Purifying polluted water through hemodialysis ﬁ lters for poor villages without electricity: the easy water for everyone approach and experience

Background: Given the need for treating polluted drinking water, our NGO Easy Water for Everyone has produced pure water in remote villages without power and achieved health bene ﬁ ts. With the goal of reaching more needy populations we report our experience and successful implementation in Ghana. Methods: In 20 villages polluted water is pumped every few days to an elevated water tank connected to a ﬁ ltration device leading to a faucet. Repurposed hemodialyzers with polysulfone membranes, having a ﬁ lter pore size of 0.003 micrometers, prevent passage of pathogens. Results: Gravity from a 3 m height pushes water through the membrane whenever the faucet is open. Back ﬂ ushing of the hemodialyzer membrane 3 times daily removes built-up of organic material and maintains ﬂ ow rates of 250 L/hour for at least 2 years. Filtered water has been culture negative. Management of problems and optimization are reported. The 5-year cost per village of < 1,500 population averaged < 2 US$ per day.


INTRODUCTION
Surface water is often polluted with parasites, bacteria and/ or viruses that can cause serious health issues (Piper et al. ). Diarrhea is a common result of drinking water polluted with pathogens and is associated with markedly elevated mortality risk, particularly in children (Levine et al. ). Pollution of drinking water is magnified near slow-flowing rivers and stagnant waters in villages where sanitation facilities are unavailable and when farm animals are nearby. Well water is also susceptible to such pathogens particularly when boreholes or wells are shallow or intermittently overcome by rising water tables. Purification of water from such contamination is feasible and highly effective when electricity or solar power is available (Peter-Varbanets et al. ). However, electricity is often not available in remote villages of developing countries, where a substantial proportion of their populations lives. Remote villages are also difficult to reach for distribution of sachet water derived from reverse osmosis treatment (Semey et al. ). In absence of power, alternative approaches currently used worldwide include biosands providing some decontamination, and the addition of antiseptic agents such as chlorine to kill remaining pathogens (Clasen et al. ). The monitoring of levels of toxic antiseptic agents may identify low levels thus indicating unsafe drinking water, while high levels may make the water unpalatable and lead individuals to choose unsafe water.
Our alternative approach particularly for areas without power is based on knowledge and experience from clinical hemodialysis for patients with kidney failure. Hemodialyzer membranes have been shown to be effective in producing sterile water that can be safely infused intravenously to patients during hemofiltration after addition of appropriate

METHODS
A particularly urgent need for uninfected drinking water was identified in Ghanaian rural villages in the estuarial region of the massive Volta River. For these villages the river has been the main source for drinking water even though it has been known to carry pathogens of fecal origin. Power has not been available in these villages, most of which are located on islands in the river. Based on the needs identified by a local physician/regional hospital director (Dr P. Narh) and driven by the interest within a given village, facilitated by a local radio station, our Non-

RESULTS
The water source is typically a lake or river that has been known to be persistently polluted with pathogens particularly coliform bacteria and fecal coliform bacteria, associated with lack of sanitation. When needed we have also purified water from wells or boreholes that have been chronically polluted. A large hose with its inlet protected by a coarse sieve is submerged in the water source, e.g.
near the edge of the river. From there it is pumped to a large elevated tank using a gasoline pump. As shown in

DETAILS AND REFINEMENTS Water storage
The water enters the large (1,000 or 4,500 liters) reservoir.
The size of the tank is chosen according to the size of the population served. This tank is elevated by 10-14 feet (3-4.3 m) to provide the gravitational force for filtration.
Since the filtration device is about 2 feet above the ground, the filtration pressure is based on a minimal height difference of 8-12 feet, i.e. when the tank reservoir is nearly empty.
Since growth of algae in the polluted water is known to be facilitated by exposure to sunlight, the tank and all lines have a black color. Water flows from the storage tank through the filter device on demand, i.e. whenever the faucet is opened. Efficiency is improved by inserting two simple commercial pre-filters at the outlet of the elevated tank as these reusable filters remove leaves, dust and particulate matter.
These prefilters are cleaned or replaced every 2-4 weeks.

The hemodialyzer filtration device
This device contains the eight commercial hollow fiber hemodialyzers that had been reprocessed after one clinical use. The reprocessing procedure has been a standard procedure for reuse on the same patient in some hemodialysis facilities (ANSI-AAMI ). The NUF500 device has been designed and patented by Y. Lass (Lass ). It is

TECHNICAL REFINEMENTS
The elevation of the water reservoir is a potential variable that deserves discussion. A greater height differential to the hemodialyzer provides greater flow rates at the faucet, but requires a stronger construction for a higher elevation of the large tank on a concrete foundation for stability. An elevation for the tank of 3 m above ground has been typical although a greater height difference may be advantageous for larger communities. Flow rates of 250 L/hour have been common particularly after each back-flushing procedure even over longer terms.
As an alternative to this gravity driven design, this same device could be set up for manual pumping, which could achieve flowrates as high as 500 L/hour.

Assigning maintenance tasks to a village:
Since backflushing needs to be done 3-4 times daily, this task had to be assigned to 2 or 3 villagers after they had been trained by our staff. Initially our NGO paid for this service by the trained villagers. Some village elders felt that this task should be the responsibility of the village and thus not depend on donated money. We welcome such independence and are working with other villages to follow this example.
Working through established village committees to provide initial and ongoing education has contributed the success of the project. Through these brief and periodic education sessions, the village population, including school children, has recognized the benefits of clean drinking water. This has brought about community empowerment, and contributed to the care, maintenance and sustenance of the devices.
Filling the water tank with the contaminated source water, e.g. river water, has been done typically 2-5 times weekly by our staff. A staff member of our NGO comes to the site by boat bringing a gasoline-driven pump to fill the tank and verify the flowrate for the drinking water. Recently some villages have indicated that they would like to perform this service themselves. Given the relatively low cost of the pump, we now provide a new pump to selected villages to allow them to become even more independent. Our plan is to maintain regular but less frequent (e.g. monthly) visits from our NGO staff member to assure that everything is working well. we arbitrarily instituted a protocol to replace the hemodialyzers every other year.
Ability to increase clean water availability for a village: We aim to have ample water available for each village so that water generated is used not only for drinking but also for handwashing and other purposes. The amount of water available per day can potentially be very high since flow is 'on demand' whenever the faucet is opened day or night.

Far greater water consumption can be easily supported by
filling the water tank more often, up to seven times weekly. The added cost would be very small. We estimate that the current design could easily serve villages up to 1,500 population, even when handwashing becomes common practice. We recommend handwashing with soap as is also strongly supported by published work (Ejemot- This was treated by a procedure that temporarily blocks the hydrostatic pressure and installs a hypochlorite solution from the faucet upstream to the hemodialyzers for a 15-minute dwell. After subsequently discarding the bleach with newly filtered water for several minutes until the taste and odor of bleach could no longer be detected (verified by undetectable levels in analyses), the system was again functioning well and levels of heterotrophic bacteria were zero or in acceptable ranges. This procedure is scheduled to be repeated on a monthly basis until we are certain that the water meets all standards consistently.

Cost of providing clean water per village:
It is difficult to accurately estimate the cost/liter; however the cost per village can be estimated and should include clean water for handwashing. We recognize that costs for the initiation of this system are quite high, but we have experience in that the life span of the system already exceeds 4 years in Ghana. Therefore, Table 1 provides the overall

Observed health benefits
In a prospective study (Port et al. ; Raimann et al. ) we collected data monthly on the incidence of diarrhea in the households of 4 villages before and after implementation of our water treatment during February through November 2018. The incidence of diarrhea for the two 5 month periods before and after implementation of the hemodialyzer filtration device was recorded.
The monthly rate per 100 villagers averaged 8.1 before versus 3.0 after initiation of the hemodialyzer filtration device. This suggests a 63% reduction in diarrhea incidence (rate ratio ¼ 0.37) when the drinking water source changed from polluted river water to hemodialyzer filtered water.
In a control group of 5 villages in the same area that had not (yet) initiated our filtered water system during the same calendar months, the same analysis suggested only a non-significant reduction in diarrhea incidence during the second 5 months.
These analyses were based on 2,605 villagers who participated in the monthly data collection. The large reduction in diarrhea has been noticed right away andy praised by mothers and teachers.

Current status
As of February 2020 there were 20 villages with functioning devices in Ghana and one in Uganda. The population served with hemodialyzer-filtered clean water consisted of approximately 8,000 villagers. Additionally, about 2,000 schoolchildren have full access to pure water for drinking and handwashing while in school.
Expansion to four additional villages was already accomplished in early 2020 in Ghana with more units planned for this year. In collaboration with two other NGOs, the same system for clean water is being prepared for installation in villages in Uganda and Senegal.

DISCUSSION
During the past 4 years, the implementation and refinement of our gravity-driven hemodialyzer filtration device have allowed us to demonstrate the success of providing clean and pathogen-free drinking water to villages where sources of water have been consistently contaminated with fecal pathogens. Since this system works well even in remote areas without any available power, we have focused primarily on villages that previously have had no other opportunities for obtaining clean water. No restrictions on water use will need to be imposed even when our encouragement to also use clean water for frequent hand washing receives full acceptance in the villages. We agree with numerous publications that emphasize the additional health benefits from handwashing (Ejemot-Nwadiaro et al. Piper et al. ). When more water is desired, we are able to increase the frequency of filling of the main water tank to several times a week, or potentially even daily. We are currently developing a method to measure the actual consumption of pure water, but we are certain that it is far below maximum capacity of water provision with the device in all studied villages.

;
In remote villages, the use of solar panels might be a useful alternative to the gravity-driven design discussed here. However, it will substantially increase the cost and the complexity of its use and perhaps also its maintenance.
Given the advantages of our gravity-driven system including simplicity and greater cost-effectiveness, we have not explored a solar energy driven alternative and have therefore not gained any experience with it. Of note, one village served by our NGO had solar panels installed two years prior to our arrival, however they had never been put to use at that village.
The polysulfone membrane used in our hemodialysisbased device is superior to a recently tested PES hollow

CONCLUSIONS AND OUTLOOK
We have demonstrated the successful use in 20 remote villages of a filtration device that is based on hemodialyzers to convert polluted water to pathogen-free drinking water.
It works well for entire villages (<1,500 population) in remote areas that have no power source. Despite a substantial initial expenditure the overall cost including maintenance is low over time with good sustainability of its use. Importantly, we have documented beneficial effects on public health in the studied villages with a clear reduction in the incidence of diarrhea after the initiation of the repurposed hemodialyzer filtration devices for drinking water. Since availability of clean water is unrestricted with our system, we are currently encouraging regular handwashing with filtered water. We conclude that the use of this device utilizing repurposed hemodialyzers in our gravity-driven system is a simple, highly effective and low-cost approach to serve remote communities in developing countries where water is exposed to fecal contamination.
We find it particularly encouraging to observe that some villages are working eagerly toward running our system independently. Considering that the dialysis industry produces millions of hemodialyzers annually, large-scale efforts to sterilize dialyzers rather than discarding them after a one time clinical use and using them for water purification could solve problems of polluted water more widely in Ghana and elsewhere. We hope to assist others in extending this useful approach for the benefit of villagers in poor regions with high incidences of waterborne disease especially in young children.

CONCLUSION
Our experience with disinfecting drinking water through gravity-fed hemodialyzer filtration shows longer-term success and acceptance in addition to our reported reductions in diarrhea. Simplicity, high effectiveness and low cost

CONFLICT OF INTEREST STATEMENT
The lead author, Dr Friedrich K. Port, affirms for all authors that they have reviewed and contributed to this manuscript.

Jochen Raimann and Seth Johnson are employees of Renal
Research Institute; all other authors declare no conflict of interest. This study has been supported solely by the nongovernment organization (NGO), Easy Water for Everyone.