Critical Review pubs.acs.org/est
Risk of Gastrointestinal Illness Associated with the Consumption of Rainwater: A Systematic Review Jonathan Dean and Paul R. Hunter* The Norwich School of Medicine, University of East Anglia, Norwich, U.K. S Supporting Information *
ABSTRACT: The collection of rainwater for human consumption is a practice well established in many parts of the world. Much of the research to date regarding this inexpensive and sustainable source has focused on its microbiological or chemical quality and there have been no reviews of epidemiological evidence regarding actual health risks associated with rainwater consumption. Electronic bibliographic databases were searched for epidemiological studies that attempted to quantify the risk of gastrointestinal disease linked to the consumption of harvested rainwater. Online databases were searched from the oldest date up to January 2011. Both observational and experimental studies were included. In addition, reference sections of key articles were searched and authors of previous studies were contacted where appropriate. Studies were assessed for relevance independently and in duplicate. Searches returned a total of 764 articles, 13 met inclusion criteria, 5 of these were outbreak reports, and 9 were studies of other design types. Pooled subgroup analysis suggests that rainwater is associated with reduced risk of illness compared to unimproved supplies (relative risk 0.57 95% CI 0.42, 0.77). There was no significant difference in risk in the pooled analysis of studies that compared rainwater to improved water supplies (relative risk 0.82 95% CI 0.38, 1.73). However, there was heterogeneity with one study showing an excess risk of campylobacteriosis. Classification of outbreak reports determined that 4 reports were “strongly associated with rainwater” while 1 report was “probably associated with rainwater”. We conclude that the evidence suggests that rainwater is safer than water from unimproved water supplies. Where feasible rainwater harvesting should be encouraged as a step toward achieving millennium development targets.
■
concluded that “Around the world, consumers of collected and stored rainwater may be at considerable risk to a variety of infectious diseases.”2 However, this review was far from systematic and did not attempt to quantify risk or to rate the quality of the studies on which the author based his conclusions. In a quantitative microbial risk assessment (QMRA) of rainwater in Queensland, Australia the authors estimated risk of both Giardia and Salmonella infections to be high (44−250 and 85−520 infections per 10,000 persons per year, respectively).3 However, the authors then went onto point out that this estimated risk was considerably greater than the actual number of infections occurring in the affected communities. So at present, the risk to health from harvested rainwater is unclear and it would appear that QMRA overestimates risk. The aim of this article was to try and improve estiumates of the health risk associated with consumption of harvested rainwater through a systematic review and meta-analysis. The focus of the review
INTRODUCTION The practice of collecting rainwater for domestic use is historically well documented.1 The collection of rainwater for human consumption is practiced globally. However there is ongoing debate regarding its quality and associated health risks.1−3 Traditionally the technique of rainwater harvesting (RWH) as a means of providing potable water has taken place in less developed nations, where alternative sources may be unavailable or of poor quality.1 However rainwater harvesting is still used in geographically isolated communities even in developed nations.3 Such practices are well documented in Brazil, China, and India for example, but also in rural and semirural regions in Australia and New Zealand.3,4 With ever increasing demands placed on limited water resources throughout the world there is growing interest in this “alternative” water procurement technique and whether it may provide a safe and reliable source of water for domestic purposes.3,4 Given that rainwater harvesting is considered to be an improved water source, wider adoption of rainwater harvesting could, in certain communities, enhance progress toward achievement of the millennium development target on access to improved drinking water.5 In a review of the literature Lye © 2012 American Chemical Society
Received: Revised: Accepted: Published: 2501
September 26, 2011 February 2, 2012 February 2, 2012 February 2, 2012 dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
Environmental Science & Technology
Critical Review
improved or unimproved.7 Where data were available to compare rainwater risk with both unimproved and improved sources within the same study, the results were presented independently in each section. Studies were assessed for methodological quality in one of three ways. Randomized controlled trials were assessed using the Jadad scoring system.8 Observational studies were assessed using the Newcastle-Ottawa scale if they were of a case-control design,9 and all other designs were given a quality rating based on a modified QATSO score (Appendix 2 in the Supporting Information).10 The Jadad scoring system produces a score of zero to five based on randomization, blinding, and participant withdrawal characteristics. The Newcastle-Ottawa scale is an eight-item instrument that provides a star system that rates studies in three broad perspectives: selection of study group, comparability of groups, and details of exposure or outcome of interest. The QATSO score provides a final rating of “good”, “satisfactory”, or “poor” to the study in question. The Jadad scoring system has been extensively reviewed by experts and is generally accepted as a valid method of assessing study quality of experimental designs in the setting of reviews.11 The Newcastle-Ottawa scale has also been critically appraised as a quality assessment instrument and has been partially validated for this purpose.8,10 The authors were unable to obtain appropriate validated methods for assessing cohort or crosssectional studies and as such modified a published but nonvalidated tool for the purposes of this assessment. These assessment instruments are not used for selection purposes but instead as a general guide to study quality. Heterogeneity, or between-study variation that cannot be explained by chance, was explored using meta-analyses. Where heterogeneity was deemed to be limited, pooled relative risk values were calculated and subjected to both random and fixed effects models. Publication bias, arising when small studies in particular are not published due to a failure to obtain a significant result, can distort results and lead to an overestimate of effect. To test for this the data sets were subjected to Egger’s test and funnel plots were generated. In addition, efforts were made to search for and include unpublished data to reduce potential bias of this nature.
was on gastroenteritis and microbial risks. Although chemical contamination of harvested rainwater may also be a risk, this issue was not addressed in this review.
■
METHODS Criteria for Inclusion. Studies were considered for inclusion irrespective of date of publication or language. Observational, interventional, or descriptive epidemiological study types were all considered for inclusion. For inclusion into the study a proportion of subjects were required to have consumed water collected from the roofs of dwellings, outhouses, or the like. If collection of rainwater was from alternative surfacesground runoff, roads, or reservoirs for instancethese studies were not included, the additional influences exerted on water quality by these situations were considered too wide ranging and unpredictable. The health outcomes of interest in this review were gastrointestinal disease or diarrheal episodes. Studies reporting waterborne illness may provide details of pathogens isolated from patient specimens or may be evidenced by symptoms. Studies were included in this review if gastroenteritis or diarrheal illness was recorded as a health outcome; where pathogens known to cause gastrointestinal disease were isolated from specimens of symptomatic patients; or where diagnoses were presented based on predictable patterns of gastrointestinal illness, i.e., cholera. Systematic Search. The following online bibliographic databases were searched using key terms (roof, rain, water, risk, epidemiology, diarrhea, diarrhea, gastroenteritis, illness, or disease); CINAHL, EMBASE, and MEDLINE. Abstracts were extracted and reviewed independently by both authors for content. Full copies were obtained of articles that met or were considered likely to meet inclusion and exclusion criteria based on abstract information. Studies were assessed for relevance using predefined selection criteria (Appendix 1 in the Supporting Information). The “ancestry approach”, whereby reference sections of key articles are systematically searched for additional studies, was also used. Where articles were found that did not present findings relevant to this review but where it was considered possible that relevant raw data were collected, study authors were contacted. Key data and setting characteristics from relevant studies were collated in a spreadsheet and later exported to statistical software (StatsDirect) for analysis. Analysis of Outbreak Reports. Outbreak reports were reviewed as a subgroup, independent of analytical study types. A classification system presented in the communicable disease report was employed to define the strength of association between water exposure and illness.6 Key features of outbreak reports are used in this system to determine association, including pathogen identification, knowledge of water treatment or water quality at the time, and type of study design used. Outbreaks analyzed using this method are ranked into one of three categories: “strongly associated with water source”, “probably associated with water source”, or “possibly associated with water source”. Analysis of Other Studies. Studies relevant to this review were further subdivided into research that compared the health risk from rainwater with that of “improved” water sources and those that compared the health risk from rainwater with that of “unimproved” water sources. To determine source type, background and methodology sections were scrutinized and relevant information was used in conjunction with WHO classification criteria to decide whether the source was
■
RESULTS The systematic searching of online bibliographic databases returned 748 articles. Additional searches including reference sections, conference reports, and publication citations returned a further 16 articles. Initial screening of abstracts in duplicate identified 71 papers considered appropriate to the review. Full scripts were obtained and 13 studies met selection criteria and were incorporated into the review (Figure 1). A further 3 studies were followed up for data: 2 authors failed to respond and the data from the third was not relevant. All papers reviewed following initial abstract screening were available in the English language. The chosen studies were conducted in 8 different countries including both developing and developed nations. The majority were undertaken in Australasia and publication dates ranged from 1978 to 2007. Studies That Investigated Health Risks Associated with Rainwater Consumption (8 Studies). Key characteristics of the 8 studies included in the review are presented in Table 1. The studies are heterogeneous in terms of design, two of them are interventional and the rest are a variety of observational types. All assess health outcomes in terms of diarrhea or gastrointestinal symptoms in a sample that in part, 2502
dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
Environmental Science & Technology
Critical Review
Rainwater vs Unimproved Sources. Three cross-sectional studies compared risk associated with rainwater with that of unimproved sources. The randomized controlled trial by Garrett and colleagues examined health outcomes in two groups of children and determined that those consuming rainwater were at lower risk of diarrheal disease than those consuming other sources (relative risk 0.70 95% CI 0.52, 0.95).14 Two other studies had similar findings although the one did not reach statistical significance (Table 3).15,16 In terms of methodological quality, the trial by Garrett and colleagues scored poorly (2/5 Jadad scoring system) (Table 2). The studies by Few et al. and Marcynuk et al., although considered a “lower level of evidence” due to design type, were determined to be of “satisfactory” and “good” methodological quality, respectively (Table 2). Once pooled, the data (Figure 2) suggest that rainwater poses no additional risk compared with unimproved sources and was associated with lower risk of diarrheal disease (relative risk 0.57 95% CI 0.42, 0.77). The data were not assessed for publication bias due to the low number of studies included and no significant heterogeneity between studies was detected. Rainwater vs Improved Sources. Four studies were identified that compared risk from rainwater consumption with that of improved sources. Three studies in this subgroup suggested that rainwater consumption may have no additional health risks compared to improved sources.15,17,18 One study found that rainwater harvesting may hold additional benefits over the alternative source (OR 0.42 95% CI 0.22, 0.80).18 However it should be noted that the comparison source was chlorinated river water which was, at that time experiencing raised cyanobacterial cell counts, and in addition this study had a low methodology quality rating (Table 2). Eberhart-Phillips and colleagues found in their case-control study that rainwater harvesting was associated with a greater risk of gastrointestinal disease compared to alternative sources (OR 3.11 95% CI 1.30, 7.41), although this finding was based on a small number of rainwater collectors.19 This study and those by Few et al. and Heyworth et al. all performed satisfactory to good when assessed for methodological quality (Table 2).15,17 Pooled subgroup data (Figure 3) suggest there is no additional health risk or benefit from the consumption of rainwater compared with improved water sources (0.82 95% CI 0.38, 1.73). Egger’s test for publication bias was applied and suggested no publication bias, and tests for heterogeneity suggested there was limited difference among studies. Outbreak Reports Potentially Associated with Rainwater Collection (5 Reports). Study characteristics of the reports can be seen in Table 4. The outbreaks included in the
Figure 1. Selection process for studies examining rainwater consumption and health risk.
at least consume rainwater. Two studies examined health effects in children of a defined age, while the rest were unspecific in the age of participant recruited. Kelly-Hope and colleagues examined enteric diseases and their associated risk factors in Vietnam using an ecological type design.12 Their findings, although not appropriate for inclusion in meta-analyses, did note there was no significant correlation between rainwater consumption and dysentery (−0.09, p-value 0.580) or cholera (−0.19, p-value 0.244). Rodrigo and colleagues recently published their findings of a double-blinded randomized controlled trial.13 This study, considered to be of high methodological quality (Table 2) compared risk for those who had filtering systems in their rainwater harvesting unit with that of those who had standard rainwater harvesting units. The findings, which could not be included in meta-analyses for lack of an alternative water source control group (i.e., all study participants used rainwater), determined that filtering systems do not reduce risk of illness (HCG incidence rate ratio (filter vs no filter) 1.05 (95% CI 0.82, 1.33). Table 1. Study Characteristics study
type
location
year
age
Eberhart-Phillips, J., et al. Few, R., et al. Garrett, V., et al. Heyworth, J. S., et al. Kelly-Hope, L. A., et al. Marcynukp et al. Rodrigo, S., et al. Saadi, O. E. L., et al.
case control cross sectional randomized controlled trial cohort ecological cross sectional randomized controlled trial case control
New Zealand Vietnam Kenya Australia Vietnam Brazil Australia Australia
1995 2010 2008 1999 2007 2009 2010 1992
all ages all ages children children all ages all ages all ages all ages
2503
comparison source improved improved and unimproved unimproved improved unimproved treated rainwater vs untreated rainwater improved
dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
Environmental Science & Technology
Critical Review
Table 2. Assessment of Methodological Quality study
type
Eberhart-Phillips, J., et al. Few, R., et al. Garrett, V., et al. Heyworth, J. S., et al. Marcynukp et al. Rodrigo, S., et al. Saadi, O. E. L., et al.
case control cross sectional randomized controlled trial cohort cross sectional randomized controlled trial case control
Jadad score
Newcastle-Ottawa scale
4/6 (67%) − satisfactory/good 2/5 4/6 (67%) − satisfactory/good 5/6 (86%) − good 5/5 selection comparability exposure *
Using the classification system (Table 5) it was determined that four out of the five outbreaks were “strongly associated with rainwater use”.20−23 In all of these incidents researchers identified the same pathogenic organisms in both case and water source, documented a water treatment failure or water quality failure of some description, and demonstrated an association using descriptive epidemiology. In another report 23 holiday resort staff suffered gastroenteritis type illness and stool samples were positive for Campylobacter spp.24 Rainwater tanks were found to contain fecal contamination but not the species identified in stools. In addition, on further investigation there did not appear to be any other water treatment or quality failure and it was concluded that this incident was “probably associated with rainwater use”. The mechanisms by which the rainwater tanks in these outbreaks became contaminated remains unclear in most cases. Various authors surmise that contamination may have come from debris, bird, or small mammal fecal matter on the rain collecting surfaces but no actual scientific evidence was found to support such hypotheses.20,21,24 In the report by Simmons et al. a water blaster from a nearby marina was thought to be aerosolizing Legionella which contaminated neighboring rainwater collecting surfaces.22 Legionella species (LP1) found in patient and rainwater tank specimens were also identified on the water blaster itself suggesting a possible link. In one incident, a faulty wastewater tank was suspected of overflowing and causing contamination of surrounding tanks, one of which collected rainwater, although such a link was not demonstrated with any certainty.23
Table 3. Individual Study Results RWH vs unimproved supply study
type
location
no.
Few, R., et al.
cross sectional
Vietnam
2519
Garrett, V., et al.
randomized controlled trial cross sectional
Kenya
960
Brazil
3747
Marcynuk, P., et al.
effect estimates (95% confidence interval) RR 0.45 (CI 0.18, 1.10)a RR 0.70 (CI 0.52, 0.95)a OR 0.47 (CI 0.33, 0.67)a
RWH vs improved supply study EberhartPhillips, J., et al. Few, R., et al. Heyworth, J. S., et al. Saadi, O. E. L., et al. a
modified QATSO score
selection **** comparability ** exposure *
type case− control
location New Zealand
no. 1242
effect estimates (95% confidence interval) OR 3.11 (1.30, 7.41)a
cross sectional cohort
Vietnam
2519
RR 0.36 (0.10, 1.26)a
Australia
1016
case− control
Australia
320
OR 0.84 (CI 0.63, 1.13)a OR 0.42 (CI 0.22, 0.8)a
Figures adjusted using multivariate analysis.
■
DISCUSSION To our knowledge this is the first systematic review that examines evidence regarding the consumption of rainwater and health. We found evidence that when compared with unimproved sources rainwater consumption is associated with fewer episodes of diarrheal disease. Although the pooled effect size did not find excess risk in people consuming rainwater compared to improved water sources, there was evidence of heterogeneity. In particular the study by Eberhart-Phillips found a significant excess risk of campylobacteriosis.19 If the findings in this study are repeated elsewhere then there may well be a risk in some contexts. It is recognized however that there are limitations to these findings which are based on few studies of variable quality and design with diverse populations, and as such further research is needed for conclusions to be drawn with confidence. It should be noted that one source of heterogeneity was the outcome measures used in the different studies. Although most studies used self-reported diarrhea, case definitions will have differed between the studies, while is was not possible in this analysis to identify how this may have impacted on the outcome measure, it is notable that the one study finding an excess risk with rainwater was of diagnosed
Figure 2. Risk of gastrointestinal illness through consumption of rainwater vs unimproved sources.
review are heterogeneous in a variety of respects: region, date, pathogens implicated, and numbers of people affect. The one commonality among them is that they were all potentially associated with rainwater consumption. All but one occurred in a public building or public gathering site and it is of note that the majority (four) of reports document outbreaks that occurred in developed nations. 2504
dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
Environmental Science & Technology
Critical Review
Figure 3. Risk of gastrointestinal illness through consumption of rainwater vs improved sources.
Table 4. Outbreak Report Characteristics authors
location
year
cases
organism
Koplan, J. P., et al. Franklin, L., et al. Merritt, A., et al. Simmons, G., et al. Kuroki, T., et al.
Trinidad Australia Australia New Zealand Japan
1978 2007 1999 2006 1994
48 27 23 4 461
Salmonella arechevalata Salmonella typhimurium 9 Campylobacter Legionella pneumophila Cryptosporidium
extra info suspected bird feces on roof
marina water blaster possible overflow from surrounding tanks
Table 5. Outbreak Report Summary and Classificationa authors Koplan, J. P., et al. Franklin, L. J., et al. Merritt, A., et al. Simmons, G., et al. Kuroki, T., et al.
location
cases
summary
Trinidad
48
Australia
23
Australia
23
New Zealand Japan
4
rural camp water supply; descriptive epidemiology showed association with water from rainwater tank, extensive bird feces on collection surface; pathogen found in stool of campers and tap from rainwater tank rural camp water supply; pathogen found in stool and 2 rainwater tanks; descriptive epidemiology showed association with activities and illness; water untreated resort rainwater supply; fecal contamination found in water tanks, but not case pathogen; association with rainwater tank demonstrated by case control study household rainwater supply; pathogen found in two of four affected households; all cases with pathogen serology; association with water blaster suggested by case control study public building water tanks; pathogen found in tank and some stool samples; descriptive epidemiology showed association; water quality failure
461
association strong [A + B + strong [A + B + probable [C only] strong [A + B + strong [A + B +
D] D]
D] D]
a
Outbreak report classification system: (A) pathogen identified in case and water; (B) water quality failure/water treatment problem; (C) analytical study demonstrates association (case-control or cohort); (D) association by descriptive epidemiology.
Campylobacter infections.19 It is possible that one of the reasons for a positive association was the more objective case definition. The extensive body of research examining microbial content of rainwater collecting apparatus consistently reports that levels of pathogenic organisms are often present in concentrations that could pose a threat to health,1−3 a feature that would appear incongruent with our findings. It is possible, as suggested by Heyworth and colleagues, that such findings may reflect the development of acquired immunity by participants.17 Chronic, low-level exposure to pathogenic organisms may provide a protective effect in the longer term and intermittent raised bacterial levels pose less of a threat to health when they occur. Conversely those consuming water from a generally “sterile” water source may be more susceptible to illness should there be sporadic contamination by pathogens.
There is an increasing body of evidence to support this suggestion, albeit not derived from populations consuming harvested rainwater.25,26 It is also important to note that the evidence presented in this review does not suggest that rainwater is without risk but rather is of the same or lower risk than alternative sources available to the population studied. This is an important distinction and as such the findings of the review should only be interpreted with the nature of these alternatives in mind. The sources deemed unimproved, based on WHO criteria, were all from studies conducted in developing countries.7,27 These sources included river water, surface water, shallow wells, and the like, all of which can be considered to be at high risk of contamination. It is perhaps unsurprising then that rainwater in comparison could be a low risk source. In one study the 2505
dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
Environmental Science & Technology
■
rainwater collecting systems were rudimentary in the extreme and subject to little or no maintenance.14 In another study it was noted that rainwater remained reasonably “pure” on initial collection but became contaminated within the home.28 These findings suggested that while rainwater was shown to be lower risk than unimproved sources this risk could be reduced further still, a concept that needs further investigation. The studies that compared rainwater harvesting with improved supplies were, with one exception, conducted in Australasia. Although the term “improved” identifies sources deemed to be of a higher quality there may be some variation in this respect. The participants in these studies were generally located in semirural or rural locations and the main water supplies identified included piped water, boreholes, or treated river water. This is not to say that comparison of rainwater with such sources is inappropriate as they represent the various alternatives for that population at that time, however it should be considered that in another setting with alternate water sources, rainwater may not compare so favorably. Inferences on the need to treat rainwater prior to consumption cannot be made based on the evidence presented in this review. Few of the studies collected data on water treatment and a range of practices from no interference to sophisticated filtering and treatment regimes may have been employed. The evidence presented by Rodrigo and colleagues answers this question in part, with their trial finding that filter systems hold no additional benefits to health over non-filtered collection devices.13 These findings however may not be generalizable to all circumstances of all rainwater collection and this needs further investigation. In addition it must be stated that recent papers have cast doubt on the public health value of in-home water treatment in low income countries.29,30 The inclusion of outbreak reports in this review provides important additional information, especially concerning the failures that led up to the outbreak. Only a small number of outbreaks were identified given the global and use of rainwater collection. This low number most likely represents underreporting of outbreaks particularly in less developed parts of the world. With systems being used on a small or individual scale “outbreaks” may involve just a few people and not be detected against the background level of endemic illness.31 Based on the findings of this review there is currently no evidence to show that rainwater holds any additional health risks over alternative improved drinking water sources. There is indeed evidence to suggest that rainwater may hold additional health benefits when compared with unimproved sources. When interpreting these results it is important to consider that these findings are based on a small number of studies of variable quality and design and further research is needed to draw conclusions with greater confidence. Classification of outbreak reports in a review setting was a novel feature of this article and one that will allow objective consideration of such material and the raising of hypotheses to be tested analytically. Where feasible, rainwater harvesting should be encouraged as a step toward achieving millennium development targets on access to improved water supplies.
■
Critical Review
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare the following competing financial interest(s):PRH was until 2010 chair of science advisory board of Suez Environment.
■ ■
ACKNOWLEDGMENTS No funding was received for this review. P.R.H. was chair of the science advisory group of Suez environment until 2010. REFERENCES
(1) Gould, J. Is rainwater safe to drink ? A review of recent findings. Proceedings of the 9th International Rainwater Catchment Systems Conference, 1999. Available at http://www.eng.warwick.ac.uk/ircsa/ abs/9th/7_4.html. Accessed 13/03/2011. (2) Lye, D. J. Health risks associated with consumption of untreated water from household roof catchment systems. J. Am. Water Resour. Assoc. 2002, 38, 1301−1306. (3) Ahmed, W.; Vieritz, A.; Gardner, T.; Goonetilleke, A. Microbial risks from rainwater tanks in south east Queensland. Water 2009, 36 (8), 80−85. (4) Gould, J. Contributions Relating to Rainwater Harvesting. Prepared for Thematic Review IV.3: Assessment of Water Supply Options, 1999. Available at http://oldwww.wii.gov.in/eianew/eia/ dams%20and%20development/kbase/contrib/opt163.pdf. Accessed 04/07/2011. (5) The World Health Organisation. Facts and figures: Water, Sanitation and Hygiene Links to Health, 2011. Available from http:// www.who.int/water_sanitation_health/publications/factsfigures04/ en/. Accessed 04/07/2011. (6) Tillett, H. E.; Louvois, J.; Wall, P. G. Surveillance of outbreaks of waterborne infectious disease: Categorizing levels of evidence. Epidemiol. Infect. 1998, 120, 37−42. (7) World Health Organisation. Statistical Information System. Access to improved drinking-water sources and to improved sanitation. 2010. Available from http://www.who.int/whosis/ indicators/compendium/2008/2wst/en/. Accessed 04/07/2011. (8) Jadad, A. R.; Moore, R. A.; Carroll, D.; Jenkinson, C.; Reynolds, D. J.; Gavaghan, D. J.; McQuay, H. J. Assessing the quality of reports of randomised controlled trials: Is blinding necessary? Control Clin. Trials 1996, 17, 1−12. (9) Wells, G.; Shea, B.; O’Connell, D.; Robertson, J.; Peterson, J.; Welch, V. The Newcastle-Ottawa Scale (NOS) for assessing the quality of non-randomised studies in meta analysis. 2001. Ottawa Health Research Institute, Ottawa. Available from http://www.ohri.ca/ programs/clinical_epidemiology/oxford.asp. Accessed 04/07/2011. (10) Wong, W. C. W.; Cheung, C. S. K.; Hart, G. J. Development of a quality assessment tool for systematic reviews of observational studies (QATSO) of HIV prevalence in men having sex with men and associated risk behaviours. Emerg. Themes Epidemiol. 2008, 5, 23. (11) Moayyedi, P.; Hunt, R. Meta-analysis: Can we mix apples and oranges? Am. J. Gastroenterol. 2004, 99, 2297−3001. (12) Kelly-Hope, L. A.; Alonso, W. J.; Thiem, V. D.; Anh, D. D.; Canh do, G.; Lee, H.; Smith, D. L.; Miller, M. A. Geographical distribution and risk factors associated with enteric diseases in Vietnam. Am. J. Trop. Med. Hyg. 2007, 76, 706−712. (13) Rodrigo, S.; Sinclair, M.; Forbes, A.; Cuncliffe, D.; Leder, K. Drinking Rainwater: A Double-Blinded, Randomized Controlled Study of Water Treatment Filters and Gastroenteritis Incidence. Am. J. Public Health 2010, 101, 842−847. (14) Garrett, V.; Ogutu, P.; Mabonga, P.; Ombeki, S.; Mwaki., A.; Aluoch, G.; Phelan, M.; Quick, R. E. Diarrhoea prevention in a highrisk rural Kenyan population through point-of-use chlorination, safe water storage, sanitation, and rainwater harvesting. Epidemiol. Infect. 2008, 136, 1463−1471.
ASSOCIATED CONTENT
* Supporting Information S
Appendix 1 and Appendix 2. This material is available free of charge via the Internet at http://pubs.acs.org. 2506
dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
Environmental Science & Technology
Critical Review
(15) Few, R.; Lake, I.; Hunter, P. R.; Tran, P. G.; Thien, V. T. Seasonal hazards and health risks in lower-income countries: Field testing a multi-disciplinary approach. Environ. Health 2009, 8 (1), 16. (16) Marcynuk, P. Preliminary Summary: Prevalence of diarrhoea among cistern and non cistern users in Northeast Brazil and further risk factors and prevention strategies [thesis]; Pan American Health Organisation: Brazil, 2009. (17) Heyworth, J. S.; Glonek, G.; Maynard, E. J.; Baghurst, P. A.; Finlay-Jones, J. Consumption of untreated tank rainwater and gastroenteritis among young children in South Australia. Int. J. Epidemiol. 2006, 35, 1051−1058. (18) Saadi, O. E. L.; Esterman, A. J.; Cameron, S.; Roder, D. M. Murray River water, raised cyanobacterial cell counts, and gastrointestinal and dermatological symptoms. Med. J. Aust. 1995, 162, 122− 125. (19) Eberhart-Phillips, J.; Walker, N.; Garrett, N.; Bell, D.; Sinclair, D.; Rainger, W.; Bates, M. Campylobacteriosis in New Zealand: Results of a case-control study. J. Epidemiol. Comm. Health 1997, 51, 686−691. (20) Franklin, L. J.; Fielding, J. E.; Gregory, J; Gullan, L.; Lightfoot, D.; Poznanski, S. Y.; Vally, H. An outbreak of Salmonella Typhimurium 9 at a school camp linked to contamination of rainwater tanks. Epidemiol. Infect. 2008, 137, 434−440. (21) Koplan, J.; Deen, R.; Swanston, W.; Tota, B. Contaminated roof-collected rainwater as a possible cause of an outbreak of salmonellosis. J. Hyg. (London) 1978, 81, 303−309. (22) Simmons, G.; Jury, S.; Thornley, C.; Harte, D.; Mohiuddin, J.; Taylor, M. A Legionnaires’ disease outbreak: A water blaster and roofcollected rainwater systems. Water Res. 2008, 42, 1449−1458. (23) Kuroki, T.; Watanabe, Y.; Asai, Y.; Yamai, S.; Endo, T.; Uni, S.; Kimata, I.; Iseki, M. An outbreak of waterborne Cryptosporidiosis in Kanagawa, Japan. Kansenshogaku Zasshi 1996, 70, 132−140. (24) Merritt, A.; Miles, R.; Bates, J. An outbreak of Campylobacter enteritis on an island resort, north Queensland. Commun. Dis. Intelligence 1999, 23, 215−219. (25) Frost, F. J.; Roberts, M.; Kunde, T. R.; Craun, G; Tollestrup, K.; Harter, L.; Muller, T. How clean must our drinking water be: The importance of protective immunity. J. Infect. Dis. 2005, 191, 809−814. (26) Swift, L; Hunter, P. R. What do negative associations between potential risk factors and illness in analytical epidemiological studies really mean? Eur. J. Epidemiol. 2004, 19, 219−223. (27) International Monetary Fund, Data and Statistics. World economic outlook report. April 2010. Available from http://www.imf. org/external/index.htm. Accessed 04/07/2011. (28) Levy, K.; Nelson, K. L.; Hubbard, A.; Eisenberg, J. N. S. Following the Water: A Controlled Study of Drinking Water Storage in Northern Coastal Ecuador. Environ. Health Perspect. 2008, 116, 1533−1540. (29) Hunter, P. R. House-hold water treatment in developing countries comparing different intervention types using metaregression. Environ. Sci. Technol. 2009, 43, 8991−8997. (30) Schmidt, W.-P.; Cairncross, S. Household water treatment in poor populations: Is there enough evidence for scaling up now? Environ. Sci. Technol. 2009, 43, 986−992. (31) Hunter, P. R.; Syed, Q.; Naumova, E. N. Possible undetected outbreaks of cryptosporidiosis in areas of the North West of England supplied by an unfiltered surface water source. Commun. Dis. Public Health 2001, 4, 136−138.
2507
dx.doi.org/10.1021/es203351n | Environ. Sci. Technol. 2012, 46, 2501−2507
