Household water treatment (HWT) can reduce the diarrheal disease burden in populations without access to safe water. We evaluated five programs that distributed biosand, ceramic, or Sawyer filters in Haiti after the 2010 earthquake and cholera outbreak. We conducted household surveys and tested Escherichia coli and turbidity in stored household untreated and treated water in ∼50 randomly selected households from each program. Across programs, self-reported filter use ranged from 27 to 78%; confirmed use (participants with reported use who also showed the filter with water currently in it) ranged from 20 to 76%; and effective use (participants who used the filter to improve water quality to international guideline values) ranged from 0 to 54%. Overall, programs that more successfully met evaluation metrics: (1) distributed an effective technology; (2) provided safe storage; (3) required cash investment; (4) provided initial training; (5) provided follow-up; (6) provided supply-chain access; (7) targeted households relying on contaminated water sources; and, (8) had experience working in the local context. These findings, similar to results of previous research on HWT, suggest that well-implemented programs have the potential to result in sustained household filter use in Haiti.

INTRODUCTION

Household water treatment (HWT) can be a cost-effective means of improving drinking water quality (Clasen et al. 2007) and reducing diarrheal disease in households where access to water and sanitation infrastructure is limited (Fewtrell et al. 2005; Clasen et al. 2006a; Clasen 2015), and is therefore recommended as part of a comprehensive strategy to prevent diarrheal disease in low-income settings without access to safe drinking water (World Health Organization/United Nations Children's Fund WHO/UNICEF 2011).

Worldwide, approximately 663 million people lack access to an improved water source and an estimated 1.2 billion (109) more rely on contaminated water sources (Onda et al. 2012; WHO/UNICEF 2015). In Haiti, 48% of rural and 65% of urban households have access to an improved water source (WHO/UNICEF 2015). However, in a 2012 study, Escherichia coli was detected in half (50.9%) of the improved water sources tested in Artibonite, Haiti (Patrick et al. 2013).

Chlorination has been widely promoted to treat drinking water in Haiti; nearly all (96%) of the 71% of Haitians who report treating their water report using chlorine-based products (Cayemittes et al. 2013). In a 2008 evaluation of a chlorine distribution program with training and follow-up, 56% of participants (versus 10% of controls) had free chlorine residual (FCR) in their drinking water and children <5 years old had 59% reduced odds of diarrhea (95% confidence interval [CI] = 0.21,0.79) (Harshfield et al. 2012). In 2010, after the earthquake, 72% and 90% of participants in this program had FCR in their drinking water 2 and 10 months post-earthquake, respectively (Lantagne & Clasen 2012b, 2013). In contrast, less than 2 years after chlorine tablets were distributed with just mass-marketing behavior change messages to rural households in Artibonite Department, Haiti, only 9.9% of households had stored water containing FCR (Patrick et al. 2013). These results suggest that when sufficient training, follow-up, and supply chain are provided, chlorine can be used to improve drinking water quality and reduce diarrheal disease in Haiti.

Before the January 2010 earthquake, filtration-based technologies had not been widely promoted in Haiti; only ∼1% of households that reported treating their drinking water reported using filtration-based systems (Cayemittes et al. 2013). In a 2005 biosand filter study in Artibonite, filters that had been used for an average 2.5 years reduced E. coli by 98.5% and turbidity by 85% (Duke et al. 2006). A follow-up study in 2011, on 55 households in this program, found that 53% of filters installed <1–12 years previously were still in use, and on average reduced E. coli by 92% (Sisson et al. 2013). These studies suggest potential for long-term use of biosand filters in Haiti.

Since 2010, distribution of filtration-based technologies has rapidly increased, and little is known about whether filter implementation programs in Haiti are resulting in improved drinking water quality at the household level. We evaluated filter distribution programs by: (1) identifying and inviting distribution programs to participate in this evaluation, and (2) visiting program households, conducting a survey, and testing water quality to evaluate the effectiveness of filter implementation programs in the post-earthquake context in Haiti.

METHODS

Program and household selection

Organizations operating filter distribution programs in Haiti between 2010 and 2014 were identified, by asking those with HWT knowledge in Haiti, and invited to participate in the evaluation. Participating organizations provided household distribution lists. We selected a distribution region from each list based on the following criteria: (1) >50 recipient households in the region; (2) ability to locate households; and (3) accessibility during the survey timeframe. After region selection, 253 households were randomly selected, 50–53 from each distribution list, for survey and water sampling.

Household surveys

Household surveys were conducted in Haitian Creole by trained Haitian enumerators in August 2014. The survey consisted of a general section of 48 questions on demographics, water knowledge, attitudes, and practices. If the participant reported receiving a filter, 46–48 technology-specific questions were asked about filter use, maintenance, cleaning, and satisfaction. Personally identifying information was not collected. Informed consent was obtained before administering the ∼40-minute survey. The Tufts University and Haitian Institutional Review Boards approved the protocol and survey tools.

Water sampling and testing

Enumerators requested a cup of drinking water; if respondents reported the water was filtered, enumerators also collected: (1) stored untreated water, and (2) water directly from the filter outlet or tap (after wiping with a disposable alcohol pad). Samples were collected aseptically in Whirl-Pak™ bags with sodium thiosulfate (Nasco, Ft. Atkinson, WI), stored on ice, and processed within 12 hours of collection.

Samples were tested for E. coli and total coliform bacteria using membrane filtration. Samples were diluted appropriately with sterile buffered water, filtered aseptically through a 45-μm Millipore filter (Billerica, MA) on a portable Millipore filtration stand, placed in a plastic Petri-dish with a pad soaked in mColiBlue24® media, and incubated at 35–37°C for 24-hours. Sterile buffered water controls were tested every 20 plates, and 10% of samples were duplicated for quality control. Plates were considered countable at <200 colonies. Turbidity was measured in excess water from paired untreated/direct-from-filter samples with a calibrated Lamotte 2020 Turbidimeter (Chestertown, MD).

Analysis

Survey and water quality data were collected on paper, entered into Microsoft Excel (Redmond, WA), and all data were analyzed using STATA/IC 10.1 (College Station, TX).

Summary descriptive statistics were calculated. Primary program evaluation metrics were: reported use, confirmed use, and effective use. Reported use was calculated as the percentage of the surveyed population that provided a drinking water sample and self-reported that it had been filtered with the program filter. Confirmed use was the percentage that met reported use criteria plus showed the filter with water in it. Effective use was calculated as the percentage of the surveyed population that met reported use criteria and used the filter to improve drinking water from ≥1 E. coli coliform forming units (CFU)/100 mL in untreated stored water to <1 E. coli CFU/100 mL in treated drinking water (Lantagne & Clasen 2012b). Effective use was also calculated using 10 E. coli CFU/100 mL as a low-risk break-point (WHO 2011).

Microbiological test results were reported using geometric mean, with values of <1 CFU/100 mL replaced with 0.5 CFU/100 mL. Additionally, where >40% of untreated water samples had ≥100 E. coli CFU/100 mL, log10 reduction values (LRVs) were calculated from paired untreated/direct-from-filter samples for E. coli as a risk reduction indicator. Where >40% of untreated water samples had ≤10 E. coli CFU/100 mL and ≥100 total coliform bacteria CFU/100 mL, LRV was calculated from paired untreated/direct-from-filter samples for total coliform bacteria as a filtration process indicator.

Cleaning was considered correct if respondents reported cleaning the tap/filter outlet and using treated (boiled, filtered, or chlorinated) water to clean their drinking water storage container, and: (1) for ceramic filters, cleaning the membrane with treated water and using a brush/cloth exclusively for cleaning the membrane; (2) for biosand filters, using the ‘swirl and dump’ method (Centre for Affordable Water and Sanitation Technology CAWST 2012) to clean the sand layer; and (3) for Sawyer filters, reporting back-flushing. Storage was considered safe if the drinking water container had a lid and tap.

RESULTS

Program and household selection

Organizations reported >140,000 biosand, ceramic, Sawyer, and Lifestraw® filters were distributed in Haiti between 2010 and 2014. Six filter distribution programs were identified and four organizations, representing five filter distribution programs, agreed to participate, including: (1) Pure Water for the World (PWW), which distributed plastic-casing HydrAid® biosand filters; (2) Clean Water for Haiti (CWH), which distributed locally manufactured concrete-casing biosand and Atabey ceramic ‘pot’ filters manufactured in the Dominican Republic; (3) L'Association Saint-Luc d'Haïti (ASSLHA), which distributed locally-manufactured FilterPure ceramic ‘pot’ filters; and, (4) a Sawyer filter distributor, who facilitated three Sawyer PointONE™(hollow fiber membrane filters) filter distributions by different organizations.

Household surveys

A total of 223 (88%) of 253 randomly selected households were visited: 44–46 households per each of the five programs (Table 1). Both PWW and CWH biosand programs have been working in Haiti since before the 2010 earthquake, distributed filters at a subsidized price, and provided on-going follow-up. The CWH ceramic program distributed filters at a subsidized price, but provided neither follow-up nor supply chain access. Both ASSLHA and Sawyer programs distributed free filters, but ASSLHA provided no follow-up or supply chain, while follow-up and supply chain varied across the three Sawyer filter distributions. All programs provided initial training on filter use and maintenance.

Table 1

Program Information

Program details PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer 
Filter brand/source HYDRAID®, plastic casing CWH, concrete-casing Atabey, Dominican Republic FilterPure, Haiti Sawyer PointONE™ 
Time working in Haiti Pre 2010 earthquake Pre 2010 earthquake Post 2010 earthquake Post 2010 earthquake Post 2010 earthquake 
Distribution timeframe March– December 2013 October 2012–November 2013 November 2011–August 2013 March–April 2014 March 2014 
Context Development Development Emergency Development Development 
Geography Rural/Semi-Rural Mountains Artibonite Delta/Rural Rural Mountains Semi-Rural Coastal/Semi-Rural Mountains 
Training Community meeting During installation Community meeting Community meeting Community meeting 
Follow-up Regular, by paid technician Regular, by paid technician None None Variable 
Household investment Subsidized Subsidized Subsidized Free Free 
Houses in region, n 92 406 70 106 98 
Houses selected, n 50 50 53 50 50 
Households surveyed, n (%) 45 (90%) 44 (88%) 44 (83%) 44 (88%) 46 (92%) 
Program details PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer 
Filter brand/source HYDRAID®, plastic casing CWH, concrete-casing Atabey, Dominican Republic FilterPure, Haiti Sawyer PointONE™ 
Time working in Haiti Pre 2010 earthquake Pre 2010 earthquake Post 2010 earthquake Post 2010 earthquake Post 2010 earthquake 
Distribution timeframe March– December 2013 October 2012–November 2013 November 2011–August 2013 March–April 2014 March 2014 
Context Development Development Emergency Development Development 
Geography Rural/Semi-Rural Mountains Artibonite Delta/Rural Rural Mountains Semi-Rural Coastal/Semi-Rural Mountains 
Training Community meeting During installation Community meeting Community meeting Community meeting 
Follow-up Regular, by paid technician Regular, by paid technician None None Variable 
Household investment Subsidized Subsidized Subsidized Free Free 
Houses in region, n 92 406 70 106 98 
Houses selected, n 50 50 53 50 50 
Households surveyed, n (%) 45 (90%) 44 (88%) 44 (83%) 44 (88%) 46 (92%) 

n = number of households.

The majority of respondents were female (68–86%) and the average respondent age was 39.4 years (Table 2). The average household size was 5.7 persons, and between 32 and 68% of households had ≥1 child <5 years old. Educational attainment of program participants was highest in ASSLHA ceramic, similar across PWW biosand, CWH biosand and Sawyer, and lowest in CWH ceramic programs. Top reported health concerns were fever (79%) and headache (63%). Respondents reported that during the prior week 9–27% of children <5 years of age had diarrhea (three or more loose or watery stools during 24 hours); this was highest in CWH biosand and ceramic programs, at 23% and 27%, respectively.

Table 2

Survey Respondent Demographics, Water Practices and Beliefs

  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer Overall 
Mean (SD) people per household 4.7 (2.3) 6.1 (2.3) 6.5 (2.4) 5.5 (2.3) 6.0 (2.3) 5.7 (2.4) 
HH has at least one child <5 years, % (n40% (45) 48% (44) 68% (44) 32% (44) 37% (46) 45% (223) 
Female respondent, % (n71% (45) 86% (44) 68% (44) 82% (44) 78% (46) 77% (223) 
Mean (SD) respondent age in years 42.4 (17.8) 38.4 (13.2) 36.7 (13.3) 37.7 (16.5) 41.6 (14.6) 39.4 (15.2) 
Respondent attended school, % (n62% (45) 68% (44)  55% (44) 91% (44) 70% (46) 69% (223) 
Mean (SD) highest grade, n7.6 (3.4) n = 28 7.2 (3.3) n = 30 5.2 (2.8) n = 24 10.3 (3.5) n = 40 7.0 (3.7) n = 32 7.7 (3.7) n = 154 
Reports female HOH can read, % (n61% (41) 59% (41) 44% (43) 76% (42) 55% (44) 59% (211) 
Reports male HOH can read, % (n71% (41) 69% (39) 68% (41) 79% (38) 70% (40) 71% (199) 
House has concrete floor, % (n53% (45) 23% (44) 59% (44) 100% (44) 72% (46) 61% (223) 
Has wired electricity, % (n0% (44) 77% (44) 0% (44) 51% (43) 2% (46) 26% (221) 
Reported diarrhea in past week, % (cases/ hh members) 6% (13/207) 8% (21/261) 17% (47/279) 2% (4/238) 2% (5/268) 7% (90/977) 
Reported diarrhea in children <5 years old in past week, % (cases/ total children) 9% (2/23) 23% (5/22) 27% (15/56) 12% (2/17) 8% (2/24) 18% (26/142) 
Primary water source is improved, % (n71% (45) 0% (44) 7% (44) 98% (44) 80% (46) 52% (223) 
Primary water source is surface water, % (n0% (45) 80% (44) 2% (44) 0% (44) 11% (46) 18% (223) 
Round-trip time (minutes) to collect water, median (lower, upper quartiles), n15 (5, 30) n = 45 7 (4, 30) n = 44 60 (30, 60) n = 44 5 (1, 15) n = 42 8 (3, 30) n = 46 10 (5, 35) n = 221 
Average times per day water is collected (SD), n3 (1.5) n = 43 4 (1.9) n = 43 2 (0.9) n = 44 4 (1.8) n = 29 3 (1.6) n = 38 3 (1.7) n = 197 
Believes water can make you sick, % (n) 100% (45) 86% (44) 89% (44) 98% (44) 100% (46) 95% (223) 
Reports received water treatment products since start of Cholera, % (n93% (45) 75% (44) 93% (44) 68% (44) 80% (46) 82% (223) 
Products received: 
 Biosand, % (n48% (42) 49% (33) 0% (41) 0% (30) 3% (37) 20% (183) 
 Ceramic, % (n0% (42) 0% (33) 32% (41) 20% (30) 3% (37) 11% (183) 
 Sawyer, % (n0% (42) 0% (33) 0% (41) 0% (30) 11% (37) 2% (183) 
 Aquatabs, % (n93% (42) 94% (33) 90% (41) 93% (30) 95% (37) 93% (183) 
Showed latrine, % (n71% (45) 41% (44) 23% (44) 79% (43) 67% (46) 56% (222) 
Observed hand washing station % (n16% (45) 7% (44) 5% (44) 21% (44) 7% (46) 11% (223) 
Observed soap, % (n7% (43) 2% (44) 0% (44) 14% (44) 2% (46) 5% (221) 
  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer Overall 
Mean (SD) people per household 4.7 (2.3) 6.1 (2.3) 6.5 (2.4) 5.5 (2.3) 6.0 (2.3) 5.7 (2.4) 
HH has at least one child <5 years, % (n40% (45) 48% (44) 68% (44) 32% (44) 37% (46) 45% (223) 
Female respondent, % (n71% (45) 86% (44) 68% (44) 82% (44) 78% (46) 77% (223) 
Mean (SD) respondent age in years 42.4 (17.8) 38.4 (13.2) 36.7 (13.3) 37.7 (16.5) 41.6 (14.6) 39.4 (15.2) 
Respondent attended school, % (n62% (45) 68% (44)  55% (44) 91% (44) 70% (46) 69% (223) 
Mean (SD) highest grade, n7.6 (3.4) n = 28 7.2 (3.3) n = 30 5.2 (2.8) n = 24 10.3 (3.5) n = 40 7.0 (3.7) n = 32 7.7 (3.7) n = 154 
Reports female HOH can read, % (n61% (41) 59% (41) 44% (43) 76% (42) 55% (44) 59% (211) 
Reports male HOH can read, % (n71% (41) 69% (39) 68% (41) 79% (38) 70% (40) 71% (199) 
House has concrete floor, % (n53% (45) 23% (44) 59% (44) 100% (44) 72% (46) 61% (223) 
Has wired electricity, % (n0% (44) 77% (44) 0% (44) 51% (43) 2% (46) 26% (221) 
Reported diarrhea in past week, % (cases/ hh members) 6% (13/207) 8% (21/261) 17% (47/279) 2% (4/238) 2% (5/268) 7% (90/977) 
Reported diarrhea in children <5 years old in past week, % (cases/ total children) 9% (2/23) 23% (5/22) 27% (15/56) 12% (2/17) 8% (2/24) 18% (26/142) 
Primary water source is improved, % (n71% (45) 0% (44) 7% (44) 98% (44) 80% (46) 52% (223) 
Primary water source is surface water, % (n0% (45) 80% (44) 2% (44) 0% (44) 11% (46) 18% (223) 
Round-trip time (minutes) to collect water, median (lower, upper quartiles), n15 (5, 30) n = 45 7 (4, 30) n = 44 60 (30, 60) n = 44 5 (1, 15) n = 42 8 (3, 30) n = 46 10 (5, 35) n = 221 
Average times per day water is collected (SD), n3 (1.5) n = 43 4 (1.9) n = 43 2 (0.9) n = 44 4 (1.8) n = 29 3 (1.6) n = 38 3 (1.7) n = 197 
Believes water can make you sick, % (n) 100% (45) 86% (44) 89% (44) 98% (44) 100% (46) 95% (223) 
Reports received water treatment products since start of Cholera, % (n93% (45) 75% (44) 93% (44) 68% (44) 80% (46) 82% (223) 
Products received: 
 Biosand, % (n48% (42) 49% (33) 0% (41) 0% (30) 3% (37) 20% (183) 
 Ceramic, % (n0% (42) 0% (33) 32% (41) 20% (30) 3% (37) 11% (183) 
 Sawyer, % (n0% (42) 0% (33) 0% (41) 0% (30) 11% (37) 2% (183) 
 Aquatabs, % (n93% (42) 94% (33) 90% (41) 93% (30) 95% (37) 93% (183) 
Showed latrine, % (n71% (45) 41% (44) 23% (44) 79% (43) 67% (46) 56% (222) 
Observed hand washing station % (n16% (45) 7% (44) 5% (44) 21% (44) 7% (46) 11% (223) 
Observed soap, % (n7% (43) 2% (44) 0% (44) 14% (44) 2% (46) 5% (221) 

n = sample size; SD = standard deviation; HOH = head of household; hh = household.

In CWH biosand and CWH ceramic programs, 0–7% of households had access to improved water sources, respectively, in contrast with >71% in other programs (Table 2). Nearly all respondents (95%) believed water can make you sick and considered water safe to drink when ‘treated’ (95%) or has ‘no bacteria’ (54%). Households largely underreported filters, in comparison with chlorine-based disinfectants, when asked to name methods for ‘treating’ drinking water. Nearly all participants (98–100%) named an HWT method, with 81–100% mentioning Aquatabs, but just 37–71% mentioning the program filter; even fewer reported receiving a filter since the cholera outbreak began. Access to latrines was lowest in CWH biosand and CWH ceramic programs (41% and 23%, respectively), whereas 67–79% of participants in other programs had access to a latrine.

The estimated time since distribution ranged from <1 to 5 years, averaging >1 year in CWH biosand and ceramic programs and <1 year in PWW biosand, ASSLHA and Sawyer programs (Table 3). Nearly all households in the PWW and CWH biosand programs reported paying for filters (96–98%), >50% in the CWH ceramic program, and few to none of ASSLHA and Sawyer participants. Of the total households surveyed, 82–100% of respondents reported receiving a filter (n = 223), of which 94% reported they have used it to ‘make the water clean’ (81%) or ‘eliminate bacteria’ (47%), they like it (86–100%), and it meets their drinking water needs (80–100%). The majority of households (82–96%) in the PWW biosand, CWH biosand, ASSLHA, and Sawyer programs reported use in the previous 24 hours; this was just 37% in the CWH ceramic program. Household members sometimes drink untreated water when ‘outside the home’ (85%) or when there is ‘no treated water’ (53%). Nearly half (19/39) of CWH ceramic filter recipients no longer used the filter because the membrane (100%), tap (5%) or storage container (11%) had broken.

Table 3

Reported Filter Use and Operation and Maintenance Knowledge

  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer Overall 
Reported receiving filter, % (n100% (45/45) 98% (43/44) 89% (39/44) 82% (36/44) 96% (44/46) 93% (207/223) 
Time since distribution, mean (min-max, years), n11 months ( < 1–2), n = 41 1.3 years (1–5), n = 39 1.2 years ( < 1–3), n = 36 <6 months ( < 1–1), n = 32 8 months ( < 1–3), n = 41 11 months ( < 1–5), n = 189 
Reported paying for filter, % (n96% (43) 98% (42) 64% (25) <1% (2) 0% (0) 51% (112) 
Reported ever using filter, % (n100% (45/45) 98% (42/43) 90% (35/39) 94% (34/36) 86% (38/44) 94% (194/207) 
Used filter in last 24 hours, % (n96% (45) 93% (42) 37% (35) 82% (34) 84% (38) 80% (194) 
Reported likes filter 100% (44) 100% (41) 86% (35) 100% (34) 97% (38) 97% (192) 
Reported meets water treatment needs 100% (45) 100% (42) 80% (35) 100% (34) 95% (38) 95% (194) 
Reported filtered water uses, % (n): 
Drinking 100% (45) 100% (42) 83% (35) 100% (34) 97% (38) 96% (194) 
Cooking 33% (45) 86% (42) 26% (35) 21% (34) 40% (38) 42% (194) 
Bathing 16% (45) 31% (42) 6% (35) 21% (34) 21% (38) 19% (194) 
Wash hands 13% (45) 10% (42) 6% (35) 15% (34) 8% (38) 10% (194) 
Wash fruits/vegetables 9% (45) 0% (42) 0% (35) 9% (34) 0% (38) 4% (194) 
Wash dishes 4% (45) 0% (42) 3% (35) 9% (34) 0% (38) 3% (194) 
Reports, % (n): 
 HH members sometimes drink untreated water 76% (45) 60% (42) 85% (39) 83% (36) 80% (44) 77% (206) 
 Received filter use training 100% (45) 100% (43) 87% (39) 97% (36) 96% (44) 96% (207) 
 Received filter cleaning training 100% (45) 100% (43) 85% (39) 97% (36) 91% (44) 95% (207) 
 Correct cleaning 40% (43) 55% (42) 33% (30) 65% (34) 32% (38) 45% (187) 
 Reports cleaning storage container with untreated water, % (n41% (44) 24% (42) 43% (30) 24% (34) 53% (34) 37% (184) 
 Knows where to buy parts 24% (45) 2% (43) 3% (39) 14% (36) 0% (44) 9% (207) 
 Knows who to contact with questions 93% (45) 35% (43) 39% (39) 56% (36) 71% (44) 59% (207) 
  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer Overall 
Reported receiving filter, % (n100% (45/45) 98% (43/44) 89% (39/44) 82% (36/44) 96% (44/46) 93% (207/223) 
Time since distribution, mean (min-max, years), n11 months ( < 1–2), n = 41 1.3 years (1–5), n = 39 1.2 years ( < 1–3), n = 36 <6 months ( < 1–1), n = 32 8 months ( < 1–3), n = 41 11 months ( < 1–5), n = 189 
Reported paying for filter, % (n96% (43) 98% (42) 64% (25) <1% (2) 0% (0) 51% (112) 
Reported ever using filter, % (n100% (45/45) 98% (42/43) 90% (35/39) 94% (34/36) 86% (38/44) 94% (194/207) 
Used filter in last 24 hours, % (n96% (45) 93% (42) 37% (35) 82% (34) 84% (38) 80% (194) 
Reported likes filter 100% (44) 100% (41) 86% (35) 100% (34) 97% (38) 97% (192) 
Reported meets water treatment needs 100% (45) 100% (42) 80% (35) 100% (34) 95% (38) 95% (194) 
Reported filtered water uses, % (n): 
Drinking 100% (45) 100% (42) 83% (35) 100% (34) 97% (38) 96% (194) 
Cooking 33% (45) 86% (42) 26% (35) 21% (34) 40% (38) 42% (194) 
Bathing 16% (45) 31% (42) 6% (35) 21% (34) 21% (38) 19% (194) 
Wash hands 13% (45) 10% (42) 6% (35) 15% (34) 8% (38) 10% (194) 
Wash fruits/vegetables 9% (45) 0% (42) 0% (35) 9% (34) 0% (38) 4% (194) 
Wash dishes 4% (45) 0% (42) 3% (35) 9% (34) 0% (38) 3% (194) 
Reports, % (n): 
 HH members sometimes drink untreated water 76% (45) 60% (42) 85% (39) 83% (36) 80% (44) 77% (206) 
 Received filter use training 100% (45) 100% (43) 87% (39) 97% (36) 96% (44) 96% (207) 
 Received filter cleaning training 100% (45) 100% (43) 85% (39) 97% (36) 91% (44) 95% (207) 
 Correct cleaning 40% (43) 55% (42) 33% (30) 65% (34) 32% (38) 45% (187) 
 Reports cleaning storage container with untreated water, % (n41% (44) 24% (42) 43% (30) 24% (34) 53% (34) 37% (184) 
 Knows where to buy parts 24% (45) 2% (43) 3% (39) 14% (36) 0% (44) 9% (207) 
 Knows who to contact with questions 93% (45) 35% (43) 39% (39) 56% (36) 71% (44) 59% (207) 

n = sample size, HH = household.

Most participants reported receiving training on filter use (96%) and cleaning (95%); this was slightly lower in the CWH ceramic program (87% and 85%, respectively) (Table 3). Correct cleaning practices were reported by 32–65% of respondents. Nearly 40% of respondents reported cleaning filtered water storage containers with untreated water. Biosand programs recommend chlorinating filtered water to compliment filtration; however, less than half (48%) reported chlorinating filtered water and 7% showed a bottle containing chlorine to the enumerator. Supply chain knowledge was low across all programs (<25%), but 35–93% of participants reported knowing whom to contact with questions about their filter.

The majority of respondents provided a drinking water sample (94%), but reported use varied as 78–80% of PWW and CWH biosand households surveyed, 50–57% of ASSLHA ceramic and Sawyer households surveyed, and 27% of CWH ceramic households surveyed both provided a water sample and reported it had been treated with the filter (Table 4). Respondents showed the filter to the enumerator in 48–100% of households, and of the observed filters, 57–100% had water in them (Table 4). Of the total households surveyed, 20–76% (n = 223) of households both met reported use criteria and showed the filter with water in it to the enumerator. PWW and CWH biosand programs had the highest (70–76%) confirmed use, ASSLHA and Sawyer had intermediate (43–54%), and CWH ceramic had the lowest confirmed use (20%). Most households had a safe storage container (67–100%), with the exception of the CWH biosand program (7%).

Table 4

Reported, Confirmed and Effective Use by Total Households Surveyed

  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic  Sawyer Overall 
Reported received filter, % (n100% (45) 98% (44) 89% (44) 82% (44) 96%(46) 93% (223) 
Provided drinking water sample, % (n93% (45) 95% (44) 91% (44) 93% (44) 96% (46) 94% (223) 
Reported sample was treated, % (n89% (45) 89% (44) 52% (44) 66% (44) 72% (46) 73% (223) 
Reported sample treated with filter, % (n78% (45) 80% (44) 27% (44) 50% (44) 57% (46) 58% (223) 
Time (hours) since treating water, median (lower, upper quartiles), n24 (5, 24) n = 35 6 (4, 24) n = 34 24 (24, 72) n = 12 24 (3, 24) n = 22 5 (2, 24) n = 26 24 (3, 24) n = 129 
Showed filter 100% (45) 98% (44) 48% (44) 82% (44) 85% (46) 82% (223) 
Safe storage observed, % (n100% (44) 7% (43) 95% (21) 100% (36) 67% (39) 73% (183) 
Water in filter, % (n100% (44) 91% (43) 57% (21) 75% (36) 85% (39) 82% (183) 
Reported use, % (n78% (45) 80% (44) 27% (44) 50% (44) 57% (46) 58% (223) 
Confirmed use, % (n76% (45) 70% (44) 20% (44) 43% (44) 54% (46) 53% (223) 
Untreated sample ≥1 and treated <1 E. coli CFU/100 mL, % (n44% (25) 25% (24) 0% (9) 57% (21) 48% (25) 39% (104) 
Effective use, 1 E. coli CFU/100 mL breakpoint, % 34% 20% 0% 29% 27% 23% 
Untreated sample was >10 and treated ≤10 E. coli CFU/100 mL, % (n48% (25) 67% (24) 33% (9) 43% (21) 40% (25) 48% (104) 
Effective use, 10 E. coli CFU/100 mL breakpoint, % 37% 54% 9% 22% 23% 28% 
  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic  Sawyer Overall 
Reported received filter, % (n100% (45) 98% (44) 89% (44) 82% (44) 96%(46) 93% (223) 
Provided drinking water sample, % (n93% (45) 95% (44) 91% (44) 93% (44) 96% (46) 94% (223) 
Reported sample was treated, % (n89% (45) 89% (44) 52% (44) 66% (44) 72% (46) 73% (223) 
Reported sample treated with filter, % (n78% (45) 80% (44) 27% (44) 50% (44) 57% (46) 58% (223) 
Time (hours) since treating water, median (lower, upper quartiles), n24 (5, 24) n = 35 6 (4, 24) n = 34 24 (24, 72) n = 12 24 (3, 24) n = 22 5 (2, 24) n = 26 24 (3, 24) n = 129 
Showed filter 100% (45) 98% (44) 48% (44) 82% (44) 85% (46) 82% (223) 
Safe storage observed, % (n100% (44) 7% (43) 95% (21) 100% (36) 67% (39) 73% (183) 
Water in filter, % (n100% (44) 91% (43) 57% (21) 75% (36) 85% (39) 82% (183) 
Reported use, % (n78% (45) 80% (44) 27% (44) 50% (44) 57% (46) 58% (223) 
Confirmed use, % (n76% (45) 70% (44) 20% (44) 43% (44) 54% (46) 53% (223) 
Untreated sample ≥1 and treated <1 E. coli CFU/100 mL, % (n44% (25) 25% (24) 0% (9) 57% (21) 48% (25) 39% (104) 
Effective use, 1 E. coli CFU/100 mL breakpoint, % 34% 20% 0% 29% 27% 23% 
Untreated sample was >10 and treated ≤10 E. coli CFU/100 mL, % (n48% (25) 67% (24) 33% (9) 43% (21) 40% (25) 48% (104) 
Effective use, 10 E. coli CFU/100 mL breakpoint, % 37% 54% 9% 22% 23% 28% 

n = sample size.

Median turbidity in source water (Table 5) was low in all programs (<5 NTU), with the exception of the CWH biosand program, where households use surface water. CWH biosand filters reduced turbidity by 98%, from a median 25 NTU in untreated water to 0.49 NTU in direct-from-filter samples (n = 6).

Table 5

Water Quality

  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer Overall 
Turbidity (NTU): 
 Untreated stored, median (lower, upper quartiles), n0.58 (0.38, 1.04) n = 10 25.44 (2.90, 50.1) n = 6 2.43 (0.83, 4.06) n = 6 0.89 (0.41, 3.78) n = 12 0.12 (0.06, 0.47) n = 13 0.64 (0.23, 2.33) n = 47 
 Direct-from-filter, median (lower, upper quartiles), n0.13 (0.10, 0.67) n = 10 0.49 (0.21, 0.73) n = 6 6.8 (0.44, 8.92) n = 6 0.70 (0.11, 1.88) n = 12 0.28 (0.03, 0.91) n = 13 0.39 (0.1, 1.37) n = 47 
Total coliform CFU/100 mL: 
 Untreated stored Geometric mean (min-max), n1,963 (220–10,485) n = 25 5,233 (800–14,250) n = 24 456 (134–1,060) n = 9 383 ( < 1–5,184) n = 22 1,532 (40–8,000) n = 25 1,451 ( < 1–14,250) n = 105 
 Direct-from-filter Geometric mean (min-max), n90 ( < 1–1,800) n = 25 13 (1–4,140) n = 24 256 (98–520) n = 7 5 ( < 1–200) n = 18 174 (3–4,036) n = 24 42 ( < 1–4,140) n = 98 
 Treated Geometric mean (min-max), n391 (2–4,602) n = 25 536 (6–8,000) n = 24 520 (179–4,006) n = 9 90 ( < 1–4,032) n = 21 608 (12–4,150) n = 25 356 ( < 1–8,000) n = 104 
E. coli CFU/100 mL: 
 Untreated stored Geometric mean (min-max), n29.3 (5–485) n = 25 691.3 (46–4,250) n = 24 78.5 (10–755) n = 9 6.6 ( < 1–250) n = 22 12.6 ( < 1–4,000) n = 25 39.2 ( < 1–4,250) n = 105 
 Direct-from-filter Geometric mean (min-max), n1.5 ( < 1–40) n = 25 1.1 ( < 1–140) n = 24 21.5 (2–260) n = 7 <1 ( < 1- < 1) n = 18 1.0 ( < 1–36) n = 24 1.2 ( < 1–260) n = 98 
 Treated Geometric mean (min-max), n2.3 ( < 1–110) n = 25 6.1 ( < 1–4,000) n = 24 16.4 (2–980) n = 9 1.2 ( < 1–400) n = 21 2.2 ( < 1–400) n = 25 3.0 ( < 1–4,000) n = 104 
  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer Overall 
Turbidity (NTU): 
 Untreated stored, median (lower, upper quartiles), n0.58 (0.38, 1.04) n = 10 25.44 (2.90, 50.1) n = 6 2.43 (0.83, 4.06) n = 6 0.89 (0.41, 3.78) n = 12 0.12 (0.06, 0.47) n = 13 0.64 (0.23, 2.33) n = 47 
 Direct-from-filter, median (lower, upper quartiles), n0.13 (0.10, 0.67) n = 10 0.49 (0.21, 0.73) n = 6 6.8 (0.44, 8.92) n = 6 0.70 (0.11, 1.88) n = 12 0.28 (0.03, 0.91) n = 13 0.39 (0.1, 1.37) n = 47 
Total coliform CFU/100 mL: 
 Untreated stored Geometric mean (min-max), n1,963 (220–10,485) n = 25 5,233 (800–14,250) n = 24 456 (134–1,060) n = 9 383 ( < 1–5,184) n = 22 1,532 (40–8,000) n = 25 1,451 ( < 1–14,250) n = 105 
 Direct-from-filter Geometric mean (min-max), n90 ( < 1–1,800) n = 25 13 (1–4,140) n = 24 256 (98–520) n = 7 5 ( < 1–200) n = 18 174 (3–4,036) n = 24 42 ( < 1–4,140) n = 98 
 Treated Geometric mean (min-max), n391 (2–4,602) n = 25 536 (6–8,000) n = 24 520 (179–4,006) n = 9 90 ( < 1–4,032) n = 21 608 (12–4,150) n = 25 356 ( < 1–8,000) n = 104 
E. coli CFU/100 mL: 
 Untreated stored Geometric mean (min-max), n29.3 (5–485) n = 25 691.3 (46–4,250) n = 24 78.5 (10–755) n = 9 6.6 ( < 1–250) n = 22 12.6 ( < 1–4,000) n = 25 39.2 ( < 1–4,250) n = 105 
 Direct-from-filter Geometric mean (min-max), n1.5 ( < 1–40) n = 25 1.1 ( < 1–140) n = 24 21.5 (2–260) n = 7 <1 ( < 1- < 1) n = 18 1.0 ( < 1–36) n = 24 1.2 ( < 1–260) n = 98 
 Treated Geometric mean (min-max), n2.3 ( < 1–110) n = 25 6.1 ( < 1–4,000) n = 24 16.4 (2–980) n = 9 1.2 ( < 1–400) n = 21 2.2 ( < 1–400) n = 25 3.0 ( < 1–4,000) n = 104 

Geometric mean total coliform bacteria ranged from 383 to 5,233 CFU/100 mL (n = 104) in untreated water and 5–256 CFU/100 mL in direct-from-filter samples (n = 98) (Table 5). Geometric mean E. coli in untreated water ranged from 6.6 to 691.3 E. coli CFU/100 mL (n = 105) (Table 5, Figure 1(a)). Geometric mean direct-from-filter samples had <2 E. coli CFU/100 mL across all programs with the exception of CWH ceramic, which had geometric mean 21 E. coli CFU/100 mL. Geometric mean E. coli concentrations in treated water ranged from 1.2 to 16.3 E. coli CFU/100 mL (n = 104). In all programs except CWH ceramic, these values were greater than direct-from-filter samples, suggesting post-filtration contamination.
Figure 1

(a) Geometric mean E. coli concentrations in household water per program; (b) Percentage of untreated water samples with ≤10, 11–100 and >100 E. coli CFU/100 mL per program.

Figure 1

(a) Geometric mean E. coli concentrations in household water per program; (b) Percentage of untreated water samples with ≤10, 11–100 and >100 E. coli CFU/100 mL per program.

More than 40% of untreated samples in the CWH biosand and ceramic programs had ≥100 E. coli CFU/100 mL (Figure 1(b)). The CWH biosand filters achieved a 2.9 LRV (99.87% reduction) in E. coli (n = 22); in contrast, the CWH ceramic filters achieved an average 0.56 LRV (72.5% reduction) (n = 3). More than 40% of untreated samples in the ASSLHA ceramic and Sawyer programs had ≤10 E. coli CFU/100 mL (Figure 1(b)). Ceramic filters in the ASSLHA program achieved a 2 LRV (99% reduction) in coliform bacteria (n = 12) and Sawyer filters achieved a 1 LRV (90% reduction) in coliform bacteria (n = 23).

The percentage of households whose water quality was improved from ≥1 E. coli CFU/100 mL in untreated to <1 in treated water samples ranged from 0 to 57% (n = 104). Effective use by the surveyed population (reported use and improved water quality with the filter), ranged from 0 to 34% (Table 4). The percentage of households whose water was improved water from >10 to ≤10 E. coli CFU/100 mL was 33–67%; thus, using 10 E. coli as the breakpoint, effective use ranged from 9 to 54%.

DISCUSSION

We evaluated five programs that have distributed biosand, ceramic, and Sawyer filters in Haiti since 2010. Our results indicate that, as measured by reported, confirmed, and effective use, program effectiveness is likely related to the extent to which programs: (1) distributed an effective technology; (2) provided safe storage; (3) required cash investment; (4) provided initial training; (5) provided follow-up; (6) provided supply-chain access; (7) targeted households relying on contaminated water sources; and, (8) had experience working in the local context (Table 6).

Table 6

Program Strategies and Characteristics

  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer 
Average time since distribution 11 months 1.3 years 1.2 years <6 months 8 months 
Technology effective − ≈ 
Safe storage container − ≈ 
Cash investment by household ≈ − − 
Received initial training ≈ 
Follow-up provided − − ≈ 
Supply chain present or respondent knows who to contact − − ≈ ≈ 
Primary water source is unimproved ≈ − − 
Program experienced in local context ≈ − − 
  PWW Biosand CWH Biosand CWH Ceramic ASSLHA Ceramic Sawyer 
Average time since distribution 11 months 1.3 years 1.2 years <6 months 8 months 
Technology effective − ≈ 
Safe storage container − ≈ 
Cash investment by household ≈ − − 
Received initial training ≈ 
Follow-up provided − − ≈ 
Supply chain present or respondent knows who to contact − − ≈ ≈ 
Primary water source is unimproved ≈ − − 
Program experienced in local context ≈ − − 

Extent to which program addressed: + high; − low; ≈ average.

It is fundamental that HWT technologies are effective at removing bacteriological contamination from source water. On average, direct-from-filter water quality was low risk (geometric mean <2 E. coli) in four of the five programs. The CWH ceramic filters, however, did not improve water quality to low-risk levels and only achieved 0.56 LRV. This contrasts both with results from the ASSLHA ceramic filters and previous research on ceramic filters (Brown et al. 2008; Abebe et al. 2014), and is hypothesized, based on factory visits, to be a result of poor quality control in manufacturing (Rayner & Lantagne 2014). The CWH biosand filters achieved an average 2.9 LRV in E. coli, suggesting these biosand filters are appropriate for the highly contaminated, turbid surface water in the Artibonite region where they were distributed. These results are consistent with, or better than, previous research results on biosand filters (Duke et al. 2006; Stauber et al. 2006; Stauber et al. 2009, 2012; Sisson et al. 2013).

With the exception of households in the CWH biosand program, safe storage containers were observed in the majority of households. In comparison with other programs, E. coli concentrations in treated samples were relatively greater than direct-from-filter concentrations. This is reflected by the increase in effective use from 20 to 54% using 10 E. coli as the breakpoint. Although post-contamination could have occurred in the household cup, the lack of safe storage containers (7%) likely allowed for contamination during storage.

Controlling for time since distribution, cash investment has been associated with long-term filter use in previous ceramic filter research (Brown & Sobsey 2006), and likely represents both perceived need and interest in using the technology. In both the ASSLHA and Sawyer filter programs, where nearly all filters were free, reported use was just 50–57%. Additionally, households in these programs had access to improved water sources, possibly influencing perceived need to treat water. Although >50% of CWH ceramic participants paid for their filters, time since distribution, high breakage rates, and the absence of supply chain likely prohibited sustained use, as discussed below.

While all programs provided initial training, the percentage of respondents reporting correct cleaning was variable, with many reporting using untreated water to clean storage containers. Household cups used to provide water for sampling may also have been washed with untreated water. These likely contributed to post-contamination, which occurred across programs, and is widely documented (Wright et al. 2004). While biosand filter users are taught to chlorinate filtered water, which would protect water after filtration, reported compliance was low. Additionally, across all programs, respondents reported drinking untreated water when outside the home or when there is no filtered water; this could potentially limit health benefits, as high adherence is necessary to realize health gains (Hunter et al. 2009; Brown & Clasen 2012). Programs are encouraged to emphasize these points in training and follow-up with households.

While CWH programs appropriately targeted high-risk populations that relied primarily on unimproved water sources, ASSHLA and Sawyer households relied primarily on improved water sources, where >40% of stored water samples had ≤10 E. coli CFU/100 mL (Figure 1(b)). This limits potential risk reduction and therefore, effective use results (Lantagne & Clasen 2013). Ceramic filters distributed by ASSLHA did however achieve a 2 LRV in total coliforms, and no E. coli were detected in any direct-from-filter samples. This suggests potential for effectiveness should water quality vary due to the use of alternate sources, an emergency, or seasonality (Kostyla et al. 2015). In contrast, Sawyer filters achieved a 1 LRV in total coliform bacteria, and 33% of direct-from-filter samples had detectable E. coli. This relatively low LRV reduction is consistent with field data on Sawyer filters as summarized by Murray et al. (2015).

Access to a market-based supply chain was not available in any of the program communities; therefore, whether households would troubleshoot, identify and replace broken parts could not be assessed. Filter technicians fulfilled this role in the PWW and CWH biosand programs, contributing to long-term filter use. The absence of supply chain and/or follow-up prohibited potential for high use rates in the long-running CWH ceramic program, where breakage was the primary reason for disuse. Time since distribution and breakage in the absence of supply chain have been associated with disuse of ceramic filters previously (Brown & Sobsey 2006; Clasen et al. 2006b).

Programs in this evaluation with the highest reported and confirmed use have been running HWT distribution projects in Haiti since before the 2010 earthquake; this is consistent with previous findings (Lantagne & Clasen 2012b). These programs have developed and modified their distribution strategies over time, and this experience likely contributed to sustained filter use.

The limitations of this work include the small sample size limiting statistical analysis, cross-sectional study design, voluntary nature of program participation, time restrictions during field research in community selection, and that stored water quality is not necessarily representative of actual filter influent water. Despite these limitations, we do not expect our conclusions would vary much as they are consistent with HWT research in Haiti and elsewhere.

CONCLUSIONS

The themes identified in this research are consistent with previous studies. Program effectiveness is likely related to the extent to which programs: (1) distributed an effective technology; (2) provided safe storage; (3) required cash investment; (4) provided initial training; (5) provided follow-up; (6) provided supply-chain access; (7) targeted households relying on contaminated water sources; and, (8) had experience working in the local context. Our results suggest potential for long-term effective use of filters in Haiti. The extent to which program strategies address these themes will likely contribute to program success in achieving health gains and reducing the burden of diarrheal disease.

ACKNOWLEDGEMENTS

This work was funded by the United States Centers for Disease Control and Prevention (CDC) as part of their ongoing work with the Haitian government to strengthen waterborne disease prevention and control in Haiti. The findings and conclusions in this report do not necessarily represent the official position of the CDC. Use of trade names and commercial sources is for identification only and does not imply endorsement by CDC, or the United States Department of Health and Human Services.

REFERENCES

REFERENCES
Brown
J.
Sobsey
M.
2006
Independent Appraisal of Ceramic Water Filtration Interventions in Cambodia: Final Report
.
Submitted to UNICEF
.
Department of Environmental Sciences and Engineering, School of Public Health, University of North Carolina
,
Chapel Hill, NC
Brown
J.
Sobsey
M. D.
Loomis
D.
2008
Local drinking water filters reduce diarrheal disease in Cambodia: a randomized, controlled trial of the ceramic water purifier
.
Am. J. Trop. Med. Hyg.
79
(
3
),
394
400
.
CAWST
2012
Biosand Filter Construction Manual
.
Centre for Affordable Water and Sanitation Technology
,
Calgary, Alberta
,
Canada
.
Cayemittes
M.
Busangu
M. F.
Bizimana
J. d. D.
Barrère
B.
Sévère
B.
Cayemittes
V.
Charles
E.
2013
Enquête Mortalité, Morbidité et Utilisation des Services, Haïti, 2012
.
MSPP, IHE and ICF International
,
Calverton, MD
.
Clasen
T.
Robert
I.
Rabie
T.
Schmidt
W.
Cairncross
S.
2006a
Interventions to improve water quality for preventing diarrhoea
.
Cochrane Database Syst Rev
3
,
104
.
Clasen
T. F.
Brown
J.
Collin
S. M.
2006b
Preventing diarrhoea with household ceramic water filters: assessment of a pilot project in Bolivia
.
Int. J. Environ. Health. Res.
16
(
03
),
231
239
.
Duke
W. F.
Nordin
R. N.
Baker
D.
Mazumder
A.
2006
The use and performance of BioSand filters in the Artibonite Valley of Haiti: a field study of 107 households
.
Rural Remote Health
6
(
3
),
570
.
Fewtrell
L.
Kaufmann
R. B.
Kay
D.
Enanoria
W.
Haller
L.
Colford
J. M.
Jr.
2005
Water, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis
.
Lancet Infect. Dis.
5
(
1
),
42
52
.
Harshfield
E.
Lantagne
D.
Turbes
A.
Null
C.
2012
Evaluating the sustained health impact of household chlorination of drinking water in rural Haiti
.
Am. J. Trop. Med. Hyg.
87
(
5
),
786
795
.
Lantagne
D.
Clasen
T.
2012a
Point-of-use water treatment in emergency response
.
Waterlines
31
(
1
),
30
52
.
Murray
A.
Goeb
M.
Stewart
B.
Hopper
C.
Peck
J.
Meub
C.
Asatekin
A.
Lantagne
D.
2015
Fouling in hollow fiber membrane microfilters used for household water treatment
.
J. Wat. San. Hyg. Dev.
5
(
2
),
220
228
.
Onda
K.
LoBuglio
J.
Bartram
J.
2012
Global access to safe water: accounting for water quality and the resulting impact on MDG progress
.
Int. J. Env. Res. Public Health
9
(
3
),
880
894
.
Patrick
M.
Berendes
D.
Murphy
J.
Bertrand
F.
Husain
F.
Handzel
T.
2013
Access to safe water in rural Artibonite, Haiti 16 months after the onset of the cholera epidemic
.
Am. J. Trop. Med. Hyg.
89
(
4
),
647
653
.
Rayner
J.
Lantagne
D.
2014
Investigations on Household Filtration in Haiti: Final Report
.
Tufts University for the Centers for Disease Control and Prevention
,
Atlanta, GA
.
Sisson
A. J.
Wampler
P. J.
Rediske
R. R.
McNair
J. N.
Frobish
D. J.
2013
Long-term field performance of biosand filters in the Artibonite Valley, Haiti
.
Am. J. Trop. Med. Hyg.
88
(
5
),
862
867
.
Stauber
C. E.
Ortiz
G. M.
Loomis
D. P.
Sobsey
M. D.
2009
A randomized controlled trial of the concrete biosand filter and its impact on diarrheal disease in Bonao, Dominican Republic
.
Am. J. Trop. Med. Hyg.
80
(
2
),
286
293
.
Stauber
C. E.
Kominek
B.
Liang
K. R.
Osman
M. K.
Sobsey
M. D.
2012
Evaluation of the impact of the plastic BioSand filter on health and drinking water quality in rural Tamale, Ghana
.
Int. J. Env. Res. Public Health
9
(
11
),
3806
3823
.
WHO
2011
Evaluating Household Water Treatment Options: Health-Based Targets and Microbiological Performance Specifications
.
World Health Organization
,
Geneva
,
Switzerland
.
WHO/UNICEF
2011
Drinking Water: Equity, Safety and Sustainability
.
UNICEF, World Health Organization
,
Geneva
,
Switzerland
.
WHO/UNICEF
2015
Progress on Sanitation and Drinking Water - 2015 Update and MDG Assessment
.
UNICEF, World Health Organization
,
Geneva
,
Switzerland
.