Making the Provisions of Water More Effective

Paper presented for Symposium on November 29, 2006 at PGIA, University of Peradeniya

"Providing water and sanitation facilities alone itself are unable to solve the problems ofcommunities effectively unless interventions do not attend to improve the hygienebehaviours. This paper is looking at the effectiveness of such an interventionThis research was funded by European Union which continued for a period of 3 years."

Sugandhi Samarasinghe & Pallitha Jayaweera,

AbstractAs water and sanitation facilities being vital factor for good hygiene behaviors, it expressesdirect relatedness in most of the disease incidences. Diarrhea, worm infestation and eyeand skin infections are diseases related to water and sanitation. Simple hygiene behaviorsare key to improving health. Hygiene promotion is therefore recognized nowadays as anessential part of water and sanitation programmes if the maximum health benefits are to begained from provision of improved facilities. This research was done in 6 countries; Kenya,Uganda, Ghana, Nepal, India and Sri Lanka targeting behaviors of hand washing, havingand using latrines, safe disposal of infant excreta and storing drinking water safely. Themain objectives were to assess the level of sustainability of behavioral change one to threeyears after a hygiene promotion intervention, to develop a methodology for simple, costeffectivelong-term monitoring of behavioral changes, to gain insight into relationshipsbetween project approaches, External conditions and sustainability of changes in hygienebehavior and to determine the policy and programming implications of the study findings asa basis for influencing future policy and increasing the effectiveness of programmes.Samples in the six-country study were done in Ghana, India and Sri Lanka. In othercountries Kenya, Nepal and Uganda, random sampling was not practical because in thecommunities concerned there were relatively few households with latrines. Therefore apurposive selection was needed. Results and observations were based on comparing withcontrol groups, using baseline data and looking for direct evidence. Effects of externalvariables as access to water, duration of the project, education and socio-economic statuswere considered. Study was observed in two ways: End dates (A comparison was made ofprojects that ended in different years) and Study Dates (comparison was made of data thatwas collected at two different times, one year apart.). The results showed that hygieneinterventions had an impact on aspects of hand washing skills/ practice and latrine use/maintenance. Sustainability of hygiene behaviors infers that, in terms of behaviorsustainability, the amount of time since the projects ended did not make much of adifference. Hygiene behaviors were seen to be similarly sustained in projects that endedfour years or only one or two years before the study.

1.0 Introduction

The challenge in water and sanitation is indeed a challenge of monumental proportion. The

figures are well known: globally over a billion people still lack access to safe and reliable

water sources; another 2.4 billion do not have proper sanitation services, and more than 5

million people die each year from water related diseases. Though access to improved

water supply and sanitation facilities has been increasing, it is just sufficient to keep pace

with population growth.

As water and sanitation facilities being vital factor for good hygiene behaviors it expresses

direct relatedness in most of the disease incidences. Diarrhea, worm infestation and eye

and skin infections are diseases related to water and sanitation. About three million

children die from diarrhea each year. Each of the three common worms (roundworms,

whipworms and hookworms) is estimated to infect more than 500 million people. Roughly 6

million people have become blind from trachoma, an eye disease. Good hygiene can help

prevent much of this, saving lives and preventing illness. For example, it is estimated that

washing hands with soap can reduce the risk of diarrhea by more than 40%. Programmes

to promote hand washing might save a million lives each year. Simple hygiene behaviors –

that is what people do, their practices for Cleanliness – are key to improving health.

Hygiene promotion is therefore recognized nowadays as an essential part of water and

sanitation programmes if the maximum health benefits are to be gained from provision of

improved facilities.

Our health-related behavior is partly determined by a complex mix of our knowledge,

beliefs, attitudes, norms and customs. Socio-economic determinants and even political

factors also play important roles. Without the resources to construct and maintain water

supply and sanitation facilities, it is difficult to attain levels of personal, domestic and

environmental hygiene conducive to health.

According to WHO they include: hand washing, having and using latrines, safe disposal of

infant excreta and storing drinking water safely. Our research therefore studied those

behaviors.

2.0 The Research

2.1 Donors : European Union

2.2 Coordinator : IRC International Water and Sanitation Centre, The

Netherlands,

2.3 Duration : 3 years

2.4 Partners : Network for Water and Sanitation, Kenya (NETWAS);

Water Aid Uganda (WAU) working in collaboration with Uganda

Association for Socio-Economic Progress (USEP);

Volta Region Community Water Supply and Sanitation Agency,

Ghana (VRCWSA);

Nepal Water for Health (NEWAH);

COSI Foundation for Technical Cooperation, Sri Lanka (COSI);

Socio- Economic Unit Foundation, Kerala, India (SEUF)

3.0 Objectives:

• To assess the level of sustainability of behavioral change one to three years after a

hygiene promotion intervention;

• To develop a methodology for simple, cost-effective long-term monitoring of

behavioral changes;

• To gain insight into relationships between project approaches, External conditions

and sustainability of changes in hygiene behavior;

• To determine the policy and programming implications of the study findings as a

basis for influencing future policy and increasing the effectiveness of programmes.

4.0 Methodology4.1 Main research activitiesInitial meetings with researchers to design the study, develop the hypotheses to be

tested, and to make drafts of the collection tools and questionnaires. At this stage too the

data input sheets were designed.

Preparatory field work, which included activities such as translating the questionnaires

into local languages, training research assistants, field testing and amending the data

collection tools. In our studies the training of the research assistants was combined with

testing of the tools. Initially the research teams needed to get permission from communities

and, in some cases, from the project principals, to carry out the research.

Field work, which included selecting the communities, sampling households, identifying

community groups. Research assistants carried out the survey activities and recorded the

data in the data sheets.

Analyzing the information, which included checking the data sheets for mistakes and

‘cleaning’ the data, making totals for each question and item, entering these into

spreadsheets for further analysis. Finally, potential associations between the results were

analyzed,

Documenting, disseminating and promoting the use of research findings at the national

and international levels.

4.2 Sampling: Samples in the six-country study were done in Ghana, India and Sri

Lanka. In other countries, Kenya, Nepal and Uganda, random sampling was not practical

because in the communities concerned there were relatively few households with

latrines. Therefore a purposive selection was needed to identify households with latrines

so that an adequate sample size could be achieved. The sanitation data from these three

countries therefore tells us about households with latrines rather than about all the

households in the communities studied.

Table 1: Samples in the six country studyCountry ResearchInstitutionSample size Remarks

Sri Lanka COSI 6 communities

2001:110 HH

2002:150 HH

In 2003, there were 4 projects

(100HH) & 2 control

communities (50 HH)

Ghana VRWSP

10 communities

2001: 220 households

(HH*), 20 schools

2002: 220 HH, 20 schools

Sample had 5 communities

where intervention ended in

1998; 5 communities ended in

2000.

India SEUF

3 communities, 346 HH

2002: 10 communities, 345

HH plus informant

interviews

Intervention ended in different

years from 1993 to 2000.

Kenya NETWAS

International

2001: 6 communities, 215

HH plus 6 women’s groups

2002: 112 HH plus 6

women’s groups plus one

control group, 29 schools

One half of 2001 households

were re-surveyed in 2002.

Individual survey of women’s

group members in 2002; group

interviews in 2001.

Nepal NEWAH 6 communities

2001: 77 HH

73 HH in 4 hill communities

were dropped from the study

2002: 150 HH

2003: 242 HH plus focus

group discussions

Because of security problems.

Two of the remaining 6 had 2-

year interventions and were

Surveyed 2 times. Four had one

year

Interventions.

Uganda Water aid -

Uganda

6 communities

2001:221 HH

2002:180 HH Plus group

and informant interviews

2 communities in each of 3

ethnic groups

5.0 Results and observations:5.1 Comparing with control groups: Comparing communities or groups that had

hygiene education/ promotion with those that did not;

In the Sri Lanka study, the intervention communities tended to perform better than the

control communities in terms of:

• Latrine cleanliness (the floor free from faecal matter): 92% (72/79) for intervention

communities versus 4% (1/22) for the control group.

• Latrine shows signs of use: 96% (75/78) for the intervention versus 77% (17/22) for the

control communities.

• Child excreta is put in latrine: 47% (14/30) versus 17% (1 out of 6 households).

Only the first of these three sets of figures is significant; however, the trend of each is in the

expected direction, with the intervention households appearing to perform better.

In the Indian study, two large communities were selected. The communities had similar

access to water supplies and both had latrine subsidies. One had a sanitation and hygiene

project intervention lasting 7 years, with a hygiene promotion campaign and education

classes. The other had no hygiene promotion or education activities. Comparing the two

communities showed:

• For hand washing skills; the demonstration of how to wash hands correctly (using

soap/ash and rubbing both hands) was performed much better by the project group: 97%

(113/117) versus 10% (10/102) for the control community.

• For reported hand washing practice; always washing both hands with soap and water

was measured through pocket voting. This showed that the project households were

significantly more likely to wash hands consistently: 86% (282/326) compared to 6%

(14/222) for the control community.

• For location of soap (for hand washing) within the household; the premise was that, if

the materials for hand washing are conveniently located, it is more likely that people will

wash their hands. In this study only the project households were likely to have soap and

water convenient for hand washing: 93% (113/121) versus 0% (0/102) for the control

community.

• For household environment; the community with the hygiene project intervention had

significantly cleaner household compounds than the control community. 97% (117/121)

versus 35% (37/105).

An interesting finding was that there was no significant difference between the project

community and control community in:

• Knowledge of critical handwashing times (before eating and after defecation): 120/120

(100%) in the project households and 81/105 (87%) in the control households. Knowledge

was clearly not related to handwashing skills or practice.

• Latrine use and cleanliness: This was at the same level for the control and project

community. Consistent latrine use was shown by pocket-voting: 95% (311/326) in the

project community and 95% (211/222) in the control community. Latrine cleanliness was

the same: 94% (117/121) and 92% (92/105 in the control households). This may indicate

that promotional activities outside the project have been important.

5.2 Using baseline data: Comparing hygiene behaviors before and later or after the

intervention.

Example: Number of households having and using a latrine before and after a project.

Baseline information for two communities in Nepal was collected during group discussions

and group interviews.

Table 2: Latrine coverage Nepal

Community >

1 2

Initial latrine coverage (%)

Final latrine coverage (%)

Rise in coverage (%)

0

43

43

1

55

54

In the baseline study from India ‘before and after’ information was collected from several

separate communities.

TTable 3: Latrine coverage in IndiaCommunity >Kal Ang Koip Mara Kavo Kap Neen Alap Puth Kaip

Initial latrine

coverage (%)

52 15 55 43 39 18 41 24 32 38

Final latrine

coverage (%)

72 41 85 75 72 55 87 71 87 100

Rise in

coverage (%)

20 26 30 32 33 37 46 47 55 62

From these examples, we can see that the project made a difference in latrine ownership.

5.3 Looking for direct evidence: Assessing whether people who participated in

project activities had better hygiene behaviors than those who did not

In our six country study, information was collected in communities where the project had

ended two or more years earlier. We compared the hygiene behaviours of people and

households that had or had not participated in certain hygiene promotion and education

activities during the project.

Personal communication: In the Kenya study, people who said they had heard about

latrines and hand washing from other trained women’s groups and neighbors had

significantly better hand washing practice (p=0.037, OR=1.5).

Attending meetings: In the Ghana study, attending small group meetings was one project

activity that made a difference. People who reported to have attended meetings where

hygiene was discussed were more likely to have better hand washing skills as shown by a

demonstration (stratified by community: p=0.0014, R=2.20, CI 1.33-3.88).

Attending required hygiene classes: The India study showed that women who

remembered hygiene education classes between 2 and 9 years later were significantly

more likely to

• have good hand washing practice (p=0.007, OR=2.04, CI 1.05 - 3.96);

• know that washing hands before eating is important for health reasons (OR 2.9, CI 1.43-

6.0); and

• have household compounds that were clean, free of faeces and other waste (OR 2.8, CI

1.22-6.6).

This was not significant for men, as they did not usually attend the hygiene classes.

Overall exposure to hygiene promotion/education: In the India study, hygiene inputs

were measured in 8 ways: participation in activities, remembering classes, video/slide

shows, drama, competitions, women involved in organization, masons giving messages,

and the number of home visits. All of these were shown to have positive links with the hand

washing practices reported by all the women of the household, although only one (health

education classes, OR 2.04, CI 1.05-3.96) was statistically significant. That these linkages

are all in the expected direction is it self significant; the probability of it arising by chance is

only about 1 in 50.

From this we can see that project activities such as hygiene classes, group meetings or

encouragement by people who had been trained, have had an impact on hand washing

behavior. We can be fairly sure of this since it was studied two or more years after the main

project activities ended.

It was interesting to note that the more personal activities (attending meetings and classes,

hearing from a neighbor) seemed to show more direct impact than the mass activities.

External variablesAccess to water

It has been thought that providing water and sanitation services, including providing water

in or close to the home, can in itself lead to better hygiene behaviours. In our study, we

compared households with good and less good access to water. Access was measured in

different ways such as the time needed to collect water (Kenya, Nepal), the distance to the

source (Sri Lanka, India), the length of queues at water points (Ghana) or the reliability of

the supply (Kenya, Ghana).

It should be noted that, in general, access to water supply was fairly good. However, in

none of the six studies was there any significant relation between access to water and

hand washing knowledge, skills or practice, or latrine cleanliness and maintenance. Only in

one country study, Ghana, did households with worse access to water tend not to have

water and soap conveniently located for hand washing (Stratified by community: p=0.046,

OR=0.57, CI 0.35-0.99). This indicates that providing a convenient water service is

probably not, in itself, a sufficient inducement to good hygiene practices.

Duration of the project

In the Indian project the duration of the intervention did not appear to be related to

behavioral outcomes in the two communities where it was measured. Here it was

suggested that the project should last as long as needed to mobilize the community, to

organize groups and to carry out the work well.

Conversely the duration of the intervention did have some effect in the Nepal programme.

The two-year intervention communities performed better than those with one-year

interventions in some elements of domestic hygiene, such as covering food (p< 0.009) and

in hand washing skills, specifically, rubbing both hands (p<0.022).

Education and socio-economic status

In two countries where it was measured, the education of women was related to hygiene

practices. Women with more education tended to have healthier behaviours. In the Kenya

study better educated women were more likely to have hand washing knowledge, skills and

practice as well as consistent latrine use. The difference between women with more and

less education was significant in all cases (p<0.02). In Nepal, women with more education

tended to demonstrate better hand washing skills and more frequently located soap

conveniently for hand washing in the household (in both cases p<0.01). The indication was

that more educated women do better in adopting hygienic practices.

Socio-economic status and behaviours were compared in two studies: India and Sri Lanka.

In the India study, the hygiene behaviours of women were found not to be related to the

socio-economic status of the community. Their behaviours were linked rather to the

hygiene classes included in the project. Those classes were positively associated with

hand washing reported by women (OR 2.04, CI 1.05-3.96), with their awareness that

washing hands before eating is important for health reasons (OR 2.9, CI 1.43-6.0), and with

their knowledge of the importance of cleanliness of household surroundings, which were

free of faeces and other waste (OR 2.8, CI 1.22-6.6). However, in the same study in India

for men, the above findings were largely reversed. Their latrine or hand washing practices

showed no significant linkages with previous hygiene promotion activities but were closely

linked to the socio-economic status of the community (as rated by the project staff and the

government). Apparently the project had little impact on the habits of hand washing or

latrine use by males.

The inference was that men who lived in richer communities were more likely to use the

latrine consistently and to wash hands consistently. In this (Indian) project women were

more involved in hygiene promotion activities than men. So it appears that there is a

gender issue. We think that if both men and women had been involved in the hygiene

promotion/education activities, there might have been a measurable impact on both men

and women. Variables such as socioeconomic status would then have become less

important. As a result of the study the researchers suggested that, in general, if a hygiene

(and community) intervention is intense, with a strong gender and poverty focus, these

linkages to education and socio-economic status would be weaker.

Other interesting findings

In two country studies, Nepal and Kenya, latrines that were considered easy to use tended

to be better maintained. (Nepal p=0.05, Kenya p=0.041). Skills are related to practice.

People who showed how to wash hands correctly also tended to have better practice. In all

three relevant studies the demonstration of good hand washing skills was associated with

reported good hand washing practice (India p<0.00004, Kenya p=0.00002, Uganda for men

only p=0.038 OR=1.93). Knowledge is not necessarily related to skills. The relation

between hand washing skills and knowledge of critical hand washing times (after

defecation) was mixed. Hand washing skills were linked with the knowledge of the need for

hand washing after defecation for health reasons in Ghana (p=0.00006) and India

(p=0.002) but not in Kenya, Nepal, Sri Lanka or Uganda.

6.0 Sustainability of behaviors

Study looked at this in two ways: End dates and Study Dates

End dates: A comparison was made of projects that ended in different years.

In five countries, behaviours were surveyed in communities where the project interventions

had ended 1 to 4 years previously. Specifically, data was collected in the Kenya, Nepal,

Uganda and Ghana studies from communities where the interventions ended in 1998 and

2000. In one of these countries, Ghana, there were some follow-up visits to the

communities by project staff so that the project did not fully ‘end’ at the specified date. In

the sixth country, India, it was possible to collect data for 10 communities where the

Paper presented for Symposium on November 29, 2006 at PGIA, University of Peradeniya

www.cosi.org.lk

interventions had ended between 1 and 9 years previously. For the studies in five countries

25 comparisons were made between a behavior and the end date of the project. The

behaviours were: hand washing skills, hand washing practices (person washes hands with

soap and water), location of soap/water in the household, latrine shows signs of use,

person uses latrine consistently, latrine is the projects compared were 1998 and 2000.

The results show that in only 2 out of 25 comparisons made did the people practice safe

hygiene behaviours more where the projects ended in 2000 than where the projects ended

in 1998. This infers that the time elapsed after the projects ended did not make much of a

difference. Hygiene behaviours were seen to be similarly sustained whether the projects

ended 4 years or only 1 or 2 years before this study.

Study dates: A comparison was made of data that was collected at two different times,

one year apart.

In our studies, surveys were made in 2001 and 2002, about one year apart. The

behaviours compared were the same as before: The data was analyzed for four countries.

In 17 cases a comparison was made of hygiene behavior changes between the two data

collection dates. In only one of the 17 comparisons was there a significant change over the

one year period, inferring that, in general, the improved hygiene behaviours were being

sustained. The one exception was in the Uganda study where hand washing skills

decreased from 49% (42/86) in the 2001 survey to 35% (76/214) in the 2002 survey. This

was significant at the 95% level (p=0.045 , OR=0.58 CI 0.33-0.99) with changes seen in

two out of three districts.

7.0 Conclusion:

Four ways to study the impact of programme interventions on hygiene practices were

examined:

• comparing results of intervention and control groups,

• showing changes over time, using baseline information,

• finding evidence of direct links between inputs during the project period in terms of

hygiene activities and outputs after the project had ended in terms of hygiene

practices,

• examining some standard external variables such as improved access to drinking

water, education and socio-economic levels.

All of this was studied after the projects had ended. This is more difficult than determining

the impact of the project during or near the end of the intervention when activities are still

fresh in the minds of people. However, evidence of impact was still found even several

years after the interventions had ended.

The results showed that hygiene interventions had an impact on aspects of hand washing

skills/practice and latrine use/maintenance. Hygiene promotion activities associated

particularly with engendering new behaviours were those involving personal contact,

attendance at group meetings, required hygiene classes.

Duration: The number of years of the intervention did not appear to be related to

behavioral outcomes in the Indian project where it was suggested that the project should

last as long as needed to mobilize the community and to carry out the work well.

Conversely, duration of the intervention did have some effect in the Nepal programme

where two-year intervention communities were better than those with one year

interventions in some elements of domestic hygiene and in hand washing skills.

Access: The results indicated that just providing water is not enough to change

behaviours, as there were no significant links between access to water and hygienic

behaviours.

Education and economic status: The results suggested that, if the hygiene promotion

and education efforts are intense and with a strong focus on reaching the poor and

reaching both men and women (poverty and gender focus), then certain external variables

may fade in importance. Specifically, the external variables that might then have less

impact on performing hygiene behaviours may be: the education level of women, the socioeconomic

status of the community, the difference between women and men in hygiene

practices.

Sustainability of hygiene behaviours infers that, in terms of behavior sustainability, the

amount of time since the projects ended did not make much of a difference. Hygiene

behaviours were seen to be similarly sustained in projects that ended four years or only

one or two years before the study. For the study in India, where the projects ended later,

women were significantly more likely to wash both hands with soap and water (p=0.004)

and were significantly more likely to use the latrine when at home (p=0.00013 ). In other

words, hand washing and latrine use practice did seem to deteriorate with time. However,

the fall-off was not very great. Even where the project had ended seven or nine years

before the survey, about four out of five (80%) of the women were reportedly still

consistently using their latrines.

"Work with other like-minded organizations – in the Water Environmental Sanitation & Hygienesector and beyond – to help communities organize themselves and build the capacity to improvetheir own livelihoods through a sustainable development process – "Community Self-Improvementand Empowerment" - COSI