Extended unstructured observations and interviews.

Two non-neighbouring households in each community were identified through a random walk. Eligible households were those with at least one child aged 1–9 years. Households were recruited around dawn on the day of observation to prevent households from cleaning or otherwise preparing for the study. Where possible, the same households were observed during both seasons. To minimise reactivity, participants were not told which specific behaviours we were interested in, rather they were informed that we were interested in learning about their daily routines. Observation took place for up to 8 hours, from around 7am until between noon and 3pm, depending on participants’ farming activities. Two female researchers were stationed in each household, moving within the house and compound as needed to monitor behaviour. Each pair comprised a foreign researcher and a local research assistant. At the end of each observation, observers compared notes. If family members were in multiple locations, emphasis was placed on observing locations where water was present so that performance of hygiene behaviours would not be missed. Whenever possible, one researcher accompanied family members to collect water. Photos were used to capture behaviour in context, and video was taken to help understand sequences of behaviour. Maps of each household were drawn and the presence of faeces and location of occurrence of the behaviours of interest (e.g. face washing, defecation) were marked.

At the end of each observation period, the female primary caregiver was interviewed and the socio-demographic questionnaire was completed. Interviews followed a topic guide and included props to explore daily routines, sleeping arrangements, soap availability and attributes, and water prioritisation and use. Participants in households observed during both seasons were interviewed twice.

Focus group discussions (FGDs).

FGDs were held across the four communities over the two seasons (Table 2). FGDs were separately conducted with mothers of children under three years-of-age, male household heads, grandmothers and 15 and 16 year-old sisters of pre-school children (rainy season only). These population groups were selected because they all play a role in caring for young children (who have the highest rates of infection). The FGDs were held with similar individuals (age and gender) and the size was kept small (5–8 individuals per group) to encourage active participation. Participants were selected by health extension workers who were instructed to invite people with different levels of wealth and education and from different areas of the communities. Multiple members of the same family were not included and there was no overlap between the populations included in the household observation and FGDs. FGDs were designed to build consensus around themes that had arisen from observation and individual interviews and to explore some additional behavioural determinants. Tools were used to encourage participants to make generalisations about the perceptions and practices of their community.

Data handling and analysis.

All FGDs and interviews were voice-recorded and transcribed into English. Observation notes were typed up. A two-step process was used for the analysis in line with the BCD approach. During data collection, the team adopted a process of ‘sequential recycling’. This is a process drawn from social marketing whereby the investigators come together at the end of each day to discuss the findings [26]. Each team member contributes anecdotes or patterns they noticed during data collection. These items are then sorted into BCD checklist categories, and discussed. When there is agreement, the item becomes a finding. Where there is disagreement or if knowledge gaps remain, then questions are developed for further exploration. This process also allowed us to identify when we had reached a point of saturation suitable to meet the study objectives. At the end of data collection a deeper thematic analysis was conducted using the BCD determinant checklist [29] as a deductive ‘top down’ coding system in relation to each of the behaviours of interest [33] (coding tree included in supplementary materials, S1 Fig). The coding was performed by two authors (KG and KS) and reviewed by a third author (SW). Where disagreements occurred the findings were discussed between all authors in order to develop a coherent thematic summary.

Ethics and consent.

The research protocol was approved in Ethiopia by the Oromia Regional Health Bureau’s Research Ethics Review and by the London School of Hygiene & Tropical Medicine Ethics Committees. Informed written consent was obtained for each household member 18 years or over. Parents consented on behalf of children under 18 years, but those aged 7 to 17 years also provided assent.

Current practices.

Body washing took place outside shortly after waking using a small bucket or pail of water filled from a larger water container. Faces were washed at this time, along with feet, legs, hands and arms in a similar manner to that depicted in Fig 2. Faces were not dried after washing. Overall, 33 of 63 (52%) individuals in the dry season and 21 of 67 (31%) individuals in the rainy season were observed washing their faces. Around two thirds of individuals used soap during face washing (64% dry season, 62% rainy season), but soap use was inconsistent within as well as between families. There were no dedicated places for face washing, but faces were typically washed outside. Face washing was only observed on one other occasion (after a child had eaten oily food), but is reportedly carried out again later in the day (after work or school, or before bed).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Girl washes with soap before school during a typical morning washing routine involving the feet, legs, hands, arms and face.

https://doi.org/10.1371/journal.pntd.0007784.g002

Face washing practices varied by age and season (Fig 3). During the dry season the majority of school children washed their faces, but there was a marked decline in face washing amongst school-going children during the rainy season. Adults washed faces less frequently in both seasons, but tended to use soap when they did wash. Pre-school children as young as three years-of-age were observed washing independently and unprompted, sometimes with soap. Only the faces of the youngest children were washed directly by their mother, on five occasions (four in the dry season) as part of a full bath which used minimal water (<1litre). No bathing was observed among older family members, corresponding with reports that school children bathe weekly and adult women may bathe as seldom as every two months.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Observed face washing behaviour amongst household members of different ages in the dry and rainy season.

https://doi.org/10.1371/journal.pntd.0007784.g003

Environmental and psycho-social determinants.

In contrast with observed practices (Fig 3), mothers state that children’s faces are washed as a priority:

“He doesn’t go without eating breakfast, so that means he can’t go without [face] washing”. (Mother, Dry Season Interview)

Face washing was perceived to be rewarding, supposedly giving children a positive start to the day and promoting sleep, responses that align with observed face washing in the morning and reported face washing before bed.

“It feels uncomfortable if their face is not washed; it refreshes them and gives them a brighter day and mind and they can also sleep easily.” (Mother, Dry Season Focus Group)

Although participants associated unclean faces with flies and disease transmission, pre-school children’s dirty faces were never spontaneously washed, so it is likely that face washing is driven more by habit and less by presence of discharge. This may also explain why all but the youngest pre-school children (who received the prioritised attention of their mother), did not always automatically wash their faces in the morning. The observed independence of body washing by age three differs from the collective opinion of mothers that children need assistance with their personal hygiene until they are around seven years old.

School attendance appears to be a strong driver of face washing (with soap). It is a requirement to have a clean face for school, and children are reportedly disciplined and even sent home from school if their faces are dirty. It is unclear whether dirty children are judged negatively by others in the community as women were divided about the extent to which people would gossip about them or their ‘dirty’ children.

“People gossip about you even if they don’t tell you. They may say: ‘Why did she leave her child with a dirty face?’” (Mother, Dry Season Focus Group)

“Everyone thinks it is up to the mother, they don’t interfere in someone else’s business.” (Mother, Dry Season Interview)

The observed reduction in face washing among school-age children during the rainy season may be due to a perceived reduction of this school-related social pressure during the school holidays. However, the rainy season is also a busy time agriculturally and it is not considered necessary or feasible to remain clean when time is predominantly spent working in the fields and you are not seen socially.

“During school [students] …want to be seen as attractive, but now [during the rainy season] we all go directly to the fields each day instead, so the frequency of washing has declined.” (Mother, Rainy Season Focus Group)

“We go home when it gets dark, we don’t even think of what we eat during those seasons, we don’t even wash our legs.” (Mother, Rainy Season Focus Group)

Water collection reflects planned activities for the day, and water is used conservatively: mothers were seen to chastise children for wasting water whilst washing. Soap use requires additional water use, so water scarcity may limit face washing with soap. As soap is kept inside, but washing takes place outside, lack of soap at the location of use may prove more limiting than the overall lack of soap. In these communities only two of the families had no soap present in the household in each study. Participants only considered soap to be necessary for face washing when a face is really dirty, which was not a cited key reason for face washing. This may further explain why soap is inconsistently used.

Current practices.

Many pre-school children in the study households had nasal or ocular discharge on their faces during the observation period; flies on faces were also abundant. Discharge was removed from children’s faces on only 14 occasions across both data collection periods. Children only occasionally wiped their own faces (on their sleeves or using their fingers). Faces were instead wiped by female caregivers using hands which were then wiped on clothing or a nearby surface. We observed a child’s garment being used for wiping on one occasion only, but the use of clothing is reportedly more common when mothers are working in the fields and have dirty hands.

Handwashing is important in this context because it is plausible that Chlamydia trachomatis is transferred from faces to hands during face wiping, facilitating transfer to another face. Handwashing was seldom witnessed other than as part of the morning washing routine. Some handwashing prior to and after eating was observed, but hands were never observed to be washed after wiping facial discharge, after handling child faeces, or after returning from defecation. Soap was not observed to be used for handwashing during the rainy season and was only used on 6 of 37 (16%) occasions during the dry season—all but once after eating.

Environmental and psycho-social determinants.

The presence of flies and discharge on children’s faces rarely cued face wiping. However, requesting to take a photo prompted discharge to be removed from children’s faces. This suggests that while discharge is tolerated, it is still considered socially undesirable. Participants felt that their environment constrains them from acting any differently:

“We usually wipe [a child’s face] with our hands and put [the discharge] on the ground. We don’t have anything else to use because we are in a rural area, but it is not enough to be clean.” (Mother, Dry Season Focus Group)

Handwashing after face wiping is not the social norm. Soap use patterns also reflected opinions that soap is only necessary when hands are sticky or visibly dirty. The lack of handwashing after face wiping suggests that discharge is not viewed to be sufficiently dirty or disgusting to cue handwashing.

Current practices.

Study participants slept communally on raised beds, straw mattresses, or goat skins. Generally, young children slept together with their parents or grandparents, whilst older children slept apart. Blankets were generally shared and although pillows were not commonly used, old garments or grain bags were sometimes used by older family members. Skins, mats and blankets used for sleeping were tidied away each morning or were put outside to air. Overnight guests are provided with clean bedding that is reserved for special occasions.

Environmental and psycho-social determinants.

Young children sleep together with their mother because they can cry at night, because it is often cold and because of the limited number of blankets available. The cost of purchasing a new blanket is seen to be prohibitive (250–300 Ethiopian Birr, around US$10). Guests are given preferential treatment:

“We don’t give them [guests] the blankets we always sleep on and use because we respect them.” (Grandmother, Dry Season Focus Group)

Current practices.

Clothing was washed in three of the ten households observed in the dry season (twice involving only an infant’s clothing, once a full load of laundry) and not at all during the rainy season. Observed behaviour contrasts with reports that the clothing of children under two is prioritised for washing and this occurs at home on a daily basis. School clothes are reportedly washed weekly, but adult clothing may be washed infrequently (bi-monthly or less), while blankets are rarely washed. Some respondents reported doing laundry at their home while others did it at nearby rivers. Rivers or other water bodies tend to be used to wash large items such as adult clothes and bedding which use more water.

Environmental and psycho-social determinants.

Participants thought that bedding should be washed or aired when it becomes “itchy”, but were unaware of any connection between dirty clothing or bedding and disease transmission. Laundry is considered taxing in terms of water volume required (70 litres was observed to be used to wash a load of both adult and child clothing), soap used and time. Effort is made to clean clothes reserved for special occasions (Timkat celebrations in the dry season) as they are worn in public. Given the limited availability of water at home and effort required to wash at the water source, laundry of clothing of infants and school children are prioritised because the former become soiled frequently and the latter need to look clean and smart for school. Laundry is deprioritised in the rainy season, when clothing takes longer to dry, water sources are muddy and life is too busy:

“We don’t usually wash our clothes regularly during the farming season [rainy season]. We work a lot in the field and go home when it gets dark. We don’t want to lose time by washing clothes; we don’t even have time for sitting.” (Mother, Dry Season Focus Group)

Current practices.

If human faeces are not safely contained in a latrine they may become a breeding site for M. sorbens flies [34]. Half of the households under observation in each season had a simple pit latrine, but these latrines were poorly constructed, with shallow pits, limited or no superstructure and small holes which made them difficult to use hygienically. Latrines were reported to be washed away in the rainy season, but this was not observed. Although many latrines were visibly in use, the presence of human faeces within the same compounds suggested that latrines were not used exclusively by the whole family. The age at which children were reportedly able to use a latrine varied from seven to 12 years, depending on whether an adult’s assistance was required. Women expressed concern over young children falling into the hole or playing with faeces. The youngest children were permitted to defecate freely within the immediate compound, including adjacent to the living structure and child defecation within the compound was witnessed on six occasions during the study, after which the faeces were left in situ until dry. Hours later it was swept up and disposed of with other rubbish in a pile in a specific area of the compound. Adults and older children who open defecated reportedly did so before dawn (behind the house or in a nearby field), or later in the day whilst working in the fields. These faeces were not reportedly covered or moved. Animal faeces may also be used for breeding by M. sorbens flies [34]. Animal faeces were also scattered throughout all of the compounds studied, as well as within the home or kitchen in houses where animals slept in the family’s living quarters. Cow dung was reused for household purposes, such as relaying the floor.

Environmental and psycho-social determinants.

Latrine building is encouraged by Health Extension Workers for health reasons, but is principally motivated by the desire to avoid a fine. Poorly constructed latrines with limited superstructure and small holes offered little privacy or enticement. Experiences of latrine use were often negative, with reports of bad smells and not feeling good after using a latrine. As such there was little demand for sanitation and few examples of hygienic public latrines (the schools and health centres visited had well-constructed, but poorly maintained or non-functional facilities). In contrast, open defecation was viewed as acceptable, comfortable, convenient and beneficial for the land:

“Faeces helps trees to grow if it is thrown in the field and it helps the land to become fertile; It [faeces] does not have a benefit at home [in a latrine].” (Mother, Dry Season Focus Group)

Little effort was made to try to remove animal faeces from compounds, possibly because animal faeces were viewed to be less disgusting than human faeces and because it was an important source of fuel.

The presence of flies was linked to “disease”. However, HIV was listed as a disease that could be spread by flies, along with diseases such as diarrhoea and trachoma. The presence of flies was thought to depend on the weather, but was also linked to fresh milk, cattle, uncovered food, faeces, dead animals, the latrine, and other dirty things in the environment:

“Flies do not land on a clean thing, on a clean face, so we have to clean our environment.” (Father, Rainy Season Focus Group)

That said, participants did not believe that there was much they could do to control flies.