressed concern, however, about children playing with, and losing or wasting, soap if it was left accessible. This finding is consistent with previous research showing mothers’ willingness to sacrifice convenience to avoid children playing with and wasting soap3. Behavior trials in this study found that hand washing with soap involved a rather complex and time-consuming sequence of actions to bring together and arrange the soap, water and vessels for pouring and collecting water, making hand washing with soap inconvenient. Research has shown that the presence of a household washstand or water dispenser is associated with increased hand washing rates (Biran, Tabyshalieva and Salmorbekova, 20054) and reduced fingertip contamination (Pinfold, 19905).

I put the soap in front so everybody can see it and remember [to wash their hands] - Mother, Shimelba behavior trial Recommendation 2: Hygiene programs in refugee camp programs should review the equipment commonly used to facilitate hand washing in the household. This assessment should include: getting an estimate of the prevalence of different methods, and assessing their effectiveness in making hand washing convenient and efficient, and in preventing loss or wastage of soap. If existing equipment is not effective, humanitarian actors should support the development of improved hand washing stations, working initially with interested ‘early adopters’, to develop and test prototypes that can later be promoted in the wider population.

100

Proportion of household surveys (%)

FIGURE 2 –HAND WASHING STATIONS IN HOUSEHOLDS IN SHIMELBA 80 Bought outside camp

Free distribution X

40

Soap availability

20

0

6

Bought in camp

60

Ban Mai Nai Soi

Shimelba

Kakuma

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

7

Increased knowledge of the five critical times for hand washing does not correlate with increased hand washing with soap at these times. In all three camps, the prevalence of knowledge about the critical times was much higher than the prevalence of observed hand washing. As figure 3 illustrates, there was no correlation between knowledge of a critical time for hand washing, and observed hand washing at that time. On the other hand, water consumption increased by 2.2 liters for every additional critical time a household’s respondent to the questionnaire survey could name. More investigation is needed about the relationship between knowledge and practice, ideally through a study that measures both hand washing behavior (by structured observation) and knowledge of critical times in households of mothers with young children. A larger sample size would be required than the 40 households per site used for structured observation in the present study.

& of critical times where hand washing with soap observed

FIGURE 3 – KNOWLEDGE VS. OBSERVED BEHAVIOR FOR CRITICAL TIMES FOR HAND WASHING. THERE WAS LITTLE CORRELATION BETWEEN KNOWLEDGE AND BEHAVIOR. 35

Critical time key (colors) After defecation

30

Before eating

The research found evidence to suggest that emotional and sensory levers, such as disgust and the desire to properly nurture children, can be effective motivators for hand washing with soap, particularly after fecal contact events. Hand washing was observed more frequently after fecal contact events as opposed to other events. ‘After defecation’ and ‘after cleaning a child’s bottom’ – the critical times linked to ‘disgusting’ events - were among the most commonly cited critical times for hand washing in the questionnaire-based household survey in the three camps. During focus group discussions, removing the smell from hands after fecal contact was one of the most commonly cited motivations for hand washing with soap. Hence disgust at the smell of feces might be an appropriate emotional lever to incorporate into hand washing promotion.

When you go to the latrine and you wash your hands with water only the smell remains...but when you wash your hands with soap there is no smell. - Dinka mother, Kakuma behavior trial

Before cooking

25

Before feeding a child

20

After cleaning child’s bottom

15

Camp key (shapes)

10

Ban Mai Nai Soi Kakuma

5

Shimelba 0

households in all camps were not reached by hygiene promotion (approximately half in Kakuma and onethird in the other two camps).

0

20

40

60

80

100

% respondents knowing critical time Pearson correlation coefficient for combined data points was -0.24 (R² = 0.0615), which indicates no association in this study between knowledge of when one should wash hands, and actual practice.

Recommendation 3: Hand washing promotion programs in refugee camps should experiment with and document alternative approaches. Programs should measure the effectiveness in changing the behavior of those reached and the efficiency (resources required or cost per person reached) of programs using these approaches. Hygiene promotion teams should consider including at least one emotional or sensory-based motivation for hand washing with soap. Focus group discussions can be used to identify the most appropriate motivation to use. The majority of the hygiene promoters in all camps had little or no prior experience with hygiene promotion or social work and were selected largely for their literacy and/or education level. For hygiene promotion programs in refugee camps to employ more diverse methods, especially more innovative methods like the use of emotional and sensory levers, a greater investment needs to be made in training, mentoring and monitoring hygiene promotion staff. Future selection procedures for new hygiene promoters will need to assess their communication skills better and training will need to focus on improving and measuring facilitation skills.

HYGIENE PROMOTION IN REFUGEE CAMPS Raising awareness and sharing messages are unlikely, on their own, to lead to changes in behavior. Hygiene promotion in the camps was largely based on educational messages about diarrhea transmission routes and diarrhea-preventing behaviors, using visual aids such as flash cards. Whilst the number of hygiene promoters was below that of the Sphere guideline for two of the three camps, the research did not find conclusive evidence that simply increasing the number of hygiene promoters or intensity of their activities would lead to increased hand washing. A higher frequency of visits to households by IRC Hygiene Promoters was associated with a higher likelihood of soap being present at the hand washing site, but did not affect feces disposal practices or knowledge of critical times for hand washing. Ban Mai Nai Soi was the only camp with a ratio of hygiene promoters per 1000 population above the Sphere 2011 guideline value of two, yet knowledge of the critical times for hand washing there was lower than in the other two camps. Of note, a significant percentage of the

8

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

UNDERSTANDING HAND WASHING BEHAVIOR IN REFUGEE CAMP POPULATIONS IN THAILAND, KENYA AND ETHIOPIA

9

IMPACT OF EDUCATION AND ORIGIN ON BEHAVIOR A higher level of education in household survey respondents was associated with better hygiene practices. Open defecation was lower in more educated households in Kakuma and Shimelba camps. Water and soap were more likely to be found at designated hand washing sites in educated households, which were also more likely to buy soap. Educated respondents were more likely to report safe methods of disposing of children’s feces. Education made no difference to the knowledge of critical times for hand washing. Refugees of urban origin generally had better defecation practices and were more likely to purchase soap. Whilst the vast majority of refugees in Ban Mai Nai Soi come from rural areas, around half the refugees in Kakuma and Shimelba (59% and 50%, respectively) are of urban origin. People of urban origin were less likely to have practiced open defecation prior to arrival in the camp (23% versus 86%), whilst people who practiced open defecation at their place of origin were less likely to have a latrine (53% versus 73%) and more likely to practice open defecation (35% versus 12%) in the camp. All these findings were highly statistically significant (p