HOUSEHOLD WATER TREATMENT TECHNOLOGIES FOR MICROBIAL REMOVAL IN KABALE DISTRICT, SOUTHWESTERN UGANDA

Health problems associated with the consumption of untreated drinking water is one of the greatest concerns in Kabale District inspite of government’s efforts to provide safe drinking water to the people. Household water treatment and safe storage has been shown to be an effective means of reducing health problems associated with unsafe drinking water. The purpose of this study was to examine household water treatment technologies (HWT) and evaluate their ability to improve microbial quality of drinking water. The specific objectives of the study were to: (i) evaluate the different water sources, household water treatment technologies, and storage options of household drinking water, (ii) establish whether the sources of drinking water influence the type of water treatment technologies used at household level, (iii) determine whether there is significant difference between bacterial counts in household drinking water samples before and after treatment, and (iv) evaluate bacteriological effectiveness of household water treatment technologies used under laboratory conditions. The study employed both analytical and descriptive research designs utilizing mixed methodologies. A multistage sampling technique was used to select 205 respondents, who were used to obtain socio-economic data, using semistructured questionnaires. Drinking and source water samples were collected from households and sources of drinking water reported with high pathogenic bacteria concentration respectively for Escherichia coli and total coliforms analyses. World Health Organization (WHO) drinking water quality guidelines were used to categorize drinking water in terms of risk level category. Statistical package of social sciences was used for data analysis. Chi square test was used to test whether sources of drinking water influenced the type of water treatment technologies used at household level. A paired sample T-test was used to compare mean difference between bacteria counts in household drinking water samples before and after treatment. A one way ANOVA was used to compare mean differences in bacteria reductions by different HWT in experiment test water samples. Descriptive statistics were used to analyze data. Majority respondents (61.5%) were using springs as their sources of drinking water. Of 46 household treated water samples, 17.4% and 45.7% of water samples fell in no risk category (0 CFU/100 ml) for total coliforms and Escherichia coli respectively. Of 20 experiment treated water samples, 40% and 73% of samples fell in no risk category (0 CFU/100 ml) for total coliforms and Escherichia coli, respectively. Treatment by application of WaterGuard tablets achieved highest total coliforms removal with 99.5% (1.9 log10), whereas WaterGuard tablets, biosand filtration method, and aqua safe tablets achieved complete removal of Escherichia coli (100%) under laboratory conditions. Chi square test yielded no significant relation between drinking water sources and the type of HWT used at household level (P <0.05). The paired samples T-test showed a significant difference between bacteria counts before and after treatment. Significant differences were observed between mean bacteria reductions in experiment test water samples. Spring water in Kabale District was found unsafe to drink unless treated. Effective water treatment products such as WaterGuard and aqua safe tablets should be promoted at local level. Local people should always be involved in simple household testing to reduce doubts on microbial efficiency of newly introduced HWT.


OPERATIONAL DEFINITION OF KEY TERMS AND CONCEPTS
Household water treatment technologies: Household water treatment technologies are methods employed for purposes of treating and storing drinking water at household level. Sikod et al. (2015) defined household water treatment and storage as methods employed for purposes of treating water in the home. Household water treatment technologies are also referred to as point-of-use water treatment technologies. The technologies encompass a range of options that enable people or communities to remove or inactivate pathogenic microbes in drinking water. In Kabale District, the technologies include; boiling, biosand filtration, let it stand and settle, application of WaterGuard and application of aqua safe tablets.
Improved drinking water source: WHO and UNICEF (2012) define improved drinking water as ones that are by nature of their construction or through active intervention protected from outside contamination, and in particular from contamination with faecal matter. Improved drinking water sources include; piped water on premises, yard or plot, standpipes, protected springs, protected boreholes and protected dug wells (WHO & UNICEF, 2011). Other improved drinking water sources are ones except piped drinking water source on the premises that are by nature of their construction protected from outside contamination especially feacal matter (WHO & UNICEF, 2011). In Kabale District, improved drinking water sources are piped water into the dwelling, stand pipes, protected springs and protected boreholes.
Unimproved drinking water sources: WHO and UNICEF, (2011) defined unimproved drinking water sources as ones that are by nature of their construction, location and management unprotected from outside contamination. Such drinking water sources include; unprotected springs, carts with small tanks, unprotected dug wells, tanker trucks, and surface waters such as rivers, dams, lakes, pond and stream (WHO & UNICEF, 2011). In Kabale District unimproved drinking water sources include; unprotected dug wells, unprotected springs, and surface waters such as stream water, ponds and water direct from the lakes. xvii Safe drinking water: UNICEF and WHO (2012) defined safe drinking water as water with microbial, chemical and physical characteristics that meet WHO drinking water quality guidelines or national standards for drinking water quality. In this case, safe drinking water refers to that water with bacteriological characteristics that meet national and international drinking water quality standards. According to UNBS, (2008), safe drinking water should have no detectable total coliforms or Escherichia coli per 100 ml of a given drinking water sample.
Microbial removal: In this case, microbial removal refers to eliminating total coliforms from drinking water.
A household: A household is defined as persons who have eaten and slept under the same roof for at least five days of the week (Valerie, 2010).
Log 10 reduction: This is a mathematical term that shows the relative number of live microbes eliminated from a water sample after water treatment or disinfection.

Safe water sources:
Water sources with bacteriological characteristics that meet national and international source drinking water quality standards.

ABSTRACT
Health problems associated with the consumption of untreated drinking water is one of the greatest concerns in Kabale District inspite of government's efforts to provide safe drinking water to the people. Household water treatment and safe storage has been shown to be an effective means of reducing health problems associated with unsafe drinking water. The purpose of this study was to examine household water treatment technologies (HWT) and evaluate their ability to improve microbial quality of drinking water. The specific objectives of the study were to: (i) evaluate the different water sources, household water treatment technologies, and storage options of household drinking water, (ii) establish whether the sources of drinking water influence the type of water treatment technologies used at household level, (iii) determine whether there is significant difference between bacterial counts in household drinking water samples before and after treatment, and (iv) evaluate bacteriological effectiveness of household water treatment technologies used under laboratory conditions. The study employed both analytical and descriptive research designs utilizing mixed methodologies. A multistage sampling technique was used to select 205 respondents, who were used to obtain socio-economic data, using semistructured questionnaires. Drinking and source water samples were collected from households and sources of drinking water reported with high pathogenic bacteria concentration respectively for Escherichia coli and total coliforms analyses. World Health Organization (WHO) drinking water quality guidelines were used to categorize drinking water in terms of risk level category. Statistical package of social sciences was used for data analysis. Chi square test was used to test whether sources of drinking water influenced the type of water treatment technologies used at household level. A paired sample T-test was used to compare mean difference between bacteria counts in household drinking water samples before and after treatment. A one way ANOVA was used to compare mean differences in bacteria reductions by different HWT in experiment test water samples. Descriptive statistics were used to analyze data. Majority respondents (61.5%) were using springs as their sources of drinking water. Of 46 household treated water samples, 17.4% and 45.7% of water samples fell in no risk category (0 CFU/100 ml) for total coliforms and Escherichia coli respectively. Of 20 experiment treated water samples, 40% and 73% of samples fell in no risk category (0 CFU/100 ml) for total coliforms and Escherichia coli, respectively. Treatment by application of WaterGuard tablets achieved highest total coliforms removal with 99.5% (1.9 log 10 ), whereas WaterGuard tablets, biosand filtration method, and aqua safe tablets achieved complete removal of Escherichia coli (100%) under laboratory conditions. Chi square test yielded no significant relation between drinking water sources and the type of HWT used at household level (P <0.05). The paired samples T-test showed a significant difference between bacteria counts before and after treatment. Significant differences were observed between mean bacteria reductions in experiment test water samples. Spring water in Kabale District was found unsafe to drink unless treated. Effective water treatment products such as WaterGuard and aqua safe tablets should be promoted at local level. Local people should always be involved in simple household testing to reduce doubts on microbial efficiency of newly introduced HWT. ii. Does the source of drinking water influences the type of water treatment technology used at household level?
iii. Is there any significant difference between bacterial counts in household drinking water samples before and after treatment? iv.
What is the bacteriological effectiveness of household water treatment technologies used under laboratory?

Research hypotheses
This present study tested the following hypotheses: i. The source of drinking water does not influence the type of water treatment technology used at household level.
ii. There is no significant difference between bacteria counts in household drinking water samples before and after treatment.
iii. There are no significant differences between mean bacteria reductions by different HWT under laboratory conditions.
1.6 Objectives of the study

General objective
To examine household water treatment technologies in use and evaluate their ability to improve the microbial quality of drinking water at household level in Kabale District, southwestern Uganda.

Specific objectives
This study was addressed by the following specific objectives: i. To evaluate the different water sources, household water treatment technologies, and storage options of household drinking water.
ii. To establish whether the sources of drinking water influence the type of water treatment technologies used at household level.
iii. To determine whether there is significant difference between bacterial counts in household drinking water samples before and after treatment. iv.
To evaluate bacteriological effectiveness of household water treatment technologies used under laboratory conditions.

Significance of the study
Results of this study will serve to enlighten water users' safety about waterborne diseases. The study results will provide information that aid in formulating policies on safe drinking water in Kabale District. The results of this study will assist the relevant authorities in designing appropriate mitigation measures to ensure that domestic water supplies are protected. Water quality data will enable policy makers to decide and implement best practices that protect human beings from dangers resulting from drinking contaminated water. The study has added on to the existing literature in IWRM and with reliable performance data under laboratory conditions. This will speed-up and scale-up the implementation of household water treatment technologies in Kabale District.

Scope and limitation of the study
The study was conducted in Kabale District, southwestern Uganda. The study identified and described the different sources of drinking water and household water treatment and storage option.

Conceptual framework
The conceptual framework (Figure 1.1

Sources of drinking water
According to WHO (2004), water is considered safe to drink as long as it does not cause any significant health risks over life time consumption. According to WHO and UNICEF (2008), drinking water sources are classified into three main categories; piped water on premises, other improved drinking water sources, and unimproved drinking water sources. Components of each category are as described in the subsequent sections.

Piped water on premises
Piped water connection on user's premises, yard or plot is the ideal service, since it provides the most convenient supply and has positive health impacts (WHO & UNICEF, 2011). Globally good progress has been made in the use of piped drinking water on premises. A report by WHO and UNICEF (2008) revealed that 2.5 and 1.1 billion people in urban and rural areas respectively have access to piped drinking water connection on their premises.

Other improved drinking water sources
The other improved water sources are protected shared community sources such as public taps, rainwater harvesting, protected dug wells, protected springs and boreholes (Bain et al., 2014;UNICEF & WHO, 2011). Since 1990, use of other improved drinking water sources has substantially increased. In Southern Asia, the population using other improved drinking water sources increased from 54% to 65 % between 1990 and 2008 (UNICEF & WHO, 2011). In Sub-Saharan Africa, use of other improved water sources increased from 33% to 42% between 1990(UNICEF & WHO, 2008, 2011. The study did not establish the characteristics of other improved drinking water sources and the possible sources of contamination.
The current study intends to fill these gaps with information.

Unimproved drinking water sources
Unprotected springs, unprotected dug wells, tanker trucks, and surface waters such as rivers, dams, lakes, pond and streams are considered "unprotected" drinking water sources UNICEF & WHO, 2008). WHO and UNICEF (2008) further revealed that 13% of the world's population without access to safe drinking water sources gets drinking water from unimproved sources. Sub-Saharan Africa has the largest population using unimproved water sources, though its figures dropped from 51% in 1990 to 42% in 2006(WHO & UNICEF, 2008. The study did not establish the characteristics of unimproved drinking water sources and the possible sources of contamination. This research study seeks to fill these gaps with information.

Household water treatment and storage options
Household water treatment technologies have been proved to be a sustainable solution for developing countries facing challenges in providing safe drinking water to their people (Clasen et al., 2007;Fewtrell et al., 2005). Sobsey (2002) found out that simple and relatively inexpensive HWT have the potential to substantially improve the quality of drinking water by reducing microbial counts and the risks of illness and death. Clasen et al. (2006) reported 60 to 90% reduction in the incidences of diarrheal diseases when an appropriate implementation of HWT was considered in communities which had no access to safe drinking water supplies.
In countries like Pakistan, Bangladesh, Nepal, India, Guinea Bissau, Mauritania, Sierra Leone and Egypt, household water treatment is largely limited to richer quintiles, whereas in Indonesia, Lao peoples Republic, Bosnia, it is lower in richer quintiles (UNICEF & WHO, 2011). Rosa and Clasen, (2010) reported that use of HWT is most common among countries within the Western Pacific WHO region with 66.8% and least common in the Eastern Mediterranean and Africa with 13.6% and 18.2% respectively.
In a study by Proto et al. (2014) on one-year surveillance of the chemical and microbial quality of drinking water shuttled from the main land (Naples, Campania, Italy) to the Eolian Islands indicates that water delivered to the islands met drinking water quality standards. However, during summer when the demand for drinking water increased Eolian Islands population increased too, the quality of the distributed water decreased most likely due to either use of vessels that were not specifically built for drinking water supply or could have used older drinking water supply containers. Such results suggest an implementation of water treatment at household level to manage the potential risk of waterborne diseases more effectively.

Examples of household water treatment technologies and storage options
Several studies have revealed that HWT such as boiling, filtration, chlorine addition, However, these methods are used in some parts of the developing world.

Boling method
Boiling method is the oldest method used to remove pathogenic bacteria from drinking water at household level. WHO (2004) recommends bringing water to a rolling boil point temperature (100 o C) to ensure any pathogenic bacteria in water are killed. Agrawal and Bhalwar (2009) found out that heating water to as little as 55 o C for several hours has proved to dramatically reduce non-spore forming bacterial pathogens in drinking water. The overall use of boiling method in Western Pacific region is 58.7%, least in the Eastern Mediterranean region (4.0%), and Africa (4.5%), where Uganda takes a lead in use of the method with 39.8% followed by Zambia with 15.2% (Rosa & Clasen, 2010). The overall global use of other adequate technologies such as chlorine is 5.6%, filtration is 4.3%, and solar disinfection is 0.2% (Rosa & Clasen, 2010). Brown and Sobsey, (2012) found out that boiling reduces the level of Escherichia coli by 98.5%. In addition, of all tested water samples, 44% never had any Escherichia coli at all. The study did not establish the reasons for use of boiling method of water treatment at household level, neither the effectiveness of other popular household water treatment technologies such as filtration, chlorination, and solar disinfection.

Chlorination
Chlorination is widely used means of water purification and very effective against most bacteria and some viruses (WHO, 2006). Examples of chemical products used for chlorination in the study area are; WaterGuard and aqua safe tablets. The effectiveness of chlorination depends on the correct dose of chlorine solution being added to the volume of water to be treated. After chlorine is added to water, it then is mixed and allowed to stand for 30 minutes before water is safe for consumption.
However, chlorination is not effective in highly turbid water (Nath et al., 2006).
A study by Boisson et al. (2013) reported that use of chlorine in treating water at household level significantly reduce faecal contamination and improves microbial quality of drinking water. However, the extent to which the population continues to regularly use chlorine to treat drinking water is not clear. A study by Levy et al. (2014) noted inconsistency of chlorination on microbial removal in a controlled setting and in household setting. According to Halder et al. (2014), chlorine was used to sanitize water transport and storage containers by 32% and 50% respectively of all surveyed individuals in the department of Yoro in rural Honduras. The study did not quantify the percentage of total coliforms eliminated as a result of disinfection by chorine. A study by Mwabi et al. (2012) in rural communities of Southern Africa reported that biosand filters have the capacity to produce 0.81-6.84 liters per hour, an indicator that it can produce approximately 25 liters per day, the recommended average water a person can use on average. The study also reports a reduction in bacteria by 99% -100% when using biosand filter. However, the study did not consider effectiveness of popular household water treatment technologies in southern Africa such as boiling, solar disinfection, let it stand and settle method among others.

Biosand filtration
Furthermore, the study didn't establish the proportion of the people using household water treatment technologies in Southern Africa.
In a study conducted by Tellen et al. ( 2010), 154 households water samples were tested. The median CFU/100 ml of Escherichia coli from water source, filter spout and storage vessel were 313, 72, and 144, respectively. In addition, Tellen et al. ( 2010) reported 98% and 99% reductions in total coliform, fecal coliform and fecal Streptococci for traditional biosand filtration and improved biosand filtration, respectively. This study did not establish the reasons for the discontinued use of the of biosand filtration method, possible factors for increase in CFU in treated stored water based on scientific evidence.

Solar disinfection (SODIS)
Solar disinfection is a simple method used to improve the quality of household drinking water by using sunlight to inactivate pathogens. The method involves filling transparent plastic bottles with water and exposing them to full sunlight. Exposure times vary from 6 to 48 hours depending on the intensity of sunlight. Use of solar disinfection can inactivate 97% of bacteria and 99% of virus (UNICEF, 2008).
Improving microbial quality of drinking water using solar radiation is well known to inactivate bacteria and its effectiveness depends on local conditions (Dessie et al., 2014). Nalwanga et al. (2014) asserted that use of solar disinfection led to satisfactory bacterial inactivation (log 10 reduction values >6 units for 11 of 13 experiments. Rainfall and cloudy conditions were the factors responsible for incomplete inactivation of bacteria that was observed (Nalwanga et al., 2014).

Household water storage options
Regardless of whether the collected or treated household water is initially of acceptable microbiological quality, it often becomes contaminated with pathogens of fecal origin during transportation and storage due to unhygienic storage and handling practices (WHO, 2014). Use of narrow-mouthed water storage vessels with hard cover or lid together with good hygiene practices reduce recontamination of drinking water by hands especially after boiling (WHO & UNICEF, 2011). Storing household drinking water in a covered container was associated with production of safer drinking water than storage in an uncovered container (Brown & Sobsey,

2012).
A research study by Adade et al. (2014) revealed that bacteria count in water stored in earthen pots water exceeded the World Health Organisation and Ghana Standard Board specified drinking water quality limits. Total coliforms ranged from 9 to 5.84 X 10 2 CFU/100 ml with a mean of 2.47 x 10 2 CFU/100 ml. Earthen pot stored drinking water from the various communities was contaminated due to unhygienic handling practices such as dipping of hands and utensils into the storage earthen pot.
However, the study neither established the level of bacteria before it was stored nor established the reasons for using earthen pots than other water storage options. The current study intends to fill these gaps with information.

Reasons for use of different household water treatment technologies
Several studies have reported various reasons for use of different water treatment technologies at household level. According to UNICEF and WHO (2011) A research study by Green (2008), reports that both rural and urban households expressed their preference to traditional HWT since they are long lasting and durable. Time factor was ignored by many respondents though a few of them revealed their preference to some methods that required short time to treat drinking water. Urban respondents reported that technologies that require long treatment time are difficult and limit their use. Cost factor also influenced the decision to use different household water treatment technologies though it did not have significant impact in the rural areas. The study did not establish the proportion of people using different household water treatment technologies. The study intends to fill the identified gaps with information.
A research study by Kakulu (2012) reported that 25% respondents were ignorant about other forms of household water treatment technologies besides boiling method.
The study further noted that 20% and 19% of respondents attributed the use of boiling method to cost-effectiveness and microbial removal effectiveness respectively. Further, four out of five (80%) households reported that chlorine was effective in disinfecting drinking water, whereas six out of nineteen (33.3%) respondents favored a cloth to strain drinking water. The study did not determine the average level of colony forming units as per WHO guidelines for drinking water quality. This study intends to fill this gap with information.

Bacterial parameters
This section outlines the parameters used to determine household water treatment efficiency, ease of analysis, some implications and responses related to finding a positive sample. Table 2.1 shows detectable levels of indicator organisms based on WHO and UNBS drinking water quality guidelines.

Total coliforms
Total coliforms are Gramma-negative, non-spore forming rod-shed bacteria capable of growing in the presence of bile salts, or other surface-active agents with similar growth-inhibiting properties, oxidase-negative, fermenting lactose at 34 -37 o C with the production of acid, gas and aldehyde within 24 − 48 hours (Payment et al., 2003).
Due to the fact that some total coliforms of non-feacal origin can be present in natural water, their presence in untreated water can be tolerated. However, when used as an indicator organism for treatment efficiency, their presence should not be detected in treated drinking water. Their detection in treated water provokes an immediate investigation.
Total coliforms provide basic information on the quality of drinking water though not an index of feacal pollution. They are preferred indicator organisms because they are easy to detect and enumerate in a water sample by simple, inexpensive cultural methods that require basic routine bacteriological laboratory facilities with welltrained laboratory technicians. The presence of total coliforms in drinking water sources may be as a result of surface water infiltration or seepage from a septic system (Anwar, Lateef & Siddiqi, 2010).

Escherichia coli
Escherichia Verhille (2013) described Escherichia coli as the best up to-date microbial indicator available to determine the health risks associated with the consumption of poor quality water. A research study by Odonkor and Ampofo, (2013) reported that two key factors made Escherichia coli a popular preferred indicator organism for feacal pollution in drinking water. Some faecal coliforms are non faecal in origin unlike Escherichia coli. The study further reported that Escherichia coli provide the best bacterial indication of faecal contamination in drinking water.
WHO drinking water quality guideline for Escherichia coli is 0 detectable per 100 ml (Table 2.1). This implies that in order to comply with the guideline, for every 100 ml of drinking water tested, no Escherichia coli should be detected in the water sample.

Study area
This study was carried out in Kabale District located in Southwestern Uganda.

Research design
The study employed analytical and descriptive research designs utilizing mixed methodologies in which both quantitative and qualitative approaches were used in data collection and analysis.

Sample size
The sample size was calculated using simple random sampling formula in order to get a representative number of households to use in the study. The formula developed by Israel (1992) was used to quantify the minimum sample size because it is most appropriate when using simple random sampling design, and yields a good sample size necessary for impact evaluations.

Sampling procedure
The study employed probability sampling with multistage sampling technique ( Figure 3.3). At the first level of sampling, all household heads in Kabale District were targeted during the study. In this case, the sample population was divided into four clusters based on county and municipal political boundaries. At the second level of sampling, each cluster formed was disaggregated into small groups (sub-counties) called strata. Four strata each one from the disaggregated clusters were randomly selected. The up-to-date number of households from each selected stratum (Table   3.2) obtained from respective sub-county chiefs was used to determine the proportionate number of households selected using a simple formula illustrated in equation 3.3. Households from each selected stratum were picked randomly during field study visits.

Field study survey
The purpose of the field survey was to establish the socio-economic and demographic characteristics of households, identify sources of drinking water, HWTS and to establish the reasons for use of different HWT. Semi-structured questionnaires (Appendix I) were used to collect socio-economic and demographic data, data on drinking water sources, HWTS and the reasons for use of different HWT. The questionnaire was divided into three sections. Section A assessed the socio-economic and demographic characteristics of respondents, section B assessed sources of drinking water and section C assessed HWTS.

Water sampling
For each household visited, both treated and untreated water samples (if the household head reports them available) were collected aseptically in sterilized 500 ml bottles during unannounced visits. For purposes of quality, field samples were collected in duplicates. The bottle corks were shielded with aluminum foil in order to avoid any form of hand contamination and adhere to aseptic techniques. The researcher assigned identification numbers to each water sample and recorded the time of sampling, type of the sample (whether treated or untreated) and the technology used to treat it. The number of days the treated water had stayed up to sampling time was recorded. All water samples were stored at 4 o C before analysis.
Experiment test water samples were collected from four unprotected water springs namely; Sapato, Hamwaro, Mukakyenkye and Kirigime water springs reported with high bacterial contaminants during our field study visits. Thereafter, each sample unit was treated by all the five treatment technologies under study that were identified during the field study survey. These include; biosand filters, aqua safe tablets, boing method, WaterGuard tablets and let it stand and settle method. All the water samples were tested for both total coliforms and Escherichia coli before and after treatment in accordance with standard methods for the examination of water and wastewater (APHA, 1981).

Pre-testing of the research instruments
Research instruments for this study were tested in the northern Kabale municipality prior to data collection. Questionnaires were tested to check whether they generate the intended data. Errors identified were noted and corrected before the actual field study.

Data management
Filled questionnaires were checked for completeness at the end of each data collection day within the field to identify any missing data before leaving the field.
Water samples were diligently labeled in the field and transported at room temperature to the laboratory for safe storage and analysis. Water samples were stored at room temperature for about 2 to 3 hours before they were analyzed.

Quality assurance and quality control
When assessing effectiveness of HWT for microbial removal, quality control and quality assurance are absolutely important. Standard methods and procedures were employed with all relevant parameters such as detection limits, repeatability and storage conditions of samples. Upon collection, samples were immediately placed on ice in coolers and then transported to the laboratory at 4°C. Upon arrival at the laboratory, samples were refrigerated at 4°C, and analyzed within six hours of collection. Previous studies indicate that such a holding time has little effect on measured total coliforms and Escherichia coli concentrations at temperatures less than 10°C, although analysis should always be conducted as rapidly as possible (Pope et al., 2003). All the bacterial analyses were carried out at NWSC-Kabale area water laboratory.

Laboratory analysis
Membrane filtration method was used in analysis of water samples in accordance with standard methods for the examination of water and wastewater. 100 ml of water were aseptically drawn from each unit of the samples and filtered through a 0.45 μm millipore filter membrane. The membrane was aseptically removed from the filtration unit by using sterile forceps and placed on the medium in the petri-dish in a rolling motion to avoid entrapment of air. Total coliforms and Escherichia coli counts were determined by incubating the membrane filter on Hichrome media at 37 o C and 44 o C for 24 hours, respectively. In order to ascertain the number of total coliforms and Escherichia coli contained in each incubated sample, colony forming units (CFU) developed on the membrane filter after incubation were counted.
Effectiveness of HWT was determined by testing the bacteriological quality household water samples (before and after treatment) and extended laboratory testing for bacterial reduction in source water by HWT identified during unannounced field study visits. Reduction of total coliforms and Escherichia coli indicator organisms in water samples before and after treatment was key performance outcome measured.
Enumeration of total coliforms and Escherichia coli before and after treatment was done following standard methods for the examination of water and wastewater (APHA, 1981(APHA, , 1995Rice, Bridgewater, & APHA, 2012;WE Federation, 2005).

Statistical analysis of data
Descriptive statistics were used to analyse socio-economic, demographic and laboratory data. Pie charts, bar graphs, and line graphs, were used to explain the bacterial characteristics of water samples before and after treatment in relation to WHO standards for drinking water quality, establish the different sources of water, HWTS, and the reasons for use of different household water treatment technologies in Kabale District, southwestern Uganda.
Chi square test was used to test the association between socio-economic and demographic factors and the type of water treatment technologies used at household level. A Paired samples T-test was used to compare mean difference between bacteria counts in household drinking water samples before and after treatment. One way ANOVA was used to compare mean differences between bacteria reductions by different HWT in experiment test water samples. All tests were compared at 95% confidence level. If the probability (p) value was less than 0.05, the test was significant but p-value greater than 0.05 showed no significant difference between the variables compared. The statistical tests were performed in SPSS version 17.0.

Ethical considerations
Before the actual study, the researcher requested for permission from respective subcounty administrators two weeks prior to the actual field study. Informed consent from the respondents was obtained during the field study and participation was voluntarily. The identity and information given by the respondents was confidential.
Only questionnaire numbers helped to link respondents' answers to water testing results in the laboratory.

Introduction
This chapter presents the results and cover discussion of findings from the field survey and laboratory analyses conducted in Kabale District southwestern Uganda.
Results on sources of drinking water, household water treatment and storage options used and the reasons for use of different household water treatment technologies are discussed in this chapter. Bacteriological quality of treated drinking water is evaluated to determine the effectiveness of water treatment technologies at household level.

Socio-economic and demographic characteristics of respondents
Socio-economic and demographic characteristics evaluated in this study include; gender, age, occupation, level of income, marital status, level of education and household size. These characteristics were selected because they influence the use of different household water treatment technologies.

Gender of respondents
Out of 205 household heads that were randomly sampled, 81.5 % were female respondents and 18.5% were male respondents (Figure 4.4). Gathering household information from both sexes on the same issues provided gender perspectives and helped in assessing the reliability of the responses from the field study survey.
Women in rural areas were more involved in regular farm activities than men thus, with limited time to undertake household activities such as household water treatment. Schmidt and Cairncross (2009) reported that household water treatment contributes less to economic and educational activities of women since it does not affect water supply. Instead, household water treatment increase work load of household members, especially women, although probably not to a very substantial extent.

Occupation of respondents
As shown in figure 4.6, majority of respondents were farmers (54.1%) followed by daily labourers (10.7%), business people (8.3%) and salaried workers (6.3%). It was further noted that 20% did not identify their sources of income. Majority

Marital status of respondents
As shown in figure 4.7, 66.8% respondents were married, 13.2% were unmarried, 2.9% were divorced and 3.9% were single parents. It was revealed that 13.2% had lost either of their spouses.

Education level of respondents
The study findings revealed that majority of respondents had attained education to a certain level. The study revealed that 46.3% had attained primary level education, 24.9% had attained secondary level education and 3.9% had attained education equivalent to a diploma. Additionally, 1% had attained education equivalent to advanced certificate and university degree ( Figure 48). Additionally, 23.9% of respondents did not have any formal education (Table 4.

Sources of household drinking water
The study findings revealed that 61.5% of respondents get their drinking water from water springs, 19% from public taps, 4.4% from private taps, 8.8% from neighbors' taps, 1.5% from shallow wells and 0.5% from boreholes (Table 4.4). 3.9 0.5 0.5

Education Level
The taps and boreholes were both privately and publicly owned. Water springs were publically owned. Shallow wells were constructed by local people faced with limited access to safe drinking water. Most toilets in the study area were constructed in less than 20 meters away from water springs (Plate 4.1). This explains why majority of water samples from springs were associated with high total coliforms and Escherichia coli concentrations. In addition, crops were seen grown in less than 10 meters away from water sources.
Constructing toilets in close proximity to drinking water source puts it at risk of feacal contamination. In related study by Sivaraja and Nagarajan (2014) found out that River Cauvery in India was loaded with coliform bacteria attributable to raw sewage. Abdulkadir et al. (2015) reported high concentration microbes in drinking water sources located in close proximity to pit latrines. Establishing crop farms in close proximity to drinking water sources puts it at risk of chemical contamination especially when fertilizers, organic manure and pesticides applied on crop farms end up in drinking water sources by runoff water after heavy rains. Nassar (2015) reported high nitrate concentration in the majority of wells with typical values of 100 -300 ppm, and exceeding 600 ppm in some other areas of the Gaza Strip.

Plate 4.1: Water springs (Source: Field survey)
Shallow wells were unprotected and their mode of formation and location exposed them to high contamination. One would first step in water in order to get well positioned to collect it. High turbidity and sediment levels characterized wells observed (Plate 4.2). Runoff from crop farms in the shallow wells' upstream was responsible for high turbidity and sediment levels. In such situation, chemical compounds such as nitrogen and phosphorus from fertilizers applied on crop farms, metals and toxins from farm tools may end up in drinking water sources as a result of runoff water after rains.
Toilet Toilet

Plate 4.2: Unprotected shallow wells (Source: Field survey)
Regarding water availability at the source, 182 (88.8%) respondents reported that water was always available whereas 23 (11.2%) reported that it was not always available for use (Figure 4.9). In rural Kabale District, irregular water supply is linked to two major reasons; when cleaning water supply tanks, water supply may be limited for a short time. When there is a technical problem in water distribution system, people do not get water at their taps in some places. In urban Kabale District, especially, in areas supplied by national water and sewerage corporation (NWSC), irregular water supply is linked to three major reasons; Power outages: When power goes off for some hours or days, water pressure goes down and water flow in the supply pipes suddenly stops. Mechanical problems such as pipe busting lowers water pressure and thus water fail to reach to some water supply points.
Cleaning water supply tanks sometimes also affects water supply especially when

28/03/2015
water demand is much higher and reserve tanks cannot handle the overwhelming water demand.

Figure 4.9: Pie chart showing percentage response of respondents on availability of water from source
Regarding frequency of collecting water from the source, the present results show that 20.5%, 42%, and 24.6% were collecting drinking water from the source once in a day, two times in a day, three times in a day respectively. Additionally, 6.8%, 2.9% and 3.4% were collecting drinking water from the source, four times in a day, five times in a day and more than six times in a day respectively (Table 4.5). The study revealed that respondents whose drinking water sources were water springs were collecting drinking water from the source more frequently than those whose sources were private or public taps. This means that type of drinking water source influences the number of times water users can collect water from source per day.

Household water treatment
The current study reveals that 75.6% of households were using any one of the following HWT: boiling, biosand filtration, let it stand and settle method, application of WaterGuard tablets and application of aqua safe tablets. Additionally, 24.4% were not using any HWT to purify their drinking water. Research has shown household water treatment as the most effective way of preventing waterborne diseases such as diarrhea, typhoid, and cholera among others. Schmidt and Cairncross (2009) found out that a part from potential to reduce diarrhea, household water treatment improves aesthetic appeal of drinking water and reduces indoor air pollution.
Of all HWT reported, boiling was noted the most common method both in rural and urban areas, and was used by both educated and uneducated people in the study area.
Let it stand and settle is common in rural areas, practiced by people with low education level and the elderly. Biosand filtration method is latest of all HWT however; its uptake has considerably been low since its inception.
The study findings revealed that, out of 155 respondents who reported treating water with any one technology, 67.3% reported they were using boiling method, 1% reported they were using aqua safe tablets, 4.4% reported they were using let it stand and settle method, 2.4% reported they were using biosand filtration method and 1.5% reported they were using WaterGuard tablets (Table 4.6). The results of this study concur with research findings by Kakulu (2012) that reported boiling method as the most popular household water treatment method with (43.6%).
The present study findings do not support Wright and Gundry (2009) who found out that water filtration and let it stand and settle methods were in widespread use. In addition, the results of this study do not support Sreenivasan et al. (2015) who found out that majority respondents (75%) were using WaterGuard as their most preferred water treatment method.  Table 4.7 shows the reasons why respondents were using different HWT. It was found that the majority who were using boiling because it was easy to use (23.2%), cheap to use (20.6%), and makes water safer to drink (7.7%). The perception that boiling is easy to use has been a result of sensitization from the ministry of health through VHTs. The study further revealed that respondents were using boiling because it had positive health impacts (9%). This indicates that awareness of the health benefits of household water treatment by VHTs has been really understood.
Other respondents revealed they were using boing because they did not about others (18%). The study results support Sodha et al. (2011), who found out that 83% respondents were using boiling because it was practical, 73% said it was easy to use, 90% believed it was cheap and only 4% felt that boiling was difficult. The study findings also support Kakulu (2012) who found out that majority of respondents were using boiling method because the method was cheap to use and did not know about other treatment options. Respondents were using let it stand and settle method because it was easy to use (66.7%), cheap to use (33.3%).
Some respondents reported they were using biosand filtration because it was easy to use (40%), cheap to use (40%), and required less time to produce ready to drink water (20%) as indicated in table 4.7. In addition, 20% of respondents reported they were using biosand filtration method because it was able to produce much volume of drinking water in a short time.
It was found out that respondents were using WaterGuard and aqua safe tablets because they believed the tablets were easy to use, cheap to use and required less time to produce drinking water (Table 4.7). Stockman et al. (2007) found out that 68% were using WaterGuard because they believed it makes water safe, 21% believed it prevents diarrhea and 10% didn't justify why they were using WaterGuard. Freeman et al. (2009) found out that WaterGuard which was commercially launched less than one year before the evaluation had a lower level of use because it was perceived to be more expensive water treatment product at US$ 0.02 per 100 litres than other water treatment products such as Klorin and PUR. In a related study, Jain et al. (2010) reported that respondents were using NaDCC tablets because they believed the tablets would improve human health (63%), prevent diseases (45%), make drinking water safer (43%), be easy to use (21%) and improve the taste of their drinking water (16%). It was further noted that 2% complained of bad smell and 1% bad taste from drinking water treated by NaDCC tablets. As shown in table 4.8, research findings established from respondents revealed that majority households were not practicing household water treatment because; treated water had bad taste and smell (6.3%), they were used to drinking untreated water (5.9%) and others believed that their drinking water sources were safe (7.8%).
The study findings concurs with Kakulu (2012), who found out that respondents were not treating their drinking water because they believed that it is safer from the source, did not have knowledge on existing household water treatment methods and cost of household water treatment was high. These results imply that people are not well sensitized about the importance of household water treatment and the dangers related to consumption of untreated water. In a related research study by Sodha et al. (2011) reported that among 28 respondents who did not treat their drinking water, 36% said consumption of untreated drinking water was their habit while 21% believed their source water was already safe.

Household water storage practice
As shown in figure 4.11, 72.2% respondents reported they were storing drinking water in any water storage container, whereas 27.8% reported they were not storing drinking water. Of those who reported they were storing drinking water, only 27.8% of respondents had treated drinking water whereas 54.8% didn't have it during field study visits. Household water storage options that were reported include; 5-litre jerricans (33.7%), plastic buckets (13.7%), 20-litre jerricans (13.2%), jugs (7.8%) and 2.9% were using plastic bottles (Table 4.9). Other household water storage options such as clay pots (0.5%), kettles (1%), beakers (0.5%) and flasks (0.5%) were reported as shown in Table 4.9. 5-liter jerricans were preferred because they are generally cheap light and durable. Plastic buckets were preferred because are they easy to clean; 20-liter jerricans were preferred because they were storing much drinking water for quite a number of days and plastic bottles were preferred because are cheap and in some cases free of charge. The current study supports Harris et al. (2009), who reported that household water storage containers such as clay pots (62%), jerricans (21%), barrels (7%), buckets (5%), jugs (3%), and bottles (1%) were used by mothers before KiBS (Kisumu Breastfeeding Study) clay pots were introduced.    (Figure 4.11). Additionally, 67.4% household water storage containers covered with a lid, 2.2 % were covered with soft covers and 30.4% were not covered with anything ( Figure 4.12). Storing drinking water in wide-mouthed containers is not recommended for storing drinking water because a wide-mouth opening greatly increases the risk of contaminating stored drinking water (Banda et al., 2007;Eshcol et al., 2009). The fact that people cannot dip their hands into drinking water stored in narrow mouthed water storage containers lowers the risk of feacal contamination. On the other hand, WHO (2013) reported that wide-necked containers such as buckets fitted with tight fitting lids are the best household water storage containers since they are easy to clean. Regarding the style of drawing drinking water from the storage container, 60.5% reported that they were pouring it into drinking cup, 11.7% reported that they were dipping the cup into water in the storage container and 1.5% reported that they were dipping with a ladle into stored drinking water and 1% reported that they were drinking directly from the storage container (

Boiling method
Total coliforms and Escherichia coli in household water samples treated by boiling method ranged from 0 to 336 CFU/100 ml and from 0 to 79 CFU/100 ml, respectively. Log 10 total coliforms reduction ranged from -0.5 log 10 to 2.4 log 10 , which corresponded with -115.6% to 99.6% removal efficiency. Log 10 Escherichia coli ranged from -0.4 log 10 to 2.0 log 10 which corresponded with -200% to 100% removal efficiency. Gupta et al. (2007) found out that stored boiled drinking water was microbiologically unsafe for human consumption in the study of Indonesian tsunami survivors in Indonesia. High total coliforms and Escherichia coli concentrations in household treated water samples could have been due to prolonged storage of household drinking water and regrowth of total coliforms and Escherichia coli. Inappropriate water storage, dipping contaminated hands and cups into stored drinking water also increase chances of high total coliforms and Escherichia coli concentrations in stored drinking water (Sobsey, 2002).

Biosand filtration method
Total coliforms and Escherichia coli concentrations in household water samples treated by biosand filtration method ranged from 9 to 106 CFU/100 ml and from 0 to 20 CFU/100 ml respectively. Log 10 total coliforms reduction ranged from -0.1 log 10 to 0.4 log 10 , which corresponded with -11.6% to 59.1% removal efficiency. Log 10 Escherichia coli reduction ranged from 0 log 10 to 0.6 log 10 , which corresponded with 0 to 75% removal efficiency. The present study does not support Stauber et al (2006), who found out that effectiveness of biosand filters on field tested water samples ranged from 0 log 10 to 2.5 log 10 (99.7%). The present study also doesn't support Earwaker (2006) Poor performance of biosand filters could be due to their irregular use and poor maintenance. Most biosand filters looked not to have been in use for some time.
Perhaps users re-started to use them when they heard about our research activities in their villages. For a biosand filter to effectively remove bacteria from drinking water, it should have been regularly used for more than 22 days. This explanation is supported by Sobsey et al. (2008) who reported that the efficacy of biosand filters in removing microbes in drinking water varied with filter maturity, dosing conditions, flow rate, pause time between doses, grain size, and filter bed contact time.
Inadequacies in biosand filter construction, operation and maintenance perhaps could further explain their poor performance in total coliforms and Escherichia coli removal. Earwaker (2006) reported that poor performance of biosand filters and low usage rates was attributed to the quality of maintenance, lack of reinforcement of educational messages and low support provided to filter users. Frisell et al. (2011) found out that poor performance of biosand filters were due to the fact that they had only been running for only two days at the time of testing, which is not adequate time for a schmutzdecke to form.

Let it stand and settle method
Total coliforms and Escherichia coli concentrations in household water samples treated by let it stand and settle method ranged from 12 to 360 CFU/100 ml and from 0 to 13 CFU/100 ml respectively. Log 10 total coliforms reduction ranged from -0.5 log 10 to 0.6 log 10 , which corresponded with -100 % to 73.5%. Log 10 Escherichia coli ranged from -0.6 log 10 to 0.0 log 10 , which corresponded with -300 to 0.0% removal efficiency. The present study supports UNCEF and WHO (2011) that found that treatment methods such as straining water through a cloth or let it stand and settle were inappropriate methods for household water treatment. Clasen and Boisson (2006) found out that total coliforms reduction treatment by let it stand and settle method resulted in total coliforms reduction mean 0.6 log 10 (95% CI= -0.4 -1.5).

Aqua safe tablets
Total coliforms concentrations in household water samples treated by application of aqua safe tablets ranged from 0 to 13 CFU/100 ml. Log 10 total coliforms reduction ranged from 0.0 log 10 to 0.2 log 10 which corresponded with 0 to 32% removal efficiency. Log 10 Escherichia coli were 1.2 log 10 , which corresponded with 100% removal efficiency. Lule et al. (2005) found out that majority of chlorine household treated water samples in the comparison study site had high Escherichia coli contamination than in the intervention study site. Ercumen et al. (2015) found out that majority of households had higher Escherichia coli counts in their stored water than their source water, indicating 77% in the controlled study site, 56% in the study site supplied with safe storage containers, 38% in the study site supplied with suboptimal chlorine (<0.2 mg/l) and 14% in the study site supplied optimal chlorine (≥0.2 mg/L). Poor total coliform and Escherichia coli removal efficiency by application of aqua safe tablets could be due to users' ignorance about the right procedure when disinfecting drinking water.
Overall, 23.9% and 19.5% of paired household water samples yielded negative log 10 total coliform and Escherichia coli reductions respectively. Negative log reduction occurs when bacteria concentration in a treated water sample is higher than it was before treatment (Brown & Sobsey, 2012). In the present study, negative log reduction perhaps resulted from poor handling of drinking water during storage, use of unsafe water storage options, inadequate cleaning of household water storage vessels and prolonged storage of drinking water. Mwabi et al. (2012) reported that negative log reductions may result from regrowth of injured bacteria which at a later time their metabolism get reconstructed and recover their growth. Mellor et al. (2013) reported that improper handling of drinking water during storage, dipping dirty hands and cups into stored drinking water and ineffective cleaning of water storage containers increase chances of storage water contamination.
Research study by John et al. (2014) found out that earthen pots (water storage options) that had dirty rims and consumers rarely washed either their hands before taking water from the storage vessel were associated with high total coliforms concentration. Additionally, some coliforms such as Escherichia coli can enter a dormant state but viable thus when raw water is tested, they may not be detected or detected in low numbers. After their dormancy period, they are detected in large numbers than before. This interpretation is in line with that of Wu et al. (2002) which state that Escherichia coli and V. cholerae can enter a dormant state, in which they are viable but not culturable in media used for their detection.

Testing hypothesis II
The paired samples T-test was used to test the null hypothesis of no significant difference between bacteria counts in household drinking water samples before and Therefore we reject the null hypothesis which stipulates, there are no significant differences between colony forming units in household water samples before and after treatment and accept the alternative hypothesis. These results indicate that household water treatment really reduces bacterial contaminants in household drinking water.      Sobsey (2002) reported that several HWT have been developed to improve drinking water quality and reduce waterborne diseases. Nevertheless little has been reported with regards to their efficiency in removing total coliforms and Escherichia coli from drinking water sources. An ideal treatment technology should be able to remove or reduce all microbial contaminants to acceptable levels recommended by authorized bodies such as WHO and UNBS for the case of Uganda. In this study, the efficiency of HWT (boiling, biosand filtration, WaterGuard tablets, aqua safe tablets and let it stand and settle) in removing total coliforms and Escherichia coli from water was determined under laboratory conditions as indicated in the subsequent subsections; Table 4.19 illustrates the performance of boiling method in removing total coliforms and Escherichia coli. Total coliforms and Escherichia coli ranged from 0 to 3 CFU/100 ml and from 0 to 1 CFU/100 ml, respectively. Log 10 total coliforms reduction ranged from 1.7 log 10 to 2.3 log 10 , which corresponded with 98.4 to 100% removal efficiency. Log 10 Escherichia coli reduction ranged from 1.1 log 10 to 1.3 log 10, which corresponded with 95 to 100% removal efficiency. High removal efficiency by boiling method could be due to heating water to a relatively high temperature (100 o C), which has the potential to kill microorganisms in the heated water. Clasen et al. (2008) found out that heating drinking water to even 55 o C has been shown to kill or inactivate most pathogenic bacteria, viruses and protozoa that are commonly waterborne. Kazmi and Khan (2013) found out that heating water at 80 o C kills pathogenic bacteria in water. Kazmi and Khan (2013) further noted that none of pathogenic bacteria including Escherichia coli were observed when water boiled up to 80 o C.  Table 4.20 illustrates the performance of biosand filtration method in removing total coliforms and Escherichia coli. Total coliforms ranged from 0 to 40 CFU/100 ml after treatment. Log 10 total coliforms reduction ranged from 0.6 log 10 to 1.8 log 10 , which corresponded with 79.1 to 100% removal efficiency. There was no

Biosand filtration method
Escherichia coli detected per 100 ml after filtration by biosand. Log 10 Escherichia coli reduction ranged from 1.1 log 10 to 1.4 log 10, which corresponded to 100% removal efficiency. The present study partially supports Vanderzwaag et al. (2009), who reported that average log reductions were 1.7 (98%) for total coliforms and 1.4 (96%) for Escherichia coli. Stauber et al. (2011) reported that when analysis of biosand filters was restricted to samples that had higher Escherichia coli concentrations, removals greater than 99% were measured. Mahmood et al. (2011) found out that the mean Escherichia coli and total coliforms after biosand filtration was nearly 96% reductions. In addition, the findings of the current study do not support Baumgartner (2006) who found out that the average bacteria removal by biosand filtration method under laboratory conditions was 96.5%. Stauber et al. (2006) found out that the geometric mean reductions of Escherichia coli by the biosand filter were 97% and 91% in laboratory experiments 1 and 2, respectively. In both experiments, the lowest Escherichia coli reductions were found during initial days of filter dosing. The minimum Escherichia coli reduction in experiment 1 was 1.2 log 10 (93%) measured on day 4 and in experiment 2, it was 0.4 log 10 (or 63%) measured on day 3. Maximum log 10 reduction of Escherichia coli in experiment 1 and 2 were 2.0 log 10 (99%) and 1.9 log 10 (98.9%). Generally, total coliform and Escherichia coli removal efficiency by biosand filtration method might be due to the biological layer formed on the top of the filter. Tellen et al. (2010) report that after 65 days, average percentage reductions in total coliform, feacal coliform and fecal streptococci were 98.9% for traditional biosand filters and 99% for the improved biosand filters. The study further reported that both modifications showed statistically significant improvements.  and Escherichia coli. The concentration of total coliforms ranged from 0 to 4 CFU/100 ml after treatment. Log 10 total coliforms reduction ranged from 1.7 log 10 to 2.3 log 10 , which corresponded with 97.9 to 100% removal efficiency. There was no Escherichia coli detected per 100 ml after treatment. Therefore, log 10 Escherichia coli were just a function of the measured Escherichia coli in untreated water samples only.
In a related study, WaterGuard tablets proved significantly more effective than other water treatment methods such as PUR and filters to produce drinking water under the detection limit (<1 CFU/100 ml) across all water sources. The study further reported that when homes were assigned WaterGuard tablets, 51% of stored drinking water samples had Escherichia coli <1 CFU/100 ml (95% CI: 46−56%). When the same households were provided PUR, it was noted that 33% had Escherichia coli concentrations <1 CFU/100 ml, and when provided filters, 39% had Escherichia coli <1 CFU/100 ml (Albert, Luoto, & Levine, 2010).  Table 4.22 illustrates the performance of aqua tablets in removing total coliforms and Escherichia coli. The concentration of total coliforms ranged from 1 to 2 CFU/100 ml after treatment. Log 10 total coliforms reduction ranged from 1.5 log 10 to 2.3 log 10 , which corresponded with 96.4 to 100% removal efficiency. There was no Escherichia coli detected per 100 ml after treatment. Therefore, log 10 Escherichia coli reductions were just a function of the measured Escherichia coli in untreated water samples only.  Table 4.23 illustrates the performance of aqua tablets in removing total coliforms.

Let it stand and settle method
The concentration of total coliforms ranged from 40 to 62 CFU/100 ml after treatment. Log 10 total coliforms reduction ranged from -0.0 log 10 to 0.7 log 10 , which corresponded with -12.5 to 78.9% removal efficiency. The concentration of Escherichia coli ranged from 4 to 32 CFU/100 ml after treatment. Log 10 Escherichia coli reductions ranged from -0.2 log 10 to 0.5 log 10, which corresponded with -60 to 70.4% removal efficiency.
Negative log 10 reductions observed were as a result of regrowth of bacteria in stored drinking water during the process of treatment. The study findings concur with the study by UNCEF and WHO (2011) which reported that HWT such as straining water through a cloth or let it stand and settle are considered inappropriate for household water treatment. In both household and experiment water testing, the calculation of log 10 reduction was limited by non-detection of total coliforms and Escherichia coli in test water samples whose concentration was <1 CFU/100 ml. This means that log 10 reduction value was just a function of the measured total coliform and Escherichia coli in test water sample before treatment.

Classification of test water samples after treatment
Classification of total coliforms and Escherichia coli in drinking water is critical since it fore tells the health risk level associated with drinking water that is microbiologically contaminated. WHO (2006) and UNBS (2008)  32.6% fell within medium risk (11 to 100 CFU/100 ml). In related study, it was found that 54.1% of biosand filtered water at household level achieved the WHO guideline value of 0 CFU per 10 ml (Earwaker, 2006). Duke et al. (2006) found out that 80% of filtered water samples had no Escherichia coli per 100 ml and 17% were within low risk category (1 to 10 CFU/100 ml). In a related study, Stevenson (2008) found out that all test water samples that were treated by WaterGuard and aqua tabs had bacteria counts < 10 CFU/100 ml (low risk category).

Figure 4.14: Bar graph showing the percentage of test water samples under different risk level categories in household testing
In contrast, majority of experiment treated water samples (40%) were free of total coliforms. A small proportion (25%) fell within the low risk category (1 to 10 CFU/100 ml), 32.6% fell within medium risk category (11 to 100 CFU/100 ml) and 35% fell within high risk category (101 to 1000 CFU/100 ml). In addition, a large (73%) of water samples were free of Escherichia coli and 20% fell within the low risk category (1 to 10 CFU/100 ml). A small proportion (5%) fell within medium risk (11 to 100 CFU/100 ml). Figure 4.17 clearly illustrates this distribution according to risk level category. Brown and Sobsey (2012) reported that 44% of boiled samples had no detectable Escherichia coli (< 1 CFU/100 ml) and 73% had Escherichia coli < 10 CFU/100 ml.

Risk category (Bacteria count)
Total coliforms Escherichia coli stand and settle method and 0.6 log 10 (50%) for application of aqua safe (Table 4.20).
Although treatment by application of aqua safe tablets and boiling showed a slight higher Escherichia coli and total reduction than other treatment technologies, the observed differences in effectiveness were significantly small. Results showed a marginally greater performance by two treatment methods in removing total coliforms and Escherichia coli thus did not strongly indicate that one of the tested methods is more effective than others.
In contrast, water treatment by household technologies during experimentation recorded a remarkable decrease in both total coliform and Escherichia coli concentrations though a few of them achieved complete removal of Escherichia coli as required by WHO (2006) and UNBS (2008) guidelines for drinking water quality.
On the basis of average percentage removal efficiency of tested HWT under laboratory conditions, the present study concluded that the order of effectiveness against total coliforms by HWT was observed to be; (i) application of WaterGuard tablets, (ii) boiling method, (iii) application of aqua safe, (iv) biosand filtration method, and (v) let it stand and settle method. The order of effectiveness against Escherichia coli by HWT was observed to be; (i) application of WaterGuard tablets = biosand filtration method and application of aqua safe, (ii) boiling method, and (iii) let it stand and settle method.

Testing hypothesis III
A one-way analysis of variance was conducted to test the null hypothesis of no significant differences between mean bacteria reductions by different HWT under laboratory conditions. The independent variable was HWT with five levels; boiling method, biosand filtration method, let it stand and settle method, application of WaterGuard tablets and aqua safe tablets. The dependent variables were mean log 10 total coliforms and Escherichia coli reductions per 100 ml. ANOVA tests yielded significant variations among mean log 10 total coliforms reduction, F (4, 15) = 14.1, P = 0.000 and mean log 10 Escherichia coli reduction, F (4, 15) = 5.7, P = 0.006 (Table   4.26). Therefore, we reject the null hypothesis which stipulates, there are no significant differences between mean reductions by different HWT under laboratory conditions and accept the alternative hypothesis. Post hoc Tukey tests were conducted across log 10 total coliforms and Escherichia coli reductions by different HWT to determine whether there were no significant differences between mean log 10 total coliforms and Escherichia coli reductions in experiment test water samples (Appendances IV and V). Results are discussed in the subsequent section.
Analysis showed significant differences in log 10 total coliforms reduction between boiling and the other four treatment technologies (P < 0.05), except when log 10 total coliforms reduction by boling was compared with log 10 total coliform reduction by application of WaterGuard tablets and application of aqua safe tablets (P > 0.05).
Significant differences in log 10 total coliforms reduction were observed between treatment by biosand filtration and other four HWT (P < 0.05), except when log 10 total coliforms reduction by biosand filtration was compared with log 10 total coliforms reduction by let it stand and settle and application of WaterGuard tablets (P > 0.05).
Additionally, results indicated significant differences in log 10 total coliforms reduction between treatment by let it stand and settle method and other four water treatment technologies (P < 0.05), except when log 10 total coliforms reduction by let it stand and settle method was compared with log 10 total coliforms reduction by biosand filtration method (P > 0.05). Significant differences were observed in log 10 total coliforms reduction between application of WaterGuard tablets and other four HWT (P <0.05), except when log 10 total coliforms reduction by application of WaterGuard tablets was compared with log 10 total coliforms reduction by boiling and application of aqua safe tablets (P > 0.05). Furthermore, results indicate significant differences in log 10 total coliforms reduction between treatment by application of aqua safe tablets and the other four HWT (P < 0.05), except when log 10 total coliforms reduction by application of aqua safe tablets was compared with log 10 total coliforms reduction by application of WaterGuard tablets and boiling method (P > 0.05).
Additionally, there was no significant differences in log 10 Escherichia coli reduction between boiling and the other four HWT (P > 0.05), except when log 10 Escherichia coli reduction by boiling method was compared with log 10 Escherichia coli reduction by let it stand and settle method (P > 0.05). The study results showed no significant differences in log 10 Escherichia coli reduction between biosand filtration and other four HWT (P > 0.05), except when log 10 Escherichia coli reduction by biosand filtration was compared with log 10 Escherichia coli reduction by let it stand and settle method (P < 0.05). Additionally results indicate significant differences in log 10 Escherichia coli reduction between let it stand and settle method and other four HWT (P < 0.05). No significant differences were observed in log 10 Escherichia coli reduction between application of WaterGuard tablets and other four HWT (P > 0.05), except when log 10 Escherichia coli reduction by application of WaterGuard tablets was compared with log 10 Escherichia coli reduction by let its stand and settle method (P < 0.05). Additionally, results indicated no significant differences in log 10 Escherichia coli reduction by application of aqua safe tablets and other four HWT (P > 0.05), except when log 10 Escherichia coli reduction by application of aqua safe tablets was compared with log 10 Escherichia coli reduction by let it stand and settle method (P < 0.05).

Introduction
This chapter presents the summary of the main findings, conclusion, recommendations of the study and areas for further research. The general objective of the study was to examine HWT and evaluate their ability to improve the microbial quality of drinking water at household level in Kabale District, southwestern Uganda. The specific objectives of the study were to: (i) evaluate the different water sources, household water treatment technologies, and storage options of household drinking water, (ii) to establish whether the source of drinking water influences the type of water treatment technology used at household level, (iii) to determine whether there is significant difference between bacterial counts in household drinking water samples before and after treatment and (iv) to evaluate bacteriological effectiveness of HWT used under laboratory conditions. Tools used to conduct the study were semi structured questionnaires, field observations and water analysis.
Data analysis tools were mainly descriptive statistics. Parametric tests such as paired T-test, one way ANOVA and chi-square were used.

Summary of findings
A total of 205 respondents were interviewed during unannounced field study visits.
Majority of the households (61.5%) reported they were using water springs as their primary source of drinking water. Other water sources were public and private water taps, shallow wells and boreholes. At least 75.6% of the households were practicing household water treatment. HWT reported were; boiling, application of aqua safe tablets, let it stand and settle method, biosand filtration method and application of WaterGuard tablets. The study further reported that 72.2% households were storing drinking water using any storage container. 5-liter jerricans, plastic buckets, 20-liter jerricans, jugs, kettles and flasks were repeatedly reported household water storage options during the field study visits. Of the 46 households that had stored treated drinking water, 71.7% were using narrow-mouthed water storage containers whereas 28.3% were using wide mouthed water storage containers respectively. The study revealed no significant relationship between drinking water sources and the type of household water treatment technologies (P > 0.05).
In household water testing, boiling method recorded a slightly higher percentage removal of total coliforms with a mean value of 28.2% (0.6 log 10) , followed by biosand filtration method with 23.8% (0.2 log 10 ), application of aqua safe tablets with 16% (0.1 log 10 ) and let it stand and settle method with -18.1% (-0.01 log 10 ). The study further revealed that treatment by application of aqua safe tablets recorded the highest percentage removal of Escherichia coli with a mean value of 50% (0.6 log 10 ), followed by biosand filtration method with 37.5% (0.3 log 10 ), boiling method with -23.3% (0.4 log 10 ) and let it stand and settle method with -200% (-0.4 log 10 ). It was further noted that majority of treated household water samples were associated with high bacteria counts compared to untreated household water samples.
In experiment water testing, application of WaterGuard tablets recorded the highest percentage removal of total coliforms with a mean value of 99.5% (1.9 log 10 ), followed by boiling method with 98.9% (2 log 10 ), application of aqua safe tablets with 98.1% (2 log 10 ), biosand filtration method with 84.8% (0.9 log 10 ) and let it stand and settle method with 38.3% (0.5 log 10 ). Treatment by application of WaterGuard tablets, biosand filtration method, application of aqua safe tablets recorded high percentage removal of Escherichia coli with a mean value of 100% (1.2 log 10 ), followed by boiling method with 98.8% (1.2 log 10 ) and let it stand and settle method with 32.8% (0.5 log 10 ). HWT significantly reduced total coliforms and Escherichia coli with exception of let it stand and settle method.

Conclusions
This study confirmed that majority households were using water springs as their sources of drinking water at household level. The observed homesteads and toilets located in close proximity to water springs were responsible for high total coliforms and Escherichia coli concentrations in water samples from water springs. Run offs from crop farms located in the shallow wells' upstream was responsible for the increased turbidity and sediment levels in most shallow wells.
This study also confirmed that majority respondents were practicing household water treatment with any one of the following technologies: boiling, biosand filtration, let it stand and settle, application of WaterGuard tablets and application of aqua safe tablets. Treatment by boiling method was reported the most popular water treatment method, unlike treatment by biosand filtration, use of WaterGuard tablets, aqua safe tablets and let it stand and settle methods. Wide use of boiling method was due to availability of fuel wood.
From this study, it was found out that majority respondents who were boiling their drinking water were doing so because the method was cheap and easy to use. This was because of efforts by the government of Uganda to promote household health through access to safe drinking water. Though the uptake of biosand filtration method was significantly low, those who were using it believed that the method was easy to use, cost effective once installed and produces water ready for consumption.
Majority respondents who were using let it stand and settle were doing so because the method is cheap. Use of aqua safe and WaterGuard tablets is significantly low more especially in rural areas. The efforts to promote wide use of WaterGuard and aqua safe tablets were hardly recognised on the ground.
Majority of household treated water samples were associated high bacteria contamination (>10CFU/100 ml). High total coliforms and Escherichia coli concentrations in household treated water samples resulted from improper handling of stored drinking water, regrowth of bacteria, and accumulation of biofilm layer inside the storage containers.
A significant decrease in total coliforms and Escherichia coli concentrations occurred after treatment of contaminated source water samples by different HWT with exception of let it stand and settle method. Although none of the HWT achieved complete removal (100%) of total coliforms, application of WaterGuard tablets was found to be the most efficient method in removing total coliforms from contaminated drinking water. Although some HWT did not significantly remove Escherichia coli from source contaminated drinking water, application of WaterGuard tablets, biosand filtration method and application of aqua safe achieved complete removal of Escherichia coli.

Recommendations
Based on the findings of this study, the following are the recommendations to the policy makers, NGOs and households in Kabale District to ensure access to safe drinking water.
i. Spring water should be treated before drinking due to high total coliforms concentration which makes it unsafe for consumption. The responsibility to improve household drinking water must generally fall on individual households.
ii. The government and NGOs should take an active role to promote effective household water treatment methods such as application of WaterGuard, application of aqua safe and biosand filtration method to replaces less effective local methods such as let it stand and settle method, which is currently practiced in some communities.
iii. Nowadays, there are available cheap water testing methods to evaluate the microbiological effectiveness of water treatment technologies. The local people should always be involved in testing for microbiological effectiveness of recently introduced water treatment technologies in Kabale District.

Areas for further research
Further research is needed in the following areas in Kabale District southwestern Uganda.
i. Assessment of the impacts of household water treatment and storage on community health.
ii. Effect of household water storage options on microbiological quality of drinking water.
iii. Role of gender in household water treatment and promotion in Kabale District southwestern Uganda.

APPENDICES APPENDIX I: Semi-structured questionnaire
Greetings! My name is Alex Saturday; I am a student of Kenyatta University pursuing Msc.
Integrated Watershed Management. Currently I am conducting a study on: Examination of household water treatment technologies for microbial removal in Kabale District, south western Uganda. I plan to administer this questionnaire to 205 household heads or adults above the 18 years of age in this district. You have haven selected by chance. I kindly request you to take part in this study.
The information you will give shall be confidential, neither shall I publish it in any newspapers nor be read on any radio or television. Only the study staff and investigators will know your answers to the questions.