Implementation of defluoridation filters in Ethiopia

From GapWiki
Jump to: navigation, search

NOTE: Article from the Geogenic Contamination Handbook


It is estimated that more than 8 million people live in fluoride-affected areas in Ethiopia (Rango et al., 2012 and references therein). The main sources of fluoride are basaltic rocks, which have both elevated fluoride content and low soluble calcium concentrations. In the Ethiopian Rift Valley, over 40% of deep and shallow wells are contaminated, and fluoride levels are often significantly higher than the present international WHO guideline value of 1.5 mg/L (Tekle-Haimanot et al., 2006). As a result, dental and skeletal fluorosis is widespread among the population of the Rift Valley. The mitigation of this health problem has been hampered mainly by the lack of suitable, inexpensive removal methods and technical support. A switch to treated surface waters for drinking is being discussed, but it is accepted that fluoride removal systems for rural communities are required, at least until longer-term solutions can be put in place. Therefore, in 2009, in collaboration with Addis Ababa University, Eawag launched the research project, “Optimization and acceptance of fluoride removal options in rural Ethiopia”, funded by the Swiss National Science Foundation and the Swiss Agency for Development and Cooperation.


The aim of the project was to combine technical and social research with field implementation to find a suitable and acceptable solution for the problem of fluoride contamination in drinking water in rural Ethiopia:

  • To compare and optimise the removal efficiency of two different filter materials in the laboratory and subsequently to test the performance of these technologies in the field
  • To assess the personal, social and situational factors that influence the continuous use of fluoride removal systems by the rural population
  • To investigate the institutional settings and identify stakeholders’ interests and preferences for the implementation of fluoride removal
  • To investigate fluoride uptake pathways via food and water
  • To strengthen the institutional capacity for research and implementation in Ethiopia

Intensive interaction between physical and social sciences was indispensable in this project, because even the best technical solution is useless when it is not accepted by the population. Another important goal of this collaborative project was capacity building and human resource development in Ethiopia. It included a south-south knowledge transfer between Kenya and Ethiopia that was aimed at strengthening the research capacity of the Addis Ababa University. The participation of NGOs consolidated the ties between research and implementation. Furthermore, the results should be applicable not only to Ethiopia but also to other fluorosis-affected developing countries.


  • Addis Ababa University (AAU): Main research partner of Eawag. The Chemistry Department developed an aluminium-based filter medium (AO). The institutional analysis of the Ethiopian water sector was conducted through the Department of Political Science and International Relations.
  • Nakuru Defluoridation Company Limited (NDC): Producer of high-quality bone char and calcium phosphate pellets in Nakuru, Kenya. Eawag and NDC have been working jointly on optimising the Nakuru Technique since 2006. NDC provided bone char and pellets to the research project.
  • Oromia Self-Help Organization (OSHO): Local NGO and field implementation partner of Eawag. Since 2007, OSHO has been introducing bone char household filters funded by Swiss Interchurch Aid (HEKS), with technical support from Eawag and NDC.
  • National Fluorosis Mitigation Project Office (NFMPO): The office is currently located at the Ministry of Water and Energy and took up work in 2009. Information was exchanged regularly with other project partners.
  • Key stakeholders that were involved in the project included water offices at national, regional, zone and district levels, development partners interested in fluoride mitigation, Ethiopian research institutions, water committees and beneficiaries in the project villages. A number of workshops with stakeholder participation were held during the course of the project to strengthen stakeholder involvement in decision-making and to disseminate results.

Integrative approach

Fig. 9.1 Technical details of the community filter in Wayo Gabriel (Terms of use: Cite original source from Handbook)
Fig. 9.2 Project planning overview (Terms of use: Cite original source from Handbook)

Two community fluoride removal filters were constructed in the Ethiopian Rift Valley for detailed field testing, one using the Nakuru Technique and one using aluminium (hydr)oxide (“AO”, a filter media developed by Addis Ababa University). The filter sites were selected during a workshop in Addis Ababa in November 2009, in consultation with representatives from regional, zone and district (woreda) water offices. For the sake of convenience, only the results from the community filter using the Nakuru Techniques in Wayo Gabriel village are discussed in this chapter (see Fig. 9.1).

A separate meeting was held with the local water committee, village administrators and a district representative to set the tariff for treated water in time for the opening of the filter. The water committee is an elected group of people from the village responsible for managing a water scheme. Individuals usually have to pay for drinking water in the Ethiopian Rift Valley, so paying for water was not a new concept. The intention was to set a water tariff that covered the operator’s salary plus more than 50% of the costs during filter media replacement during the three-year project, and to explore the potential of 100% medium-cost coverage in collaboration with OSHO for the long term.

The filter design was adapted from the one-tank systems used by NDC in Kenya to a more sophisticated version that guarantees optimal utilisation of the filter medium. The system consists of two filtration tanks in series (2 m3 each), of which either can be used as the main tank (first) or the polishing tank (second). When fluoride breakthrough occurs, the filter medium in the main tank is replaced and the flow reversed (the main tank with the fresh filter medium then becomes the polishing tank, and vice versa). A storage tank for the treated water (5 m3) allows a slow and continuous water flow, while providing sufficient reserves for times of greater demand. (From previous laboratory experiments, it is known that the fluoride uptake capacity of the Nakuru Technique increases with reduced flow rates). The system is sufficiently simple to handle, so that the operator, usually someone from the village, does not need special skills except for some basic training. The operator is also in charge of collecting the water fee from the users.The different components of the integrated study are shown in Figure 9.2.

Fluoride uptake through food and water

To determine the amounts of fluoride ingested through food and water, interviews were conducted with 20 families on their daily diet over the previous seven days and on recipes for the most common dishes. Based on this information, the most commonly consumed food ingredients were collected in nine households around Wayo Gabriel and analysed for their fluoride content. The selected households collected drinking and cooking water from three different water sources with fluoride concentrations of 0.75 mg/L (average of water treated at the community filter 0–1.5 mg/L), 3 mg/L and 10 mg/L. Using the information from the interviews and the results from the analysis to estimate mean daily consumption, the mean daily fluoride uptake through food and water was calculated.

Filter performance

Data on the quality and consumption of the treated water were collected in order to analyse the fluoride removal performance of the filter and to guarantee safe drinking water for the consumers. Weekly measurements of fluoride concentrations were conducted and water meter readings taken to find out how much water had been consumed. On a monthly basis (and more frequently during the first few months), water samples were taken from all four sample taps (raw and treated water after Tank A and after Tank B), and a complete chemical analysis was carried out. Fluoride measurements were generally conducted every week.

Behavioural change

In both villages with community filters, a baseline survey to determine the psychological factors that influence the desired behaviour (using fluoride-free water for drinking and cooking) was conducted using structured questionnaires in 100 randomly selected households in each village. Three different behavioural change campaigns (interventions) were then undertaken to promote the use of the filtered water. Surveys were conducted after each intervention and at the end of the 18-month promotion period. A team of ten local college students were recruited and trained to conduct the interviews. The duration of one interview was, on average, one hour per household; the questionnaires were translated into Amharic and Oromifa.

Institutional analysis

This task was performed in four steps.

  • Step 1: Stakeholders involved in fluoride mitigation in Ethiopia were identified through literature review and contacts with experts in the field.
  • Step 2: Seventy end-users in the field (35 of them in Wayo Gabriel) were interviewed personally by the PhD student about affordability and access to safe drinking water, using semi-quantitative questionnaires.
  • Step 3: Representatives from water offices at different levels, development partners and members of the National Fluorosis Mitigation Technical Advisory Committee were selected and interviewed, using a qualitative questionnaire, about sustainability, preferences, opportunities and the threats of different fluoride mitigation options.
  • Step 4: A Multi-criteria Decision Analysis was carried out during the final project workshop to compare stakeholders’ preferences which referred to different fluoride removal technologies.

Cost and Affordability

One of the most important aspects that needs to be addressed in order to achieve sustainable and successful fluoride mitigation in Ethiopia is the issue of cost. When donor funding runs out, can the costs of fluoride removal still be covered? Within this case study, an analysis of the expenditures that need to be taken into account when installing and managing a fluoride removal filter was carried out. More details about the individual cost components can be found in IRC (2011).

  • Capital Expenditure (CapEx): These are the funds that need to be invested in fixed assets, such as filter tanks and pipes, in initial awareness raising campaigns and in training operators, in the water committee and the district water office. CapEx can pose a significant investment at the start of a project.
  • Capital Maintenance Expenditure (CapMEx): Occasional cost of renewing (replacing, rehabilitating, refurbishing) essential parts of the system (e.g. filter material) in order to ensure that services continue at the same level of performance that was first delivered.
  • Operation and Minor Maintenance Expenditure (OpEx): Cost of daily operation and light maintenance (e.g. power, salary of operator). OpEx does not include costs of major repairs.
  • Expenditure on Direct Support: Pre- and post-construction support activities directed at local stakeholders and users. This could include monitoring, technical advice, administrative or organisational support, conflict resolution, capital maintenance, training and refresher courses, the provision of information and resource mobilisation.


Fluoride uptake through food and water

The total average fluoride intake of an adult consuming treated water for drinking and cooking from the community filter was estimated to be around 6 mg/day. (For comparison, the daily fluoride intake of an adult consuming water with 10 mg/L is 25 mg/day.) This is close to the tolerable upper intake level.

Filter performance

Fig. 9.3 Results of fluoride monitoring from 20.05.2010 to 28.02.2013 in the Nakuru Technique filter. CP: contact precipitation (“Nakuru Technique”), BC: bone char. At around 800 m3, the concentrations of fluoride and salts in the raw water rises (Terms of use: Cite original source from Handbook)

The Nakuru Technique community filter was not saturated at the end of the research project, mainly due to an initially low water consumption. Based on the experience of NDC in Kenya, it was expected that the filter could treat at least another 750 to 1000 m3 until fluoride breakthrough, if not more, because of the improved design compared to the NDC filters. Nevertheless, the field test in Ethiopia revealed two major challenges remaining for this optimised and more sophisticated filter design:

Slow and continuous flow: The stainless steel tanks could not be pressurised as planned because of leaks in the lid seal. Instead, the operator needed to turn the main water line on and off manually. As a result, water passed rapidly through the filter tanks for only a few hours instead of the intended slow and continuous flow over 24 hours. This is the reason for the fluoride level fluctuation after tank A (see Fig. 9.3). Nevertheless, the bone char layer was still able to remove all remaining fluoride.

Two interchangeable filter tanks: The operation proved to be more complicated than expected. After 50 m3 of water had been treated, a wrong valve was opened, and raw water bypassed the system. After this incident, another valve was not completely closed, and some water passed only through tank A but not through tank B. This was noticed and rectified after 450 m3 of water had been treated. These problems were finally resolved by installing float valves in the two filter tanks to control the water flow automatically. A detailed operation manual for the filter is being developed in participation with OSHO, district water officers and community filter operators.

Behaviour change

Fig. 9.4 Behaviour change interventions (Terms of use: Cite original source from Handbook)

The baseline survey in Wayo Gabriel revealed that the consumption of fluoride-free water was hindered mainly by (i) high perceived costs but also by (ii) perceived taste, (iii) perceived ability and (iv) commitment (Section 8.3, Huber and Mosler, 2012; Huber et al., 2012). Furthermore, the behaviour of others had a strong influence on individual households (people who think that many others are also collecting water from the community filter are more likely to collect water from the same source themselves). Based on this understanding of psychological factors, the following interventions were conducted to increase the consumption of fluoride-free water and to keep consumption sustainably high (Huber et al., 2014; Fig. 9.4):

  • Phase 1: Persuasion campaign. Households were visited by a health promoter who was trained in persuasion techniques to tackle perceived costs (determined by the baseline survey to be an important factor) and perceived vulnerability (conventional wisdom holds that raising awareness about the severity of health effects may stimulate behaviour change).
  • Phase 2: Photo promotion. People that fetched water at the community filter had their picture taken, and they received these with a reminder slogan added below the picture (Fig. 9.4). The promotion aimed to motivate new users to try filtered water and to help people (with the picture as a reminder) to remember fetching water at the community filter.
  • Phase 3: Flag promotion. Households were again visited by promoters and asked to commit themselves to consume only fluoride-free water in the future. A blue flag was installed on the household’s roof to make their commitment public. The aim of this was to increase people’s commitment and at the same time, to inform other villagers that the people in that particular household were consuming treated water.

An increase in fluoride-free water consumption by people who had received the photo reminder and had put it up on the wall where it was visible was observed in Phase 2. The flag promotion in Phase 3 resulted in an increased average fluoride-free consumption of all households in the area. People who had committed themselves and had a flag on their roof increased consumption significantly more than others. However, an increase in fluoride-free water consumption was also observed by those who had not received the commitment intervention. They probably saw the flags all over the village and therefore realised that many of their neighbours were using fluoride filtered water.

After a 6-month break during which no surveys or interventions were carried out, the long-term effectiveness of the behaviour-change activities was evaluated. Most photos and flags were still in place, and all of the households that had switched to the consumption of fluoride-free water were still buying this water. The overall consumption was still high, even though people without any intervention slightly decreased their consumption. In general, it can be concluded that the promotion strategies were very successful in increasing and maintaining the consumption of fluoride-free water within the community, with the exception of the “conventional wisdom” intervention targeting awareness of the risk of contracting fluorosis.

Institutional analysis

Fig. 9.5 Map of stakeholders involved in fluoride mitigation in the Ethiopian Rift Valley (Terms of use: Cite original source from Handbook)

Figure 9.5 shows stakeholders that were identified as being active in fluoride mitigation in Ethiopia. Many are supportive, and a few stakeholders are neutral or unsupportive. The National Fluorosis Mitigation Project Office (NFMPO) was found to have established a reasonable basis for coordination and communication between different stakeholders. An important point to be addressed in the near future is the location of the NFMPO; i.e. whether it should be embedded in the existing institutional structure of the Ministry of Water and Energy (MoWE) or whether it should be set up independently in a research institute or a university. There are several organisations that could become more involved. Of these, the Ministries of Health and Education would be important partners.

Cost and affordability analysis

Under the current situation, when all types of costs are considered, rural communities in Ethiopia are not (yet) able to afford fluoride removal activities without significant subsidies from other sources such as governments or NGOs. This is especially the case when fluoride concentrations in the raw water are high, and the filter material needs frequent replacement or regeneration. There is a remaining need for fluoride mitigation options to be developed or adapted in order to achieve higher cost-effectiveness. Organisations that are implementing fluoride removal units need to assess carefully the willingness and ability of stakeholders (beneficiaries, government, NGOs) to cover certain types of costs sustainably. Cost indicators should be included in the monitoring procedure.


The results of the intake analysis show that a high percentage of fluoride is taken up via water used for drinking and cooking. If fluoride-contaminated water is treated with a removal technique, a significant reduction in the risk of developing skeletal fluorosis can be expected. The Nakuru Technique fluoride removal community filter in Wayo Gabriel can reduce fluoride concentrations to below the WHO guideline of 1.5 mg/L, although the fluoride uptake capacity should be increased further to make the system more cost-effective (and reach 100% cost coverage by the local community). The adapted filter design could contribute to achieving this goal, but only if the operation is carried out properly. It was shown that the Nakuru Technique is well accepted by consumers. This contradicts previous studies that stated that bone char is generally culturally not acceptable in Ethiopia. It was also shown that simply providing a filter is not sufficient; in Wayo Gabriel, it was only after well-designed promotional campaigns that the majority of consumers used fluoride-free water for drinking and cooking purposes. However, fluoride exposure through food remained at levels high enough to cause dental, and possibly also skeletal, fluorosis. While reducing fluoride exposure through water is necessary to mitigate fluorosis, it is not sufficient. The results of the Ethiopian case study were communicated to the major stakeholders during a two-day workshop in April 2012 in Addis Ababa.


  • More focus on the “software” components. Capacity building for local authorities and NGOs in effectively promoting behaviour change in communities, combined with close monitoring of the consumption of fluoride-free water.
  • Close monitoring and documentation of newly installed fluoride removal options during the first few years to further optimise filter design and to obtain information on real filter performance and costs.
  • Reduction in the overall costs of defluoridating drinking and cooking water. This could include optimising the production of the filter media, regeneration or reuse in agriculture and testing of newly developed, low-cost filter media in the field.
  • Increased involvement of health authorities in fluoride mitigation by supporting a combination of fluoride removal with microbiological drinking-water treatment, sanitation and hygiene promotion. Health impact studies could complement further fluorosis mitigation activities.
  • Food intake represents a significant source of fluoride exposure. Strategies need to be developed to reduce fluoride exposure through foodstuffs through changes in either agricultural or cooking practices.


For references, please visit the page References - Geogenic Contamination Handbook.