Access to handwashing with soap facility: a post-sensitisation investigation of drivers

The promotion of hand hygiene and safe water is an effective intervention measure in reducing the spread of infectious diseases, including the recent deadly COVID-19 (WHO & UNICEF 2020). Even before the COVID-19 pandemic, considerable evidence had been discovered about the effectiveness of hand hygiene in reducing the incidence of many diseases, especially diarrhoea, respiratory infections, and eye infections (Pandve et al. 2016). A 2020 United Nations report acknowledged that hand hygiene is about ‘the cheapest and easiest’, yet the most effective strategy to avert the spread of COVID-19 and other diseases (United Nations 2020). Hand hygiene is an essential component of the Water, Sanitation, and Hygiene (WASH) agenda and therefore a fundamental human right that remains central to achieving Sustainable Development Goal 6 (SDG 6). Because of this, only hand hygiene facilities, particularly for handwashing with soap (HWWS), are monitored by WHO/UNICEF, despite the fact that the SDG hygiene target also covers facilities for addressing other hygiene issues like menstruation and food hygiene (Mara & Evans 2018).

Unfortunately, the challenge of inaccessibility to hand hygiene facilities (Mwachiro 2014) has constrained interventions towards improving hygiene and achieving the SDGs (Global Handwashing Partnership [GHP] 2020). Globally, it is estimated that about 3 billion people lack access to hand hygiene facilities (UNICEF/WHO 2019). In 2015, households having facilities for HWWS ranged from 15% in Sub-Saharan Africa to 76% in Western Asia and North Africa (Mara & Evans 2018; United Nations 2020). This translates into an average of only 19% of the world's population practising hand hygiene, particularly after visiting the toilet (GHP 2020).

The consequences of inaccessibility to hand hygiene facilities are enormous and include the loss of about 300,000 lives annually, a majority of whom are children under age 5 who die due to poor hygiene-related diarrhoea (George et al. 2017). In Ghana, over 10,000 children die each year from diarrhoea and pneumonia, both of which are preventable by hand hygiene, yet the practice of hand hygiene remains low in the country (Dajaan et al. 2018).

Towards a resilient effort to increase handwashing practice in Ghana, UNICEF has over the past 5 years initiated several HWWS interventions across all Metropolitan, Municipal and District Assemblies (MMDAs), especially the poor districts in the Upper East Region (UER) of Ghana, in partnership with Environmental Health and Sanitation Units (EHSUs) and local non-governmental organisations (NGOs). One of the cheapest and yet most effective forms of HWWS facilities known as ‘Tippy-Tap’¹ (see Mbakaya et al. 2020) has been the main technology adopted by UNICEF for hand hygiene promotion. This facility has been promoted in many communities across rural Ghana, especially in the UER, as part of the Community-led Total Sanitation (CLTS) programme. The type of Tippy-Tap recommended to the communities is made of local materials with approximately zero production cost. Unfortunately, lack of awareness, adoption, and use of the facility may leave a huge sustainability gap. This has heightened the need to conduct research into establishing potential factors that could lead to the adoption and use of the emerging handwashing technology.

Admittedly, there is a paucity of empirical evidence on Tippy-Tap adoption and use in developing countries. Nonetheless, Naughton et al. (2015) examined the factors that lead to the use of Tippy-Tap in two communities in Mali, which is located in one of the world's poorest regions. Overall, six factors were identified and studied as potentially critical for the lasting use of handwashing stations. These factors include gender, educational training, water proximity, seasonality, wealth, and station adoption. In a related study in Kenya, Christensen et al. (2015) applied randomised controlled trials to investigate the adoption of WASH interventions. As part of the findings, the study showed that for intervention households, there is a 71–85% point increase in the likelihood of their having a place for handwashing and a 49–66% point increase in the likelihood of having soap available. In Ghana, the few available studies have mainly been school-level adoption and analysis, paying little or no attention to household-level analysis (e.g., Dajaan et al. 2018).

The current handwashing challenges are associated with behavioural change, which is more closely linked to personal, household, and community-level characteristics than the indicators used in these previous studies. People in the same community or country behave differently when it comes to handwashing, which implies that personal and household-level characteristics may influence behaviour (WHO 2009) more than other indicators as used by previous studies. In addition to theoretical underpinning, the selection of indicators used in the present study was based on empirical evidence (e.g., Amoah 2017; Wolde et al. 2022). These indicators such as personal (gender, education, age, occupation), household (household size), and community (social status) generally differ from what has been used in previous studies by Naughton et al. (2015), such as gender, training, water, seasonality, wealth, and monitoring.

This study, therefore, seeks to investigate the key factors that drive the uptake or otherwise of ‘Tippy-Tap’ in Binduri District after the sensitisation programme. The first objective of this study is to determine the proportion of households that have adopted and used the Tippy-Tap facility. The second objective is to investigate the post-sensitisation programme determinants of Tippy-Tap adoption and use by households. Given that the intervention is a continuum initiative, the findings of this study will inform future intervention strategies within the community and other communities with similar characteristics across the world. Again, it will influence the decision on the intensity of the sensitisation programme, which can influence a further increase in the rate of adoption and use of the Tippy-Tap facility.

During the first phase of the project intervention, the baseline data on households having and practising HWWS in 127 communities were collected in December 2017 through a household survey (results = 4%). In March 2018, the sensitisation programme was carried out in all the communities as the implementation phase of the project. After the implementation of the project, an end-line survey was conducted in June 2018 at the household level using a checklist for HWWS to monitor results for assessment of the project impact.

Next, a cross-sectional survey was subsequently conducted in March 2019 to evaluate the extent of adoption and associated determining factors. Using the March 2019 data, this study explores determinants of HWWS facility (‘Tippy-Tap’) adoption and use in the Binduri District.

The total number of households in all the 127 communities is 6,188. Given that all households in these 127 communities were given ‘Tippy-Tap’ sensitisation training, they formed the sampling frame with household-heads who are 18 years and above constituting the unit of analysis. Out of this, a sample size of 714 which is greater than 10% of the household population was determined at a confidence level and confidence interval (margin of error) of 95 and 3.45%, respectively, using an online sample size calculator by Creative Research Systems (2012) based on a method developed by Jerrold (1984). This is statistically representative of the district's household population. Furthermore, a stratified random sampling technique was used in selecting the respondents. Because of the homogeneous nature of the communities, each community was treated as a unique stratum, after which the listed sample in the strata was randomly selected using a simple random number generator. Owing to the fact that the sample was determined with replacement from the population, missing data were accordingly replaced. Based on the sample size and sampling technique approaches, we argue that generalisation of the findings to the population is possible.

A structured questionnaire was designed and administered as the data collection instrument. Two experts on WASH, HWWS, and questionnaire construction were made to review the questionnaire to ensure it captured the topic under investigation effectively and also check for confusing, double, and leading questions for the purpose of validity checks. The instrument was first piloted before it was finally used for its purpose. The interviewers administered the questionnaire in-person using the face-to-face method. The interviewers were personnel of the EHSUs of the Binduri District who understood the cultural practices of the communities and could write, read, and speak English and the predominant local language (Kusal) fluently. In a house of more than one household, only one household head was interviewed to avoid bias. The questionnaire had four main sections and covered information on demography, socio-economic characteristics, determinants of HWWS, and household water treatment and safe storage (HWTS) availability, and perceived impacts of climate change on HWWS and HWTS implementation. The data were analysed using STATA 15 software.

is the dependent variable showing the probability that a household will adopt and use the facility or otherwise. This is dependent on gender which is a dummy variable (Male = 1, Female = 0), Edu measures educational status of the respondent (Educated = 1, Uneducated = 0), age of the respondent in years (⁠⁠), age squared (⁠⁠), household size is the number of people in a household (⁠⁠), Social Status (SS) is a dummy variable which identifies people with social responsibility in their respective communities, Occupation (Occ) is a dummy variable with those who are engaged in daily economic activity (e.g., farming/petty trading) for the past month were coded as 1 and otherwise, coded as zero (0). The error term is captured as u.

In this section, we first present the descriptive statistics of the variables and then the results obtained from the regression model. The results are discussed to unearth the relevance and implications for sustaining handwashing behaviour.

The adoption and use rate of 89.8% (≈90%) as shown in Table 1 refers to the functional Tippy-Taps and the remaining 10.2% (≈10%) include the non-functional Tippy-Taps and households who have not installed any at all. With regard to the critical times at which they normally wash their hands using Tippy-Taps, respondents indicated they do so before eating (18.4%); after visiting toilet and before breastfeeding (46.3%); after visiting toilet and before eating (19.7%); and others (15.6%) that include after visiting a sick person/hospital; and after attending a funeral.

From Table 1, the dummy nature of the variable adoption & use is revealed from the maximum value of 1 and a minimum value of 0. The mean value is 0.898 implying that a greater proportion of the respondents adopted and used the ‘Tippy-Tap’ facility. This suggests that the sensitisation programme had a significant impact on the attitudes of the households in Binduri. Again, approximately 46% of the respondents are males while 54% are females reflecting female dominance in WASH-related issues in the district. Traditionally, education has been generally free in the northern part of Ghana since the post-independent era; hence, it was not surprising when the majority of the respondents were found to be educated. The average age of the respondents was approximately 39 years with the minimum and maximum being 23 and 73 years, respectively. Household sizes range from a minimum of 2 to a maximum of 10 with an average household size of approximately 7. Furthermore, about 61% of respondents have some social responsibility in the community. These are mainly religious, economic and market, transport, cultural, and political community responsibilities. Finally, it is evident that approximately 76% of the respondents found themselves in a form of occupation which is predominantly farming or petty trading. Although this is high, it is an undoubted fact to acknowledge that the proportion of respondents who are not engaged in any occupation is also relatively high and may raise concern for the attention of policy makers. Interestingly, most of those without occupation and a section of the farmers were identified as yearly travellers to the south during the dry season (November–February) to work and return home to enjoy their meagre earnings between rainy seasons (March–October). Observably, the few who choose not to work end up struggling to fend for their families as compared to those who reinvest their earnings into other economic activities.

For validity of results, we show in Table A1 (Appendix, Supplementary material) that the covariates are not severely serially correlated. The models are estimated with robust standard errors, whereas community differences and bias from interviewers are controlled by introducing fixed effects. This is achieved by creating dummies for each interviewer and including them in the model. We acknowledge that, with or without interviewer dummies, the results did not change. The Walt test is highly significant, suggesting an overall model fit. With these diagnostics, we are confident in the results.

In Table 2, two main Logit models are estimated with their associated marginal effects. In models 1 and 2, we estimate the covariates by controlling for only community fixed effects. However, in models 3 and 4, the same covariates are estimated while we controlled for both community fixed effects and interviewer fixed effects. The results for models 1 and 2 and models 3 and 4 are approximately the same across variables. This lends credence to the robustness of the results.

Galasso et al. (2020) have posited that Gender differences exist in attitudes and behaviours in all countries. They argue further that relative to males, females are more likely to respond positively to a public health concern; however, according to Amoah et al. (2019), males are more likely to respond positively provided they are in charge of the household budget. Similarly, Amoah & Addoah (2021) have shown that female household-heads who are financially empowered are more likely to promote good hygiene/environmental practices. Given the ambiguous nature of existing findings regarding gender, we further examined gender and access to HWWS. Using females as the reference category, we find a negative and statistically significant relationship between gender and access to HWWS. That is, with a marginal effect of −0.1529, we show evidence that being a male decreases the probability of accessing HWWS by approximately 15.3%. That is, female household-heads in Binduri are more likely to purchase and use the HWWS relative to male household-heads. This finding supports the earlier claim that regarding public health and good hygiene practices, the role of females as drivers in the household cannot be overemphasised.

Education in this study is used as a measure of one's ability to know, learn, and adopt a new technology. As posited by Amoah et al. (2021), education correlates with attitudinal change. So, using uneducated as the reference category, the results reveal a positive and statistically significant relationship between respondents who are educated and have access to a HWWS facility. By implication, with a marginal effect of 0.2194, we argue that being educated increases the probability of accessing the HWWS facility by about 23%. This means that educated households in Binduri have a higher probability of purchasing and using the new HWWS facility as compared to the uneducated. This evidence is consistent with Vicente-Molina et al. (2013) and Amoah & Addoah (2021) who show that knowledge influences behaviour. Nonetheless, it is important to note that the association between formal education and handwashing practices could depend on the level of education. For instance, some previous studies posited that a higher level of education could have a significant impact on handwashing practices than a lower level of education (e.g., Wolde et al. 2022).

Age has been used in the social science literature as a measure of a person's ability to manage resources sustainably (see Amoah & Addoah 2021). That is, owing to health implications, older people based on their experience in life may have a relatively higher tendency to purchase health facilities with sustainable health implications. Indeed, we expect older people to access the HWWS facility more than younger people. Evidence of a negative and statistically significant difference between age and access to HWWS is found. We report an associated marginal effect of 0.0731, which implies that a 1-year increase in age decreases the probability of accessing HWWS by approximately 7.31%. Similarly, we introduced the squared term of age, and found a positive and statistically significant relationship with HWWS. With a marginal effect of 0.0007, we argue that as age doubles, respondents are more likely to access HWWS. This means that, relative to the younger cohorts, older people have a higher probability of accessing the HWWS facility. A further investigation of the turning point using the nlcom command in STATA shows that below 52 years, respondents are less likely to have the facility; however, at 52 years and above, respondents are more likely to have the HWWS. This corroborates the finding by Jong-Gyu (2019) that demonstrated differences in handwashing behaviour based on age, with older people exhibiting better handwashing practices than younger people. It further supports the finding in Amoah (2020), where with respect to averting water behaviour, a turning point in age is found. These findings lend credence to behavioural changes due to different phases a person is confronted with.

Again, large numbers are important in determining how people behave as a result of peer or cohort effect. In communal living environments as in the present case, we expect larger household sizes to behave differently from smaller household sizes. That is, because of peer or cohort effect, we expect respondents from larger household sizes to have access to HWWS. Interestingly, the results show a negative and statistically significant effect which is counter intuitive. We report a marginal effect of 0.1484, which implies that if a household size rise by one person, the probability to have access to HWWS will decrease by 14.84%. Though counter intuitive, given the characteristics of the study area, free riding could plausibly account for this behaviour. In a broader sense, the communal living effect in these communities makes it legitimate for households sharing the same compound to engage in free riding of common facilities such as HWWS.

In poor communities, social status is considered an important factor in influencing community behaviour. People with somewhat high social status are generally seen as ‘community figures’ who are expected to exhibit leadership for others to follow. Again, their influence as role models cannot be overemphasized as they demonstrate leadership worth emulating. Previous research has revealed a significant relationship between a leader's and community members' (residents') hand hygiene compliance, with the community members' compliance being significantly correlated with the leader's compliance (Shim et al. 2019). Respondents who believe they possess such leadership qualities either in church, work, traditionally, etc., are socially responsible, and people look up to them. In communities or households with such role models owning the HWWS facility, we expect such leaders to influence the behaviour of the community or households to also own the facility. From our results, we observe a positive and statistically significant relationship between social status and access to HWWS. The evidence from the marginal effect is reported as 0.1427, which indicates that respondents who are socially responsible have a 14.3% probability of having an HWWS facility. That is, respondents who are held in check by their social commitment have a higher tendency to comply as compared to those who are without such responsibility.

Occupation is used in this study to represent the ability to afford where cost implications may deter respondents from patronising the facility, although the facility is designed for both non-poor and poor to afford. However, we expect the non-poor to have an advantage in accessing the facility as compared to the poor assuming a reasonable cost implication is associated in buying the gallon to store the dispensing water. That is, respondents who have an occupation are expected to spend a fraction of their income to access and maintain the HWWS facility. The results show a positive and statistically significant relationship between occupation and access to HWWS. With a marginal effect of 0.1545, it suggests that respondents with occupation are about 15.5% more likely to access HWWS as compared to respondents without occupation. This finding corroborates the basic demand theory, where all other things held constant, respondents with purchasing power are more likely to buy the facility than those who lack the purchasing power. Several studies (Odo & Mekonnen 2021) have shown that purchasing power is key to ownership decisions.

Putting the results in perspective, we acknowledge that according to the Ghana Statistical Service (2014), the district under study is economically highly vulnerable, susceptible to the spread of epidemics due to the non-availability of toilet facilities to about 84% of the population. Additionally, residential pipe-borne water supply is practically non-existent with the majority of households relying on boreholes. The recent outbreak of COVID-19 and other public health concerns made it necessary for the introduction of a pro-poor yet innovative facility to facilitate handwashing. A HWWS facility, ‘Tippy-Tap’ was introduced to the community for the first time. Interestingly, regardless of the outbreak of the COVID-19 pandemic, not all households adopted the ‘Tippy-Tap’ facility – although a majority did. From the public health perspective, concerns of why total adoption rate has not been achieved may be raised in such a vulnerable community. In fact, it goes without saying that any outbreak of an epidemic or a pandemic in such a vulnerable community may sweep away precious lives. Indeed, this study considers an investigation into the determinants of ‘Tippy-Tap’ adoption in the district as a worthwhile exercise. Overall, this new evidence shows that the determinants of adoption of ‘Tippy-Tap’ go beyond personal demographic characteristics to include household and community-level characteristics. That is, towards designing strategies for ‘Tippy-Tap’ adoption as a panacea to COVID-19 and other epidemic spreads, the findings of this study posit that the focus should go beyond personal demographic characteristics to include household and community-level determinants.

Also, we show in Figure 3 that women and female children are predominantly responsible for household water supply burden. This converges with earlier claims that women and female children bear the household water supply burden in Ghana (Amoah 2020). Also, where the woman or the female child bears the household water supply responsibility, they are more likely to adopt the Tippy-Tap. In cases where all (man, woman, and children together) and only the man is responsible, which is rare, there is only a 1 and a 3% chance, respectively, to either adopt the Tippy-Tap HWWS facility or otherwise. By implication, this evidence attests to how the intervention has helped the cause of women and female children in the district.

The present study assessed the post-implementation of community-level promotions of HWWS in the Binduri District to identify the drivers of households' responses to such interventions so that successive initiatives and policies could be informed and guided by those influential factors. The study concludes that personal (gender, education, age, occupation), household (household size), and community (social status) level characteristics are the key drivers of HWWS adoption and practice in the study area. Adoption of HWWS practice is typically associated with personal and community-level characteristics, while household factors were adversely correlated. There is relatively low adoption of HWWS by people below 52 years, and those uneducated compared with those above 52 years, and the educated. These findings reinforce the need for increased health awareness through HWWS sensitisation with a due emphasis on the younger generation, uneducated households, and those without occupation and no generative income. HWWS should be made a public policy issue, and intensive public education is implemented to emphasise its health benefits so that households will adopt and maintain HWWS behaviour as a habit and a priority. Staff of the EHSUs at the district level should be incentivised to identify households which do not practice HWWS to be educated and encouraged to own and use them. The study recommends further studies in the areas of empirical understanding and verification of whether or not the HWWS facilities are used, as well as whether or not their widespread adoption has had a backstopping effect on reducing hygiene-related diseases in the communities. In addition, it is important that further studies explore factors influencing the low adoption of HWWS by people aged below 52 years.

The project implementation was carried out by UNICEF through the Regional Environmental Health Units (REHU) of the UER in accordance with the required international ethical standards. The cross-sectional survey was conducted through the guidance and staff of REHU. Thus, the necessary ethical protocols by WHO/UNICEF were followed. Anonymity and confidentiality were observed for respondents who did not want to be tagged.

We are grateful to Mr Charles Awuni and Mr Juventius Awinsone Asayuure, all of REHU, Bolgatanga, UER.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.