Abstract
Introduction Behaviour change interventions have the potential to improve sanitation and hygiene practices in urban settings. However, the evidence on which behaviour change interventions are effective is unclear. This scoping review assesses the effectiveness of behaviour change interventions on sanitation and hygiene practices in urban settings.
Methods We performed electronic searches across five databases and one grey literature database to identify relevant studies published between 1 January 1990 and 20 November 2023 in English. Eligible study designs included randomised and non-randomised controlled trials with a concurrent control. Studies were eligible for inclusion if they reported a behaviour change intervention for improving sanitation and/or hygiene practices in an urban setting. Individual behaviour change intervention components were mapped to one of nine intervention functions of the capabilities, opportunities, motivations, and behaviour (COM-B) framework. Risk of bias was assessed for each study using an adapted Newcastle-Ottawa scale.
Results After de-duplication, 8,249 documents were screened by abstract and title, with 79 documents retrieved for full-text screening. We included 13 studies ranging from low- to high-quality. The behaviour change interventions had mixed effects on sanitation and hygiene practices in urban settings. Specifically, interventions improved latrine quality but not safe child faeces disposal. Interventions often improved handwashing with soap at key times and sometimes increased the presence of soap and water at the handwashing facility. There is limited evidence on the effect on food hygiene practices. Most study outcomes were measured between 6 and 12 months after intervention implementation, which may undermine the sustainability of behaviour change interventions.
Conclusion Despite mixed effects on sanitation and hygiene outcomes, behaviour change interventions can improve certain practices in urban settings, such as latrine quality improvements and handwashing with soap at the household or compound level. More ambitious behaviour change interventions are needed to reduce disparities in sanitation and hygiene access in urban areas globally.
Introduction
Addressing sanitation and hygiene in urban areas, particularly in informal settlements, is essential for achieving Sustainable Development Goal (SDG) targets on water, sanitation, and hygiene (WASH) (SDG target 6.2). Lack of access to sanitation and hygiene is associated with enteric (1) and respiratory infections (2). In 2022, an estimated 36% of urban residents did not have access to safely managed sanitation, 25% of whom had access to basic sanitation, as defined by the WHO and UNICEF Joint Monitoring Programme (JMP) (3). It was also estimated that 17% of urban residents did not have access to basic hygiene services, with 10% having access to a limited hygiene facility and 7% with no access at all in 2022 (3). However, urban coverage of safely managed sanitation and basic hygiene services varied between countries and regions. Access to basic hygiene services was lowest in sub- Saharan Africa and Oceania and lowest in sub-Saharan Africa for access to safely managed sanitation (3). There are also marked sub-national disparities in access to sanitation and hygiene between high- income and low-income urban areas. Informal settlements, where approximately one-quarter of the global urban population resides, often lack formal WASH services (4). In the urban United States, in 2019, almost one million persons lacked access to at least basic sanitation, especially among people experiencing homelessness and substandard housing (5).
Safe sanitation and hygiene practices, such as latrine use, safe handling and disposal of faeces, handwashing with soap at key moments (e.g., after using the toilet, before food preparation), and hygienic food preparation and storage practices, are important for the prevention of communicable diseases, especially in urban areas where population densities are high. Two previous reviews found that behaviour change interventions achieved mixed results for improving WASH behaviours, such as handwashing with soap (6,7). However, these reviews did not disaggregate results between urban and rural settings. An understanding of behaviour change interventions that have been implemented specifically in urban settings can inform more effective future interventions, as well as identify areas for future research.
Behavioural frameworks and theories have been used to develop and design interventions to target sanitation and hygiene behaviours. Specifically, they can be used to understand the various factors that drive behaviour. For example, Michie et al. (8) developed a behavioural framework, known as the capabilities, opportunities, motivations, and behaviour (COM-B) framework (8), widely used in behaviour change programming. The COM-B framework identifies nine intervention functions that can be used to identify determinants of behaviour (8). In this scoping review, we identify and map the behaviour change intervention components used in the included studies to the nine intervention functions of the COM-B framework and assess their effectiveness, individually or in combination, on targeted sanitation and hygiene behaviours.
The aim of this scoping review is to evaluate the effectiveness of behaviour change interventions on household-level sanitation and hygiene practices in urban settings. The objectives are to: 1) identify household-level behaviour interventions targeting sanitation and hygiene practices in urban settings, 2) map the intervention components to the COM-B framework intervention functions, 3) assess their effectiveness, and 4) identify evidence gaps for future research.
Methods
This scoping review follows the five steps for scoping reviews outlined by Arksey and O’Malley (9). The five steps are summarised below in relation to this review. A scoping review, as opposed to a systematic review, was selected to explore the breadth of available literature and to iteratively search and review documents and extract relevant data. The protocol for this scoping review was pre-registered on OSF registries (https://osf.io/qghl). We used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMAScR) guidelines (10). A PRISMAScR checklist is included in the Supplementary Information (Table A1).
Step 1: Specify the research question
This scoping review seeks to answer the following question: “what is the effectiveness of behaviour change interventions targeting household-level sanitation and hygiene practices in urban settings?” The research question is deliberately broad to allow for a comprehensive mapping of behaviour change interventions to sanitation and hygiene outcomes.
Step 2: Identify the relevant literature
We searched five databases to identify peer-reviewed literature: 1) PubMed, 2) Medline, 3) Global Health, 4) Cochrane Library, and 5) Web of Science. The grey literature search was conducted in the World Bank e-library database. Search terms related to behaviour change, sanitation and hygiene, and urban settings were combined with Boolean operators to search the databases, with search strategies adapted for each database. An example search strategy is published in the Supplementary Materials (Table A2). Searches were conducted on 20 November 2023. The search was limited to studies published in English and from 1 January 1990 onwards. The publication cut-off date was selected based on the introduction of the Millennium Development Goals in 1990 (MDG7: “To ensure access to drinking water and sanitation for all”). The reference lists of included studies and similar systematic reviews (6,7,11) were also hand-searched to identify additional relevant references.
Step 3: Study selection
Only studies with interventions that sought to change behaviour evaluated against a concurrent control group were included. The comparison is the use of a programme with no promotional approach or other promotional approaches. Eligible study designs were randomised controlled trials (RCTs), including cluster-RCTs, quasi-RCTs, and non-randomised controlled trials and pre-post studies. There was no restriction on target population. We included individual, household, and community-level interventions with the specific aim to improve sanitation and/or hygiene practices in urban settings.
An urban setting is broadly characterised by high population density, the concentration of administrative bodies, infrastructure, and services, and income generation activities (12). The specific criteria for what constitute an urban setting varies by country and is usually defined by national governments. Urban areas also include informal settlements (13), slums (13), as well as people experiencing homelessness in unsheltered urban locations (5). We relied on author-reported definitions of an urban setting.
In this scoping review, we used the COM-B framework intervention functions to identify and classify the behaviour change approaches used in the included studies. The COM-B framework includes three domains and nine intervention functions, as defined by Michie et al. (8). The first domain, capability, is defined as having the necessary physical ability, stamina, skills (physical capability), or knowledge (psychological capability) to engage in the activities involved in performing a behaviour. Second, opportunity relates to factors that lie outside the individual and that influences one’s ability to perform a behaviour, such as physical opportunity (i.e., resource availability) or social opportunity (i.e., social norms). Third, motivation refers to the “brain processes that energize and direct behaviour” and can be triggered by fear or disgust, for example. The nine intervention functions, or broad categories of things one can do to change behaviour, include: i) education, ii) persuasion, iii) incentivisation, iv) coercion, v) training, vi) restriction, vii) environmental restructuring, xiii) modelling, and ix) enablement (Table 1) (8).
The two primary outcomes for this review were household sanitation and hygiene practices. Relevant sanitation outcomes included latrine use, latrine construction or rehabilitation, building a septic tank, lining a pit, safe faeces handling and disposal (including child faeces), connecting to a piped sewer network, formal safe pit emptying, and latrine cleanliness. Hygiene outcomes included those related to handwashing with soap, handwashing with soap at key times (e.g., before eating, before food preparation, after visiting the toilet, after children’s faeces disposal or cleaning the baby’s bottom, or other key times defined in the studies), and handwashing facility construction. Food hygiene outcomes included boiling or reheating food before eating, using safe drinking water to prepare food, and hygienic storage of food (e.g., food covered with lid or refrigeration).
All documents retrieved from electronic searches were transferred to Endnote for de-duplication. To identify relevant documents, three reviewers (CM, KD, and MM) screened documents by title and abstract, excluding only clearly irrelevant documents, i.e., not related to sanitation and hygiene behaviour change interventions and urban settings. Full texts of all potentially eligible documents were then retrieved and independently double-assessed for inclusion by three reviewers (CM, KD, and MM). Any disagreement between reviewers about eligibility following title and abstract screening was resolved through discussion to build consensus. Disagreement was resolved through discussion with a fourth reviewer (RD) where consensus could not be reached.
Step 4: Extract, map, and chart the data
Study characteristics and results from included studies were double-extracted independently by three reviewers (CM, KD, and MM) using a standardised data extraction template in MS Excel and then cross-checked for accuracy. As with inclusion, a fourth reviewer (RD) provided arbitration if agreement on data extraction could not be reached. The data extraction form included information on study characteristics, such as author, publication date, study design, study dates, study location and urban setting, target population, and sample size. We also extracted data on behaviour change approaches and sanitation and hygiene outcomes (Supplementary Materials Table A4). Only results for intervention arms targeting sanitation and hygiene behaviours were extracted.
Step 5: Summarise, synthesise, and report results
First, intervention components for each intervention arm were mapped to one of the nine intervention functions of the COM-B framework (8). Second, we recorded the measure of effect, 95% confidence interval, and p-value for each outcome. Third, we summarised the results of the included studies to describe the effect of the behaviour change interventions on the sanitation and hygiene outcomes. Finally, we identified evidence gaps.
Risk of bias (quality) assessment
We assessed risk of bias in individual studies using an adapted Newcastle-Ottawa scale, as used in previous systematic reviews (1,2,8). The scale considered seven areas of bias: selection, response, follow-up, misclassification, outcome assessment, outcome measurement, and analysis bias. Each study received a score of up to nine, a higher score indicating a smaller risk of bias. Risk of bias was assessed by one reviewer (CM) for each study with a subset of scores reviewed by a second reviewer (KD). Any discrepancies between the two reviewers were resolved by a third reviewer (RD).
Results
Search results
Electronic searches were conducted on 20 November 2023, identifying 9,771 records. After removing duplicates, 8,218 records were screened by abstract and title. Thirty-one additional documents were identified through reference screening. Seventy-nine documents were sought for retrieval for full text screening. Thirteen documents were included in the review (Figure 1). The 65 documents excluded during full text screening are listed with reasons for exclusion in the Supplementary Materials (Table A5). Most studies excluded during full text review were either conducted in rural settings or did not have a control group.
PRISMA flow chart.
Description of included studies
The 13 included studies consisted of 12 peer-reviewed studies and 1 grey literature report (Table 1). Study designs included five cluster-randomised controlled trials (RCTs), three RCTs, three non- randomised trials, and two quantitative process evaluations. The two process evaluations were a controlled before-and-after study (14) and a cRCT (15). The studies were published between 2002 and 2022, though most studies were published after 2015. Almost all studies were conducted in sub- Saharan Africa (46%, n= 6) (1 in Côte d’Ivoire, 1 in Tanzania, 1 in Zambia, 1 in Uganda, 1 in Mozambique, and 1 in Kenya) or South Asia (46%, n= 6) (Bangladesh n= 3, Pakistan n= 2, and Nepal n= 1). One study was conducted in Latin America (Peru, n= 1). Eleven studies were in lower-middle-income countries (Bangladesh, Côte d’Ivoire, Tanzania, Kenya, Zambia, Pakistan, and Nepal), while two studies were in low-income countries (Mozambique and Uganda) and one in an upper-middle-income country (Peru). Seventy percent (n= 9) of studies were implemented in low-income urban areas, also called informal settlements, slums, or shanty towns among the included studies. One study was conducted in a commune within the city, while three were in peri-urban areas.
Types of interventions
We identified 16 intervention arms across the 13 studies. The most frequently used COM-B framework intervention functions were environmental restructuring (n= 10), education (n= 10), and persuasion (n= 8) (Table 2). Most environmental restructuring was implemented in combination with other intervention functions (n= 10), such as education (n= 4), persuasion (n= 1), modelling (n= 1), and more than one intervention function (n= 4). Environmental restructuring alone was implemented in only one study (17). Two studies developed an intervention based on psychosocial theory, such as RANAS (24,25). Most studies compared interventions to a control group only (n= 10), while three studies compared two interventions to one another and to a control group. The interventions mainly targeted the individual level (54%, n= 7), such as users of shared sanitation or caregivers of children, while the other studies targeted the household level, for example, those with unimproved sanitation facilities (46%, n= 6). Thirty-one percent of studies (n= 4) included intervention components that were non- sanitation or hygiene related that were excluded from the scoping review analysis, e.g., received a cholera vaccine (18), water treatment (19,23), and water quality test results (20).
Outcomes
We identified 36 sanitation and hygiene outcomes across the 13 included studies. Fourteen outcomes related to sanitation and 21 to hygiene. Of the 21 outcomes related to hygiene, 19 were on hand hygiene and two were on food hygiene. The sanitation outcomes specifically related to latrine quality (n= 13) (14,16,24,25) and safe child faeces disposal (n= 1) (26). Hand hygiene outcomes included handwashing with soap and at various key moments (e.g., after using the toilet or before eating) (n= 13) (15,17,19,21–23) or the presence of soap and water at the household handwashing facility (n= 6) (14,18–21). The two food hygiene outcomes included using clean utensils for infant feeding and hygienic storage of leftover food (15).
Outcome definitions and measurement timepoints varied across the studies. Latrine quality was measured as either latrine cleanliness, such as no visible faeces in pan (16), having a rotation cleaning system (14,24), and cleaning frequency (25), or latrine privacy, measured as having an indoor and outdoor lock (14,24). Sanitation-related outcomes relied on both fieldworker observations (n= 8) and self-reporting (n= 5). Hygiene outcomes included either observed handwashing with soap (n= 12), self- reported handwashing with soap (n= 1), observed presence of a handwashing facility with soap and water (n= 6), and observed food hygiene practices (n= 2). However, each study measured handwashing with soap differently and often before or after different key moments. For example, one study reported daily handwashing rates (20), while another study reported the proportion of occasions participants washed their hands with soap after using the toilet (17). The observed presence of a handwashing with soap, often used as a proxy measure for handwashing, was measured consistently across studies, though measured at different timepoints after intervention implementation. Over half of the outcomes were measured between 6 months and 12 months after baseline (54%, n= 7), with two studies that measured outcomes less than six months after baseline (15,17). Two studies measured the medium and long-term effects of a behaviour change intervention on hygiene 18 months and 5 years post- intervention (19,23). Two studies did not state when they measured outcomes (20,25).
Intervention effectiveness
There are mixed results on the effect of behaviour change interventions on sanitation and hygiene outcomes in urban settings. Interventions were associated with improved sanitation and hygiene practices in 27 out of the 36 outcomes, while there was no effect on 9 out of 36 outcomes. Interventions had a positive effect on 13 out of 14 outcomes related to sanitation. However, intervention effects on handwashing were mixed as 13 out of 20 outcomes had a positive effect and 7 out of 20 outcomes had no effect. The only two food hygiene outcomes included in this review had mixed results.
Effectiveness of sanitation interventions
Behaviour change interventions improved latrine quality (14,16,24,25), but did not improve safe child faeces disposal (26) (Table 3). Alam et al. (16) found that compounds that received a latrine cleaning intervention were significantly more likely to have cleaner toilets (e.g., no visible faeces on latrine pan) after the intervention compared to controls. Tidwell et al. (24) found intervention households were more likely to have cleaning rotas, inside and outside latrine locks, and toilets with simple covers or water seals. Tumwebaze et al. (25) found that shared toilet users that received either one of two latrine cleaning interventions were significantly more likely to clean their shared latrine. Bick et al. (14) found that compounds receiving an improved sanitation intervention were much more likely to be private, almost twice as likely to be observably clean, and twice as likely to be well-maintained. In addition, individual cleaning frequency was significantly higher and frequent collective cleaning was reported more often by intervention respondents. However, intervention compounds were unlikely to have and adhere to a formal rota for cleaning shared latrines. Yeager et al. (26) found that intervention respondents were not more likely to practice safe child faeces disposal than control respondents.
Each circle represents one outcome reported in an individual study. A red circle indicates a negative or no effect on the reported outcome, while a green circle indicates a positive effect on the reported outcome.
Effectiveness of hygiene
Behaviour change interventions had mixed results on improving hygiene practices in urban settings (Table 2).
Observed presence of soap and water at the latrine
Behaviour change interventions sometimes increased the observed presence of soap and water at the handwashing facility. In Dhaka, Bangladesh, Biswas et al. (18) found that intervention respondents were significantly more likely to have water and soap or soapy water present at the hand washing place than control respondents. In another study in Dhaka, Bangladesh, one intervention arm (environmental restructuring and modelling) increased the presence of soap and water at the latrine, while another intervention arm (education, persuasion, coercion) did not (21). In Karachi, Pakistan, Luby et al. (23) followed up participants 18 months after a behaviour change intervention ended (27) and found that mothers in households originally assigned to the intervention were 1.5 times more likely to have a handwashing facility with soap and water. Five years after the same intervention followed up by Luby et al. (23), Bowen et al. (19) reported that intervention households were 3.4 times more likely to have a handwashing station with soap and water than control households. Bick et al. (14), found that few intervention latrines had signs of soap use at the household handwashing facility with no difference from control latrines in Maputo, Mozambique.
Observed hygiene practices
Most behaviour change interventions improved observed handwashing with soap at key moments. Amon-Tanoh et al. (17) found that the ‘environmental restructuring, persuasion, and enablement’ intervention was effective at increasing handwashing with soap after toilet use, while environmental restructuring alone (provision of handwashing stations) had little effect. In Langford et al. (22), mothers in the intervention group were significantly more likely to wash hands with soap in four out of five key junctures compared to the control group: after cleaning a baby’s bottom, before cooking, feeding the baby, and eating. In Simiyu et al. (15), the intervention improved caregiver handwashing with soap before food preparation but had no effect on caregivers’ handwashing practices before infant feeding. In Bowen et al. (19), intervention households more commonly reported handwashing before cooking and before meals than control households five years after the intervention. In another study, one intervention arm (environmental restructuring and modelling) increased handwashing after last defecation, while another intervention arm (education, persuasion, and coercion) did not (21). The only study that investigated food hygiene practices found that the intervention improved observed hygienic feeding of infants (i.e., using a utensil) but had no effect on caregiver hygienic food storage practices (15) (Table 1).
Reported hygiene practices
Only one study reported on handwashing behaviour. The study found no significant differences between the intervention and control groups were observed in reported handwashing behaviours (20).
Risk of bias
The risk of bias ranged from high to low for sanitation and hygiene outcomes. In most studies, participants and those collecting the data could not be blinded to intervention allocation, but randomised study design, low loss to follow up, and blinding of data analysts contributed to higher Newcastle-Ottowa Scale scores. The self-reported or observation of outcomes, as well as high rates of loss to follow up, led to lower scores. Three studies were not randomised and seven had a loss to follow up >10%. Most sanitation and hygiene outcomes were observed. The full assessment is in the Supplementary Materials (Table A5).
Discussion
This scoping review included 13 studies evaluating the effect of behaviour change interventions on sanitation and hygiene practices in urban settings. The results suggest that behaviour change interventions can improve certain sanitation and hygiene practices, such latrine quality and handwashing with soap at key moments. There is mixed and limited evidence on the effectiveness of behaviour change interventions on other outcomes, such as safe child faeces disposal and food hygiene practices. The 13 studies were implemented in 10 countries and primarily in urban informal settlements in sub-Saharan Africa and South Asia. No studies were conducted in high-income countries where disparities in access to sanitation and hygiene remain in urban areas. We also note that two excluded studies were implemented in both urban and rural settings but did not disaggregate results by setting (28,29). Compared to sanitation and hygiene interventions targeting rural areas, the evidence base is much more limited for the urban setting. Evidence specific to behaviour change interventions in urban settings is important for addressing the sanitation and hygiene challenges in this context.
Most studies included at least two intervention functions in their behaviour change interventions, thereby limiting the ability to tease out the specific effect from each intervention component. Almost all studies related to hand hygiene relied on environmental restructuring in combination with more traditional forms of interpersonal communication (e.g., persuasion, education, modelling). Among the environmental modification interventions, the majority focused on hardware provision, specifically providing an improved handwashing station to households. Nudges or environmental cues, which have generally shown to improve behavioural outcomes (30–32), warrant further exploration in urban settings. Other behaviour change approaches, such as community mobilisation, social marketing, advocacy, and financial incentives also warrant further exploration. For example, several studies have reported that willingness-to-pay for sanitation products and services is well under market prices in low-income urban areas (33–35). Financial incentives, such as income- or area-based subsidies, may bridge the gap between cost and willingness to pay for improved sanitation and safe emptying services (35).
The interventions targeted a narrow range of sanitation and hygiene behaviours. Latrine quality improvements were the most targeted sanitation-related behaviour. While sanitation quality can be an important predictor of sanitation use (36), latrine quality improvements alone may have limited impact for reaching SDG targets. Only one study evaluated safe child faeces disposal, and no studies targeted the use of latrines, safe pit emptying, or faecal sludge management. In addition, no studies evaluated the use of novel sanitation technologies designed for the urban marketplace, such as container-based sanitation (CBS). Handwashing with soap was the most targeted hygiene behaviour. Only one behaviour change intervention targeted food hygiene behaviours, which highlights an important evidence gap.
Outcome definitions and measurement timepoints varied significantly across the studies, thus making it difficult to compare results. For example, handwashing with soap was either measured via structured observation, self-reported behaviour, or proxy measures. Most included studies used structured observations of handwashing behaviour, often considered the gold standard for measuring behaviour., though more resource-intensive (37). Alternatively, some studies used the presence of a handwashing facility with soap and water as a proxy measure for handwashing behaviour. While this method allows for rapid and low-cost data collection, it’s accuracy may be limited (38). In addition, one study relied on self-reporting handwashing, which is prone to recall bias (37). Outcome measurement time points ranged from 6 to 12 months post-intervention, with only one study investigating long-term intervention effect. The findings suggest that behaviour change interventions were overall effective at improving certain sanitation and hygiene practices, but it is unclear whether they are effective long- term.
We note the limited scope of robust, large-scale interventions addressing sanitation and hygiene at the municipal or community-level. All interventions included in this review focused on household- or compound-level improvements. We considered compound-level interventions as household interventions as the interventions were delivered to a relatively small sample of households within clusters of compounds. No studies explicitly addressed community-level behaviour or behaviour change nor did they focus on connecting to municipal water or sewerage systems. With recent emphasis of urban sanitation programmes on Citywide Inclusive Sanitation (39), rigorous evaluation of efforts to improve urban sanitation are needed. While consistent water supply is necessary for safe sanitation and effective hygiene behaviours, only one included study adjusted for water supply in their analysis (24) and only one study provided information on water supply at baseline (16).
This scoping review has several limitations. First, our search was limited to English and may have missed relevant documents published in other languages. Second, we searched one grey literature database and may have missed additional relevant grey literature published elsewhere. Third, the outcomes amongst the included were too heterogeneous to conduct a meta-analysis. We also did not evaluate publication bias. Fourth, due to the nature of the interventions, blinding of participants and enumerators was often not possible, which may lead to outcome measurement bias. Results included in the review may also have been biased due to self-reporting of sanitation and hygiene outcomes. Finally, only one study evaluated food hygiene, which limits the generalisability of results.
Conclusion
Our results suggest that behaviour change interventions have the potential to improve sanitation and hygiene practices in urban settings, such as latrine cleanliness and handwashing with soap at the household or compound level. However, more ambitious interventions should be evaluated to increase their impact. Opportunities for future interventions include evaluating community-level behaviour change interventions, connecting households to water or sewerage networks where available, CBS acceptability, uptake and use, and food hygiene practices. Nonetheless, this review highlights that behaviour change is an important component of interventions for sanitation and hygiene in urban settings.
Conflict of interest
The authors have declared that no competing interests exist.
Funding statement
This research was funded by the United Kingdom’s Foreign, Commonwealth and Development Office (grant code: 301186). The funder had no role in study design, data collection and analysis, decision to publish, nor preparation of the manuscript.
Data Availability
All data produced in the present work are contained in the manuscript
Supplementary materials
References
