Sanitation-Hygiene Practices and Social Norms in Community-led Total Sanitation for Sustainability of Open Defecation Free Status: A Survey of Suna West Sub-County, Migori County, Kenya

11

Community-led total sanitation (CLTS) has been used to stir sanitation -related 12 behaviour change and attain open defecation free (ODF) status. CLTS interventions suffer high 13 rates of reversion such that their gains are unsustainable in most contexts. 14

This study aimed to determine the role of sanitation hygiene practices and social 15 norms on open defecation free status in Suna West Sub County. 16 Methodology: A cross -section study design was employed using question naire and 17 observation checklist to collect data from 384 households. 18

Results revealed that 66.1% households had partially reverted to non-ODF status. The 19 sanitation-hygiene practices associated with being ODF includes: use of elevated racks 20 (AOR=0.81; CI= 0.34-1.90; p=0.625), use of treating water (AOR=2.81; CI= (0.97-8.06); 21 p=0.055), regularly clean latrines (AOR=2.96; CI= 0.63-13.89; p=0.17), pouring of ash over 22 the pit of the latrine (AOR=4.08; CI= 1.73-9.62; p<0.001) and use of dug out pits for waste 23 disposal (AOR=2.41; CI= 1.02-5.68; p<0.045). On social norms, the study found that 24 laws/penalties (AOR=4.15; CI= 2.20-7.80; p<0.001) and rewards/incentives (AOR=0.17; 25 CI=0.096-0.306; p<0.001) had less odds of being ODF. Moreover, odds of being ODF was low 26 for households who reported that construction/maintenance materials were expensive 27 (AOR=0.17; CI=0.84-2.79; p=0.169), that it was embarrassing to people defecate in the open 28 (AOR=0.060; CI0.19-1.23=; p<0.129) and that it is okay to defecate in bushes/r ivers/dams 29 (AOR=0.623; CI=;0.329-1.179 p<0.146). 30

The results of this study show partial reversion to non -ODF status in previously 31 certified villages. However, households that sustained ODF status had several sanitation 32 hygiene practices. Interestingly, households that displayed social norms were less likely to be 33 ODF. Overall, the findings of the present study demonstrate that the CLTS process failed to 34 instil social norms around proper sanitation to inspire community collective action thus little 35 influence on sustainable behaviour change. The findings of this study therefore highlight the 36 need to enhance good hygiene sani tation practices, while instilling social norms to inspire 37 community collective action. 38 KEY WORDS : Community -led total sanitation (CLTS) ; Open defecation free (ODF) ; 39 sanitation hygiene practices; social norms. 40 41 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. 2

42 According to the World Health Organisation (WHO), roughly 842,000 lives are lost in low - 43 and middle -income countries annually as a consequence of inadequate water, hygiene and 44 sanitation (1). Indeed, poor sanitation has been connected to infections such as (2) diarrhoeal 45 diseases, nematode infections and environmental enteropathy (EE). In Kenya, for instance, 46 diarrhoea claims the lives of roughly 3,100 children annually and trachoma, schistosomiasis 47 are health problems linked to poor sanitation. In part, the burden of these diseases is attributed 48 to open defecation that exposes a large part of the population to sanitation-related diseases (3). 49 It is in light of such negative impacts of poor sanitation that the Government of Kenya adopted 50 Community-led Total Sanitation (CLTS) as a strategy to improve sanitation. Community -led 51 Total Sanitation was introduced by Plan International Kenya in 2007 and was approved by the 52 then Ministry of Public Health and Sanitation (MOPHS) as a key framework for promoting 53 hygiene and sanitation at the household level. In 2011, MOPHS established CLTS as the 54 national strategy for ensuring rural sanitation and set a national target to reduce open defecation 55 (4). 56 Community led total sanitation involves three main steps: pre -triggering, triggering and post-57 triggering(5). At the pre-triggering stage the community members are mobilised for triggering 58 process in which community members are able to see the reality of mass defecation and how it 59 affects the community negatively (6). Finally, the post -triggering phase involves follow -up 60 aimed at verification and certification of the villages as open defecation free(7). 61 There has been a lot of research on why people do what they do, what influences their actions. 62 In the recent sanitation interventions and community -led total sanitation, key emphasis has 63 been on establishing social norm around unacceptability of open defecation (6). The process of 64 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 3 CLTS has been known to cause peer pressure by invoking emotions such as shame and disgust 65 and change perceived social norms to establish open defecation as being socially 66 unacceptable(8). Social norms and how strong they are, has been found to greatly influence 67 sustainable behaviour change in relation to sanitation practices. Where positive social norms 68 were well rooted, the chances of achieving sustainable behaviour change were higher(9). 69 It is important to understand why people engage in open defecation for failure to do so has led 70 to failed interventions(10). Open defecation has been known to be an independent action, that 71 is, people engage in open defecation because they believe it meets their needs and that it is not 72 harmful to them and others, it is a custom rather than a social norm (11). However, to eliminate 73 open defecation in a particular group, there is need to create a social norm promoting latrine 74 use and maintenance. People will need to think that their reference network think they should 75 use and maintain latrines and them believing in that will motivate them to engage in open 76 defecation-free behaviour(10). 77 The results of a study on the sustainability of ODF status in Kenya conducted by (12) revealed 78 that the sustainability of ODF achievements remained a major concern with over 70% of 79 villages that had received partial or full OD F status reverting to non -ODF status. Among the 80 factors that demotivate community members from using a latrine after becoming ODF relates 81 to physical aspects of the latrine (such as lack of privacy and fear of the latrine collapsing) and 82 sharing a latrine with other people (14). In addition, slippage from ODF status has also been 83 linked to collapse or poor structural integrity of latrines as well as unsustainable behaviour 84 change following sanitation -related interventions (14). The objective of this study was to 85 investigate the ODF status of households in ODF certified villages of Suna West Sub County 86 and to determine association between ODF status with sanitation-hygiene practices and social 87 norms in households of Suna West Sub-County, western Kenya. 88 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 4

89 Study Area 90 The study was done in Suna West Sub -County in Migori County which has a population of 91 117,539 with a density of 406 persons per km 2 and 29,251 households. The sub -county, one 92 among eight others, has four wards; Ragana-Oruba, Wasweta II, Wiga and Wasimbete wards. 93 It is bordered by Kuria West sub-county to the south-east, Nyatike sub-county to the west, Suna 94 East Sub-County to the north-eastern side and Tanzania to the south -west. Migori county has 95 poverty level measured at 32.0 (2016) and is listed among the counties with the most income 96 inequality as measured by the gini coefficient. Migori county’s a gini index is 0.464(15). 97 Study Design 98 A cross-sectional study design was used across two wards, Ragana-Oruba and Wasweta II, that 99 were purposively chosen for having attained ODF status at least one year to the study which 100 was carried out from 17th to 21st July, 2020. 101 Target population 102 The unit of analysis was the household with the targeted participants being the household 103 heads. The sample size for households featured in the study was determined using the statistical 104 formula as used by Fisher .(16). The households that had participated in the initial verification 105 were included and any child -headed household was ex cluded. The data collection was done 106 across 384 households identified through simple random sampling of the households as found 107 in MOH 513 which is the household register maintained by the Community Health Volunteers 108 (CHVs). 109 110 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 5 Sampling Design 111 The sample size for households in the village featured in the study was determined using 112 Fisher’s formula; the standard deviation set at 1.96, which corresponds to 95% CI. The 113 proportion of desired population was assumed at 0.50, the margin of error 0.05. We u sed the 114 initial sample size of 384 to capture the true variability of the population. 115 Data Collection Tools and Procedure 116 The survey tools were developed in line with the constructs of the key questions that we sought 117 to answer with reference to previously validated instruments that were reviewed by my 118 supervisors and peers. Once that was done, we pre -tested the tool and adjusted any questions 119 that we sought to answer. Cronbach’s alpha was computed to test for reliability of data 120 collection instrument and output yielded result of an internal consistency 0.7. 121 Validated structured questionnaire and observation checklist were used in this study. 122 Observation checklist was used to collect information to corporate or refute claims made by 123 respondents in the questionnaires. 124 The researcher provided scientific oversight of this study including training and technical 125 support for the research team and oversaw the development of coding frame and data analysis. 126 Under supervision, the research assistants (RAs) helped in the recruitment and obtaining 127 informed consent from the participants to take in the part in the study. 128 The tools were provided in English. 129 Study Variables 130 Dependent Variables 131 Dependent variables were measured as follows: a) access to a latrine - availability of Individual 132 latrine, shared latrine/neighbours; b) privacy - availability of door or some form of barricade 133 provided for each superstructure; c) Squat hole cover – a barrier to prevent movement of flies 134 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 6 in and out of the latrine hole d) hand washing facility - availability of tap/leaky tin near latrine 135 with water inside, soap/ash available; e) no exposed faeces – via simple transect walk, no 136 visible faeces should be found within the surrounding of the home. For a household to be ODF, 137 all the 5 parameters must be met and the absence of any one would be an indication of partial 138 reversion to non-ODF status while absence of all the five indicators would reflect full reversion 139 to non-ODF status of the household. 140 Independent Variables 141 For independent variables, the frequency of variables such as treating water (boiling or use of 142 chemicals), covering food using lid over cooking pots when cooking and during storage, using 143 elevated racks to hold utensils off the ground while drying, regular cleaning of latrine, 144 application of ash around & the squat hole of latrine, and using dugout pit for waste disposal 145 were measured using Likert scale: always, most of the time, sometimes, rarely, not at all. Scores 146 were aggregated into two - Yes (always, most of the time, sometimes) and no (rarely, not at 147 all). 148 While variables such as Care of the family - Empirical & normative expectation regarding 149 health of the family, Shame/disgust/fear/pride Regrettable occurrence/ unpleasant emotion that 150 cause a feeling of resolution, Cultural/social/religious beliefs - Person’s belief alignment as 151 pertaining culture, society and religion, Laws/penalties - Rules within a given set up and 152 punishment imposed for breaking the set rules, Need to improve things in the family - 153 Empirical & normative expression of obligation to make things better, Follow ups and support 154 - The subsequent actions following CLTS and material assistance for the same, 155 Rewards/incentives - Some form of payment given in recognition of work done or to stimulate 156 greater output, and Peer pressure - The empirical &normative expectation regarding consistent 157 latrine use were measured using Likert scale: strongly disagree, disagree, neutral, agree and 158 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 7 strongly agree. The scores were aggregated into two; Yes (agree and strongly agree), No 159 (strongly disagree, disagree, neutral). 160 Statistical Analysis 161 Summation for the observed non-negotiable indicators was done to determine the ODF status. 162 Wilcoxon signed-rank test was used to determine if there was a significant difference in ODF 163 status as at the time of the study and verification. Chi -square test of independent was used to 164 determine association between sanitation and hygiene practices, social norms a nd ODF status 165 and binary logistic regression was done to determine the relationship between sanitation 166 hygiene practices, social norms and ODF status. 167 While the outcome variable was highly prevalent, raising concerns about the potential for odds 168 ratios to over-estimate the true association compared to risk ratios, we chose to use logistic 169 regression as it offers a robust framework for adjusting for multiple confounders and provides 170 a straightforward method for hypothesis testing. Nevertheless, we interpret ed the results with 171 caution, to effectively utilize established statistical methodologies and maintain consistency 172 with previous research in similar contexts. 173 Ethical considerations 174 Ethical approval for the study was granted by the Maseno University Ethics Review 175 Committee; Ref: MSU/DRPI/MUERC/00821/99 and National Commission for Science 176 Technology and innovation, Kenya. To protect the privacy and anonymity of participants, their 177 personal identifying information was not recorded in the questionnaires or checklists and 178 households were reported as numbers. The data was coded to secure the privacy and anonymity 179 of participants. Further, there was disclosure regarding the nature of the stu dy, its purpose, 180 what it involves, the procedures to be done, risks -if any by taking part in the study and since 181 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 8 the study did not involve experimental drugs or procedures used in the study, the study posed 182 no significant risk to participants. 183 Informed consent. 184 Written informed consent was gotten from the participants. They were also informed that 185 taking part in the study was out of free will and that they were free to withdraw from the study 186 at any time. 187

188 Socio-demographic characteristics of study participants 189 Of the 384 participants, 62.8% were females, 73.4% were aged 25 -59 years, while 58.6% and 190 31.8% had primary education and secondary education respectively. On socio-economic status, 191 three-quarter (75.3%) of households had $0 -38.9 monthly income. A half (53.1%) of the 192 households had 0 -5 years old child and 27.1% had at least one member with a disability or 193 chronic illness. Socio-demographic characteristics of the study participants are summarized in 194 Table 1. 195 Table 1: Socio-demographic characteristics of study participants from Suna West Sub-196 County, Kenya 197 Variable Frequency (n) Percentage (%) Gender Male 143 37.2 Female 241 62.8 Respondents’ Age 18-24 years 58 15.1 25-59 years 282 73.4 60 and above 44 11.5 Level of Education No education 26 6.8 Primary education 225 58.6 Secondary education 122 31.8 Tertiary education 11 2.9 Level of Income 0-38.9 289 75.3 39-77.9 72 18.8 78-155.8 6 1.6 155.9-233.7 3 0.8 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 9 233.8-311.6 6 1.6 311.7 and over 8 2.1 Household composition With persons 0-5 years. 204 53.1 With persons 6-12 years. 278 72.4 With persons 13-24 years. 299 77.9 With persons 25-59 years. 354 92.2 With persons above 60 years. 72 18.8 Persons with disability or chronic illness 104 27.1 Table legend 1: The average exchange rate as at the time of the study was KES 128.37 for a dollar Open Defecation Free Status 198 On ODF status, only 33.9 % (n=130) were found to be ODF one year after certification. When 199 the indicators were analysed singly, it was observed that access to latrine and no exposed faeces 200 were at 100%; with 95.3% (n=366) owning individual latrines while the remaining 4.7% 201 (n=18) reporting to use shared latrines, Table 2. 202 Table 2: Results on ODF indicators 203 Indicator Median percentage (%) P value No. of villages reporting 100% Access to latrine 100 1.0 13 (100%) Squat hole cover present 63 0.002 0 (0%) Privacy 82.4 0.002 0 (0%) Hand washing facility 82.4 0.004 2 (15%) No exposed faeces 100 0.056 8 (61.5%) 204 Association between sanitation and hygiene practices and open defecation free status 205 This analysis determined the association between various household sanitation practices and 206 the likelihood of being Open Defecation Free (ODF). The results show that household that pour 207 ash into latrines had higher odds of ODF status compared to households that do not 208 (AOR=4.082; 95% CI:1.730 – 9.615; p value = 0.001), while households that dug out pits for 209 waste disposal had higher odds of ODF status compared to household s that do not 210 (AOR=2.410; 95% CI:1.018 – 5.682; p value = 0.045). Other practices such as treating water, 211 using elevated racks, and regularly cleaning of latrines were not significantly associated with 212 ODF as shown in Table 3. 213 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 10 Table 3: Association between sanitation and hygiene practices and open defecation free status 214 in Suna West Sub-County, Kenya 215 Characteristic ODF n (%) OR (95%CI) P value AOR (95%CI) P Value Treating water Yes 123 (35.3) 3.226(0.912–8.547) 0.019 2.809 (0.978-8.065) 0.055 No 5 (14.7) Ref Ref Using elevated racks Yes 117 (35.7) 2.203 (1.106-4.444) 0.027 0.808 (0.343-1.901) 0.625 No 11 (20.4) Ref Ref Regular cleaning of the latrine Yes 128 (35.2) 4.878 (1.122-13.513) 0.035 2.959 (0.629-13.889) 0.170 No 2 (10.0) Ref Ref Pouring of ash Yes 121 (38.5) 4.348 (2.083-9.091) <0.001 4.082 (1.730-9.615) 0.001 No 9 (12.9) Ref Ref Dug out pit for waste disposal Yes 122 (38.4) 4.504 (2.083-9.804) <0.001 2.410 (1.018-5.682) 0.045 No 8 (12.1) Ref Ref 216 Association between social norms and open defecation free status 217 This study found association in a number of the social norms and ODF status. For instance, 218 individuals who were not subjected to laws and penalties were considerably more inclined to 219 be ODF compared to those who were (AOR = 4.145; 95% CI: 2.203 –7.802; p < 0.001), 220 indicating that the implementation and enforcement of legal frameworks are not essential for 221 encouraging ODF behaviours. Furthermore, the lack of follow-ups and support was correlated 222 with increased odds of open defecation (OR = 1.985; p = 0.025), although the adjusted effect 223 was not statistically significant (AOR = 2.111; p = 0.118). While follow-up support appears to 224 have some effect, its impact weakens when accounting for other variables. Interestingly, those 225 who did not anticipate receiving incentives were notably more likely to retain ODF status 226 (AOR = 0.172; 95% CI: 0.096 –0.306; p < 0.001). The reliance on intrinsic motivation and 227 communal norms could prove to be more sustainable than dependence on external incentives 228 for fostering sanitation practices. Additionally, it is important to highlight that individuals who 229 did not consider the construction or maintenance of latrines to be costly were more likely to be 230 ODF (OR = 1.934; p = 0.003), albeit this was not significant after adjustment (AOR = 1.526; 231 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 11 p = 0.169). Perceived costs may initially influence the uptake of latrines, but other factors like 232 norms and enforcement may take precedence in determining long -term behaviour. Finall y, 233 respondents from communities where the majority felt embarrassed about not having a latrine 234 were more likely to be ODF (AOR = 1.412; p = 0.046), suggesting that collective social 235 pressure and shared values regarding sanitation can serve as a significant motivator in 236 achieving and maintaining ODF status, as illustrated in Table 4. 237 Table 4: Association between social norms and open defecation free status Suna West 238 Sub-County, Kenya 239 Social Norms ODF n (%) OR (95% CI) p value AOR (95% CI) p value Subjection to laws and penalties Yes 60 (24.3) Ref Ref No 70 (51.1) 3.256 (2.090-5.074) <0.001 4.145 (2.203-7.802) <0.001 Follow-ups and support Yes 115 (33.3) Ref Ref No 15 (50.0) 1.985 (1.089-3.620) 0.025 2.111 (0.827-5.387) 0.118 Expectation of rewards/Incentives Yes 60 (22.6) Ref Ref No 70 (58.8) 0.239 (0.150-0.380) <0.001 0.172 (0.096-0.306) <0.001 Construction/maintenance expensive Yes 70 (28.5) Ref Ref No 50 (49.5) 1.934 (1.251-2.991) 0.003 1.526 (0.835-2.788) 0.169 Embarrassing to see people defecate in open Yes 91 (32.0) Ref Ref No 39 (41.9) 1.587 (0.982-2.567) 0.06 0.489 (0.194-1.232) 0.129 Majority ashamed for not having latrine Yes 86 (29.8) Ref Ref No 41 (45.6) 1.927 (1.188-3.126) 0.008 1.412 (1.014-5.739) 0.046 Okay to defecate in rivers/bushes/dams Yes 84 (28.2) Ref Ref No 46 (56.1) 0.254 (0.151-0.427) <0.001 0.623 (0.329-1.179 0.146 240 241 242 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 12

243 Open defecation status 244 This study found 66.1% reversion one -year post -ODF in previously certified households. 245 Similarly, an earlier study had shown that 70% of villages reverted back to open defecation, 246 three years after certification among seven sub -counties featured in the stu dy (12). However, 247 low reversion rates have been observed (8%) in Ethiopia and Ghana after one year of CLTS 248 implementation and (14.5%) in Indonesia after two years of ODF certification (4; 10). In these 249 studies, latrine presence - latrine presence and usage – was used as the measure for 250 sustainability, the possible reasons for recording lower reversion rate. Nevertheless, reversion 251 has been found to be common in villages within sub -Saharan Africa where it has been 252 associated with several factors (17 ,9 ) Moreover, sustainability of ODF achievements has been 253 previously found to be a major challenge in Kenyan communities. 254 In a study by Tyndale-Boscoe et al. (2013) in Uganda, Kenya, Ethiopia and Sierra Leone two 255 years after CLTS, a 13% reversion was reported when latrine presence was used to measure 256 sustainability. However, the reversion rate would have drastically increased to 92% had the 257 study used the 5 indicators used during the initial verification process which included 258 functional latrine, means of keeping flies away (water seal or squat hole cover), absence of 259 faecal matter, presence of hand washing facility with soap/ash and evidence of latrine use in 260 the re-verification (18). This, is a much higher reversion rate like the 66.1% observed in this 261 particular study when all the indicators are used to measure sustainability. 262 While governments and most organisations have been very successful in getting households to 263 build and retain latrines, less success has been achieved in improving sanitation behaviour 264 change which is the major aim of CLTS (18). Overall, the findings of this study suggest that 265 there is need to harmonise or standardize indicators of ODF status. Furthermore, although the 266 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 13 current protocol is very clear on the non-negotiable indicators, there is need to re-look at their 267 role in defining ODF status. Doing so will help in defining concepts upfront, in developing any 268 kind of monitoring tool of post-ODF status. 269 Association between sanitation hygiene practices and open defecation status 270 We found significant association between sanitation hygiene practices and ODF status. Further, 271 demonstrated that households that complied with the sanitation hygiene practices were more 272 likely to be ODF. This resonates with studies done in Indonesia in which participant s from 273 better performing villages on ODF outcomes reported that messages around sanitation 274 promotion and good hygiene had been constantly promoted through mosques and churches. 275 Further, local groups carried out monitoring after CLTS implementation in an effort to promote 276 hand washing with soap, treating of drinking water, proper food handling, solid and liquid 277 waste management by households(10). 278 That households that poured ash in the pit latrines were found to be ODF is not surprising 279 because pouring of ash in the pit latrines manages smell from latrines and therefore encourages 280 consistent latrine use by all members of the household. This finding supports findings of 281 previous study in which it was reported that smelly and unimproved latrines turned people back 282 to open defecation in Ethiopia (19). Further, latrine usage by women was tampered with 283 negatively as a result of perceptions around latrine cleanliness and smell inside (10). While this 284 particular study found that 21% of households presented no evidence of the use of a hand -285 washing facility and 46% households did not wash their hands with soap and water always 286 after using a latrine, in a study done in four African countries, there was an overall reversal rate 287 of 17% for signs of use of a handwashing facility and 75% for consistent handwashing with 288 soap and water. Reversion for consistent hand washing with soap and water in Homa Bay and 289 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 14 Kilifi stood at 83% and 67% (18). Thus, this study recorded much lower reversal rate on 290 consistent hand washing with soap as compared to previous studies. 291 Training on hand washing with soap, water treatment, preparing food in a hygienic way and 292 proper storage and solid wastes disposal are standard parts of sanitation program (20). 293 According to Lilje, in a study done in Chad, the individual perception to treating water was 294 rated high. Respondents thought positively about the issues of water treatment and did not 295 perceive it to be taking much effort, time or cost (21). This mirrors the findings of this present 296 study, households that treated water were high. Water treatment commodities were available 297 in public health offices and distributed by CHVs at household level during dry seasons, other 298 commodities were offered at health facilities to mothers attend ing clinics and further, there 299 were chlorine dispensers strategically situated in communal water points. 300 Association between social norms and open defecation status 301 The study found association between a given number of normative and empirical expectations 302 and ODF status a reflection of the existence of social norms within Suna West Sub County. 303 Indeed, the existence of social norms is predicated on their functions, which in turn are driven 304 by ecological expectation (11). This notion is supported by the results of this study. However, 305 a potential limitation of the current study is whether such norms have broader functions needed 306 to support programmes such as CLTS beyond their active phase. 307

of this study found that household that responded that the health of the family motivates 308 them to be ODF were less likely to be ODF. This is in spite of previous studies like that done 309 by UNICEF that reported that the most prominent motivator towards ODF status was concern 310 for the health of the family(17). Households believed that stopping open defecation resulted in 311 reduction in diar rheal diseases thus motivating them to stop open defecation (14). In another 312 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 15 study, Moran, (2017) reports that health, even though may not have been a driver for the initial 313 defecation behaviour change, people do continue to make effort to maintain and use latrine 314 because of their health and that of the family(22). 315 Further, the study found no association between care for the family, latrine accessibility and 316 ODF status a reflection of no prudential personal normative belief and no association was found 317 in privacy/security offered by latrine and even peer pressure and ODF status. In a previous 318 study, provision for privacy for superstructure, pride and the convenience of using of latrines 319 were found to be important drivers for women in respect to building latrines in Indonesia (10). 320 This particular study found an correlation between those that were embarrassed about not 321 having a latrine and ODF status. This is in tandem with previous studies that reported that 322 shame/ disgust motivated households into behaviour change(14). 323 This study found that individuals who did not consider the construction or maintenance of 324 latrines to be costly were more likely to be ODF, however, this was not significant after 325 adjustment to indicate that perceived costs may initially influence the upta ke of latrines, but 326 other factors may take precedence in determining long -term behaviour change. This is 327 consistent with previously studies that reported high cost of building, maintenance and repair 328 of latrines were among the reasons for reversion to non -ODF status (19). In their study, 329 Bongartz et al. (2016) suggested that though CLTS was a zero-subsidy strategy, there was need 330 for incorporation of sanitation marketing to CLTS to help those who can afford make informed 331 choice even though this could pose challenge of interfering with behaviour change process 332 (23). 333 Those that said rewards and incentives motivate them to be ODF were found to be less likely 334 to be ODF. This resonates with the findings of a study done in East Java in which households 335 that received some form of subsidy did not become ODF, it was discovered that subsidy was 336 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 16 divisive since it was never enough for all households and thus hampered collective action, also, 337 incentives has been found to have the capacity to corrupt intrinsic motivation ( 19; 11). In the 338 study done on sustainability by UNICEF (2014), some of the enablers of sustainability were 339 natural leaders working together and post-ODF follow up by CHVs (14) 340 Novotný et al. (2017), in their study, concluded that so cial norms were important 341 instrumentally as sanitation outcomes depended on the level to which social influences were 342 able to shape the perceptions of benefits or risks on sanitation -related awareness in positive 343 ways(8). Similarly, Odagiri et al. (2017) found that in addition to economic levels and lack of 344 reliable access to water, weaker social norms were significantly associated with the reversion 345 to open defecation practices (10). When looked at singly, latrine usage and open defecation 346 were sustained meaning the social sanctions played out well. However, in the other areas of 347 hand washing with soap, provision of privacy and use of squat hole cover, there was significant 348 reversion registered meaning the social sanctions weren’t applied across all the non-negotiable 349 indicators. Suna West sub county had no deep -rooted social norms neither did the CLTS 350 process inculcate new norms to bring about the overall change in sanitation hygiene practices 351 desired and to sustain it after the pressure of certification was off. 352

353 There was partial reversion to non-ODF status in households one year after certification. This 354 was mainly attributed to 3 major indicators: provision of hand washing facility, squat hole 355 cover and privacy. Moreover, there was sustained ODF status in househ olds with good 356 sanitation hygiene practices. It was, however, evident that social norms were not embedded on 357 the CLTS process, thus failing to create social norms around sanitation and hygiene practices 358 to enhance community collective action towards ODF st atus sustainability. Therefore, it is 359 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 17 important to enhance good hygiene sanitation practices, while instilling social norms to inspire 360 collective and sustainable action needed in programmes such as CLTS. 361

362 We acknowledge the people of Suna West Sub-County without whom this research would not 363 have been done. 364 AUTHORS CONTRIBUTION 365 N.R.A. conceived the presented idea. N.R.A. developed and performed the study under the 366 supervision of C.O.A. and P.O.O. All authors discussed the results and contributed to the final 367 manuscript. 368 CONFLICT OF INTEREST 369 Authors declare no conflict of interest. 370

371 1. World Health Organization. (2018). Sanitation [Fact sheet]. http://www.who.int/news-372 room/fact-sheets 373 2. UNICEF. (2015b). The impact of poor sanitation on nutrition. UNICEF. 374 3. Njuguna, J., & Muruka, C. (2017). Open defecation in newly created Kenyan 375 counties: A situational analysis. Journal of Health Care for the Poor and 376 Underserved, 28(1), 71–78. https://doi.org/10.1353/hpu.2017.0007 377 4. Crocker, J., Saywell, D., & Bartram, J. (2017). Sustainability of community-led total 378 sanitation outcomes: Evidence from Ethiopia and Ghana. International Journal of 379 Hygiene and Environmental Health, 220(3), 551–557. 380 https://doi.org/10.1016/j.ijheh.2017.02.011 381 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 18 5. Cavill, S., Chambers, R., & Vernon, N. (2015). Sustainability and CLTS: Stock 382 taking. Frontiers of CLTS: Innovations and Insights, 4. Institute of Development 383 Studies. 384 6. Kar, K., & Chambers, R. (2011). Triggering: An extract from the handbook on 385 community-led total sanitation. University of Sussex. 386 7. Sarah, L. (2016). Certification of open defecation free status: Emerging lessons from 387 Kenya. In P. Bongartz, N. Vernon, & J. Fox (Eds.), Sustainable sanitation for all: 388 Experiences, challenges, and innovations (pp. 211–226). Practical Action Publishing. 389 8. Novotny, J., Kolomaznikova, J., & Humňalová, H. (2017). The role of perceived 390 social norms in rural sanitation: An explorative study from infrastructure-restricted 391 settings of South Ethiopia. International Journal of Environmental Research and 392 Public Health, 14(7), 794. https://doi.org/10.3390/ijerph14070794 393 9. Mukherjee, N. (2012). Achieving and sustaining open defecation free communities. 394 Action Research. 395 10. Odagiri, M., Cronin, A. A., Rah, J. H., & Thomas, A. (2017). Enabling factors for 396 sustaining open defecation-free communities in rural Indonesia: A cross-sectional 397 study. International Journal of Environmental Research and Public Health, 14(12), 398 1572. https://doi.org/10.3390/ijerph14121572 399 11. Bicchieri, C., & Noah, T. (2017). Applying social norms theory in CATS 400 programming. University of Pennsylvania. 401 12. UNICEF. (2015a). Sustainability of ODF practices in Kenya. 402 https://www.unicef.org/esa/sites/files/2018-09/UNICEF-Kenya 403 13. Singh, S., & Balfour, N. (2015). Sustainability of ODF practices in Kenya. UNICEF 404 Kenya. 405 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint 19 14. UNICEF. (2014). Evaluation of the WASH sector strategy ‘Community Approaches to 406 Total Sanitation’ (CATS). https://www.unicef.org/evaluation 407 15. Kenya National Bureau of Statistics. (2020). Kenya population and housing census. 408 https://www.knbs.or.ke 409 16. Fisher, A. A., Laing, J. E., Stoeckel, J. E., & Townsend, J. W. (1991). Handbook for 410 family planning operations research design (2nd ed.). Population Council. 411 17. UNICEF. (2013). Community-led total sanitation in East Asia and Pacific: Progress, 412 lessons and directions. UNICEF East Asia and Pacific. 413 18. Tyndale-Biscoe, P., Bond, M., & Kidd, R. (2013). ODF sustainability study. Plan 414 International. 415 19. Mukherjee, N. (2016). Building environments to support sustainability of improved 416 sanitation behaviours at scale: Levers of change in East Asia. In P. Bongartz, N. 417 Vernon, & J. Fox (Eds.), Sustainable sanitation for all: Experiences, challenges, and 418 innovations (pp. 195–210). Practical Action Publishing. 419 20. Magala, J. M., & Roberts, B. (2009). Evaluation of strategy for scaling up 420 community-led total sanitation in Ghana. UNICEF. 421 21. Lilje, J., Kessely, H., & Mosler, H.-J. (2015). Factors determining water treatment 422 behavior for the prevention of cholera in Chad. The American Journal of Tropical 423 Medicine and Hygiene, 93(1), 57–65. https://doi.org/10.4269/ajtmh.14-0613 424 22. Moran, H. (2017). ODF sustainability in Timor-Leste. Australian Aid. 425 23. Bongartz, P., Vernon, N., & Fox, J. (Eds.). (2016). Sustainable sanitation for all: 426 Experiences, challenges, and innovations. Practical Action Publishing. 427 428 429 . CC-BY 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted August 5, 2025. ; https://doi.org/10.1101/2025.08.01.25332764doi: medRxiv preprint