Interventions to address unsafe child feces disposal practices in the Asia-Pacific region: a systematic review

In 2017, an estimated 2 billion people lacked access to basic sanitation services, with 673 million practicing open defecation (UNICEF & WHO 2019). Each year, the diarrheal disease causes over 800,000 deaths due to a lack of access to water and sanitation, of which nearly 300,000 are children (Prüss-Ustün et al. 2019). Eliminating open defecation is one of the Sustainable Development Goals (SDGs) and has been a major water, sanitation, and hygiene (WASH) sector priority, particularly in the Asia-Pacific region where rates of unsafe feces disposal can exceed 50% (UNICEF & WHO 2019). Most interventions to reduce open defecation have focused primarily on adults, even though the safe disposal of child feces is just as essential from a public health standpoint (Chakma et al. 2008; Majorin et al. 2014; Rand et al. 2015). The United Nations Children's Fund and World Health Organization's Joint Monitoring Programme (JMP) defines safe child feces disposal (SCFD) as a child defecating directly into an improved latrine or a caregiver depositing or rinsing child stools into an improved latrine (UNICEF & WHO 2019). Unsafe CFD (unsafe CFD), therefore, involves open defecation by children, failure to remove child feces from the open, or co-disposal of child feces with other solid wastes if that solid waste is not disposed of in a sanitary manner (Majorin et al. 2019; UNICEF & WHO 2019).

The Asia-Pacific region has particularly high rates of unsafe CFD; in India, for example, the JMP estimates that 60% of child feces are disposed of unsafely (UNICEF & WHO 2019). This percentage, however, is estimated based on latrine access rather than actual CFD practices, so the true prevalence of unsafe CFD is likely underestimated. Majorin et al. (2014) and Preeti et al. (2016) found rates of unsafe CFD to be 81 and 73% in India, respectively – substantially higher than the JMP estimate. Households practicing unsafe CFD show a higher prevalence of childhood diarrheal disease compared to those practicing safe disposal (Bawankule et al. 2017; Islam et al. 2020). Unsafe CFD is associated with environmental enteropathy in children, a disease of chronic intestinal inflammation, in addition to impaired growth and development (George et al. 2016).

Development of effective interventions to address unsafe CFD requires an understanding of the drivers behind this behavior. One of the most significant determinants of CFD practices is the use of a toilet by adults in the household (WSP 2015). Bauza et al. (2019) found that CFD was more likely to be unsafe if female adults in the household did not defecate into a toilet, and Islam et al. (2018) found unsafe CFD to be more common in rural areas, which may be due to the higher prevalence of open defecation by adults in these areas (UNICEF & WHO 2019). A study conducted by Beardsley et al. (2021) evaluated factors associated with unsafe CFD in India and found that the poorest wealth quintile was significantly more likely to practice unsafe CFD, as well as households in which females did not attend primary school. Preeti et al. (2016) and Ayele et al. (2018) found parents’ educational attainment to be significantly associated with feces disposal practices, with less educated households more likely to practice unsafe CFD. Lower-income households’ association with unsafe CFD is potentially due to poor access to child defecation materials or the lack of income to purchase nappies or latrines that children can use (Miller-Petrie et al. 2016; Preeti et al. 2016; Ayele et al. 2018).

In addition to household income and educational attainment, access to a sanitation facility, whether improved or unimproved, was also significantly associated with SCFD (Beardsley et al. 2021). The presence of water at the handwashing site and child age have also shown associations with SCFD (Azage & Haile 2015; Islam et al. 2018; Bauza et al. 2019). The prevalence of unsafe CFD was found to be higher when there was no water present for handwashing, as well as when the children in the household were younger (Azage & Haile 2015; Islam et al. 2018; Bauza et al. 2019). In terms of child age, Azage & Haile (2015) and Islam et al. (2018) reported that the feces of children younger than 18 months and younger than 12 months of age, respectively, were more likely to be disposed of unsafely compared to that of older children. This may be due to the perception that younger children's feces are less harmful or are less likely to cause sickness than those of older children or adults (Brown et al. 2013; Ayele et al. 2018).

Though clear evidence of the adverse health impacts of unsafe CFD continues to arise, there is little evidence on the effectiveness of interventions targeting the improvement of CFD practices in the Asia-Pacific region (Rand et al. 2015). Morita et al. (2016) and Majorin et al. (2019) conducted systematic reviews of interventions to improve CFD practices to prevent the disease, but both had a global focus and found inconclusive evidence on the effectiveness of these interventions. Morita et al. (2016) and Majorin et al. (2019) both identified only one intervention whose sole focus was to improve CFD behaviors (Yeager et al. 2002). This sole study did not result in a significant reduction of unsafe CFD behaviors. All other studies assessed by Morita et al. (2016) and Majorin et al. (2019) evaluated packaged WASH interventions in which CFD was only a small component, suggesting that the improvement of CFD behaviors has not been a research priority. Majorin et al. (2019) suggested that children should be encouraged to use latrines and that child feces should be disposed of into latrines; however, the evidence reviewed was of low certainty, and the authors were unsure about the interventions’ true effects.

The objectives of the present study are (1) to identify both behavior change and hardware interventions targeting the improvement of CFD practices in the Asia-Pacific region, (2) to assess the strengths and weaknesses of the identified behavior change and hardware interventions, and (3) to compare the effectiveness of behavior change versus hardware interventions at improving CFD practices in order to provide recommendations for promoting SCFD.

A systematic review of published literature was conducted using PubMed, Embase, and the Global Health database within EBSCO. Search terms were related to children, feces disposal, and Southern and Eastern Asian countries. Medical subject heading (MeSH) terms in PubMed and Emtree terms in Embase were used to facilitate searches and identify the full body of literature. Terms for feces disposal were limited to titles and abstracts by utilizing the [tiab] search feature in PubMed and its equivalents in Embase and the Global Health database. Search terms used in this our study are presented in the Supplementary Materials. Initial searches including the Southern/Eastern Asia terms yielded few results; searches were modified to exclude these terms, and articles were manually screened to include only studies that took place in this geographic region.

Database searches were limited to articles published in English in 2011 or later, and the final search was conducted on May 31, 2021. Only studies conducted within this 10-year period were included to reflect new evidence of interventions to improve CFD practices. All studies were uploaded to Covidence, a systematic review production online tool, and duplicate studies were removed. Two reviewers screened the remaining studies by title and abstract for relevance, with a third reviewer acting as a tie-breaker. The final selection of studies occurred after full-text review.

Studies selected for inclusion must have included interventions where changes in CFD practices were measured as an outcome. These included interventions with a specific CFD component as well as hardware interventions that would enable CFD in latrines and toilets. Studies on interventions that only measured open defecation by adults and review articles were excluded. Results were then synthesized by study characteristics such as age group, sample size, intervention time frame, intervention description (hardware or behavioral, general summary), and behavioral outcomes and findings. Quality assessment was performed by two reviewers and involved describing the level of evidence (Ackley et al. 2008; Brownson et al. 2009; Manga et al. 2021) and risk of bias (Sterne et al. 2016; Higgins et al. 2022) that were then used to rate overall certainty.

The level of evidence was assigned to studies based on the methodological quality of their design and applicability. Levels were ranked as A, B, C, or D, with A being the highest level of evidence. Because systematic reviews were excluded from this review, Level A was reserved for randomized controlled trials (RCTs) (Ackley et al. 2008). Other peer-reviewed research such as cohort and cross-sectional studies were rated Level B. Level C included government reports, and Level D was assigned to pilot studies and formative research (Brownson et al. 2009).

The risk of bias was evaluated using the risk-of-bias tool (RoB 2) for RCTs (Higgins et al. 2022) and the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) tool for other study types (Sterne et al. 2016). RoB 2 analysis involved the evaluation of bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in outcome measurements, and bias in the selection of the reported result (Higgins et al. 2022). Domains in ROBINS-I that are not included in RoB 2 include bias due to confounding, selection bias, and bias in the classification of interventions (Sterne et al. 2016).

Overall certainty ratings imply confidence that the true effect lies close to the estimated effect determined in the study. Studies at Level A were initially given a high certainty, studies at level B were rated intermediate, and studies at Level D were initially rated as low. Overall certainty was then either increased or decreased based on the risk of bias estimates. Based on the information obtained through this review, recommendations for future programs and interventions were provided.

The 29 studies excluded after full-text screening either did not include interventions where changes in CFD practices were measured as an outcome (i.e., ‘Wrong study design’), were conducted outside of the Asia-Pacific region, were outside of the 10-year time frame, or were duplicates of another article.

For example, Bawankule et al. (2017) assessed associations of unsafe CFD with childhood diarrhea in India but did not describe an intervention to address those factors, and this study was therefore excluded. Majorin et al. (2019) and Igaki et al. (2021) both conducted systematic reviews of interventions that included improving CFD practices among other factors, but reviews were not an included reference type, so these studies were excluded.

Gimaiyo et al. (2019) and Ellis et al. (2020) assessed CFD practices in Kenya, which is not in the Asia-Pacific region. Studies conducted by Stanton & Clemens (1987) and Baltazar & Solon (1989) were also excluded as they were outside of the 10-year time constraint.

The 15 studies selected for inclusion were evaluated for level of evidence, risk of bias, and overall quality. A more detailed summary of the study findings is shown in Table 1. Quality assessment results are shown in Table 2.

India's Total Sanitation Campaign consisted of government subsidies to improve sanitation access for those below the poverty line. MANTRA was another program in rural Odisha, India delivered by a non-governmental organization that aimed to increase latrine ownership by encouraging all households in a village to build their own latrines, with piped water provided through the program. These two interventions required participants to construct their own pour-flush latrines for household use. While these hardware interventions led to increased access to improved sanitation, high rates of unsafe CFD still persisted (Majorin et al. 2014; Bauza et al. 2019; Reese et al. 2019). Only 35% of villages reported SCFD following the MANTRA program (Bauza et al. 2019; Reese et al. 2019), and less than 30% reported SCFD following the Total Sanitation Campaign (Majorin et al. 2014; Patil et al. 2014; Freeman et al. 2016). Bauza et al. (2019) and Reese et al. (2019) are included together as one study reported in two papers, as Bauza et al. conducted a deeper dive into the CFD aspects of the same program. These study findings suggest that although improved sanitation is a necessary component of SCFD, its sole implementation does not guarantee SCFD as other components are equally important, particularly behavior change interventions directed toward caregivers of young children.

The WASH Benefits Bangladesh trial and Sanitation, Hygiene Education and Water Supply in Bangladesh program (SHEWA-B) were two other large-scale interventions implemented in rural Bangladesh. In contrast to the Total Sanitation Campaign and the MANTRA program, WASH Benefits and SHEWA-B also included behavioral components. WASH Benefits used community promoters to encourage mothers to teach their children to use potties, to safely dispose of feces in latrines, and to regularly remove human and animal feces from the compound. SHEWA-B also involved home visits by community promoters, in addition to community meetings and social mobilization activities. WASH Benefits showed decreases in the presence of visible feces around the household (Luby et al. 2018; Parvez et al. 2018). Observed safe disposal of feces from child potties was moderate (Parvez et al. 2018). Neither WASH Benefits nor SHEWA-B showed statistically significant improvements in CFD behaviors or child defecation in potties or latrines (Huda et al. 2012; Parvez et al. 2018).

In addition to large-scale improved sanitation programs, small-scale sanitation hardware was a component of some interventions. These included the provisioning of tools to improve CFD practices, such as mini-hoes, sani-scoops, and child potties (Ashraf et al. 2012; Sultana et al. 2013; Hussain et al. 2017). Ashraf et al. (2012) and Hussain et al. (2017) conducted small-scale hardware interventions that also included a behavioral component. Ashraf et al. (2012) utilized Community Health Promoters, who visited households three times during the study duration and used cue cards, images, and text messages to encourage sanitation use. Hussain et al. (2017) conducted group discussions and individual interviews throughout a 30-day period to gather data on potty likes and dislikes, feces disposal practices, barriers to use, and recommendations for improvement. Ashraf et al. (2012) reported an increase in SCFD from 25% at baseline to 73% at 3 months post-intervention, and Hussain et al. (2017) reported an increase from 4 to 85%. These two combined hardware and behavior change interventions reported the greatest increases in SCFD post-intervention of all studies included in the review.

Biswas et al. (2021) conducted a solely behavior change intervention to improve SCFD practices. Mobile voice and text messages were sent to households to promote SCFD behaviors, and home visits were conducted which emphasized the importance of cleaning child feces immediately after a defecation event, noting that child feces were just as dangerous as adult feces. Behavior change messages were well received by participants, and the intervention was adapted to include pictures as those were easier for participants to understand. This formative research study identified that most young children's feces were disposed of unsafely, while older children tended to defecate directly into latrines.

Because of the potential improvements in SCFD practices following provisions of child potties/tools and behavior change messages, scenarios involving combinations tools/potties, presence of toilets, and behavior change components were evaluated.

Sultana et al. (2013) designed and piloted hoe and shovel tools for the removal of feces from rural Bangladeshi courtyards and modified the tools based on community feedback. The most preferred product was a lightweight mini-hoe with a curved edge for containing and removing feces. Mothers also expressed a desire to purchase plastic potties as a place for children to defecate, though few families could afford them.

After scooping feces, mothers reported either disposing of the feces in the nearest bush or ditch, or inside the toilet (Sultana et al. 2013). However, dry leaves and straws were also scooped up with the feces, and mothers did not want to dispose of it in the toilet for fear of blocking the toilet pipe.

Bauza et al. (2019) assessed the impacts of India's MANTRA campaign on SCFD, with breakdowns by defecation location and ultimate disposal location. The feces of children who defecated into a potty were either disposed of in a drain/ditch or left in the open, and none of the feces from potties were disposed of safely in a toilet or latrine.

Majorin et al. (2014) evaluated impacts of India's Total Sanitation Campaign on SCFD and also stratified by defecation location and ultimate disposal location. Among feces defecated into a potty, 14% were disposed of safely in a latrine and 86% were disposed of with other garbage.

None of the studies in this review fell under the category of tools/potties with no toilets present. However, studies conducted by Sultana et al. (2013), Majorin et al. (2014), and Bauza et al. (2019) showed that feces collected by scoops and potties were often not disposed of into a latrine, even when latrines were present.

Ashraf et al. (2012) provided potties and customized hoe-like tools to households, with visits from community health promoters to encourage their use. From baseline to 3 months post-intervention, children's potty use increased from 15 to 84% (p < 0.001), and disposal of feces into latrines increased from 25 to 73% (p < 0.001).

Hussain et al. (2017) provided three options of potties to caregivers of children and conducted semi-structured interviews and group discussions to encourage their use. At baseline, caregivers reported that 4% of children defecated into potties, and 4% of households reported safe disposal of feces into latrines. Post-intervention, 86% of children defecated into potties, with 85% of households safely disposing of feces into latrines.

Caruso et al. (2022) conducted a larger-scale combined hardware and behavioral intervention among 3,723 latrine-owning households in India and assessed its effects on SCFD. In addition to community meeting, household visits, and latrine assessments, mothers of young children were provided with plastic scoops and potties and encouraged to practice SCFD. At 4 months post-intervention, latrine use increased by 6.4 percentage points, and SCFD increased by 15.2 percentage points.

Parvez et al. (2018) and Rahman et al. (2018) evaluated impacts of the WASH Benefits Bangladesh trial on SCFD, which provided sani-scoops, child potties, and improved latrines to households in addition to home visits from community health promoters. Rahman et al. (2018) spot-checked intervention fidelity and found that 80% of intervention households had sani-scoops and potties easily accessible to the mother. However, additional support was required to encourage and teach rural mothers how to train their children to consistently use the potty.

Parvez et al. (2018) monitored child defecation in potties and safe disposal of feces among intervention and control groups following the WASH Benefits trial but did not find significant improvements in potty use or SCFD. The use of child potties during child defecation was moderate, but the use of sani-scoops for cleaning human and animal feces was low.

Biswas et al. (2021) conducted a behavior change intervention with bi-weekly voice and text messages and home visits to promote SCFD behaviors. While most households had access to a latrine, latrines were often shared among multiple households, with many having to travel far distances to reach them. Feces in potties were primarily disposed of in latrines, though some were disposed of in ditches or neighborhood garbage piles particularly during the rainy season when household members had to travel far distances to latrines.

Most households in the combined intervention implemented by Hussain et al. (2017) had access to a latrine, though 37% shared latrines with other households. Those who did not own a latrine but shared with other households generally disposed of the child feces either into the latrine or a designated pit.

The majority of included studies (53% (n = 8)) were hardware interventions only, meaning they provided some sort of sanitation technology to participants without any behavior change component. An additional six studies (40%) included both hardware and behavioral components, while one study (7%) focused solely on behavior change.

Of the 15 included studies, seven (47%) were RCTs rated as Level A, as shown in Table 2. The review also included three cohort studies (20%), one quasi-experimental study (7%), and one cross-sectional study (7%) that were rated as Level B. No government reports were included, so no studies were given Level C. The remaining three studies (20%) were formative research or pilot studies and were evaluated as Level D.

According to the Cochrane guidelines, certain study limitations can increase the risk of bias and therefore decrease the overall certainty rating. A low risk of bias score implies confidence that there were no major or minor sources of bias that could have influenced results. An intermediate risk of bias indicates the presence of one major or several minor study limitations, and a high risk of bias indicates the presence of more than one crucial limitation that may seriously compromise the validity of study findings (Sterne et al. 2016; Anthoj et al. 2020; Higgins et al. 2022).

Seven studies (47%) were scored a low risk of bias. Among these, five were RCTs that had large sample sizes, randomization via random number generator or lottery, little to no missing outcome data, or other outcome indicators measured to mitigate reporting bias that led to the low risk of bias score. Bauza et al. (2019) and Reese et al. (2019) conducted cohort studies with large sample sizes and adjustments for potential confounders that led to their low risk of bias judgment.

Four studies (27%) were scored an intermediate risk of bias. Freeman et al. (2016) and Rahman et al. (2018) conducted RCTs that used self-reported CFD data, one data collection point, and had no control arm for fidelity checks, leading to the intermediate risk of bias score. Majorin et al. (2014) and Lamichhane et al. (2018) conducted studies subject to courtesy, recall, and social desirability bias.

Four studies (27%) were rated a high risk of bias. This rating was due to small sample sizes, the use of convenience sampling, self-reported CFD, potential for courtesy bias, and the lack of follow-up.

Of the seven RCTs, six retained their high certainty score due to a low-to-intermediate risk of bias. The remaining study was ranked down to intermediate certainty as it had more than one major limitation that may have compromised study findings (Rahman et al. 2018).

Of the five Level B studies, two moved upwards to a high certainty rating (Bauza et al. 2019; Reese et al. 2019). Lamichhane et al. (2018) and Majorin et al. (2014) retained their intermediate rating, while Ashraf et al. (2012) moved down to a low rating.

The three formative research and pilot studies all showed a high risk of bias and therefore retained their original low certainty ratings (Sultana et al. 2013; Hussain et al. 2017; Biswas et al. 2021).

Our systematic review sought to identify both behavior change and hardware interventions addressing unsafe CFD practices in the Asia-Pacific region, to assess the quality of identified studies, and to compare the effectiveness of each intervention type. We confirmed the lack of evidence on behavior change interventions to improve CFD practices. Of the 15 studies selected for inclusion, only four focused specifically on improving SCFD practices. The remaining 11 studies evaluated the impacts of large-scale interventions, such as the Total Sanitation Campaign, MANTRA programs, WASH Benefits Bangladesh program, and SHEWA-B, on SCFD practices. These interventions targeted improving child feces management as only a small part of larger water, sanitation, and hygiene interventions. Only one study implemented a solely behavioral intervention (Biswas et al. 2021), and six studies evaluated a combination of hardware and behavioral approaches.

Strictly hardware interventions, such as India's Total Sanitation Campaign and MANTRA program, showed limited improvements in SCFD practices. Only 35% of villages reported SCFD following the MANTRA program (Bauza et al. 2019; Reese et al. 2019), and less than 30% reported SCFD following the Total Sanitation Campaign (Majorin et al. 2014; Patil et al. 2014; Freeman et al. 2016). While the Total Sanitation Campaign and the MANTRA program increased household latrine access, younger and pre-ambulatory children are not able to use latrines directly (Majorin et al. 2014). Other sanitation hardware components, such as child potties and sani-scoops, may be important tools for facilitating SCFD. However, provisioning these tools without a behavior change component may not be sufficient to improve SCFD practices, particularly when there are no latrines present (Sultana et al. 2013; Majorin et al. 2014; Bauza et al. 2019).

Luby et al. (2018), Parvez et al. (2018), and Rahman et al. (2018) also evaluated the impacts of a large-scale intervention, the WASH Benefits Bangladesh trial, on SCFD practices. WASH Benefits aimed to improve water quality, sanitation, handwashing, and child nutrition and feeding – alone and in combination – among rural communities in Bangladesh. In addition to providing sanitation hardware such as latrines and potties, WASH Benefits also included behavioral components, with visits from community promoters to encourage mothers to teach their children to use potties, to safely dispose of feces in latrines, and to regularly remove human and animal feces from the compound. Luby et al. (2018) found significant improvements in latrine access following the trial, and Rahman et al. (2018) found that 80% of intervention households had sani-scoops and potties easily accessible to the mother. However, the use of child potties during child defecation was moderate, and the use of sani-scoops for cleaning human and animal feces was low (Parvez et al. 2018). The studies did not find statistically significant improvements in SCFD after the WASH Benefits trial (Luby et al. 2018; Parvez et al. 2018).

Sultana et al. (2013) also conducted a solely hardware intervention in which they designed a mini-hoe for scooping and disposing of feces. However, caregivers reported that dry leaves and straws were also scooped up with the feces, and they feared disposing of it in toilets and blocking pipes. Parvez et al. (2018) also found low use of sani-scoops for collecting and disposing of feces, in comparison to child potties. This suggests that potties may be a better tool for promoting SCFD than mini-hoes or sani-scoops, as long as support is given to encourage and teach rural mothers how to train their children to consistently use the potties.

The combined behavioral and hardware approaches conducted by Ashraf et al. (2012) and Hussain et al. (2017) reported the greatest increases in rates of SCFD from pre- to post-intervention. These small-scale trials used tailored messaging and provided multiple tool/potty options to households, which were positively accepted by communities. While these potties may be important for facilitating SCFD, they require latrine access for safe disposal of the collected feces (Sultana et al. 2013; Hussain et al. 2017). This indicates the need for further research on combined interventions as a method of promoting SCFD. The success of these small-scale combined interventions also suggests that sanitation tools should be tailored to community preferences in order to increase uptake and sustained use.

The studies included in this review identified the age of the child, ambulatory status of the child, and the availability of latrines as determinants for SCFD. As children age, latrine use and other safe feces disposal methods increase (Sultana et al. 2013; Bauza et al. 2019). Whether a child was ambulatory or not was an important marker of latrine use, where latrine use was almost exclusively tied to ambulatory children (Majorin et al. 2014). Latrine use by parents was another important marker for SCFD; caregivers that practiced safe sanitation methods were more likely to safely dispose of children's feces (Hussain et al. 2017). Latrine access was largely identified as a necessary component of SCFD, but as evidenced by the strict hardware MANTRA and Total Sanitation Campaign programs, it is not sufficient to improve SCFD practices without a behavioral component (Majorin et al. 2014; Lamicchane et al. 2018; Bauza et al. 2019; Reese et al. 2019).

Previous studies have also indicated open defecation by adults in the household as one of the most significant determinants of unsafe CFD practices (WSP 2015; Islam et al. 2018; Ellis et al. 2020); however, solely prioritizing reducing open defecation by adults is not necessarily effective in also reducing unsafe CFD (Rand et al. 2015). Bhatt et al. (2019) identified personal and cultural factors contributing to the burden of open defecation in Nepal as based on either choice or compulsion. Here, motivation by choice includes the use of open defecation as a form of socialization and a method to get outside and exercise, as well as its motivation by habit or convenience. Factors contributing to open defecation by compulsion include the absence of a latrine at home as well as issues with privacy and hygiene (Bhatt et al. 2019). Thus, future studies should consider these cultural factors when designing behavior change interventions targeting open defecation by adults, while also explaining the importance of safely managing the feces of their children.

Beardsley et al. (2021) evaluated factors associated with unsafe CFD in India and found that higher household income, higher educational attainment, and access to a sanitation facility, whether improved or unimproved, were also significantly associated with SCFD. Those using improved facilities had 6.55 times the odds of practicing SCFD, and those using unimproved facilities had 5.32 times the odds of SCFD, as those with no sanitation facility present (Beardsley et al. 2021). However, the authors found no significant association between unsafe CFD and handwashing, access to an improved water source, or use of shared sanitation facilities. Aliyu & Dahiru (2019) and Sahiledengle et al. (2021) reported similar findings to Beardsley et al. (2021) in Nigeria and Ethiopia, respectively, though the use of an improved water source in these studies was significantly associated with SCFD. This suggests that, while sanitation access is a major component of better SCFD practices, unsafe CFD may be an indicator of larger-scale societal problems such as poverty and lack of education for women. Tackling not only behavior change and sanitation access in the WASH sector, but also poverty and education as broader sectoral issues, may be necessary to significantly improve both SCFD and public health in the future. The inclusion of only articles written in English is one limitation of this study. With so few studies focusing on behavior change as a necessary component of improving SCFD practices, and instead focusing on sanitation hardware, there is much need for the inclusion of behavioral components in future interventions. Large-scale governmental hardware programs in the Asia-Pacific region have increased access to improved sanitation and reduced the need for open defecation and unsafe CFD by compulsion, yet both open defecation and unsafe CFD remain prevalent. Eliminating both open defecation and unsafe CFD cannot be achieved without the incorporation of behavior change interventions in government policies and practices. Community-level combined interventions, such as those conducted by Ashraf et al. (2012) and Hussain et al. (2017), have shown promise for improving SCFD practices. Scaling up behavior change interventions should be the focus of future governmental programs.

Access to sanitation is a defining determinant of SCFD. Large-scale programs and hardware interventions are important for providing communities with the infrastructure necessary to improve SCFD practices, but such interventions may be improved by the addition of a behavior change component. Hardware interventions tend to be ineffective at promoting SCFD if implemented alone. With so little evidence available on the effectiveness of behavioral interventions in reducing unsafe CFD in the Asia-Pacific region, future interventions should focus on addressing the determinants of unsafe CFD practices, especially in poorer households with younger, non-ambulatory children. While the combined behavior change and hardware interventions reported the greatest increases in SCFD, they were of lower quality due to small sample sizes, the use of self-reported data, and the use of observed feces as an indicator of open defecation. This demonstrates a need for future research on how behavior change models combined with hardware interventions impact SCFD in the Asia-Pacific.

This study was primarily limited by available literature that met the selection criteria. Many of the selected studies were not designed to report quantitative results and instead used qualitative methods such as interviews and focus groups to draw conclusions. This variety of methods and result presentations made comparing interventions difficult. Only one behavior change intervention met inclusion criteria, so the true differences between impacts of behavioral and hardware interventions were difficult to ascertain.

We would like to thank Stacy Torian and Karen Grigg of the University of North Carolina at Chapel Hill Health Sciences Library for their assistance with database search strategy. We thank Ayu Siantoro, Mahesh Nathan, and Nong Davith of World Vision Indonesia. We also thank Katharine Conaway of the University of North Carolina at Chapel Hill for assistance with the creation of figures in this manuscript.

Conceptualization (M.M., L.S., A.L.); Methodology (L.S., A.L., M.M., B.T., R.C.); Software (L.S.); Formal Analysis (L.S., M.M.); Investigation (M.M., L.S., A.L.); Resources (L.S., A.L., M.M.); Data curation (L.S., A.L., M.M.); Writing – original draft preparation (L.S.); Writing – Review and Editing (M.M., B.T., V.B., R.C.); Visualization (L.S., M.M.); Supervision (M.M.); Project Administration (M.M.). All authors read and approved the final manuscript.

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

The authors declare there is no conflict.