Background
Handwashing with soap has been viewed as one of the most cost-effective ways of reducing the global infectious disease burden [
1]. The benefits associated with handwashing with soap largely stem from reductions in diarrhoeal diseases [
2‐
4], a common cause of morbidity and a leading cause of death among children under-five [
5]. Despite the irrefutable evidence in favour of handwashing, it is rarely practiced at times when pathogen transfer could be interrupted, and even more rarely involves the use of soap [
6].
Interventions that promote handwashing with soap are therefore vitally important to public health, yet changing behaviour is notoriously difficult. It may be possible to optimise the effectiveness of an intervention by concentrating efforts during a “teachable moment”, a concept from education describing a naturally-occurring event or life-stage that motivates a person to acquire new behaviours [
7]. In 2002, McBride and colleagues examined the evidence for teachable moments using the example of smoking cessation [
8]. They developed a model that characterised a teachable moment as an event which i) increases a person’s risk perception, ii) triggers a strong affective or emotional response, and iii) redefines their social role or self-concept. Disease diagnosis, hospitalisation and pregnancy were identified as teachable moments.
It is also well known in marketing circles that major changes in life-status such as getting married, divorced, having a child or reaching pensionable age are associated with significant changes in the pattern of consumption [
9,
10]. Markov chain analysis of large datasets available on household consumption in the USA suggests that the first major change in such patterns takes place when a household gains its first child [
11].
As well as being a life-changing, emotionally-charged event, pregnancy is a time when women willingly adopt new behaviours to minimise health risks to both the mother and child; evidence suggests that other lifestyle change interventions can also be successfully implemented antenatally [
12,
13]. New motherhood could be viewed as a continuation of this teachable moment, particularly for primiparous women: they have an innate concern for the wellbeing of their baby and they desire to be - and to be viewed to be - a “good” mother.
As well as hypothesising that new and expectant mothers are particularly susceptible to behaviour change interventions, there are compelling public health reasons for focusing handwashing behaviour change interventions at this time in a woman’s life: 1) a mother’s hygiene behaviour directly impacts the health of her child, therefore early adoption of better hygiene practices that continue as the child grows could reduce morbidity and mortality from common diseases such as diarrhoea and respiratory infections [
2,
14‐
16]; 2) children learn important life-skills from their mother [
17‐
19]; and 3) if intervention takes place early enough, it is plausible that handwashing could have an important impact on neonatal survival by reducing sepsis and tetanus [
20]. A trial to quantify the impact of handwashing during the perinatal period on neonatal morbidity and mortality is currently underway (
http://clinicaltrials.gov/ct2/show/NCT01309321?id=10036&rank=2) which should strengthen the evidence base for this third assertion.
Standard approaches to hygiene promotion - focused on educating people about germs and the value of using soap - have rarely resulted in positive, sustained behaviour change [
2,
6,
21‐
23]. It is now increasingly recognised that hygiene behaviour is determined by a range of factors and is deeply-rooted in the environment in which the behaviour takes place [
6]. Although the perinatal period could be an opportune moment to change behaviour, little is known about handwashing behaviour in this population with regards to how, when and where a new mother fits handwashing into her daily activities and what barriers to optimal practice exist (e.g. fatigue, disruptions of baby to daily routine etc.). To design effective hygiene promotion interventions targeting new and expectant mothers it is therefore important to carry out formative research [
21,
24‐
27]. The formative research described in this paper was conducted to gain insight into the daily lives and handwashing practices of new mothers in Indonesia, and to identify factors facilitating and hindering handwashing with soap in this population.
Methods
Study setting
Indonesia has more than 230 million inhabitants spread over 17,000 islands [
28]. The formative research was conducted in Serang District, 72 km west of the capital Jakarta in Banten Province, Java, and an area where the local study partner (MCHIP – Maternal & Child Health Integrated Program) operates. Serang District includes the city of Serang and 28 sub-districts and has a total population of around 1.5 million. Two sub-districts (one urban and one rural) were randomly selected for the study: urban Kramatwatu and rural Pamarayan.
In Indonesia the infant mortality rate is 34 per 1000 live births. Infant mortality is higher in Banten Province than the national average, at 46 per 1000 live births [
29]. Common childhood killers such as acute respiratory infections and diarrhoeal diseases are prevalent: 11% and 14% of children under five were reported to have experienced an episode of ARI or diarrhoea respectively in a two-week period [
29]. In Indonesia, 88% of urban and 70% of rural households can access water on their premises, 21% and 5% of which (urban and rural respectively) have a piped water source. Data from the Indonesian Demographic and Health Survey indicates that 75% of urban households and 43% of rural households have a private toilet, although national figures mask great disparity between Provinces [
29].
Participant selection
Village midwife records from Community Health Centres in Kramatwatu and Pamarayan were used to identify a purposive sample of women who had given birth within the last year, including both primiparous and multiparous women, and infants as well as neonates. Potential participants were told the study aimed to learn how their life has changed since having children.
Informed consent was obtained from all participants and their husbands whenever possible. Ethical approval was granted by LSHTM Independent Ethics Committee. The Maternal and Child Directorate of the Ministry of Health Indonesia also supported and approved the study.
Data collection
A range of tools can be employed for formative research. Due to our focus on how to introduce new hygiene behaviours into this teachable moment, we required good information about how new mothers currently engage in child-care and personal hygiene behaviours. This is very difficult to investigate through questioning. On the other hand, measuring behaviour accurately is also challenging; relying on self-report often produces implausible data [
30,
31], while structured observation is expensive and intrusive [
32,
33]. For these reasons, we opted to capture this information using direct observation (video recording).
Video recording
We used small hand-held video cameras (Panasonic SDR-S50) to directly observe behaviour. We piloted this method in India and Bangladesh and found it acceptable to participants and a valuable data source. Three local women were trained to carry out unobtrusive video recording. They were instructed to film the mother continuously during two periods: when she first woke (around 4 am) for five hours, and from 4 pm until after the evening meal (typically another three hours). Filming at these times maximised the opportunity to observe periods of intense activity in the household such as food preparation, hygiene routines and eating at least one meal. To ensure filming began when the respondent woke up, the field worker stayed in a respondent’s home the previous evening, something found to be culturally acceptable in this setting. Although this was the primary reason for arriving the night before, gaining familiarity with the respondent in this way may possibly reduce reactivity among mothers when filming actually began (it has been proposed that reactivity may be higher in the initial period of observation) [
34].
Participants were not followed when they left the household compound, and any visitors to the home were not filmed unless they consented. Field workers were requested to keep the mother’s hands on camera whenever possible and to keep a record of any periods of absence from the compound to aid interpretation of video footage. Participants’ wishes for privacy were respected at all times, for example, by turning the camera away during breastfeeding and while in the bathroom.
Interviews
Semi-structured interviews were conducted the day after filming by four researchers (three female) working in two teams, each consisting of one LSHTM and one Indonesian researcher (the latter also acted as translators). Participants were questioned about their handwashing practices and motivations for handwashing, how their life has changed since having children, their knowledge of health risks and their sources of information about child-care.
If we want to insert or change a particular behaviour in a person’s routine, it is first important to understand the particular sequence of activities that take place currently – the “script”. In a process called “script elicitation” (used in cognitive psychology [
35]), we asked primiparous women to describe their day from the time they wake in the morning to when they go to bed. Cards with simple drawings, made by the authors in advance, and depicting everyday activities and chores such as breastfeeding, doing laundry or cooking a meal, were laid out in front of the respondent as she talked, each card being laid down depicting the activity she had just mentioned. Once she had finished describing her day, the respondent was prompted about any obvious omissions (e.g. eating lunch) and the gaps in the routine were filled in. Inclusion of scripting framed discussion about
how the young mothers’ daily lives had changed and suggested a variety of lines of questioning in interviews. The picture cards were also used as visual aids to help understand the flexibility of the daily routine, order of activities and to describe a typical day before having children.
Data analysis
Videos were reviewed daily so that unclear portions of film could be immediately discussed and interpreted together with the field worker and with the participant (in interview the following day). A detailed sequence of events was recorded using an Excel database. The database was used to identify occasions when hands were washed with and without soap and to document details of the physical environment. Previous work in this field has shown us that hands may well be washed with soap, but the activities that prompt handwashing (and soap use) are not usually times considered important from a public health perspective. We decided to categorise handwashing behaviour according to whether or not soap was used to wash hands and whether or not hands were washed at key times so we could better describe handwashing behaviour. Each participant was categorised into one of five categories: Washer (washes hands with soap at least once at a critical time); Reactive Washer (washes hands with soap at least once but at a non-critical time, i.e. motivated by reasons such as dirt or smell on hands); Rinser (washes hands at least once at a critical time but using water only); Reactive Rinser (washes hands with water only, never at a critical time); and Avoider (never observed to handwash). These categories best described their hand-washing behaviour based on the times when they washed their hands and whether or not they were ever seen to use soap to hand-wash on this occasion. Categorising individuals in this way allows the data to be viewed differently, illustrating the range of behaviours that take place and the differences between individuals in terms of whether and when soap is used to wash hands.
We performed thematic analysis of interview transcripts to provide insights into practices and underlying motives and drivers of behaviour. This analysis allowed unexpected themes and concepts to emerge from the information provided by informants. Scripts (daily routines) from the different women were aggregated to find the most frequent sequence of daily events, thus generating a “master” routine that illustrates the course of life before and after giving birth to a first child. All aspects of this study conform to the RATS guidelines for qualitative research.
Discussion
This formative research was designed to capture detailed information on the lives and handwashing practices of new mothers in Serang, Indonesia. Handwashing with soap occurs at a low level, but is not constrained by water or soap availability. Hands are typically washed after eating, cooking, doing household chores and after cleaning a child: times when hands are visibly dirty, oily, smelly, sticky or otherwise uncomfortable. We divided the participants into handwashing ‘types’ (such as those who use soap only when they are motivated by hands which are visibly dirty or smell – so-called ‘reactive washers’) to illustrate the range of hygiene behaviours that take place with respect to hands, and the differences between individuals in terms of whether and when soap is used to wash hands. We believe this categorization is insightful and may be true of other populations, and so could usefully guide program development on handwashing behaviour in future projects. This classification may also have implications for health promotion efforts, as it is possible that each of these types will require different kinds of intervention, since they appear to be motivated to use soap in different kinds of circumstances. While it is difficult to make specific recommendations for other handwashing programs without knowing their objectives, we believe that any ability to predict the types of handwashers likely to be found in a population can assist other programs, especially if they are not able to conduct research themselves. Knowing such types has program implications – for example, reactive washers are likely to be motivated by disgust at visible contamination on their hands, while avoiders seem not to respond to such cues to handwash, and so will require other means to change their behaviour.
Emotional drivers of behaviour are also important determinants of handwashing [
36]. Observation of when hands are washed, supported by interview responses about why hands are washed, lead us to hypothesise that handwashing at these times is most likely to be driven by feelings of “disgust” directed at substances or smells perceived to have contaminated hands; and “comfort”, desire for hands to feel clean [
6]. “Disgust” is therefore also the probable motivator for urban women who wash hands after returning home, ridding hands of dirt from the environment or other people they have contacted, while rinsing hands after doing laundry could be to remove the harsh feeling of detergent (“comfort”). Conversely, washing hands that look and feel clean derives little benefit, which fits with the low levels of handwashing observed at relevant food-hygiene junctures (i.e. before cooking or eating). Observed handwashing behaviour is consistent with other low-income countries [
6,
37] and other studies in Indonesia [
38].
Both the video footage and interviews indicate that handwashing is infrequent and does not seem to be prompted by having a new child, although “nurture” (desire to care for one’s offspring) has been previously demonstrated to drive maternal handwashing behaviour [
6]. Failure to wash hands is not because of a lack of time; these mothers have considerable time on their hands, being almost exclusively concerned with child-care, and having been relieved of other responsibilities by others in the household. Rather, hand washing may not be seen as a necessary part of being a good mother in this society. Sporadic instances of handwashing/using baby wipes before breastfeeding are likely to be an attempt to practice a behaviour desired by the midwife: information provided in interviews about the advice midwives gave concerning handwashing before breastfeeding thus matched the behaviours observed on film. The same mother was observed to respond differently to different child defecation events, possibly due to whether or not hands were contaminated with faeces. In a previous study in Burkina Faso, the stools of young children were regarded as less offensive than the stools of older children [
39]. It is possible that mothers do not find their infant’s faeces disgusting and for this reason “disgust” may not be a strong driver of handwashing behaviour at this time. It would have been interesting to have collected more information on this. However, encouraging the mother to handwash after clearing up a baby’s stool would hopefully translate into correct hygiene behaviour later in life. Although hands are allegedly washed after cooking with chilli for the protection of the child, it is more probable that handwashing after cooking is habitual.
The automaticity of existing handwashing behaviour [
36] is one reason why it is hard to change. A “habit” can be defined as a behaviour that is performed frequently in a constant context [
40]. The context winds up cueing the behaviour so it occurs automatically in that situation. This means that habits are context-dependent. When the context changes existing habits are disrupted, providing an opportune time to insert new behaviours and form new habits [
41]. Our findings confirm that new motherhood results in many changes to a woman’s daily routine and diet, giving us reason to be optimistic about the potential for behaviour change at this time. They also confirm that new motherhood is an appropriate teachable moment as defined by McBride [
8]: women are aware of and concerned about health risks their child faces; their social role has changed; and they adore their new baby, undoubtedly a strong emotional response. New mothers are likely to prove a receptive audience if they can be convinced of the benefit of an intervention: although handwashing behaviour at present does not appear to have changed as a result of the new baby, “nurture” motives could be stimulated if women see handwashing as a trait of a “good mother” or they believe it will be beneficial to their child’s health. This could be an important campaign angle. Primiparous women are particularly open to new advice and would be a relevant target. Furthermore, as women willingly follow the advice of health professionals and family members, it could be relevant for midwives – who frequently contact women during the peri-natal period – to be involved in delivery of a community-based intervention to improve hand hygiene in this population, and for that intervention to also target influential family members. The frequency of contact with the health system at this time could be useful for reinforcing handwashing messages.
Further, rather than attempt to introduce handwashing with soap before breastfeeding (one of the current recommendation from midwives), which happens too frequently to be constantly interrupted by trips to a handwashing location, it is contact with faeces that should be the primary concern if the desire is to set a handwashing habit that will benefit the child’s health at a later date. In addition, we know from the videos that considerable time and care is already invested in cleaning and dressing babies in some households after a defecation event. The videos also showed us how variable handwashing practices were, even within the same individual, partially due to the presence of cues such as the physical setting, presence of an object, or visually dirty hands. Baskets for the various clothes and ointments needed at this time are a part of every new mother’s ‘kit’ for newborn childcare in Indonesia. It would probably be relatively easy to insert handwashing with soap soon after cleaning the baby’s bottom and before extensive further contact with the child. This is likely to be especially true if there is a handwash stand or other visual reminder present in a relevant location within the household as well. As we observed these cleaning and changing rituals being explicitly taught to young mothers by their mothers, it is natural to target these mothers-of-mothers as the appropriate channel for communicating the need to include handwashing with soap as part of normal child-care operations.
This study had some obvious limitations. After every interview a discussion took place to clarify any points of confusion. However, we cannot exclude the possibility that the presence of “foreigners” during interviews may have influenced respondents, nor the fact that bias can also stem from local researchers’ perceptions of their own culture. The collection of video footage provided a rich data source, although the challenge of filming at first light and the inability to capture activity occurring inside a closed bathroom remained drawbacks, the latter a perennial difficulty in studies of handwashing behaviour [
34,
42]. Although some degree of reactivity might be expected in any study of behaviour [
34,
42], it was only evident in three films where the women in question were clearly performing tasks they had not done previously in an attempt to look good. In one particular film, her mother could be heard issuing instructions in the background. It is difficult to know what other activities women did or did not do because they were being filmed, but we did not see obvious posing for the camera, women performed multiple activities many times in the same way on film, and they did not know exactly what behaviour we were interested in capturing. As participants were identified using the village midwife’s records, we have not included women who do not access antenatal services or those who give birth at home without a skilled birth attendant. As accessing antenatal services and choosing to be helped by a skilled attendant are strongly correlated with income, education, and living in an urban or rural locality [
43], and handwashing practices in this study and other studies are associated with the same factors, it is possible that we would have seen different behaviour if these women had been included.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
RA designed the study with contributions from EI. KG performed analysis and drafted the manuscript. All authors were involved in data collection and read and approved drafts of the manuscript, including the final version.