Water System Publications
water storage vessels and in situ chlorination in a Bolivian community: a simple
method to improve drinking water quality
Quick R, Venczel L, Gonzalez O, Mintz E, Highsmith A, Espada A, Damiani E, Bean
N, De Hannover R, Tauxe R
Epidemiologic investigations of the Latin America cholera epidemic have repeatedly
implicated untreated drinking water and water touched by hands during storage
as important vehicles for disease transmission. To prevent such transmission,
we provided a new narrow-mouthed, plastic, water storage vessel and 5% calcium
hypochlorite solution for home disinfection of stored water to a Bolivian Aymara
Indian community at risk for cholera. We evaluated acceptance of this intervention
and its effect on water quality. Each of 42 families in the study obtained water
from a household well; fecal coliform bacteria were found in water from 39 (93%)
of 42 wells and 33 (79%) of 42 usual water storage vessels. One group of families
received the special vessels and chlorine (group A), a second received only the
special vessels (group B), and a third served as a control group (group C). Water
samples collected every three weeks from group A special vessels had lower geometric
mean fecal coliform colony counts (P < 0.0001) and lower geometric mean Escherichia
coil colony counts (P < 0.0001) than water from group B or C vessels. Adequate
levels of free chlorine persisted in these vessels for at least 5 hr. The special
vessels and chlorine solution were well accepted and continued to be used for
at least six months. Use of the vessel and chlorine solution produced drinking
water from nonpotable sources that met World Health Organization standards for
cholera epidemic that began in Peru in January 1991 and swept through Latin America,
causing more than 1.000,000 reported cases and 10.000 deaths, highlighted serious
deficiencies in water quality and sanitation in this region. Investigations of
cholera outbreaks in Latin America have implicated the consumption of fecally
contaminated surface and municipal water sources as a major risk factor for disease
transmission.1-5 In one investigation, cholera transmission was specifically
associated with domestically contaminated water stored in open, wide-mouthed containers.
The cost of providing
potable water to each household and sewage treatment to each community in Latin
America has been estimated at $200 billion, and these improvements would take
at least 12 years to implement (de Macedo CG, Pan American Health Organization,
unpublished data). This definitive approach to cholera prevention is not currently
feasible. There is an urgent need for preventive measures that are effective,
inexpensive, and easily disseminated and implemented.
November 1992, we initiated a pilot project in El Alto, Bolivia, to determine
the feasibility of introducing a new intervention into a community at risk for
cholera: calcium hypochlorite solution (a form of chlorine) for disinfection of
drinking water stored in the home, and specially designed, plastic, narrow-mouthed,
lidded, 20-liter water storage vessels to prevent recontamination of treated drinking
water during storage.
El Alto, Bolivia,
is a fast-growing city of approximately 400,000 persons who live on the altiplano
above La Paz. Most communities in El Alto lack such basic services as potable
water systems, sewage, disposal, and trash removal. In
the community of Aymara Indians in which we conducted this investigation, all
families obtain water from shallow wells (2-5 meters) that they dig in front of
their homes. Most households store drinking water in buckets. All families dispose
of human waste on the open ground or in a nearby river. This investigation was
conducted during the rainy season from November 1992 through March 1993.
MATERIALS AND METHODS
design. We systematically selected 42 households from a group of 55 community
volunteers. These households were randomized into three groups: group A (15 families)
received the 20-liter, narrow-mouthed, lidded water vessel (Figure 1, [Catalog
Number 420-812; Toico, Inc., Toledo, OH]), referred to as the special vessel)
plus the calcium hypochlorite solution, group B (15 families) received the special
vessel only. without the disinfection solution, and group C (12 families) served
as controls, receiving neither special vessel nor disinfectant.
conducted a baseline survey of demographic characteristics and water handling
and sanitary practices of all participating families. After baseline data were
collected, the special vessels were distributed to groups A and B, and the calcium
hypochlorite solution only to group A. Families in groups A and B received instructions
about proper use and cleaning of the special vessel. Group A also received education
about proper disinfection dose and storage for the chlorine solution. Group C
received no education. During the nine-week intervention phase, each family was
visited every three weeks for a total of three visits per family. At each visit,
the head of household or the person most involved in water handling was interviewed
regarding water handling and treatment practices. The field team replenished the
chlorine solution during each visit. At the conclusion of the study
a meeting was held with participants to solicit opinions about the intervention.
Special vessels were then distributed to all control families, and all group B
and control families received chlorine and education about its proper use. Three
months later, each family was reinterviewed about water handling and treatment
FIGURE 1. Specially-designed,
plastic, narrow-mouthed, lidded, 20-liter water storage vessel used in the El
Alto study (dimensions: 10'
X 10' X 15.5").
Calcium hypochlorite solution
(0.5%) was prepared for this study from concentrated (70% chlorine) high-test
hypochlorite (HTH) powder by trained personnel from the University of San Simon
in Cochabamba, Bolivia. The solution was packaged in 250-ml opaque containers
with a 2-ml screw cap. Participants were instructed to put four capfuls (8 ml)
of disinfectant in the 20-liter vessels each time they tilled them with water.
Instructions for disinfectant use were printed in Spanish and Aymara on the disinfectant
container, and each family was given a poster with the same instructions to put
on the wall.
of water quality. Baseline water samples were collected from each household
during the baseline survey. including one from the household well (source water)
and one from the container used for water storage in the home (original pre-intervention
vessels are referred to as usual vessels). Three rounds of sampling were conducted
during household visits in the intervention phase. In the first two rounds, serial
water samples were taken from the household water storage vessels (either the
special vessel in groups A and B or the usual vessel in group C) at intervals
of 0. 2, and 4 hr from the time the field team arrived at the household to take
the first water sample. In the third round, samples were taken from wells and
home storage vessels. During follow-up interviews conducted three months after
the intervention phase, water samples were collected only from the special vessels.
The water sample
collection method was designed to reflect the water handling practices of the
study population. The resident's usual implements were used for water sampling,
including their unsterile well water collection bucket and cup used to scoop water,
stored in the home. Water temperature was,measured with a digital probe (Hach
CO), pH with indicator papers (Mikro Insta-chek; Micro
Essential Laboratory, Brooklyn, NY), and turbidity with a Jackson turbidimeter
(Robens Institute of Health and Safety, University of Surrey, Guildford. United
Kingdom), a clear plastic tube that is filled with a water sample through which
a black ring at the base of the tube is visualized. Turbidity is expressed as
Jackson units, which measure the light path through a suspension that permits
the visualization of the black ring; clear water has a value of less than five
Jackson units. Free and combined chlorine levels were determined using the N,N-diethyl-phenylenediamine
(DPD) colorimetric method (Free and Total Chlorine Kit; Hach Co.. Loveland. CO).
Water for microbiologic testing was collected in sterile 500-ml polypropylene
bottles, which were placed in coolers with frozen ice packs, and transported to
the National Institute of Health Laboratories in La Pax for analysis within 24
hr of collection.
sample portions of 0.5, 5. and 50 ml were filtered through 0.45-µm porosity,
47-mm diameter cellulose fillers (Gelman Sciences. Ann Arbor. MI). Diluting and
rinsing were performed using 0.01 M phosphate-buffered saline (pH 7.5). Each filter
was transferred to 60-mm petri plates containing m-FC agar (Difco Laboratories,
Detroit, MI), a selective medium for fecal conforms. and incubated at 44.5°C
for 24 hr. Blue colonies were counted as fecal coliforms. All filters with any
visible fecal coliform colonies were transferred to petri plates containing nutrient
agar and 4-melhy-lumbelliferyl-beta-D-glucuronide. After 4 hr of incubation at
44.5°C. colonies were illuminated with a handheld ultraviolet light, and fluorescent
blue colonies were counted as Escherichia coli.
densities/100 ml were calculated from colony counts for sample volumes of 0.5.
5. and 50 ml. assuming, that the number of colonies per plate followed a Poisson
distribution. If the number of colonies on a single filter was in the countable
range (10-100). that count was used to estimate the density of fecal coliforms
or E. coli. If two or more filters had counts in this range, or if all
filters with colonies had counts < 10, the density was estimated by dividing
the total number of colonies on those filters by the total volume of sample filtered.
If one or more of the filters had counts < 10, and the others were too numerous
to count (TNTC), a maximum likelihood estimator7 was used to estimate
the colony counts of the sample. If all the counts were TNTC, the concentration
was estimated to be twice the upper limit of the countable range (100) of the
highest dilution (0.5 ml), or 200 colonies per 0.5 ml. If no colonies were present
on the plate, a lower limit of detection of less than or equal to 1 colony per
50 ml sample was assumed and the count was recorded as less than or equal to 2
colonies per 100 ml. For calculations of mean bacterial densities, samples whh
no detectable fecal coliform or E. coli colonies were given a value of
two colonies per 100 ml.
means were calculated for estimated bacterial densities of the water tested from
wells and vessels from groups A, B, and C for each sampling episode. Multiple
linear regression models were used to evaluate the differences between fecal coliform
and E. coli colony counts in well water samples. Generalized estimating
equations1 were used to determine the association between the three
geometric mean densities of fecal coliforms and E. coli colony counts in
(pre-intervcntion) geometric mean fecal coliform and Escherichia coli colony
counts in water samples from household wells and usual household water storage
vessels. El Alto. Bolivia. December. 1992*
coliform (colonies/100 ml)
coli (colonies/100 ml)
means were calculated from logs of bacterial colony counts and presented in the
table as amilogs. Pairwise comparisons between group A (water vessel plus disinfectant),
group B (water vessel only), and group C (control) showed no statistically significant
differences between any of the three groups for either water source.
survey. The mean age of the 42 survey respondents was 34.3 years (range 12-68);
33 (79%) were female. The 42 households selected included 213 members (mean family
size, five persons; range 3-9); 109 (51%) were male. The median age of persons
in this sample was 14 years (range 2 months to 80 years). Among 84 persons s 18
years of age, 22 (26%) had no formal education, 36 (43%) had less than six years
of schooling, and 26 (30%) had more than six years. In the four weeks preceding
the survey, 35 (18%) of 199 household members for whom data were available had
diarrhea, including 11 (32%) of 34 less than five years of age.
water source for all households was a shallow, unlined well dug on their property.
All 42 respondents stored water from the well in their homes for drinking. Well
water was obtained daily by 39 (93%) respondents; the other three (7%) obtained
water every other day. Twenty-two (52%) respondents acknowledged that someone
had put their hand into drinking water stored in the house.
(86%) respondents said they boiled their drinking water. Of these, 13 (36%) said
they always boiled their water, one (3%) almost always, and 22 (61%) sometimes.
Only one (2%) respondent had used chlorine as a water disinfectant; 33 (79%) had
no knowledge of, or experience with, chlorine. Thirty-two (76%) respondents regularly
filtered their water through a cloth to remove worms and other visible contaminants.
quality. The temperatures of well and container water samples ranged from
7°C to 23°C (mean 13°C). The water pH was uniform at 5.5 in all samples
throughout the study. Turbidity was less than five Jackson units in 69% of stored
water samples in the baseline water sampling visit, 79% of vessel samples in the
first sampling round of the intervention phase. 90% in the second round, and 72%
in the final round. Less than 5% of all samples had turbidity counts > 20 Jackson
chlorine was detected in baseline stored water samples. In the first sampling
round of the intervention phase of the study, the mean free chlorine residual
(i.e.. concentration) measured in the special vessels of group A households was
1.6 mg/L (range 0-3.0; detectable, chlorine in 14 of 15 [93%]). In the second
round, the mean free chlorine residual was
1.2 mg/L (range 0-3.4; detectable chlorine in 12 of 14 [86%]), and in the final
round, the mean was 0.7 mg/L (range 0.1-1.5; detectable chlorine in 100%). Mean
free chlorine levels in the special vessel did not change over the 4-5-hr period
during which serial water samples were taken on two separate occasions.
samples from household wells and from usual water vessels showed no significant
differences in fecal coliform and E. coli colony counts between groups
A, B, and C (Table 1). Fecal coliforms were found in water from 39 (93%) of 42
wells and 33 (79%) of 42 usual home water vessels. Escherichia coli was
isolated from water samples from 37 (88%) of 42 wells and 29 (69%) of 42 usual
home water vessels. During the intervention phase, water from group A vessels
had substantially lower geometric mean fecal coliform colony counts (P < 0.0001)
and lower geometric mean E. coli colony counts (P < 0.0001) than water from
group B or C vessels in all three sampling rounds (Table 2). There were no significant
differences between fecal coliform or E. coli colony counts in water from
group B and C vessels. Water quality in 93% of the samples from group A vessels
met World Health Organization microbiological guidelines of less than or equal
to 1 E. coli colony per 100 ml (i.e., no detectable colonies).
During a focus group
meeting of study participants at the completion of the intervention phase of the
study, all persons expressed satisfaction with the special vessels and the chlorine.
Although several persons noted that treated water had a chlorine taste, all said
they grew accustomed to the taste and continued to use the chlorine. The only
change that was suggested for the vessel was to increase its size. All 30 households
that received the special vessels at the beginning of the intervention phase were
observed to be using them throughout the study. None of the special vessels disappeared
or broke during this time.
the follow-up evaluation three months after completion of the study, water was
sampled from the special vessel in 40 of the 42 original households. Water samples
had a mean free chlorine residual of 0.5 mg/L (range 0-3.5); 32 (73%) water samples
had detectable chlorine. Twenty-four (60%) of 40 samples had no detectable fecal
coliform colonies, and 31 (78%) had no detectable E. coli colonies. One
household was no longer using the special vessel; no water sample was taken from
this household. Another family, had moved out of the community and was lost to
2: Geometric mean
fecal coliform and Escherichia
coli colony counts
per 100 ml in water samples from household storage vessels for three sampling
rounds. El Alto, Bolivia, December 1992-March 1993*
coliform (colonies/100 ml)
means were calculated from logs of bacterial colony counts and presented in the
tables as anilogs. Results of pairwise statistical comparisons between group A
(water vessel plus disinfectant), group B (water vessel only, and group C (control)
are indicated; the referent is group A. None of the pairwise comparisons between
groups B and C were statistically sifinificant and these results are not included
in the table.
American cholera epidemic has served as a stark reminder of the inadequacy of
the sanitary infrastructure in Central and South America. The high mortality rate
from diarrhea among children less than five years old in Latin America (4.2 deaths
per thousand children per year 1981-19869) underscores the urgent need
for prevention efforts. The inexpensive, simple, and easily disseminated point-of-use
water treatment and storage system described here may offer a practical method
to protect drinking water supplies in many communities until resources are obtained
to provide universal piped, treated water.
intervention was acceptable to a population of Aymara Indians, who were able to
disinfect their drinking water on a sustained basis. Although some decrease in
chlorine use and increase in water contamination was noted at the follow-up evaluation
three months after the study was completed, all but one participant continued
to use the container, and most were still using the disinfectant solution. The
decrease in use of disinfectant was probably due in pan to inconvenience in obtaining
a supply of calcium hypochlorilc solution. After the study ended, the solution
was no longer delivered to participants' homes and had to be purchased at a local
narrow-mouthed vessel was developed for several reasons. First, investigations
of the cholera outbreak in Latin American2 and elsewhere (Swerdlow
DL. Centers for Disease Control and Prevention, unpublished data) have implicated
drinking water stored in household containers as a risk factor for infection.
Second, the hypothesis that water is contaminated during storage is supported
by other studies documenting increasing levels of fecal coliforms in drinking
water stored over time in wide-mouthed vessels.10 Finally, the use
of water vessels with narrow openings has previously been shown to improve water
quality in the home, probably, by hindering the introduction of hands into the
vessel.11,12 In addition to the narrow mouth and spigot, the special
vessel had a screw-on lid, giving the further benefits of hindering contamination
and decreasing the rate at which chlorine volatilized from treated drinking water.13,14
The importance of this feature was shown in a study-recently conducted in Huaricana,
Bolivia using a similar intervention; families using chlorine in their usual wide-mouthed
household containers did not achieve a reliable improvement in microbiologic
quality of their water
(Quick RE. unpublished data). The observation that chlorine levels in water in
the special vessel were constant over a 4-hr period indicates that chlorine did
not volatilize rapidly, nor was it consumed by the plastic itself. The finding
that disinfected water in the special vessel did not become recontaminated over
time indicated that this narrow-mouthed, lidded vessel enabled households to safely
store their water.
the special water vessel used in this study was well accepted and used for several
months by all but one family, it was by itself not sufficient to improve water
quality because the source water used to till the vessels was contaminated. Water
disinfection was a critical component. The use of a vessel with a standard volume
and a standard disinfectant solution made dosing simple. The narrow range of measured
chlorine levels and the consistency of results shows that home drinking water
disinfection can be reliably and safely performed even with a relatively uneducated
population. Although chlorine treatment gives water a noticeable taste, this did
not impede its use among study participants who consistently produced and consumed
bacteriologically acceptable drinking water. Another study, however, documented
that noticeable taste, fear of toxicity. and the belief that water treatment is
not necessary can be impediments to treating household water with chlorine.15
In addition to using a simple dosing strategy, adequate promotion and education
are essential to successful use of chlorine disinfectant.
the concentrated HTH powder used to make calcium hypochlorite solution for this
study is convenient to transport, it has two serious drawbacks. First, it is caustic
and therefore hazardous to transport and to reconstitute into solution if not
handled properly. Second, because it is not produced in most developing countries,
HTH powder must be purchased and imported, which increases the expense and the
risk of an interruption of disinfectant supply. Inexpensive alternatives to HTH
exist. For example, disinfectant solution for water can be produced in rural and
periurban communities by available appropriate technology (e.g.. Sanilec On-Site
Hypochlorite Generator; Eitech International Corp., Sugarland. TX, and MIOX System:
Los Alamos Technical Associates. Inc., Albuquerque, NM). which permits self-sufficiency
in disinfectant production12 (Loret P, unpublished data). The cost
of producing disinfectant has been estimated at $0.25 per family per year.12,16
of disinfection efficiency, fecal coliforms and E. coli, were chosen for
this study. Fecal coliforms include E. coli. which is presumed to be exclusively
fecal in origin, and some species of Klebsiella and Citrohacter,
which have nonfecal sources. Little is known about the impact of fecal coliforms
on drinking water quality when there is a mixture of fecal and nonfecal organisms.17
Total and fecal coliforms have been used as indicators of water treatment efficacy
for many years, but other organisms such as E. coli, coliphage, and enterococci
may also be appropriate.18,19 Thus, both fecal coliforms and E.
coli were included in this study to compare their prevalence in natural waters
in Bolivia and their survival in disinfected water. Eschenchia coli may
be a more suitable indicator of water contamination and disinfection effectiveness
for this study site because its presence in water samples and inactivation kinetics
are similar to fecal coliforms and it is a more specific indicator of human fecal
study had several limitations. The participants were from a self-selected group
of community residents, who may have been more motivated to comply with study
requirements than the general population. Although the findings of this study
cannot be readily generalized, they demonstrate what a motivated community can
accomplish. Attrition was low, in contrast to a study of household chlorination
in Brazil in which nine (36%) of 25 participants dropped out.15 The
problem of attrition would be best addressed through the application of social
marketing techniques and improvement of disinfectant distribution to make its
use more convenient (e.g., home delivery). Study participants may have also exhibited
the Hawthorne effect; that is, their performance was improved simply by the process
of being observed.20 However, most participants continued to use chlorine
after the end of the study. This study also had the potential for information
bias. Because the water vessels are a tangible presence, field workers could not
be blinded to intervention and control groups, raising the possibility of interviewers
biasing results toward a positive impact of the intervention. This risk was reduced
by the use of chlorine measurement to provide objective verification of compliance.
The virtual absence of bacteria in intervention group water eliminated bias in
conclusion, this study demonstrated that Aymara Indian families in a periurban
area of Bolivia were able to use successfully a simple, inexpensive system of
water treatment and storage to greatly improve their drinking water quality. The
special water vessel was well accepted and durable. A reliable, sustainable system
of production, distribution, and marketing of disinfectant is critical to achieving
the goal of community self-sufficiency in safe water production. The efficacy
of this water treatment and storage system in preventing diarrheal diseases was
not tested in this pilot study because we first wanted to evaluate the acceptability
and effectiveness of the intervention in improving water quality; the
next phase, a study of the effectiveness of the intervention in preventing diarrhea,
is in progress. In acute emergency settings, people can boil their water to prevent
disease transmission. Universal, piped, treated water and sewage treatment represent
the long-term and definitive solution to waterborne disease transmission, but
the high cost of such an intervention will slow progress in providing these services.
Therefore, point-of-use disinfection with a safe water storage vessel may offer
promise as a sustainable medium-term intervention to improve water quality and
prevent disease transmission in many parts of the developing world.
We are grateful to Carmen Revollo, Victor Diaz, and Patricia Machaca for diligent
and reliable work in the field and the laboratory. For technical advice, we are
grateful to Drs. Mark Sobsey and Christine Moe. For invaluable assistance in surmounting
administrative and logistical hurdles, we thank Dr. Alvaro Munoz Reyes.
This work was supported in part by a Prevention Research and Development Grant
from the Centers for Disease Control and Prevention.
Rics AA. Vuaia DJ, Beingolea L. Palacios JM, Vasquez E. Wells JG. Baca NG. Swerdlow
DL. Pollack M. Bean NH. Seminario L, Tauxe RV, 1992. Cholera in Piura. Peru: a
modern urban epidemic. J Infect Dis 166: 1429-1433.
Swerdlow D, Mintz E. Rodriguez M. Tejada E. Ocampo C. Espejo L. Greene KD, Saldana
W. Seminario L, Tauxe RV. Wells JG. Bean NH. Ries AA. Pollack M. Vertiz B, Blake
PA. 1992. Transmission of epidemic cholera in Trujillo, Peru: lessons for a continent
at risk. Lancet 340: 28-32.
Weber JT. Mintz ED. Canizares R, Semiglia A. Gomez I, Sempertegui R. Davila A.
Greene KD, Puhr ND, Cumeron DN. Tenover FC, Barrett TJ, Bean NH, Ivey C. Tauxe
RV, Blake PA, 1994. Epidemic cholera in Ecuador: antimicrobial resistance and
transmission by water and food. Epidemiol Infect 112:1-11.
Mujica 0, Quick RE, Palacios AM. Beingolea L, Vargas R. Moreno D. Barret TJ. Bean
NH, Seminario L, Tauxe RV, 1994. Epidemic cholera in the Amazon: the role of produce
in disease risk and prevention. J Infect Dis 169: 1381-1384.
5. Cardenas V. Saad
C. Varona M. Linero M. 1993. Waterborne cholera in JRiohacha. Colombia. 1992.
Bull Pan Am Health Organ 27: 313-330.
Mates A, Shaffer M. 3989. Membrane filtration differentiation. of E. coli
from coliforms in the examination of water. J Appi Bacterial 67: 343-346.
7. Haas CN. Heller
B, 1988. Averaging of TNTC counts; Appi Environ Microbiol 54: 2069-2072.
8. Zeger SL, Liang
KY. 1986. Longitudinal data analysis for discrete and continuous outcomes. Biometrics
Marlines J, Phillips M. Feachem RGA. 1993. Diarrheal diseases. Jamison DT. Mosley
WH. Measham AR, Bobadilla JL. eds. Disease Control Priorities in Developing
Countries. New York: Oxford University Press. 91-116.
Han AM,Oo KN, Midorikawa Y, Shwe S, 1989. Contamination of .drinking water during
collection and storage. Trop Geogr Med 41: 138-140
Deb BC. Sircar BK. Sengupta PG, De SP, Mondal SK. Gupta DN Saha NC, Ghosh S, Mitra
U, Pal SC, 1986. Studies on interventions to prevent eltor cholera transmission
in urban slums. Bull World Health Organ 64: 127-131.
Mintz ED, Reiff FM. Tauxe RV, 1995. Safe water treatment and storage in the home:
a practical new strategy to prevent waterbome disease. JAMA 273: 948-953.
13. Pinfold JV, 1990.
Faecal contamination of water and fingertip-rinses as a method for evaluating
the effect of low-cost water supply and sanitation activities on faeco-oral disease
transmission. II. A hygiene intervention study in rural north-east Thailand. Epidemiol
Infect 105: 377-389.
Hammad ZH. Dirar HA, 1982. Microbiological examination of sebeel water. Appi
Environ Microbiol 43: 1238-1243.
Kirchhoff LV, McClelland KE, Pinho MDC, Araujo JG, De Sousa MA, Guerrant RL, 1985.
Feasibility and efficacy of in-home water chlorination in rural North-eastern
Brazil. J Hyg (Camb) 94: 173-180.
Wilt VM, Reiff FM, 1993. La Desinfeccion del Agua a Nivel Cascro
en Zonas Urbanas Margmates y Rurales. Serie Ambiental No. 13. Washington. DC:
Pan American Health Organization
Barrell RAE, Rowland MGM, 1979. The relationship between rainfall and well water
pollution in a West African (Gambian) village. J Hyg (Camb) 83: 143-150.
18. Cabelli VJ,
Dufour AP, McCabe LJ, Levine MA, 1982. Swimming associated gastroenteritis and
water quality. Am J Epidemiol 115: 606-616.
Moe CL, Sobsey MD, Samsa GP, Mesolo V, 1991. Bacterial indicators of risk of diarrheal
disease from drinking-water in the Phillipines. Bull World Health Organ
Adair J, 1984. The Hawthorne effect: a reconsideration of the methodological artifact.
J Appi Psychol 69: 334-345.
R, Venczel L, Gonzalez O, Mintz E, Highsmith A, Espada A, Damiani E, Bean N, De
Hannover R, Tauxe R. Narrow-mouthed water storage vessels and in situ chlorination
in a Bolivian community: a simple method to improve drinking water quality. American
Journal of Tropical Medicine and Hygiene 1996; 54: 511-516.