Tintle et al. Tropical Medicine and Health (2021) 49:1 https://doi.org/10.1186/s41182-020-00291-y Tropical Medicine and Health RESEARCH Open Access Diarrhea prevalence in a randomized, controlled prospective trial of point-of-use water filters in homes and schools in the Dominican Republic Nathan Tintle1, Kristin Van De Griend2, Rachel Ulrich3, Randall D. Wade4, Tena M. Baar4, Emma Boven1, Carolyn E. A. Cooper5, Olivia Couch1, Lauren Eekhoff4, Benjamin Fry4, Grace K. Goszkowicz4, Maya A. Hecksel4, Adam Heynen6, Jade A. Laughlin4, Sydney M. Les4, Taylor R. Lombard4, B. Daniel Munson1, Jonas M. Peterson5, Eric Schumann4, Daniel J. Settecerri4, Jacob E. Spry4, Matthew J. Summerfield4, Meghana Sunder4, Daniel R. Wade4, Caden G. Zonnefeld1, Sarah A. Brokus4, Francesco S. Moen4, Adam D. Slater4, Jonathan W. Peterson7, Michael J. Pikaart5, Brent P. Krueger5 and Aaron A. Best4* Abstract Background: Lack of sustainable access to clean drinking water continues to be an issue of paramount global importance, leading to millions of preventable deaths annually. Best practices for providing sustainable access to clean drinking water, however, remain unclear. Widespread installation of low-cost, in-home, point of use water filtration systems is a promising strategy. Methods: We conducted a prospective, randomized, controlled trial whereby 16 villages were selected and randomly assigned to one of four treatment arms based on the installation location of Sawyer® PointONE™ filters (filter in both home and school; filter in home only; filter in school only; control group). Water samples and self- reported information on diarrhea were collected at multiple times throughout the study. Results: Self-reported household prevalence of diarrhea decreased from 25.6 to 9.76% from installation to follow-up (at least 7 days, and up to 200 days post-filter installation). These declines were also observed in diarrhea with economic or educational consequences (diarrhea which led to medical treatment and/or missing school or work) with baseline prevalence of 9.64% declining to 1.57%. Decreases in diarrhea prevalence were observed across age groups. There was no evidence of a loss of efficacy of filters up to 200 days post-filter installation. Installation of filters in schools was not associated with decreases in diarrhea prevalence in school-aged children or family members. Unfiltered water samples both at schools and homes contained potential waterborne bacterial pathogens, dissolved heavy metals and metals associated with particulates. All dissolved metals were detected at levels below World Health Organization action guidelines. (Continued on next page) * Correspondence: best@hope.edu 4Biology Department, Hope College, 35 E. 12th St, Holland, MI 49423, USA Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 2 of 14 (Continued from previous page) Conclusions: This controlled trial provides strong evidence of the effectiveness of point-of-use, hollow fiber membrane filters at reducing diarrhea from bacterial sources up to 200 days post-installation when installed in homes. No statistically significant reduction in diarrhea was found when filters were installed in schools. Further research is needed in order to explore filter efficacy and utilization after 200 days post-installation. Trial registration: ClinicalTrials.gov, NCT03972618. Registered 3 June 2019—retrospectively registered. Keywords: Drinking water, Point-of-use filter, 16S rRNA community, Diarrhea, Heavy metals Background PointONE™ water filter. Laboratory tests with the Saw- Globally, diarrhea caused over 1.65 million deaths in yer® PointONE™ water filter suggest it aligns with the 2016 [1] and half a million deaths in 145 low- and United States Environmental Protection Agency stand- middle-income countries in 2012 due specifically to ard for bacteria and protozoa removal [12]. Prior studies inadequate drinking water [2]. According to the World have identified a significant decrease in diarrhea preva- Health Organization (WHO), there are 1.7 billion cases lence [12, 13]. Some studies have argued that filters in of childhood diarrheal disease annually and 525,000 the field have been fouled and under-utilized in practice children under 5 die from the disease each year, making [14, 15]; however, others have noted numerous limita- diarrheal disease the second leading cause of death in tions of these studies [16] and reasonably good perform- children under 5 [3]. Since a major source of diarrhea is ance at removing E. coli and coliforms over a 1- to 3- fecal pathogens via fecal-oral transmission [4, 5] many of year period in the field [17]. Thus, there is a continued these lives could have been saved through clean drinking need for carefully designed field trials to evaluate the ef- water [2] and proper hand hygiene [5, 6]. ficacy of Sawyer® PointONETM and other hollow fiber The Dominican Republic is considered a middle- membrane filters. income country [7], and thus is at potentially high risk In an attempt to better understand filter efficacy, for negative impact of diarrheal disease. Results of utilization, and the impact of deployment in different household surveys confirm high prevalence of diarrhea community locations, we designed a prospective, in children under 5 in the Dominican Republic (32.3% randomized, controlled trial whereby 16 villages in the treated for diarrhea with oral rehydration salts in 2002; Dominican Republic were selected and randomly 46.3% in 2007 and 52.8% in 2013) [8]. Other studies have assigned to one of four treatment arms based on the found a similarly high burden of diarrheal disease. In location of Sawyer® PointONE™ filter installation (filter 2003, the Dominican Republic Demographic and Health in both home and school; filter in home only; filter in Survey reported that an average of 14% of all children school only; control group). Village households were under 5 years suffered from diarrhea, with rates up to followed over time, monitoring self-reported health 29% in certain provinces [9]. characteristics. Drinking water samples were also obtained Other studies have found related concerns about sus- from a subsample of households to monitor water quality tainable access to clean drinking water. In a peri-urban for bacterial and chemical contamination in order to, first, district of Santo Domingo, a study of 266 households directly assess filter removal of bacteria and particulates found that although 57% of interviewees believed their and, second, to assess whether the sole use of this type of child to be at risk for diarrhea, and 90.6% believed that filter as an intervention was contraindicated due to the boiling water could prevent diarrhea, only 42% reported presence of toxic levels of dissolved heavy metals. that it was rare for their child to drink untreated water with 34.5% stating ‘insufficient fuel’ as the primary Methods barrier [10]. Another study conducted in the Puerto Sixteen villages in Dominican Republic were selected for Plata region to evaluate E. coli levels in household water inclusion in the study. Each selected village sent school- sources found that in unimproved water sources, 47% aged children to a private school within a pre-identified were of high to very high risk according to WHO water private school network. Each school draws its students quality guidelines and in improved sources, 48% were of from a unique geographic area (i.e., village) surrounding high to very high risk [11]. the school. Rural and urban villages across the country While multiple options for reducing diarrheal preva- were included (Fig. 1). The geographic distribution of lence by providing clean drinking water and using schools (North, South, East, Capital-1, Capital-2 where proper hand hygiene exist, inexpensive but potentially “Capital” indicates schools in and near Santo Domingo) highly effective, point-of-use solutions remain an under- is shown in Table 1. Donor funding was available utilized option. One such option is the Sawyer® through a non-profit organization to provide point-of- Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 3 of 14 Fig. 1 Location of 16 selected villages in the Dominican Republic. Stars denote the approximate location of villages included in this study. Villages that were close in proximity were offset on the map to provide distinct representations for each village use water filtration systems (Sawyer® PointONE™ , add- diarrhea in the region [9] in children under 5 (15% or itional details on filter construction described elsewhere more vs. 15% or less). Villages initially assigned to the [13]) to in-network, sponsored households with a control group subsequently were assigned to the school school-aged child that attended the private school. The filter or home filter treatment group. All treatment total number of sponsored households eligible for inclu- groups received filters in the home by the end of the sion at the start of the study in August 2018 was 675, study. with breakdown by village shown in Table 1. Survey and data collection Randomization procedures We administered a short demographic, health, and In order to evaluate the efficacy of filter installations in economic baseline questionnaire to all households upon both schools and homes, a modified factorial design was installation of point-of-use filter in the home [13]. Five used (Fig. 2). Each of the 16 villages was initially local data collectors were trained by the research team. assigned to one of the following treatment groups: Each data collector was assigned to a village based on control (no filter initially), home filter only, school filter pre-determined geographic areas (Table 1) and was the only, or simultaneous (school and home) filter installa- same person for pre-filter (control group 2–8 weeks tion. Assignment of villages to treatment groups (see prior to filter installation), baseline (filter installation) Table 1) followed a covariate adaptive randomization and follow-up surveys at respective households and strategy, whereby predefined covariates are balanced schools. Data collection took place from September across treatments using the method of minimization 2018 through April 2019. [18]. We performed the method of minimization includ- Filter installation included a brief training on use of ing the following covariates: (a) reported number of the filter by the data collector, mention of importance of sponsored children at each school as of August 2018 handwashing and a demonstration by the respondent of (high [51+]) vs. low [50 or fewer]), (b) geographic loca- proper filter use and backflushing. After installation of tion/watershed (five groups [19];), (c) whether the school the filter, data collectors attempted to collect 2-, 8-, and anticipated receiving a filter or not (some schools 16-week follow-up surveys. This study was approved and declined needing filters due to already having a self- monitored by Dordt University IRB and all surveys in- reported ‘safe drinking water solution’), and (d) govern- cluded a statement of consent for research use. The trial is ment public health statistics regarding endemic levels of registered with www.clinicaltrials.gov as NCT03972618. Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 4 of 14 Table 1 Household characteristics by village De-identified Total sponsored households Total households in Village Initial Geographic School village name eligible for inclusiona primary analysesb response treatment location received rate groupc filter?d A 26 25 96.2% Control North No B 30 21 70.0% Simultaneous Capital–1 Yes C 35 17 48.6% Control Capital–2 No D 32 5 15.6% Home Capital–1 No E 46 38 82.6% Control North No F 60 41 68.3% Home North No G 18 12 66.7% Home North No H 40 36 90.0% School Capital–2 Yes I 38 22 57.9% Control South Yes J 24 3 12.5% School East Yes K 72 4 5.6% Home East No L 50 16 32.0% Control East No M 47 32 68.1% Home Capital–2 Yes N 62 8 12.9% Control East Yes O 50 15 30.0% Simultaneous South Yes P 45 27 60.0% Home Capital–1 No Total 675 322 47.7% – – – a Number of sponsored households within the village as of August 2018 b Use in primary analyses means that at least one household survey pre-household filter installation and at least one household survey post-household filter installation are available, consent of the respondent was received, and at least 90% of the survey question responses were valid/not missing c Since all households and eligible schools ultimately received a filter, treatment groups listed in the table are “initial” groups with subsequent installation of school/home measurements and commensurately collected survey data d Nine schools declined receiving a filter at the school because they determined they were not in need of a filter. Note: No water samples were taken before this determination was made Eligibility for inclusion in analyses in this manuscript Survey variables required (a) at least one survey at or before the time of The primary outcome we considered was self- filter installation (pre-survey), (b) at least one survey reported, 2-week prevalence of diarrhea, including completed at least 2 weeks after filter installation (post- whether diarrhea had economic or educational conse- survey), (c) obtained consent for data to be used for quences, causing hospitalization and/or missed school research purposes, and (d) having provided answers to at or work. The household survey administration in- least 90% of the survey questions. Using these criteria, cluded questions that determine (a) the size of the the overall eligible data response rate was 47.7% (322/ household (number of adults and children living in the 675 households; see Table 1) with variation in response household), and then for each household member, rates by village (minimum = 5.6%; maximum = 96.2%) whether they have in the past 2 weeks, (b) had diar- and region (generally lower response rates in the low- rhea, and if so, whether that diarrhea (c) caused lands (East) region). missed days of work (adults) or missed days of school Across the 322 households in the primary analysis, (children), and (d) required hospitalization. Additional there were 1075 household-level survey administra- information used in analysis was (1) season (fall tions. These household surveys consisted of 72 con- [October, November], winter [December, January, trol group baseline survey responses, 322 baseline February], or spring [March, April]), (2) days since survey responses upon filter installation, and 681 filter installation, (3) region (East, West, South, Central–1, follow-up survey responses (mean number of follow- Central–2), and (4) water source (city, purchased, well or up survey responses per household was 2.11). Among other (e.g., river, catchment)). the 681 follow-up survey responses, 107 of the responses indicated not using the filter and thus were ignored in pri- Water sampling methods mary analyses. An intent to treat sensitivity analysis in- Between 2 and 7 untreated drinking water source cluded these households. samples were collected at each of the 16 locations Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 5 of 14 Fig. 2 Depiction of the assignment of schools and households to treatment groups. Depending upon whether the school in the village accepted or declined a filter, villages were initially assigned to either four or two treatment groups, respectively. In order to ensure that all households and schools that accepted a filter ultimately received one, secondary randomization of control group households was performed as depicted in the figure (56 total samples). Information on location, date of the potential waterborne pathogens [21] (Acinetobacter, sample acquisition, and whether the sample came from a Aeromonas, Burkholderia, Campylobacter, Enterobacter, home or from the school can be found in Additional file Escherichia/Shigella, Francisella, Helicobacter, Klebsiella, 1. Water samples were tested for the presence of dissolved Legionella, Leptosipra, Mycobacterium, Pseudomonas, heavy metals, heavy metals associated with particulate Salmonella, Staphylococcus, Tsukamurella, Vibrio, and matter, and potential bacterial pathogens using Sawyer® Yersinia) were assessed through 16S rRNA amplicon PointONE™ hollow fiber membrane filters to capture sequencing following the Schloss Lab MiSeq SOP [22, 23] particulates and bacteria and using a metal chelating foam for sequencing and initial data processing to produce 97% to capture dissolved heavy metals. ICP-OES was used to operational taxonomic units and assigned taxonomies measure dissolved heavy metals that have been identified using the Silva Release 132 alignment and database [24]. by the WHO as potential health hazards (Arsenic, Barium, Downstream analyses of 16S amplicon sequencing data Cadmium, Chromium, Copper, Lead, Nickel, and were performed using the R packages Phyloseq [25] and Selenium) [20]. A combination of spectrophotometry, vegan [26]. Details of water sampling, testing, analysis, and SEM-EDS, and powder X-Ray diffraction (PXRD) were sequencing quality control methods are available in used to determine the concentration of particulates Additional file 2. Sequencing data associated with this (mg/L) and whether those particulates contained study have been deposited in the Short Read Archive under heavy metals of interest. Lastly, the presence of 18 PRJNA670359 (https://www.ncbi.nlm.nih.gov/bioproject/ bacterial genera identified by the WHO as containing PRJNA670359). Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 6 of 14 Statistical analysis of households with a filter at their children’s school was Generalized linear mixed-effects models with a logistic somewhat higher at follow-up (31% vs. 18%). link function were used to predict diarrhea prevalence by filter status. Random effects were used for repeated measures, with additional fixed effects covariate adjust- Two-week diarrhea prevalence by home filter status ments for season, water source, and household size. Before filter installation, 25.6% of households reported at Generalized linear mixed effects models with random least one member of the household experiencing diar- effects for repeated measures were also used to compare rhea within the previous 2 weeks (Table 3; Fig. 3). Preva- water sampling data between pre- and post-filter lence of diarrhea dropped substantially to 9.8% after installation measurements. All statistical analyses were filter installation, a difference which remained statisti- completed using R version 3.6.2 [27] and used a signifi- cally significant even after adjusting for other variables cance level of 0.05. (adjusted odds ratio (aOR) = 0.29 (95% CI 0.16, 0.51), p < 0.001). Adjusted odds ratios were statistically signifi- Results cant for adult-only diarrhea and for children under 4 Characteristics of the villages years (0.14 and 0.09, respectively). While the aOR was Characteristics of the villages are shown in Table 2. still small (0.49) it was not statistically significant (p = Purchased water usage fell from 36 to 10% from baseline 0.07) for school-aged children after adjusting for other to follow-up, likely reflecting reliance on the filter to variables (Table 3). We also completed an intent to treat provide safe and clean water. Modest, though statistically analysis including the 107 households reporting not significant, changes in region were observed from base- using the filters (total n = 681) which demonstrated line to follow-up (e.g., Capital–1 accounted for 21% of higher post-filter installation diarrhea prevalence overall follow-up data, but only 12% of baseline data due to (10.3%), as well as in adults (4.8%) and children under proportionally more follow-up surveys conducted in this the age of 5 (7.1%, with slightly lower prevalence in region as compared to others). As expected, the percent school aged children (5.7%; Additional file 3). Table 2 Village characteristics Variable Before home filter installation After home filter installation p value for difference1 (n = 394) (n = 574) Mean (SD) or %(x/n) Mean (SD) or %(x/n) Household size 4.60 (1.72) 4.41 (1.33) 0.049* Region North 38.8% (153/394) 35.0% (201/574) 0.004** East 8.6% (34/394) 5.9% (34/574) South 15.0% (59/394) 15.5% (89/574) Capital–1 12.4% (49/394) 21.4% (123/574) Capital–2 25.1% (99/394) 22.1% (127/574) School filter status Before/declined 82.0% (323/394) 68.6% (414/594) < 0.001*** After 18.0% (71/394) 31.4% (180/594) Season Fall 67.3% (265/394) 19.2% (110/574) < 0.001*** Winter 32.7% (129/394) 57.3% (329/574) Spring 0.0% (0/394) 23.5% (135/574) Water source City 49.4% (195/394) 79.6% (457/574) < 0.001*** Purchased 36.0% (142/394) 10.3% (59/574) Well 10.4% (41/394) 8.5% (49/574) Other 4.1% (16/394) 1.6% (9/574) *p < 0.05; **p < 0.01, ***p < 0.00 1 Logistic regression model with random effects term accounting for repeated measures Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 7 of 14 Table 3 Prevalence of self-reported diarrhea by home filter status Aggregation Before filter in homea% (x/n) After filter in homea% (x/n) OR (95% CI)b aOR (95% CI)c Any member of the household 25.6% (101/394) 9.8% (56/574) 0.24 (0.16,0.37)*** 0.29 (0.16,0.51)*** Adults only 17.3% 4.4% (25/564) 0.04(0.01,0.1)*** 0.14 (0.04,0.44)*** (68/392) School aged children only 12.2% 6.0% 0.20(0.10,0.44)*** 0.49 (0.23,1.05) (46/378) (33/546) Children less than 5 11.8% 4.3% 0.15 0.09(0.011,0.73)* (12/102) (6/139) (0.15,0.15)*** *p < 0.05; **p < 0.01, ***p < 0.001 aCounts are each measurement of each household bOdds ratio accounting for repeated measures cAdjusted odds ratio accounting for repeated measures and adjusted for household size, season (fall, winter, spring), water source, school filter status, and region Prevalence of diarrhea was significantly lower than or after adjusting for covariates, there was no evidence baseline throughout this study up to 200 days post-filter of statistical interaction between school filter status and installation (adjusted p < 0.001 for all three of = 7–50 home filter status for any household member, adults, days vs. baseline; 50–100 days vs. baseline and 100–200 school-aged children, or children less than 5 (adjusted p days vs. baseline; Fig. 4). While diarrhea prevalence values of 0.63, 0.70, 0.76, and 0.99, respectively). Further- continued to decline over time (11.8% 7–50 days; 10.0% more, there was no evidence of a statistically significant 50–100 days; 6.1% 100–200 days), these differences were decline in 2-week diarrhea after installation of school fil- not statistically significant after accounting for repeated ters, for any age group (Table 4) before or after adjusting measures, seasonal, and regional differences (7–50 days for covariates, including home filter status. vs. 50–100 days (adjusted p = 0.45); 7–50 days vs. 100– 200 days (adjusted p = 0.38)). Economic and educational consequences of diarrhea Prevalence of diarrhea with economic or educational Two-week diarrhea prevalence by school filter status consequences decreased significantly across all age We also examined whether 2-week prevalence of diar- groups, and for each consequence (i.e., missed work, rhea was impacted by the installation of water filters in missed school, hospitalizations) across all age groups (all schools, irrespective of water filters in the home. Before p < 0.05 before adjusting for covariates) after filters were Fig. 3 Prevalence of 2-week diarrhea before and after filter installation by age. Unadjusted SE bars are shown. Statistical significance between groups (***p < 0.001, **p < 0.01, *p < 0.05; NS p > 0.05) represents statistical significance from repeated measures models adjusted for household size, season (fall, winter, spring), water source, school filter status, and region Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 8 of 14 Fig. 4 Diarrhea prevalence by days since installation. Unadjusted SE bars are shown. Statistical significance between groups (***p < 0.001, **p < 0.01, *p < 0.05; NS p > 0.05) represents statistical significance from repeated measures models adjusted for household size, season (fall, winter, spring), water source, school filter status, and region installed in homes (Table 5). Before filters were installed Quality of drinking water sources in homes, prevalence of diarrhea with economic or edu- In order to assess the quality of drinking water sources, cational consequences was 9.64% among any member of 56 unfiltered water samples were collected across the 16 the household, compared to 1.57% after filter installation villages (30 home water sources, 25 school water (p < 0.001). For children, missed school due to diarrhea sources, 1 unknown) throughout the course of the study. decreased from 4.23 to 0.55% (p < 0.001). After adjusting Thirty-five samples were collected at the time of filter for covariates (Table 5), prevalence of diarrhea with any installation and 21 after filter installation. Characteristics economic or educational consequence for any member of each water sample and outcomes for detection of of the household still showed a statistically significant dissolved heavy metals, particulates and bacteria are decrease after filter installation (aOR 0.07, p < 0.05), provided in Additional file 1. while other decreases were no longer statistically signifi- We analyzed bacterial communities of drinking water cant. There was no evidence of an impact of the school sources through 16S rRNA amplicon sequencing. Up to filter or interaction between school and home filters on 5 gallons (~ 19 L) of water from each drinking source diarrhea with economic or educational consequences (p was filtered through two, 0.1 micron hollow fiber > 0.05 in all cases; detailed results not shown). membrane filters linked in a tandem pair. The first filter Table 4 Prevalence of self-reported diarrhea in household by school filter status Aggregation Before filter in schoola% (x/n) After filter in schoola% (x/n) OR (95% CI)b aOR (95% CI)c Any member of the household 15.9% (114/717) 17.1% (43/251) 1.01 (0.62,1.65) 1.29 (0.69,2.40) Adults only 9.62% (68/707) 10% (25/249) 0.87 (0.40,1.88) 1.58 (0.34, 7.39) School aged children only 7.99% (55/688) 10.2% (24/236) 1.09 (0.51,2.33) 1.17 (0.55,2.52) Children less than 5 7.18% (13/181) 8.33% (5/60) 0.34 (0.018,6.33) 0.47 (0.09,2.54) *p < 0.05, **p < 0.01, ***p < 0.001 a Counts are each measurement of each household b Odds ratio accounting for repeated measures c Adjusted odds ratio accounting for repeated measures and adjusted for household filter presence (y/n), household size, season (fall, winter, spring), water source, and region Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 9 of 14 Table 5 Prevalence of diarrhea with economic or educational consequence by home filter status Aggregation Before filter in homea After filter in homea Odds ratiob Adjusted odds ratioc Any member of the household with any 9.64% (38/394) 1.57% (9/574) 0.01(0.00,0.05)*** 0.07 (0.0014,0.56)* economic or educational consequence Adult hospitalized due to diarrhea 3.59% (14/390) 0.888% (5/563) 0.02 (0.00,0.18)*** 0.18 (0.0027,11.6) Adult missed work due to diarrhea 5.38% (21/390) 0.888% (5/563) 0.01 (0.00,0.09)*** 0.02 (0.00,1.57) Child missed school due to diarrhea 4.23% (16/378) 0.549% (3/546) 0.01 (0.00,0.09)*** 0.051 (0.0016,1.59) Child less than 5 hospitalized due to diarrhea 7.84% (8/102) 2.88% (4/139) 0.01 (0.00,0.32)* 0.01 (0.00,12.0) *p < 0.05, **p < 0.01, ***p < 0.001 a Counts are each measurement of each household b Odds ratio accounting for repeated measures c Adjusted odds ratio accounting for repeated measures and adjusted for household size, season (fall, winter, spring), water source, school filter status, and region in the pair (designated “Filter A”) filtered the source p = 0.002). Similar patterns were observed for Filter B water, capturing cells and particulate matter; the second communities (Filter B: df = 32, F = 1.9172, R2 = 0.76354, filter in the pair (designated “Filter B”) captured cells p = 0.001; betadisper, F = 27.287, p = 0.005). There were and particulate matter (if any) of the filtrate of Filter A. distinct bacterial communities associated with samples This design allowed the analysis of bacterial communi- from different types of water sources. For example, each ties of both the source water and the filtered water. Two of the kits in Fig. 5c represents a different water source types of controls were also analyzed. First, on each day type and is representative of the positions in the ordin- of processing, backflush controls included (1) water ation of other kits of the same water source type sampled backflushed through an unused filter, (2) a full volume from different geographic locations. Each of the water type of water used for backflushing all filters was pelleted, bacterial communities are distinct from the other water and (3) a small volume of water used for backflushing all type bacterial communities in pairwise comparisons of all filters was added to DNA extraction tubes. Second, each samples, with all but one comparison having significant sequencing library plate included a positive control of a differences both between (PERMANOVA) and within mock community of 8 bacteria (ZymoBIOMICS Micro- (betadisper) group distances (Table 6). bial Community Standard) and a negative control con- We used the list of bacterial taxa identified in each sisting of ultrapure water (see Additional file 2 for water source to determine the presence or absence of details of quality control processes applied to the bacter- genera containing known waterborne pathogens as ial community data). Evaluation of the pairwise distances defined by the WHO [21]. A genus was designated as between bacterial communities of each filter in the study present if it was found in A filter replicates for a kit but (Fig. 5a) showed that the communities derived from not found in B filters or process controls. At least one unfiltered source water (Filter A) formed a group that is genus from the list of 18 waterborne bacterial pathogens largely non-overlapping with the communities derived was detected in 50 of 51 water sources passing quality from filtered source water (Filter B). The difference control metrics. The single water source without indica- between the two groups was statistically significant tion of potential pathogens was from a city tap water (permutational multivariate analysis of variance [PERM sample (Kit 233). Additionally, we tested for the ANOVA] [28] ; df = 1, F = 9.3361, R2 = 0.04755, p = presence/absence of heavy metals in water sources. We 0.001); however, within group dispersion may also con- detected at least 1 of 8 dissolved heavy metals in 38 out tribute to between group significance (betadisper, F = of 56 water samples (67.9%). Table 7 shows detected 4.1054, p = 0.045). Biological replicates of each water bacterial pathogens and dissolved heavy metals stratified source grouped tightly in most cases for the unfiltered by whether the sample was obtained at the time of filter water (Filter A), whereas the replicates for filtered water installation or during a follow-up visit. No significant (Filter B) from the same kit did not cluster as tightly differences were observed in the prevalence of detected (Fig. 5b). Replicate Filter B samples are generally co- bacteria or heavy metals in the source water between in- localized with Filter B samples from other kits and with stallation and follow-up across all measured bacteria and backflush negative controls (Fig. 5b). Consistent with the heavy metals measured. Similarly, no seasonal or linear patterns in Fig. 5b, Filter A replicates from the same kit changes in prevalence were observed over time. Further- were less distant from each other than Filter A commu- more, mean particulate levels were not significantly dif- nities from other kits, explaining a large fraction of the ferent at the time of filter installation (0.246 vs. 0.166 observed variance (PERMANOVA; Filter A: df = 54, F = mg/L; p = 0.57), and also did not differ significantly over 7.3144, R2 = 0.82808, p = 0.001; betadisper, F = 2.8424, time during the study (linear trend p = 0.36; seasonal Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 10 of 14 Fig. 5 NMDS ordination (k = 3, stress = 0.141) of Bray-Curtis pairwise distances of bacterial communities from drinking water sources collected in the study (either home or school locations). Each water source was sampled with a kit consisting of three tandem filters (Filter A and Filter B), yielding three biological replicates of a water source per kit. Each point represents backflushed contents from Filter A, Filter B, laboratory backflush controls, mock community controls, or negative controls. Each kit is represented by up to 3 Filter A points and 3 Filter B points. Samples with fewer than 5000 sequencing reads were not included in the analysis. The same ordination is used for all three panels. a All backflush samples included in the study and associated controls. Shape and color delineate whether a sample is from Filter A, Filter B, or a control. b A subset of four kits selected to show the relationship between biological replicates of Filter A and Filter B within and between kits. Shape delineates Filter A, Filter B, and control samples. Color delineates which kit (i.e. water source) each filter sample represents along with the type of control. c The same kits as shown in b with identical coloring showing kit membership. Shape delineates the type of water source effect p = 0.35). Detected particulates were found to metal content to determine the appropriateness of the contain Ba, Ce, Cl, Cr, Fe, P, Mn, S, Ti, and Zr. See point-of-use filter technology for long-term use by resi- Additional file 1 for details about all water quality dents. Two-week self-reported diarrhea prevalence in all results. age groups drops substantially after introduction of filters in homes, with unadjusted declines of between Discussion approximately 50 and 75%. Similarly, large reductions in The purpose of this study was to evaluate the impact of economic and educational consequences of diarrhea the introduction of point-of-use filters into different were also observed after introduction of filters in homes community settings on the health of school aged chil- across all age groups. No corresponding change in dren and their families. Drinking water sources were also prevalence was noted after introduction of filters into evaluated for bacterial, particulate and dissolved heavy schools, nor was any moderating effect found for the Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 11 of 14 Table 6 Statistical tests of bacterial communities by water source type Comparison groups Statistical test Degrees of freedom F statistic Rb p value City tap water/city water filled underground cistern PERMANOVAa 1 4.184 0.07317 0.001 betadisperb 1 19.474 – 0.001 City tap water/purchased container PERMANOVA 1 4.714 0.05699 0.001 betadisper 1 2.632 – 0.113 City tap water/pumped well PERMANOVA 1 5.603 0.08673 0.001 betadisper 1 7.365 – 0.010 City water filled underground cistern/purchased container PERMANOVA 1 4.335 0.10004 0.001 betadisper 1 48.137 – 0.001 City water filled underground cistern/pumped well PERMANOVA 1 2.750 0.12089 0.001 betadisper 1 6.022 – 0.026 Purchased container/pumped well PERMANOVA 1 5.421 0.10752 0.001 betadisper 1 29.216 – 0.001 a PERMANOVA (adonis) generates a pseudo F statistic that describes the among group differences relative to the within group distances b betadisper generates an F statistic that describes the difference in within group dispersion of samples relative to the centroid of the group school filter on the effectiveness of the home filter. Re- that were part of this study, though school aged chil- sults from water sample testing confirm that bacteria dren received health hygiene instruction as part of were prevalent in water samples, including genera that standard curricula across the study sites (personal are known to contain pathogens associated with water- communication, C. Nelson). However, despite these borne disease. Limited amounts of dissolved metals and positives, further field work is needed, especially metal-containing particulates were observed in the evaluation of filter effectiveness and utilization over drinking water sources. longer time periods. While previous research [13–15, 17] yielded mixed Interestingly, little impact of school filter installation results from the Sawyer® PointONE™ filter system in on diarrhea prevalence was observed. There are likely laboratory testing and field environments, our ran- numerous causes for this finding. First, and importantly, domized, controlled field trial provides additional evi- only school-aged children in the family receive water at dence of their effectiveness in reducing waterborne the school and, thus, any potential direct impact of disease out to at least 200 days post-installation in school filters would only be observed on school-aged real-world settings. The hollow fiber membrane tech- children. However, in our data, this direct impact was nology is designed to remove bacteria and particu- not observed. Second, follow-up conversations with lates, but not dissolved ions. Evaluation of unfiltered school and non-profit workers (personal communica- drinking water sources available to families confirmed tion, C. Nelson), suggest that, in part due to a relatively the presence of highly diverse bacterial communities short school day (4 h), children typically bring water that varied with source type. Only a single water from home or buy water (not necessarily clean and safe) source tested in this study was free of bacterial gen- from vendors near the school instead of getting water era that are known to contain waterborne pathogens, from school. Third, only seven of the 16 schools in the indicating that hollowfiber membrane filtration should study were willing to accept a filter for their school, with be an effective method for improving drinking water nine schools having determined that their water was of quality and health outcomes in these areas. None of sufficient quality that no filter was needed, despite no a the drinking water sources contained dissolved heavy priori water testing. metals above WHO action standards, suggesting that An important feature of our study evaluating the po- point-of-use filters were an appropriate solution for tential effectiveness of both school and home filters was the villages and schools targeted by this intervention. a sequential study design implementing the method of Furthermore, prevalence of reported diarrhea dropped minimization to balance covariates. As has been noted substantially, though was not eliminated completely, by others [18], this method is rarely implemented in after filter installation in line with expectations that practice but can be highly effective. In our case, the the filter eliminates bacterial and parasitic causes of method of minimization maximized statistical power by diarrhea from water while not impacting other diar- limiting the impact of potential confounding variables rheal causes (e.g., viral). Finally, it is worth noting over time and ensuring a powerful study design. How- that no WASH programs existed in the communities ever, even here, there are additional variables that could Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 12 of 14 Table 7 Summary of test results of source water (unfiltered) Type of water Bacterial Percent of samples Percent of samples p value comparing p value for p value for sample testing genus/heavy present at filter install present at follow-up install to follow-upb linear trendc seasonal metal effectd Bacteria detected Acinetobacter 53.8% (14/26) 35.3% (6/17) 0.25 0.39 0.56 Aeromonas 16.1% (5/31) 20% (4/20) 0.72 0.51 0.64 Burkholderia 6.45% (2/31) 5% (1/20) 0.83 0.31 0.54 Campylobacter 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Enterobacter 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Escherichia/ Shigella 45.5% (10/22) 50% (6/12) 0.80 0.90 0.90 Francisella 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Helicobacter 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Klebsiella 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Legionella 58.1% (18/31) 70% (14/20) 0.18 0.06 0.41 Leptospira 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Mycobacterium 30% (9/30) 57.9% (11/19) 0.06 0.19 0.08 Pseudomonas 85.2% (23/27) 94.4% (17/18) 0.35 0.62 0.40 Salmonella 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Staphylococcus 12.9% (4/31) 15.8% (3/19) 0.78 0.79 0.87 Tsukamurella 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Vibrio 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Yersinia 3.23% (1/31) 0% (0/20) 0.98 0.69 0.47 Dissolved heavy Arsenic 0% (0/34) 0% (0/22) 1.00 1.00 1.00 metal ions detected Barium 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Cadmium 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Chromium 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Coppera 64.7% (22/34) 68.2% (15/22) 0.79 0.38 0.34 Leada 8.82% (3/34) 4.55% (1/22) 0.55 0.38 0.63 Nickel 0% (0/34) 0% (0/22) 1.00 1.00 1.00 Seleniuma 5.88% (2/34) 0% (0/22) 0.96 0.16 0.21 a Detected levels were below WHO action guidelines in all samples b From Fisher’s exact test comparing baseline to follow-up prevalence c From test of linear trend over course of study d From test of seasonal differences over course of study have minimized the impact of covariates even further community buy-in and support (see previous para- (e.g., proportion of households using unfiltered water in graphs) may serve not only to improve actual interven- each community; proportion lacking sanitation services tion efficacy but also to improve statistical aspects of per region, etc.); these remain potentially confounding study design leading to more quantifiable and robust variables in our analysis. This, combined with the use of findings. a self-report diarrhea measurement, represent two areas This study also combined public health surveys with for future work and limitations of this study. evaluation of both bacterial and chemical components of Another limitation of our study worth noting is the drinking water quality using the same hollowfiber mem- widely varying response rates by village. While the im- brane filtration technology as was used for the interven- pact on our findings is minimal, due in part to the study tion in homes and schools. The tandem filtration design design, additional statistical power and reduced impacts of test kits provided a direct demonstration of removal of potential confounding would be realized through of potentially harmful bacteria and particulate matter consistently higher response rates across all villages. that could be associated with toxic metals. Dissolved Some variation and lower response rates are expected heavy metals were assessed from the filtered water when working in remote villages in countries and areas sources to determine if potential toxins might persist in with limited infrastructure; however, having additional the drinking water source. Efforts to evaluate drinking Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 13 of 14 water sources should take into account the biological Additional file 3. Prevalence of self-reported diarrhea by home filter sta- and chemical components of the water to ensure that tus: Intent to Treat Analysis. proper intervention technologies are used. Additional file 4. Metadata file for 16S rRNA amplicon data analysis. The It is important to note that this study produced pro- tab “DR_MiMarks-Metadata” contains all metadata associated with files of the bacterial communities in each water sample amplicon sequence data deposited in the SRA. The tab “Key” contains anexplanation of all column headers. through sequencing of 16S rRNA amplicons. The use of this sequencing-based evaluation of the presence/ab- Abbreviations sence and relative abundance of bacterial genera is not WHO: World Health Organization; EPA: Environmental Protection Agency traditionally applied in the evaluation of drinking water Acknowledgements quality, although this is beginning to change [29–31]. The authors thank Darrel Larson, Chad Nelson, Gary Higgins, and numerous More routinely, culture-based methods are used to others in the Child Hope Network for their efforts in administering data assess viable loads of fecal indicator bacteria as a proxy collection and their vision for the project. for the presence of pathogens, or qPCR-based molecular Authors’ contributions methods are used to detect the presence of specific path- NT, KVDG, RU, RDW, SAB, AH, and AAB co-led the development of the ogens. When interpreting results from the sequencing research questions and study design. DM, EB, OC, CGZ, and NT led data approach employed in this study, it is important to monitoring, data cleaning, and statistical analyses. AAB, SAB, MJP, JWP, BPK,ADS, RDW, FSM, TMB, CEAC, LE, BF, GKG, MAH, JAL, SML, TRL, JMP, ES, DJS, recognize that the method does not determine whether JES, MJS, MS, and DRW designed laboratory protocols and water sampling bacteria detected in the sample are viable or not (this is kits, participated in RNA extraction, metals and particulate testing, final dataset also true of qPCR approaches). The method also does creation, and reporting. NT, KVDG, RU, RDW, SAB, BPK, and AAB wrote andrevised the manuscript. All authors read and approved the final manuscript. not identify specific pathogens (this is also true of routine culture-based approaches). However, it (1) Funding allows for a single assay that characterizes the entire Portions of the authors’ time and materials were supported by a grant from Sawyer Products, Inc. Representatives of the funding agency were consulted bacterial community, including all known potential during study design and aided in training for use of point-of-use filters in pathogenic groups of bacteria, and (2) it provides an op- the field. Neither the funding agency nor its representatives contributed to portunity to study bacterial communities across different data collection, data analysis, or interpretation of data. drinking water sources from around the world as these Availability of data and materials approaches become routine. Future studies should The survey datasets used and/or analyzed during the current study are include the use of both culture-based and sequence- available from the corresponding author on reasonable request. The 16S amplicon sequencing datasets generated and/or analyzed during based approaches to better characterize the relationship the current study are available in the Short Read Archive (SRA), https://www. between the two data types. ncbi.nlm.nih.gov/bioproject/PRJNA670359. Ethics approval and consent to participate Conclusions Consent for participation in research was obtained from all participants. This Our controlled study provides compelling evidence of study was conducted under the approval of the Dordt University Institutional Review Board. the efficacy of the PointONE™ filter in homes in a field setting, providing substantial reductions in bacterially Consent for publication caused diarrhea for the entire length of our study (200 Not applicable. days post-filter installation). Water quality sampling Competing interests provided strong complementary evidence to support Portions of the authors’ time (NT, KDVG, RDW, SAB, FSM, JWP, MJP, and AAB) self-reported diarrhea information and supported the were supported by a grant from Sawyer Products, Inc. choice of this filtration technology that removes only mi- Author details crobial and particulate contaminants. While we observed 1Department of Mathematics and Statistics, Dordt University, 700 7th St. NE, no significant impact of school filters, in other contexts Sioux Center, IA 51250, USA. 2Department of Sociology, Dordt University, 700 7th St. NE, Sioux Center, IA 51250, USA. 3Department of Mathematical where water is being consumed more regularly by school Sciences, Montana State University, P.O. Box 172400, Bozeman, MT 59717, children, additional impact may be observed. Further USA. 4Biology Department, Hope College, 35 E. 12th St, Holland, MI 49423, 5 work is needed to evaluate the post-200-day utilization USA. Chemistry Department, Hope College, 35 E. 12th St, Holland, MI 49423, USA. 6Department of Biology, Dordt University, 700 7th St. NE, Sioux Center, of filters and long-term filter efficacy in a field setting. IA 51250, USA. 7Geological and Environmental Sciences Department, Hope College, 35 E. 12th St, Holland, MI 49423, USA. Supplementary Information Received: 21 October 2020 Accepted: 17 December 2020 The online version contains supplementary material available at https://doi. org/10.1186/s41182-020-00291-y. References Additional file 1. List of water quality kits reported in this study, 1. Troeger C, Blacker BF, Khalil IA, Rao PC, Cao S, Zimsen SR, et al. Estimates of associated metadata for each kit, and water quality results. the global, regional, and national morbidity, mortality, and aetiologies of Additional file 2. Supplemental Methods for water quality testing. diarrhoea in 195 countries: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Infect Dis. 2018;18:1211–28. Tintle et al. Tropical Medicine and Health (2021) 49:1 Page 14 of 14 2. Pruss-Ustun A, Bartram J, Clasen T, Colford JMJ, Cumming O, Curtis V, et al. 25. McMurdie PJ, Holmes S. phyloseq: An R Package for Reproducible Burden of disease from inadequate water, sanitation and hygiene in low- Interactive Analysis and Graphics of Microbiome Census Data. PLoS ONE. and middle-income settings: a retrospective analysis of data from 145 2013;8(4):e61217. https://doi.org/10.1371/journal.pone.0061217. countries. Trop Med Int Health TM IH. 2014;19:894–905. 26. Jari Oksanen FGB, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, 3. World Health Organization. Diarrhoeal disease. 2017. https://www.who.int/ et al. vegan: Community Ecology Package version 2.5-4. 2019. https://CRAN. news-room/fact-sheets/detail/diarrhoeal-disease. Accessed 16 Jul 2020. R-project.org/package=vegan. 4. Annette P, David K, Lorna F, Jamie B. Estimating the burden of disease from 27. R Core Team. R: A Language and Environment for Statistical Computing. water, sanitation, and hygiene at a global level. Environ Health Perspect. Vienna, Austria: R Foundation for Statistical Computing; 2020. https://www. 2002;110:537–42. R-project.org. 5. Thomas JC, Weber DJ. Epidemiologic methods for the study of infectious 28. Anderson MJ. Permutational Multivariate Analysis of Variance (PERMANOVA). disease. 1st ed. New York: Oxford University Press; 2001. In: Wiley StatsRef: Statistics Reference Online. American Cancer Society; 6. Sack RB. Bacterial and Parasitic Agents of Acute Diarrhea. In: Bellanti JA, 2017. p. 1–15. https://doi.org/10.1002/9781118445112.stat07841. editor. Acute diarrhea: its nutritional consequences in children. New York: 29. Vierheilig J, Savio D, Ley RE, Mach RL, Farnleitner AH, Reischer GH. Potential Vevey/Raven Press; 1983. p. 53–65. applications of next generation DNA sequencing of 16S rRNA gene 7. The World Bank. World Bank Country and Lending Groups. https:// amplicons in microbial water quality monitoring. Water Sci Technol. 2015; datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank- 72:1962–72. country-and-lending-groups. Accessed 16 Jul 2020. 30. Santos QMB los, Schroeder JL, Sevillano-Rivera MC, Sungthong R, Ijaz UZ, 8. UNICEF. UNICEF Data. UNICEF DATA - Child Statistics. 2020. https://data. Sloan WT, et al. Emerging investigators series: microbial communities in full- unicef.org/. Accessed 16 Jul 2020. scale drinking water distribution systems – a meta-analysis. Environ Sci 9. Ministerio de Salud Pública. Encuesta Demográfica y de Salud República Water Res Technol. 2016;2:631–44. Dominicana 2013. Santo Domingo: Centro de Estudios Sociales y 31. Roguet A, Esen ÖC, Eren AM, Newton RJ, McLellan SL. FORENSIC: an Online Demográficos (CESDEM) y ICF International; 2014. https://dhsprogram.com/ Platform for Fecal Source Identification. mSystems. 2020;5. https://doi.org/ pubs/pdf/FR292/FR292.pdf. 10.1128/mSystems.00869-19. 10. McLennan JD. To boil or not: drinking water for children in a periurban barrio. Soc Sci Med. 2000;51:1211–20. Publisher’s Note 11. Baum R, Kayser G, Stauber C, Sobsey M. Assessing the microbial quality of Springer Nature remains neutral with regard to jurisdictional claims in improved drinking water sources: results from the Dominican Republic. Am published maps and institutional affiliations. J Trop Med Hyg. 2014;90:121–3. 12. Hydreion LLC. Microbiological testing of the Sawyer 7/6B Filter. Report No. S05-03; 2005. 13. Tintle N, Heynen A, Van De Griend K, Ulrich R, Ojo M, Boven E, et al. Evaluating the efficacy of point-of-use water filtration units in Fiji. Trop Med Health. 2019;47:48. 14. Kohlitz J, Hasan T, Khatri K, Sokota A, Iddings S, Bera U, et al. Assessing reported use and microbiological performance of a point-of-use household water filter in rural Fiji. J Water Sanit Hyg Dev. 2013;3:207–15. 15. Murray A, Goeb M, Stewart B, Hopper C, Peck J, Meub C, et al. Fouling in hollow fiber membrane microfilters used for household water treatment. J Water Sanit Hyg Dev. 2015;5:220–8. 16. Lindquist ED, Norman WR, Soerens T. A review of: Fouling in hollow fiber membrane microfilters used for household water treatment (2015) Murray, A., Goeb, M., Stewart, B., Hopper, C., Peck, J., Meub, C., Asatekin, A. & Lantagne, D. J. WASHDEV 5 (2), 220–228 doi:10.2166/washdev.2015.206. J Water Sanit Hyg Dev. 2015;5:229–31. 17. Murray AL, Stewart B, Hopper C, Tobin E, Rivera J, Mut-Tracy H, et al. Laboratory efficacy and field effectiveness of hollow fiber membrane microfilters used for household water treatment in Honduras. J Water Sanit Hyg Dev. 2017;7:74–84. 18. Scott NW, McPherson GC, Ramsay CR, Campbell MK. The method of minimization for allocation to clinical trials: a review. Control Clin Trials. 2002;23:662–74. 19. USGS. USGS EROS Archive - Digital Elevation - Shuttle Radar Topography Mission (SRTM) 1 Arc-Second Global. doi:10.5066/F7PR7TFT. 20. Chapter 12: Chemical fact sheets. In: Guidelines for drinking-water quality. Fourth edition incorporating the first addendum. Geneva: World Health Organization; 2017. p. 307–442. https://www.who.int/water_ sanitation_health/publications/gdwq4-with-add1-chap11.pdf?ua=1. Accessed 1 Oct 2020. 21. Chapter 11: Microbial fact sheets. In: Guidelines for drinking-water quality. Fourth edition incorporating the first addendum. Geneva: World Health Organization; 2017. p. 231–306. https://www.who.int/water_sanitation_health/ publications/gdwq4-with-add1-chap11.pdf?ua=1. Accessed 1 Oct 2020. 22. Kozich JJ, Westcott SL, Baxter NT, Highlander SK, Schloss PD. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol. 2013;79:5112–20. 23. Patrick D. Schloss. MiSeq SOP. MiSeq SOP. https://mothur.org/wiki/miseq_ sop/. Accessed 13 Oct 2020. 24. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web- based tools. Nucleic Acids Res. 2013;41:D590–6.