BACTERIA COLONIZING POINT-OF-USE,
GRANULAR ACTIVATED CARBON FILTERS AND
THEIR RELATIONSHIP TO HUMAN HEALTH
BY
REBECCA L. CALDERON AND ERIC W. MOOD
DEPARTMENT OF EPIDEMIOLOGY AND PUBLIC HEALTH
YALE SCHOOL OF MEDICINE
NEW HAVEN, CONNECTICUT 06510
CR-811904-01-0
PROJECT OFFICER
ALFRED P. DUFOUR
HEALTH EFFECTS RESEARCH LABORATORY
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
ABSTRACT
An epidemiological study on the health effects associated
with bacteria found in granular activated carbon (GAC) point-
of-use filters was conducted by the Department of Epidemiology
and Public Health, Yale School of Medicine. The study
population consisted of Navy families that lived in off-base
Navy housing in Groton, Connecticut. Two types of filters were
used during the study. The first type was a filter that
attached at the end of the kitchen faucet. Participating
families used a filter with either a cartridge containing
granular activated carbon or a blank cartridge. The second
type of filter was a dual cartridge bypass filter that had its
own faucet at the kitchen sink. Bacterial data and health data
were collected during monthly visits to the subjects' homes.
All water samples were analyzed for heterotrophic bacteria
using two media. The carbon filters eluted significantly
higher densities of bacteria than unfiltered water (tap water
or water from the blank cartridge). Of the two media used for
heterotrophic bacterial analyses, R2A recovered significantly
more bacteria than did the standard plate count agar. Mean
Standard Plate counts from the faucet carbon filters and the
bypass filters were 689 and 1049 per ml respectively compared
to 198 and 53 per ml for blank filter effluents and the tap
water respectively. The R2A medium yielded roughly twice these
levels for each of the four categories of waters. Bacterial
isolates from a random sample of homes with carbon filters were
identified. The majority of the bacteria belonged to the
Flavobacterium and Pseudomonas genera. In particular, the
bypass bacteria were primarily P. stutzeri. All families were
monitored for grastrointestinal and dermatologic symptoms.
Despite the high bacterial densities there were no differences
in gastrointestinal illness or dermatologic illnesses between
the control group (blank faucet filter) and the two test groups
(carbon faucet filter and bypass carbon filter). An exposure
analysis estimated an average daily intake of 10° organisms
per day per person for carbon filter users.
This report was submitted in fulfillment of CR-811904-01-0
by Yale University under the sponsorship of the U.S. Environ-
mental Protection Agency. This report covers a period from
January 1985 to August 1986, and work was completed as of
February 1987.
ii
-------
contents
Abstract ii
Figures iv
Tables v
Acknowledgment vi
1. Introduction 1
2. Conclusions 4
3. Materials and Methods 5
4. Results and Discussion 11
Bibliography 18
Figures and Tables 20
Appendices 39
A. Questionnaires and Health Diary 39
B. Marketing Report 49
iii
-------
Figures
1 Schematic diagram of faucet filter as it is installed
(courtesy of P. Regunathan) 20
2 Schematic diagram of bypass filter as it is installed
(courtesy of P. Regunathan) 21
3 Heterotrophic bacterial densities as measured by
standard plate count agar and R2A agar by month for
tap water and bypass carbon filters 22
4 Heterotrophic bacterial densities in tap water and
carbon faucet filters as measured by standard plate
count agar and R2A agar by month 23
5 Heterotrophic bacterial densities in tap water and
blank faucet filters as measured by standard plate
count agar and R2A agar by month 24
6 Heterotrophic bacterial densities in tap water and
water temperature by month 25
7 Heterotrophic bacterial densities in carbon faucet
filters and water temperature by month 26
8 Heterotrophic bacterial densities in blank faucet
filters and water temperature by month 27
9 Heterotrophic bacterial densities in bypass carbon
filters and water temperature by month 28
IV
-------
Tables
Number Paae
1 summary of water sample types collected over a 17
month period 29
2 Geometric mean and standard deviations of heterotrophic
bacterial densities from installation water samples — 29
3 Geometric means and standard deviations of heterotrophic
bacterial densities discharged from two granular
activated carbon filters, blank filters and tap water as
measured by Standard Plate Count agar and R2A agar 30
4 Correlation coefficient of water temperature with
bacterial densities from bypass filters, carbon and
blank faucet filter and tap water using Standard Plate
Count and R2A agar 30
5 Types of bacteria isolated from tap and carbon faucet
filter water samples 31
6 Types of bacteria isolated from tap and bypass carbon
filter water samples 31
7 Person years of surveillance for each filter group in
the point-of-use filter study by sex 33
8 Person years of surveillance for each filter group in
the point-of-use filter study by age group 33
9 Symptom rate in the point-of-use filter study per person
year by filter group 34
10 Incident density of gastrointestinal illnesses in each
study group by age and sex 35
11 Percentages of reported uses of GAG filtered water by
filter type 36
12 Percentages of how water was collected by subjects from
a filter by filter type 36
13 Densities of heterotrophic bacteria discharged with
volume flow from bypass and faucet point-of-use granular
activated carbon filters 37
14 Consumption of filtered water by point-of-use filter
type 38
v
-------
ACKNOWLEDGMENTS
The project benefited from the help of several organizations
and individuals. We wish to acknowledge the following:
From U.S.E.P.A. for their expert advice:
Frank Bell, Office of Drinking Water
Alfred Dufour, HERL and project officer
From the U.S. Navy, Sub-base, New London, Ct. for their cooperation:
Commander Christopher Holmes -Head, Occupational and Preven-
tive Medicine
R.C. Clark, Housing Director
Residents of Navy Housing
From the filter manufacturers for their equipment and advice:
John Jiambalvo, Director Research, Associated Mills
Dr. P. Regunathan, Vice President, R & D, Everpure, Inc.
From the Water Quality Association for supplemental fundings and
expertise on selection of filters:
Douglas Oberhamer, Executive Director
Lucius Cole, Technical Director
VI
-------
SECTION 1
INTRODUCTION
There has been an increased awareness of the chemicals
contained in drinking water. Many of these chemicals affect
the aesthetic properties of water such as taste, color and
odor. Other chemicals such as trihalomethanes and other
volatile organics may be potentially harmful. It is estimated
that only one half of one percent of water produced by a
municipal water treatment facility is actually used for cooking
and drinking (Regunathan, et al.. 1983). Some water consumers
wish that this small fraction of potable water could meet more
stringent requirements for chemical contamination whether for
health effects or aesthetic effects. In some communities the
water quality can be more effectively controlled at the point
of use than at the central treatment plant. Many taste, color
and odor complaints are due to microbial regrowth, corrosion
products in the water main and by-products of chlorination.
These events occur after leaving the treatment plant. In many
households, the installation of point-of-use (POU) treatment
devices are used to provide specially treated water for
drinking and cooking purposes.
Many different kinds of POU devices are available for
consumer use. The most commonly used devices contain granular
activated carbon (GAC). Filters with GAC remove common tastes
and odors, some turbidity, chlorine, and many organic contami-
nants. Charcoal has been used for water purification for
centuries. The need for higher adsorptive capacities led to
the- development of activated carbon in the early 1900's.
Activation provides the carbon with a high degree of porosity
and an associated high surface area. The high surface area
renders activated carbon an excellent adsorbent for a wide
range of chemicals. This high surface area also provides an
excellent surface for biomass attachment and a favorable
environment for the growth of microscopic organisms (Rice and
Robson, 1982). Several papers have reported the colonization
and subsequent discharge of higher densities of bacteria in
laboratory studies on GAC filters (Geldreich et al., 1985,
Brewer and Carmichael, 1979, Bell, et al^., 1984). The effects
of such growth have been reported as both advantageous (Rice
and Robson, 1982) and disadvantageous (Wallis, Stagg and Melnick,
1974) . Of special concern is the health hazard that may be
created if pathogenic organisms adhere to and grow on carbon
surfaces. This concern was categorized succinctly by Dufour
(1985) into five questions:
-------
1. What are the possible health risk concerns?
2. What bacteria are associated with granular activated
carbon?
3. What are the origins of these health risk concerns?
4. Are there reasons to question the health risk concerns?
5. How can the assumed health risks be verified?
The types of health risks of concern are related to the
use of water from POU-GAC filters. The primary use of water
from POU-GAC filters is for drinking and culinary purposes. It
follows that the primary health concern would be gastrointestinal
infections associated with bacteria in water from GAC-POU
filters. A secondary use of water from POU filters would be for
handwashing (contact) and hence there is a secondary risk of
dermatologic infections associated with the bacteria discharged
from GAC-POU filters. To better understand the health concerns,
the second question: "What are the bacteria associated with
activated carbon filters?" must be asked.
The types of bacteria associated with granular activated
carbon are diverse and depend on three factors; 1) the types of
bacteria found in the influent water; 2) the ability of a
species of bacteria to attach to GAG; and 3) the ability of a
species, once attached, to multiply in the carbon media and
slough off in sufficient densities to be detected in the filter
effluent. One study by Geldreich and coworkers (1985) reported
that four out of seven tested species including Pseudomonas
aeruainosa readily colonized GAC filters in the laboratory.
Other studies report that many of the bacteria that colonize in
GAC filters are native to aquatic environments and normally are
not considered to be pathogenic. This leads to the third
question: "What are the origins of the health risk concerns?"
A majority of the concerns arise out of the types of bacteria
that cause nosocomial (hospital acquired) infections. Many
bacteria found in potable water have been found in increasing
numbers as causes of nosocomial infections (Karnad, e£ aJL.
1985). A secondary concern arises over the higher densities of
these bacteria eluted from the GAC filters. Geldreich and
coworkers reported densities of heterotrophic bacteria exceeded
1000 cfu per ml. An 8 ounce glass of water (237 ml) would
contain approximately 2 x 105 organism. Since little is known
about infective doses for the majority of heterotrophic bacteria,
the high densities suggest a dose may be sufficiently high
enough to cause disease. However, this poses the fourth question:
"Are there reasons to question the health concern?"
The few studies done on the effect of ingesting high
densities reported contrary data (Dufour, 1985). Two bacteria,
Klebsiella pneumoniae and Pseudomonas aeruainosa both of which
were shown to colonize GAC filters in laboratory studies
(Geldreich, et al., 1985) were used in human feeding experiments.
Doses of up to 108 organisms showed no illnesses in the human
-------
volunteers. Both of these organisms are common causes of nosocomial
infections but healthy volunteers did not develop gastrointestinal
infection when challenged.
While there is evidence to suggest that bacteria associated
with GAG filters do not cause disease, the overall question of
a health effect is still a concern. There are two major reasons
for this continued concern. The first is the diverse nature of
bacteria found in distribution systems and hence able to colonize
GAG filters. Little is known about these organisms and what if
any effect they would have on people who may be exposed to
higher densities of these bacteria. The second reason is that
these filters will be used in a wide variety of homes where the
spectrum of people exposed will range from infants to senior
citizens. Previous studies on gastrointestinal illnesses have
focused on families with young children (Cabelli, 1982; Monto
and Koopman, 1980) . It is believed this age group is most
susceptible to gastrointestinal infections. This leads to the
last and final question: "How can the assumed health risks be
identified?"
One way to answer the question would be to conduct an
epidemiological study on the use of POU-GAC filters and relate
it to illnesses in the user populations. The objective of this
study was to do just that, namely conduct an epidemiological
study which would evaluate the health effects, if any, of
bacteria from a distribution system that colonized POU-GAC
filters. In the years of 1984 to 1986, a study on the health
effects of higher bacterial densities in effluent waters from
POU-GAC filters was conducted by the Department of Epidemiology
and Public Health, Yale School of Medicine under the terms of
a cooperative contract with the Health Effects Research Labora-
tory, U.S. Environmental Protection Agency.
-------
SECTION 2
CONCLUSIONS
1. Point-of-use granular activated carbon filters (by-pass and
faucet) are colonized by heterotrophic bacteria from the
water supply distribution system. Once colonized, the
filters discharge significantly higher densities of hetero-
rophic bacteria than unfiltered tap water.
2. Exposure to these higher densities of heterotrophic bacteria
(over 1000 cfu per 1.0 ml) did not cause an increase in
acute symptomatology (gastrointestinal and dermatologic) as
compared with those people who were exposed only to
unfiltered water.
3. Gastrointestinal infections reported to a physician by study
participants were not associated with bacteria colonizing GAG
filters.
-------
SECTION 3
MATERIALS AND METHODS
A. MICROBIOLOGY
Sample Collection
Two water samples were collected from each house during the
initial visit by one of the field investigators. The first
sample, approximately 100 milliliters in volume, was collected
from the cold water faucet in the kitchen after the water had
run for about two minutes. After the faucet filter or bypass
filter had been installed, a second 100 milliliter sample of
the filtered water was collected. This second sample was
collected as soon as the effluent from the filter was clear of
all air or any minute carbon particles. All homes, i.e. both
test and control residences, were sampled at least once per
month. At that time, two water samples were collected. The
procedure in collecting these two samples was as follows:
1. the mouth of the discharge orifice for unfiltered water was
wiped with a prep swab saturated with 70% isopropyl alcohol
and allowed to dry thoroughly; the faucet was turned on
and the first 50 milliliters were discarded and the next
100 milliliters were collected in a sterile bottle which
contained sodium thiosulphate.
2. the discharge orifice for filtered water was treated the
same as above and a sample collected as outlined.
Bacterial Analysis
Two media were used on all water samples to measure the
total heterotrophic bacterial count. The first was a Standard
Plate Count Agar(Difco) (APHA, 1981) and the second was a media
called R2A (Difco) developed by the U.S. Environmental Protection
Agency (Reasoner, 1985) that is being evaluated currently as a
possible replacement for the SPC medium.
With the exception of samples collected on Sunday nights,
all samples were plated on the same day of collection. The
Sunday samples were done on Monday morning after being stored
at 4°C. Plates (R2A and SPC) were predryed at room temperature
to allow the plating of 1.0 milliliter onto the agar. All "A"
samples (unfiltered tap water) had 1.0 milliliter of water
examined on R2A and 1.0 milliliter on SPC plates. The "B"
samples (filtered water) were either diluted 10 fold or 100
-------
fold in sodium phosphate buffer (pH=7.2). From these final
dilutions a 1 ml and a 0.1 ml aliquot were plated in duplicate
on R2A and SPC media.
The SPC plates were incubated at 35°C for five days and
the R2A plates were incubated at 22°C for five days. Although
Standard Methods (1982) recommends a three day incubation period,
Reasoner and workers (1985) reported increased recoveries on
both SPC and R2A with extended incubation. Space and time
constraint in the laboratory allowed a five day incubation.
Identification of Isolates
A random sample of households with POU-GAC filters was
chosen to have their bacterial isolates identified. Representa-
tives of each colony with similar morphology were picked from
"A" and "B" samples and were streaked for isolation on a
modified R2A agar that contained an additional 1% dextrose.
This modified R2A medium was found to be superior to SPC and
plain R2A media in laboratory passage of isolates.
The isolated colonies were Gram stained and all Gram
negative organisms were inoculated into glucose oxidative
fermentation tubes. These tubes were incubated at 30°C for
21 days. This extended procedure was recommended for water
bacterial isolates (Spino, 1985). All oxidative positive and
nonreactive isolates were further identified using API, NFT
strips. All fermentative isolates were speciated using API 20E
identification strips. These identification strips were further
supplemented with motility and oxidase testing. Motility was
tested using MIO medium (Difco) and the isolates were incubated
at room temperature for 48 hours.
Flushing Experiments
• A series of flushing experiments were done to determine
the effect of flushing the filter on heterotrophic bacterial
counts in filter effluent. The faucet filter studies were
conducted in the laboratory on a faucet filter containing carbon
catridge attached to the laboratory sink. For one week prior
to the experiment the filter was run for five minutes twice a
day to seed the filter. The bypass filter used for the flushing
experiments was one of the filters installed in the subjects
home. The subject was asked not to use the filter in the previous
twelve hours prior to our visit. A series of 250 ml (faucet
filter) or 500 ml (bypass filter) sample bottles containing
sodium thiosulfate were filled subsequently with filter effluent.
Samples were then plated on R2A agar only and incubated at 22
for five days.
-------
B. FILTERS
Faucet Filter
The faucet filter used was a Pollenex WP-100 GAG water
filter. This filter is listed by the National Sanitation
Foundation (NSF) for Taste and Odor Reduction under standard
No. 42 which is for "Aesthetic Effects". The Pollenex filter
was attached at the end of the kitchen faucet. The faucet filter
households were divided into test and control groups. The test
group had a filter cartridge with carbon and the control group
had an empty cartridge. A schematic flow diagram of the unit is
shown in Figure 1. The carbon cartridge contained approximately
50g.
Bypass Filter
The second type of POU filter was an Everpure QC4-THM dual
cartridge bypass filter. This filter has a NSF listing under
Standard 42 and an additional listing for Turbidity Reduction,
Cyst Reduction and Total THM Reduction under Standard 53 which
is for" Health Effects". This filter was installed under the
sink with its own faucet in the upper right corner of the sink
(Figure 2). The amount of GAG in the first cartridge was
approximately 500g, while the second cartridge included about
lOOg of Powdered Activated Carbon.
Installation
The Polenex faucet filters were installed by research
personnel during the initial visit to the subjects' house. All
filters were painted with a number in a hidden place on the
filter. This number identified whether the cartridge was a
blank or a test cartridge. After the installation, the subjects
were given a small demonstration on proper use of the filter,
such as using the switch to get filtered or unfiltered water.
This information was also summarized on a paper given to them.
The Everpure bypass filters were installed by a licensed
plumber. After the plumber had completed the installation, the
use of the filter was demonstrated by a member of the research
team and an explanation was given on how water could be turned
off in the case of any water leaks.
Maintenance
The faucet filters were changed every three to four months.
At the time of the change, the entire unit was changed so that
the subject would not know whether they were getting a filter
with a carbon or a blank cartridge. In most cases filters were
changed from carbons to blanks and blanks to carbons. On
occasion filters would break or leak and filters would have to
be repaired before the next monthly visit.
-------
The bypass filters required very little maintenance. The
filter cartridges were changed when the flow became impeded or
subjects complained of taste problems. Approximately five sets
of cartridges were changed during the study. The cartridges
were changed by the researcher during the monthly visit.
C. EPIDEMIOLOGY AMD STUDY SITE
City of Groton Water Supply
The present source of Groton 's water supply is predominantly
derived from the Great Brook watershed and a portion of the
Billings Avery watershed. The mouth of Great Brook empties
into the Poquonamock River and its source is located in the
southerly portion of the Town of Ledyard. This watershed is
15.9 square miles in area and develops a surface water supply
which has been estimated to have a safe daily yield of 12.1
million gallons per day.
The City of Groton 's Filtration Plant, where the water is
treated, consists of three units; the original built in 1939,
has four, one-half million gallons per day filter beds; an
addition built in 1950 has four one million gallons per day
filter beds; and an addition built in 1961 having three, two
million gallons per day filter beds, which results in a total
of twelve million gallons per day nominal capacity rating of
the plant. In 1971, one of the two million gallon per day sand
filter beds was rebuilt utilizing a mixed filter media of
ilmenite and anthracite, which has doubled the capacity to four
million gallons and has increased the nominal rating of the
plant to fourteen million gallons per day. The current daily
usage averages 11.18 million gallons per day. In 1986, the City
of Groton, Department of Utilities reported an average free
chlorine residual of 1.1 ppm at the filtration plant and 0.9
ppm in the Navy housing area distribution sites. Their quarterly
heterotrophic bacteria plate counts in the plant effluent
averaged <1 per ml. There were four distribution sampling
locations in the study area. These locations averaged <1 cfu
per ml. (Report furnished by City of Groton, Department of
Utilities.)
Description of
The pool of subjects used in this study was taken from
families who resided in off base U.S. Navy housing in Groton,
Connecticut. This population was chosen because 1) they all
had the same water supply; 2) they had access to free medical
care and; 3) the majority of housing residents are families
with children under the age of ten.
From a Navy housing mailing list, with over 5000 addresses,
800 families were initially selected randomly to receive the
first questionnaire asking them to participate in the study.
8
-------
From the pool of returned questionnaires, people were selected
for the study if 1) they indicated they wished to participate
on the questionnaire and 2) they had a personal rotation date
that exceeded the length of the study. These people were contacted
by phone and an appointment was scheduled to install a filter
in their home. Six months after the start of the study, the
bypass filter was added to the study. An additional 700 question-
naires were sent out to solicit additional families. These
households were selected based on the criteria described above.
The first questionnaire was a general survey on how
residents of the house felt about the taste, odor and color of
their water. Additional questions were asked about hot water,
dishwashers and the use of any water treatment devices. The
final section asked questions on demographics, personal rotation
date and willingness to participate in the study. This question-
naire was accompanied by an information letter stating the
purpose and requirements of the study (See Appendix A for letter
and questionnaire) .
At the first installation visit all subjects were asked to
read and sign a consent form (Appendix A) . The use of the health
diary was explained to them. In this study, the effect of
bacteria associated with carbon filters and their impact on
health was measured in terms of acute symptomatology. The
health diary used was a revised diary used previously in EPA
sponsored studies. The diary method is a variation on self-
administered questionnaires. Self -administered questionnaires
were preferred for many reasons. The first is that this avoids
problems of interviewer error. Secondly, since information on
all household members was collected, the self-administered
questionnaire allowed time to ascertain the status of all
household members . Most importantly, self -administered question-
naires have been shown to collect information more accurately
(Cannell and Fowler, 1963). The diary consisted of a list of
gastrointestinal and dermatological symptoms. The severity of
the symptoms was measured in terms of affect on subjects daily
living such as did they stay home or in bed and did they seek
medical care. If they sought medical care they were questioned
subsequently to determine what the diagnosis was. This diary
was exchanged for a new one at the monthly visit and the symptoms,
if any, were reviewed by the researcher with an adult member of
the house. This supplemental interviewing is recommended for
self -administered questionnaires (Moser and Kalton, 1972) .
Marketing Survey
Questions were submitted by both filter manufacturers to be
included at the bottom of the monthly diary. Questions concerning
taste, odor and color reduction were asked of both Pollenex and
Everpure filter users. In addition, Everpure had questions that
-------
related to the design and function of their filter. The results
of the marketing questions used are summarized in the Appendix B.
Filter Use Survey
At the end of the study an exit survey on use of filters
and amount of filtered water consumed by subjects was done.
The adult member of the household present at the last monthly
visit was interviewed concerning the intake of water by all
members of the household (See Appendix A) .
Human Investigation
All questionnaires, health diaries, consent forms, informa-
tion letters, went to the Yale University Human Investigation
Committee (HIC) before use for clearance and approval. No
questionnaire was used until an HIC approval was received.
D. MEDICAL CLINIC VISITS AND CLINICAL LABORATORY PROCEDURES
The health survey approach to establishing an association
between the microbial quality of GAC filtered water and illness
in water consumers was supplemented by examining the direct
relationship between bacteria isolated from patients who visited
a physician because of a gastrointestinal illness and bacteria
colonizing the GAC filter used by the study participant. This
direct approach was implemented in the following manner. The
medical clinic records of all individuals participating in the
study were flagged so that (1) the examining physician would be
aware that the patient was a study participant and (2) a
bacterial work-up on the patient's specimen would be requested
if the diagnosis indicated a possible bacterial etiology. Any
bacterial isolates from patient specimens were to be preserved
for further laboratory examination. The participants were
requested at the beginning of the study to consult a clinic
physician if they experienced severe gastrointestinal illness
or skin infections. In addition, the participants were
requested to notify the study coordinator if they consulted a
physician so that a follow-up investigation could be conducted
to determine how the illness was diagnosed and if a clinical
specimen was obtained. In the event that a bacterial isolate
was obtained from a clinical specimen, it was to be characterized
biochemically and identified to the species level if possible.
Subsequent to the identification of the clinical bacterial
isolate, the GAC in the filter serving the patient would be
removed from the filter housing and examined for the presence
of a bacterial isolate identical to the patient isolate. Although
the finding of a bacterial strain in a GAC filter that is similar
or identical to one isolated from a patient using the filter is
not unequivocal evidence of cause and effect it does present a
fairly strong indication that the filter may have been the
source of the causative organism.
10
-------
SECTION 4
RESULTS AND DISCUSSION
A. MICROBIOLOGY
Approximately 1800 paired samples were collected over the
seventeen month duration of the study. This does not include
approximately 220 pairs of filter effluent and tap samples
collected during the installation visit to the homes. There
were four sample categories. The first category was the tap
water sample collected from each home on every monthly visit.
The remaining three categories were filter effluent samples.
The water samples collected from faucet filters were either
from filters containing a cartridge with granular activated
carbon (GAG) or from filters containing a cartridge that was
empty. The fourth category of water samples came from the
bypass GAC filters. The total number of samples analyzed for
each group is summarized in Table 1.
Geometric means and standard deviations of bacterial counts
of the filters at installation are summarized in Table 2. The
majority of the carbon filters (bypass and faucet) eluted no
(<1 cfu/ml) bacteria upon installation indicating lack of
contamination of carbon filters during manufacturing or installa-
tion. The blank cartridges may have been somewhat colonized
but since these were empty cartridges some of the bacteria may
have been from the tap. The tap sample at installation is
different from the tap samples collected in subsequent months
in that the water ran for two minutes before collection during
the installation sampling compared to the specific procedure
used during the rest of the study period as outlined in Section 3.
Initially the number of blank filters installed was smaller
because the blank filters were added to the study two months
after the study began.
The overall geometric mean and standard deviation of
bacterial counts are given in Table 3. There were two
heterotrophic bacteria media used on each sample, R2A and
Standard Plate Count (SPG) . In each of the four sample types
the R2A medium detected more heterotrophic bacteria than did
the SPG. This increased detection was statistically significant
(p<.05, paired t-test). The bacterial densities were different
for each of the four types of water samples. As expected, the
highest densities of bacteria were found in the two carbon
containing filters. The bypass filter discharged the highest
densities of bacteria. The lowest densities from a filter
were discharged by the blank filter. The higher densities found
11
-------
in the bypass filter versus the facuet filter may be due to a
higher amount of carbon found in the bypass filter, hence providing
more surface area. An analysis of variance was performed on
the four water categories and their bacterial densities and
each of the category mean densities of bacteria was significantly
different from each other.
Each of the two carbon filters had significantly higher
bacterial densities in their study effluent samples than the
installation samples. However, the overall geometric study
mean of bacterial densities in blank fil ter effluent was not
statistically different from those found upon installation.
The mean for blank filters was higher indicating a possible
colonization of bacteria on the walls of the blank cartridges,
but these cartridges had a certain density of bacteria associated
with them at installation.
The bacterial densities were grouped together by calendar
month. A geometric mean for each month was calculated for each
of the four categories and plotted. The graphs illustrate
the magnitude of the differences between the tap water and
bypass filters (Figure 3) , the tap water and carbon faucet filters
(Figure 4) and -the tap water and blank faucet filters (Figure 5).
The graphs also pointed out seasonal trends and the influence
of ambient temperature on bacterial densities. To illustrate this
association better, the average monthly raw water temperature
at the Groton water filtration plant was plotted along with
bacterial densities for each of the water sample categories.
Graphically, there were temperature trends associated with the
tap water samples (see Figure 6), the carbon faucet filter (see
Figure 7) and the blank faucet filter water samples (see Figure 3).
The fluctuation in bacterial densities eluted from bypass
filters (see Figure 9) were not as great. Statistically all
bacterial mean densities in the four water sample categories
were highly correlated with water temperature (Table 4). This
correlation was true for bacterial densities whether measured
by K2A or SPC agar.
Paired water samples (tap and filter) from either bypass
or faucet carbon filter households were randomly chosen for
bacterial species identification. A total of six faucet filter
households and nine bypass filter households were examined.
The identification of bacterial isolates was performed on both
the tap water and filter effluent samples. The results of the
tap/carbon faucet water samples are in Table 5. The majority
of the identifiable isolates were Flavobacterium or Pseudomonas.
In comparison to the bypass filters (Table 6), the tap/carbon
faucet pairs seem to be more diverse. The most notable feature
of the bypass filters was the predominance of Pseudomonas stutzeri.
This organism was predominant in seven out of nine bypass
filter effluents examined. The table notes the number of isolates
that subsequently did not grow or grew very slowly in the
laboratory. These isolates were not identified. One such
12
-------
isolate was a small coral pigmented isolate that routinely took
up to 15-21 days to grow at room temperature. Repeated passage
of this isolate did not change its growth rate.
B. EPIDEMIOLOGY
The study period for each of the filter groups was slightly
different. The faucet filter portion of the study began with
installations in February and March of 1985 and continued until
June of 1986 (approximately 15 months) . In this time period
the average length of an individual household participation was
six months. The bypass filter study began with installations
in June of 1985 and continued until August of 1986 (approximately
13 months). The average length of participation of an individual
household with a bypass filter was seven months. This difference
is due partly to the different drop out rates. In general the
drop out rate for bypass filters was much less. This is partly
due to the tighter screening of subjects and the emphasis upon
the cost of filter installation. Another explanation for the
two differences is that households in the faucet filter study
often dropped out after receiving a blank. These families
complained that they liked a carbon filter and would not participate
if they did not have a carbon filter unit. This was despite
the fact that the researchers never confirmed whether a household
had a blank or carbon filter.
The number of person-years for the study by sex and filter
groups is presented in Table 7. The number of person-years for
each filter group by age group may be found in Table 8. The
sex and age distributions for each filter group were similar.
Approximately 40% of the persons in each filter group classified
by age were in the highly susceptible population (less than 10
years of age).
There are two kinds of error which may occur in analyzing
data from an epidemiological study. The first error, called
the Type I error, consists in declaring that the difference in
population studied is real when in fact no difference exists.
This is often called the alpha significant level and is typically
set at .05 or smaller.
The second type of error, called Type II, consists of
failing to declare populations significantly different when in
fact they are different. To control for a Type II error the
power of a test is determined. The power of a test is denoted
as 1-beta, where beta is the probability of failing to detect a
specified difference as being statistically significant.
The smallest number of person-years was observed in the
blank filter group. This group had 180 person-years. With
this as a limiting factor, an example of the smallest proportional
difference that could be detected with an alpha = .05 and power =
95 is .20-.05 or a proportional difference of .15 between two
13
-------
groups. However, the proportional difference significantly
detected is dependent also on the magnitude of the proportions
and the number of comparisons being made. In Table 9, the
incident density per person-year is given for each of the
symptoms. There were no statistically significant differences
detected between the three study groups for any of the symptoms.
The sample size was sufficiently large enough to detect any
statistically significant differences.
A stratified analysis of the combined gastrointestinal
symptoms was done. The incident density or proportion with
symptoms stratified by age and sex for each filter group is in
Table 10. The stratified data did not show any significant
differences between study groups for gastrointestinal illnesses.
There was a trend for the highest disease proportion to be found
in the under six years of age groups. This was not unexpected
as studies have reported higher gastrointestinal symptom rates
in preschool children (Monto and Koopman, 1982).
In examining sex differences there did not appear to be
any differences between study groups. However, within each
study group there were three statistically significant differences
detected between males and females. Two of these occurred in
the by-pass group and one was in the blank filter group. This
difference entailed a significantly higher gastrointestinal
rate in females over males in the adult age group. There were
two phenomena in the population to explain this difference.
First many of the adult female subjects became pregnant during
the course of the study. This would increase the level of
gastrointestinal symptoms. Secondly, this study was carried
out on many families where the male adult typically was gone
(out to sea) for many months during the study. Hence, the
adult male typically would appear healthy when in reality he
was absent. The other group that showed significant differences
was the by-pass <6 years group. In this group the females
reported a higher incidence of gastrointestinal illness than
males. Explanations for this phenomena would be speculation.
C. CLINICAL FINDINGS
A total of 3333 symptomatic illnesses were reported during
the course of the study. Two-hundred and sixty-nine of the
illnesses were considered by the study participants to be
severe enough that they consulted a physician. Forty-two
percent of the patients who were seen by a physician were less
than six years of age, 10% were ages 6 through 10 years, 6%
were ages 11 through 18 years and individuals over the age of
18 comprised 36% of this group. A follow-up investigation of
all reported physician visits revealed that none of the patient
illnesses were diagnosed as having a bacterial etiology and,
therefore, clinical specimens were not collected for laboratory
work-up. The nonoccurrence of illnesses or infections that
might be attributable to bacteria associated with GAG filters
14
-------
indicates that the ingestion or use of GAC filtered water was
not a risk factor for disease in the populations examined in
this study.
D. EXPOSURE
In order to estimate an actual exposure level, all households
remaining in the study by May of 1986 were interviewed as to
consumption of filtered water and filter use habits. A single
person (usually the mother) was interviewed by a Yale researcher
for members of her family and their drinking water habits. A
series of questions were then asked about how water is taken
from the filter and how they used filtered water. The categories
of use for filtered water is summarized in Table 11. For cold
drinks, the majority of both filter users reported using
filtered water almost exclusively. The two filter groups
reported different use patterns for each of the remaining
categories. The by-pass filter users reported: using filtered
water for hot drinks (98%), cooking (88%) and ice cubes (88%).
This was compared to faucet filter users that reported 62% for
cooking, 58% for hot drinks and 83% for ice cubes. Some of
these differences can possibly be explained by filter design or
filter flow. A further report on likes, dislikes and other
consumer type questions is in the marketing report attached
hereto (Appendix B).
The subjects were then asked how they collected water for
use from their filter. The collection process was slightly
different for each filter group (Table 12). The by-pass filter,
which has its own faucet, were divided between "immediately" or
"after a few seconds". The faucet filter group let the filter
run a few seconds (62%) or longer (21%). The slight difference
in use pattern may have had an effect on the possible dose
received. The faucet filter exhibited a flushing phenomena of
reduction in effluent heterotrophic bacterial densities after
approximately 1.25 liters of water were withdrawn (Table 13).
This was in contrast to the by-pass filters which showed equiv-
alent reduction in bacterial cell densities after three liters.
The final question analyzed in order to estimate exposure
was amount of water consumed daily. The amount of filtered
water used by subjects is summarized in Table 14. The distribution
for each of the ounce categories were slightly different for
each filter group. The by-pass filter group consumed slightly
higher volumes of water than did the faucet filter group. This
was further evidenced by comparing the mean ounce consumption
reported by the by-pass households (31.5 ounces per person) with
the mean reported by the faucet filter households (26.4 ounces
per person). If the ounces are converted to milliliters, an
estimate of the number of organisms ingested on daily basis can
be made. The average milliliters consumed per day is multiplied
plied by the geometric mean for the carbon filters to estimate
15
-------
an average heterotrophic bacterial density per day. The results
are as follows:
BY-PASS FILTER
31.5 oz/day/person x 29.57 ml/oz x 2000 org/ml =
1.9 x 106 organisms/day
CARBON FAUCET FILTER
26.4 oz/day/person x 29.57 ml/oz x 1000 org/ml =
7.8 x 105 organisms/day
The estimated average daily intake for the subjects who used
GAC-POU filters was approximately 106 organisms per day.
E. DISCUSSION
The colonization of filters was not unexpected as many
laboratory studies reported on colonization of GAC filters by
bacteria (Geldreich, et al.. 1985; Bell, et aT.., 1983). The
magnitude of colonization and the variation in filter effluent
bacterial densities due to filter type and temperature was new.
Of particular interest was the high incidence of colonization
of the by-pass filters by Pseudomonas stutzeri. There appears
to be no apparent reason for such high incidence of this organism
in these filters.
In general, the primary genus colonizing both types of
POU-GAC filters was Pseudomonas. Pseudomonas aeruainosa has
been implicated in many water contact diseases such as hot tub
folliculitis and swimmer's ear (Calderon and Mood, 1982). In
this study, there was only one isolation of P. aeruainosa. The
majority were other pseudomonads such as P. diminuta and
P. stutzeri.
The second most common bacterial colonizer was the genus
Flavobacterium. This geneus, like the non-P. aeruqinosa
pseudomonads, is a common soil organism and is often found in
drinking water. A literature search indicated that this genus
has not been associated with outbreaks of gastrointestinal or
dermatologic disease. The amplification of their densities by
POU-GAC filters did not result in any increased health risk in
our study.
The gastrointestinal rate in all categories was higher
generally than what has been reported in other studies (Monto
and Koopman, 1980). However there was no significant difference
between the group that had unfiltered water versus the group
that had carbon filtered water. Tests on the statistical power
of the data analysis indicates a sufficient sample size to detect
a difference. Other reasons for no health effects associated
16
-------
with consumption of water containing high density of bacteria
relate to either the organism, the exposure or the host.
Considering the ubiquitous nature of the organisms that
colonized the POU-GAC filters, it may well be that these organisms
just do not cause infections or disease in man. The exposure
route often determines whether an organism causes disease or not.
In this study the primary route of exposure was the gastrointestinal
or dermatologic route. It may be that these bacteria do not
lead to diseases in man by these routes of exposure due to
host barriers such as stomach acid and normal flora of skin and
intestinal tract.
Finally the lack of a health effect may center on the
host. It is well documented that Leaionella. a common soil and
aquatic organism, only causes Legionnaires' Disease when it
comes into contact with an immunocompromised population. The
organisms found on POU-GAC filters may only be of concern to a
very select small population considered to be immunocompromised.
This study does not address the problems that may be experienced
by an immunocompromised population. Furthermore, this study
does not address other exposure routes such as respiratory.
There are GAC water treatment devices that filter the entire
house water supply. These filters are called point-of-entry
because they are installed at the point where water enters the
house. The extent of colonization of the whole house filters
by bacteria has not been reported in the literature. If one
assumes some colonization of point-of-entry filters, there is a
potential for total body exposure to bacteria via shower
generated aerosols. Hence, the question of health effects
associated with bacterial colonization of granular activated
carbon filter still needs to be addressed with point-of-entry
water filters.
17
-------
BIBLIOGRAPHY
American Public Health Association. Standard Methods for the
Examination of Water and Wastewater. 16th ed., American Public
Health Association, Inc., New York, 1985.
Bell, F.A., D.L. Perry, J.K. Smith and S.C. Lynch. "Studies on
Home Water Treatment Systems", JAWWA 76:126-130, 1984.
Brewer, W.S. and W.W. Carmichael. "Microbiological Characteriza-
tion of Granular Activated Carbon Systems". JAWWA 71:738-740,
1979.
Cabelli, V.J., A.P. Dufour, L.J. McCabe, M.A. Levin. "Swimming-
associated gastroenteritis and water quality". Am. J. Ecidem-
iol. 115:606-616, 1982.
Calderon, R. and E.W. Mood. "An Epidemiological Assessment of
Water Quality and 'Swimmer's Ear"1, Arch. Environ. Hlth. 37:300-
305, 1982.
Cannell, C.F. and F.J. Fowler. "Comparison of a self-enumerated
procedure and a personal interview: a validity study."
Public Opinion Quarterly. 21:250-264, 1963.
Camper, A.K. M.W.LeChevalier, S.C. Broadaway and G. McFeters.
"Evaluation of procedures to desorb bacteria from granular
activated carbon", J. Microbial. Methods 1:187-198, 1985.
Dufour, A.P. "Epidemiological Concerns of Bacterial Growth in
Activated Carbon Filters" Proceedings Water Quality Association,
March 1985.
Geldreich, E.E., R.H. Taylor, J.C. Blannon and D.J. Reasoner.
"Bacterial Colonization of Point-of-Use Water Treatment Devices"
JAWWA 71:72-75, 1985.
Karnad, A., S. Alvarez and S.L. Berk. "Pneumonia caused by
gram-negative bacilli", Am. J. Med. 791f1A);61-7. 1985.
Mayer, K.H. and S.H. Zinner. "Bacterial Pathogens of Increasing
Significance in Hospital-Acquired Infections", Rev. Infect.
Pis. 1:S371-S379, 1985.
Monto, A.S. and J.S. Koopman. "The Tecumseh Study: XI Occurrence
of Acute Enteric Illness in the Community", Amer. J. Enidemiol.
112:323-333, 1980.
18
-------
Moser, C.A. and G. Kalton. Survey Methods in Social Investigation.
Basic Books, Inc., Publishers, New York 2nd edition pp. 256-
279, 1972.
Perry, D.L. , J.K. Smith and S.C. Lynch. "Development of Basic
Data Knowledge Regarding Organic Removal Capabilities of
Commercially Available Home Water Treatment Units Utilizing
Activated Carbon, Final Report - Phase 2", Gulf South Research
Institute, New Orleans, La. (July 1980).
Reasoner, D.J. and E.E. Geldreich. "A New Medium for the Enumera-
tion and Subculture of Bacteria from Potable Water", Appl.
and Environ. Microbiol. 49.: 1-7, 1985.
Regunathan, P, W.H. Beauman and E.G. Kreusch. "Efficiency of
point-of-use treatment devices", JAWWA 75:42-50, 1983.
Rice, R.G. and C.M. Robson Biological Activated Carbon-Enhanced
Aerobic Biological Activity In GAG Systems. Ann Arbor Science
pp 19-22, 1982.
Spino, D.F. "Characterization of Dysgonic, Heterotrophic Bacteria
from Drinking Water", Appl. and Environ. Microbiol. 50;1213-
1218, 1985.
Wallis, C., Stagg, C.H. and Melnick, J.L. "The Hazards of
Incorporating Charcoal Filters into Domestic Water Systems",
Water Res. 8.:11, 1974.
19
-------
FAUCET ATTACHED
Vi/DIVERTER
Figure 1. Schematic diagram of faucet filter as it is installed
(courtesy of P. Regunathan).
20
-------
^-SEPARATE
FAUCET
LINE BYPASS
UNDER THE SINK
FILTER SYSTEM
Figure 2. Schematic diagram of bypass filter as it is installed
(courtesy of P. Regunathan).
21
-------
10 s-.
LJ
0- 10 *-
ID
L^
o
10 -d
Bypass Filters
TAP WATER
SPC= Solid Line
R2A= Dashed Line
—i 1 1 1 1 1 1 1 1 1 1 1 1 1—r— —r
MAMJJASONDJFMAMJJA
MONTH
Figure 3. Heterotrophic bacterial densities as measured by
standard plate count agar and R2A agar by month for
tap water and bypass carbon filters.
22
-------
104-.
103-
LJ
10
2-
O
10 -
,\ Carbon Faucet
Filters
TAP WATER
SPC= Solid Line
R2A= Dashed Line
MAMJJASONDJFMAMJJA
MONTH
Figure 4. Heterotrophic bacterial densities in tap water and
carbon faucet filters as measured by standard plate
count agar and R2A agar by month.
23
-------
10 S
10 3^
UJ
CL
O
10 -
TAP WATER
SPC= Solid Line
R2A= Dashed Line
—i 1 1 1—i 1—i 1—i 1 1 1 1 1 r- —r
MAMJJASONDJFMAMJJA
MONTH
Figure 5. heterotrophic bacterial densities in tap water and
blank faucet filters as measured by standard plate
count agar and R2A agar by month.
24
-------
UJ
Q-
CJ
10 -
SPC= Solid Line *
R2A= Dashed Line *
Temperaiure =Triangles
-20
30
LJ
-10
LU
CL
2
LJ
LJ
L0
Figure 6. Heterotrophic bacterial densities in tap water and
water temperature by month.
25
-------
10
SPC= Solid Line *
R2A= Dashed Line *
Temperatdr«, =Triangles
Carbon
Filters
Lo
MAMJJASONDJFMAMJJA
MONTH
Figure 7. Heterotrophic bacterial densities in carbon faucet
filters and water temperature by month.
26
-------
10 *T
LJ
0-102j,
o
10 -
Blank Faucet
Filter^ _.^ \
SPC= Solid Line *
R2A= Dashed Line *
Temperature=Triangles
-30
-20
LJ
CL
2
LJ
-10
MAMJJASONDJFMAMJJA
MONTH
-0
Figure 8. Heterotrophic bacterial densities in blank faucet
filters and water temperature by month.
27
-------
10 +q
Bypass Filters
10 3-
SPC= Solid. Line *
R2A= DashfeH, Une *
T9mperati/re=
Ld
LL.
o
10 d
-.—«L0
-30 Q
LJ
CE
20 <
LJ
Q_
-10
Ld
01
MAMJJASONDJFMAMJJA
MONTH
Figure 9. Heterotrophic bacterial densities in bypass carbon
filters and water temperature by month.
28
-------
Table 1. Number of water sample types
collected over a 17 month period
(1985-1986).
Sample N
Bypass Filter 722
Carbon Faucet Filter 559
Blank Faucet Filter 486
Tap 1766
Total 3533
Table 2. Geometric mean and standard deviations of heterotrophic
bacterial densities from installation water samples.
MEDIA
SPC1 R2A2
Sample
Bypass Filter (62)
Carbon Faucet Filter (110)
Blank Faucet Filter (43)
Tap Water (215)
Mean
0.2
6
85
6
STD
34
5
9
17
Mean
0.4
9
98
11
STD
41
6
7
20
135°C, five days.
222°C, five days.
29
-------
Table 3. Geometric means and standard deviations of heterotrophic
bacterial densities discharged from two granular
activated carbon filters, blank filters and tap
water as measured by Standard Plate Count agar and
R2A agar.
MEDIA
SPG1 R2A2
Sample
Bypass filter
Faucet filter
Blank filter
Tap Water
Mean
1049
689
198
53
STD
5
16
17
12
Mean
2042
1035
289
92
STD
4
17
19
13
^sOc, five days.
222°C, five days.
Table 4. Correlation coefficient of water temperature with
bacterial densities from bypass filters, carbon and
blank faucet filter and tap water using Standard
Plate Count and R2A agar.
MEDIA
Water Sample SPC1 R2A2
(Correlation coefficient for water temperature)
Bypass filter .617 .663
Carbon faucet filter .743 .761
Blank faucet filter .931 .883
Tap water .923 .956
T-ISOQ, five days.
222°C, five days.
30
-------
Table 5. Types of bacteria isolated from tap and carbon faucet
filter water samples from six homes.
Organism
Alcaliaenes spp.
Acinetobacter spp.
Actinobacillus spp.
Flavobacterium SPP.
Pseudomonas spp.
Pseudomonas aeruainosa
Pseudomonas diminuta
Pseudomonas fluorescens
Pseudomonas paucimoblis
Pseudomonas vesicular is
Not identified
Number of Isolates
Tap Only Filter Only
1
1
1
2
1
0
0
0
0
1
12
1
0
0
0
0
1
1
1
1
1
7
Both
0
2
0
5
1
0
0
0
0
2
31
-------
Table 6. Types of bacteria isolated from tap and bypass carbon
filter water samples from 9 homes.
Organism
Number of Isolates
Tap Only Filter Only
Both
Acinetobacter spp.
Flavobacterium spp.
Pseudomonas spp.
Pseudomonas diminuta
Pseudomonas paucimoblis
Pseudomonas stutzeri
Pseudomonas vesicular is
Yeast
Not identifiable
1
3
2
1
1
0
1
1
12
1
3
2
0
1
7
3
0
8
0
1
0
0
1
0
1
0
32
-------
Table 7. Person years of surveillance for each filter group
in the point-of-use study by sex (1985-1986).
Filter Group
person years - (%)
Sex
Male
Female
Total
Bypass
116.5
113.8
230.3
(50.
(49.
(100.
6)
4)
0)
Faucet Carbon
117.8 (51.
111.8 (48.
229.6 (100.
4)
6)
0)
Faucet Blank
94.6 (52.
87.3 (48.
181.9 (100
0)
0)
.0)
Table 8. Person years of surveillance for each filter group in
the point-of-use study by age group (1985-1986).
Age
Filter Group
person years - (%)
Bypass
Faucet Carbon
Faucet Blank
<6 years
6 to 10 years
11 to 18 years
>18 years
59.0 (25.6)
34.8 (15.1)
15.3 (6.7)
121.2 (52.6)
43.1 (18.1)
43.8 (19.0)
24.2 (10.5)
120.2 (52.0)
43.1 (23.9)
32.4 (17.9)
15.9 (8.8)
89.3 (49.4)
Total
230.3 (100.0)
231.3 (100.0)
180.7 (100.0)
33
-------
Table 9. Symptom rate in the point-of-use filter study per
person year by filter group (1985-1986}.
Sex
Bypass
Filter Group
person years - (%)
Faucet Carbon1
Faucet Blank
Vomiting
Nausea
Diarrhea
Fever
Body aches
Neck pain
Skin rash
Infected wound
.38
.54
.72
.67
.55
.17
.16
.01
.43
.61
.90
.56
.55
.19
.15
.03
.41
.61
.93
.78
.70
.28
.13
.03
1No signifcant difference between the three study groups for
any of the symptoms.
34
-------
Table 10. Incident density of gastrointestinal illnesses in
each study group by age and sex.
Age Group
(per person year)
Group <6 years
Bypass
Female 2.8*
Male 1.8*
Carbon faucet
Female 2 . 3
Male 2 . 3
Blank faucet
Female 1 . 9
Male 2 . 5
6-10 years 11-18 years
1.2 0.7
1.4 0.6
1.4 1.3
1.6 1.2
1.7 1.2
1.7 1.3
>18 years
1.9*
0.6*
1.6
1.0
1.5*
0.8*
*Rate statistically higher in females than males.
35
-------
Table 11. Percentages of reported uses of GAC
filtered water by filter type.
Use Bypass Faucet
Cooking 88% 62%
Hot Drinks 98% 58%
Cold Drinks 98% 100%
Ice Cubes 88% 83%
Table 12. Percentages of how water was collected
by subjects from a filter by filter
type.
Procedure Bypass Faucet
Immediately 41% 17%
After a Few Seconds 40% 62%
Run a Long Time 19% 21%
Total 100% 100%
36
-------
Table 13. Densities of heterotrophic bacteria discharged with
volume flow from bypass and faucet point-of-use
granular activated carbon filter.
Faucet
Bacteria
(cfu/ml)
382
231
217
143
78
40
44
46
40
22
filter
Volume
(ml)
250
500
750
1000
1250
1500
1750
2000
2250
2500
Bypass
Bacteria
(cfu/ml)
412
247
175
150
133
93
78
54
51
53
filter
Volume
(ml)
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
37
-------
Table 14. Consumption of filtered water by point-of-use filter
type.
Ounces1
16 oz or less
17-32 02
33-64 oz
>64
Total
Mean
Bypass
35%
34%
27%
4%
100%
31.5 ounces
Faucet
47%
29%
22%
2%
100%
26.4 ounces
1Ounces per person per day.
38
-------
APPENDIX A
QUESTIONNAIRES AND HEALTH DIARY
39
-------
Yale University
SCHOOL OF MEDICINE
Department of Epidemiology
aid Public Health
P.O. Box 3333
60 College Street
Dear Resident: N«" «~' CmnKti"" °^
The Department of Epidemiology and Public Health, Yale School of Medicine,
is conducting a study on point-of-use granulated activated carbon filters. These
filters are used in homes at the kitchen sink to filter water used for drinking,
cooking and handwashing. These filters remove many water impurities as well as
improve taste and odor problems. We are interested in determining if these fil-
ters do improve the quality of the water and if there is an improvement in health
associated with better water quality. We have received permission from the Navy
base authorities to conduct this study on residents «fi Navy housing.
Your household has been randomly selected for the first part .of the study.
Attached is a questionnaire on basic background information and water use. If
you wish to participate in the second part of the study, please indicate so on the
questionnaire. The second part of the study will require the following:
1) Completion on a monthly basis of a health diary for the household;
2) Notify researchers at Yale if any of the listed symptoms require medi-
cal care;
3) Allow a researcher from Yale on a monthly basis to collect a water sam-
plexor analysis when a member of the household notifies us as in 02; and
4) Allow Yale researchers with trained personnel at no cost to the subjects,
to install a filter at the kitchen sink.
The study will be conducted from February 1985 to November 1985. Not every family
will participate for the entire period as households will be phased in and phased
out of the study. For every family that completes the study a $10.00 donation
will be made to the Navy Relief Fund. If you think you would like to participate
but need more information please indicate so on the questionnaire or call us as
indicated. Even if you do not wish to participate further please fill out the
questionnaire and return it in the self-addressed stamped envelope. Thank you
for your time and cooperation.
Sincerely,
Rebecca L. Calderon
Dept. of Epidemiology & Public Health
Yale School of Medicine
40
-------
YALE SURVEY
Please fill in the blank or circle your answer.
1. Which of the following best describes how your water tastes:
a. Acceptable taste all the time
b. Acceptable taste most of the time.
c. Unacceptable taste most of the time.
d. Unacceptable taste all the time.
e. No opinion
2. Which best describes the amount of odor in your water:
a. Never has an odor
b. Occasionally has an odor
c. Has a constant odor
d. No opinion
3. Which best describes the color and/or turbidity in your water:
a. Never has any turbidity and/or color
b. Occasionally has turbidity and/or color
c. Is constantly turbid and/or colored.
d. No opinion
4. Do you have a dishwashing appliance in use in your home? YES NO
5. Are there any of the following water treatment devices being used in your home?
a. Water softener YES NO DON'T KNOW
b. Faucet filter YES NO DON'T KNOW
c. Iron removal filter YES NO DON'T KNOW
d. OTHER (specify)
6. What type of water heater is used in your home?
a. Electric
b. Gas
c. Oil
d. OTHER (specify)
7. What is the temperature of the hotwater in your house?
(If you don't know, leave it blank)
Is there a storage tank with your hotwater heater? YES NO DON'T KNOW
8. Which of the following materials is your kitchen sink made of:
PORCELAIN STAINLESS STEEL
Is there a sprayer/hose apparatus on the kitchen sink? YES NO
41
-------
PAGE 2
9. Which best describes your home?
a. A mobile home or trailer
b. A one bedroom house
c. A two bedroom house
d. A three bedroom house
e. A four bedroom house
f. Other (specify)_
10. How many persons are there living in your household?
11. How many persons are there who are:
a. Less than 2 years old
b. 2-5 years old_
c. 6-10 years old
d. 11-15 years old
e. 16-21 years old__
f. greater than 21 years old_
12. Does everyone in your household use the Navy clinic and/or hospital for minor
health problems such as cold, diarrhea, fever or rash?
YES NO
If NO, how many members of your household would seek care some place else
for such minor health problems?_
13. What is the Personnel Rotation Date for the active duty member of the household?
(If more than one active duty member in the household, use the earliest PRD)
If you think your family would like to participate in the second part of the study,
(please see accompany letter) please fill in your name and address.
NAME
ADDRESS
If you have any questions please fill in the phone number and a convenient calling
tine and we will contact you.
NUMBER —
TIME
Or if you wish, you can call us at 1-785-2881, 9:00 an to 5:00 pm, Monday - Friday.
Please return this questionnaire in the postage paid envelope.
42
-------
CONSENT FOR PARTICIPATION IN A RESEARCH PROJECT
YALE UNIVERSITY SCHOOL OF MEDICINE - YALE NEW HAVEN HOSPITAL
You and your family are invited to participate in a study on the improvement
to health and quality of drinking water by use of residential point-of-use granu-
lated activated carbon filters. These filters are used to remove some impurities
and to improve odor and taste problems. You have been chosen for this study because
you live in housing provided by the U.S. Navy.
In this study each home will have a filter installed at the kitchen sink. One
half of the homes will receive a filter without carbon, and the other half will have
filters with carbon. The homes with carbon in their filters will be selected at
random (by the luck of the draw). These homes will change over the course of the
study so that every home during the study will have carbon in their filter. This
is a blind study so it should not be apparent to you at anytime whether you have
a filter with carbon or without carbon. The study will last for up to nine months
during which time we will ask you to keep a monthly diary on intestinal and skin
ailments occurring in all members of the household. In addition, we would like to
arrange for monthly visits to collect water samples for analysis and to insure
proper maintenance of the filters. This may require up to 30 minutes of your time
every month, including water sampling and questionnaire answering.
If you or any member of the household should experience one of the above classes
of illnesses and have decided that medical care is needed, we ask that you immed-
iately let us know. We would like to collect additional water samples for analysis.
This study may be of no direct benefit to you individually, but will improve
our knowledge of whether use of water filters improve the quality of drinking water
and the health of people who use them. You will not be paid for participation, but
all filters, their maintenance and water testing will be provided for free during
the study. In addition, for every family that fully participate in the study, a
$10.00 donation will be made to the Navy Relief Fund in their name.
In all records of the study you will be identified by a number and your name
will be known only to the researchers and the Navy clinic/hospital. Your name will
not be used in any scientific reports of the study. There will be complete confid-
entiality of all records.
You are free to choose not to participate. Due to expense and effort of in-
stalling filters, we ask that you try to commit to participating for the duration
of the study. If for some reason you can not complete the study or must withdraw
or choose not to participate, it will not adversely affect your relationship with
Yale or the Navy.
43
-------
We have used some technical terms in this form. Please feel free to ask
about anything you don't understand and to consider this research and the consent
form carefully before you agree to participate.
Authorization: I have read this form and decided that
(name of head of household)
will participate in the project described above. Its general purposes, the parti-
culars of involvement and possible hazards and inconveniences have been explained
to my satisfaction. My signature also indicates that I have received a copy of
this consent form.
Signature
Relationship {self, parent, etc)
Date
Signature of Person Obtaining Consent
If you have further questions about this project or your rights as a research
subject or if you have a research related problem, please contact the project
director, Rebecca L. Calderon, 785-2831, Department of Epidemiology and Public
Health, Yale School of Medicine, New Haven, Ct. 06510.
THIS FORM IS NOT VALID UNLESS THE FOLLOWING
BOX HAS BEEN COMPLETED IN THE HIC OFFICE.
THIS FORM IS VALID ONLY UNTIL
(date).
HIC PROTOCOL
INITIALED:
44
-------
YALE HEALTH SURVEY
This is page one of the health diary which we are asking you to keep over
the duration of the study. Notice that the second page is the actual diary of
health symptoms. Every month a water sample will be collected and at that
time we will collect your health diary for the past month and will give you a
new second page for the. ensuing month. At the end of your participation in the
study we will collect the front page. The front page is to assign everyone in
the household a number for identification and to collect basic information. Be
sure to keep everyone's number the same throughout the study. For convenience,
we suggest you place your diary in a prominent place such as the refrigerator
door or family bulletin board.
Please review the list of symptoms on the next page. Note that we are
interested in only when these symptoms occur and their duration. Please check
or fill in only the boxes that apply. If you or a member of your household
decide to seek medical help for any of the symptoms we ask that you call us
immediately. We may want to collect a water sample. To call New Haven,
dial 0 and then dail our number 785-2881. Wait for either a tone or an
operator to ask you for your number. The number for the study is
620-185-8977-9916. If you have trouble getting through, call us collect. When
you or any member of the household vist the doctor, be sure to tell him/her that
you are a study participant.
PERSON
FIRST NAME
(optional)
DATE OF BIRTH
SEX
#1
#2
#3
#4
#5
#6
45
-------
PAGE 2 FAMILY ID
MONTH
PERSON NUMBER
DATE OF ONSET OF SYMPTOMS
SYMPTOMS
A. Vomiting or throwing up
B. Nausea or feeling nauseous
C. Diarrhea or loose bowels
D. Fever (state temperature)
E. Body aches
F. Neck pain
G. Rash
H. Infected wound (location e.g. knee)
How long did the symptoms last? (in days)
Did the peraon stay home? •
Did the person stay in bed?
Did the person seek medical help?
Where did you go for medical help?
Navy hospital/clinic?
Private physician?
9ttier?
I
At the end of the month, please answer the following questions?
Please rank the benefits the system provides in order of Importance to you and your family.
(1 •= moat important; 6 = least important)
Taste and odor removal
Improved clarity and color of water
Health benefits of cyst and asbestos removal
Health benefits of organic chemical removal
Appliance protection provided by lime-scale
inhibition
Chlorine removal
-------
HOV7 TO USE YOUR POLLE1IEX PURE WATER "99"
FOR CLEAR, CLEAN FILTERED WATER MOVE SWITCH TOWARD YOU.
IMPORTANT: RUN WATER THROUGH THE REPLACEABLE FILTER FOR TEN
SECONDS BEFORE FIRST USE OF DAY. WHEN FIRST USING A NEW FILTER
CARTRIDGE FLUSH FOR SEVERAL MINUTES. FOR REGULAR UNFILTERED
TAP WATER HOVE SWITCH AWAY FROM YOU.
The Pure Water "99" Water Filter is designed to restrict the water
flow. The slower the water flows through the Activated Charcoal
bed, the more "contact time" there is, allowing for better adsorp-
tion of chemical contaminants. Filter has a rated service flow
of .5 gallons per minute; a minimum operating pressure of 10 PSIG
and maximum operating pressure of 125 PSIG. For greater filtration,
you may wish to slow down the water to allow even more "contact
time" than provided by the built-in flow restrictor. Do not run
hot water through the filter.
We will replace the filter every three months. Do not attempt to
take the filter apart. If it is not working properly, you can
remove the entire unit from the faucet. If you call us we will
replace the filter.
47
-------
As pact of the final stage of this study we would like to ask a few
questions on how you use your filter and the filtered water. Can
we take a few moments of your time to ask the questions.
ID
1. Did you use filtered water for any of the following:
a) Cooking YES NO
b) Drinking
Boiled YES NO
Not Boiled YES NO
c) Making Ice Cubes YES NO
2. Which of the following best represent how you collect water
from the filter:
a) Take the water as it immediately comes out of the filter
b) Let the filter run a few seconds and then collect the water
c) Let the filter run MORE than a few seconds before collecting
the water
3. For each of the people in the house how many 8 oz. glasses of
filtered water would you estimate they drink every day? Do NOT
include water that has been cooked/ e.g./ coffee/ tea/ formula.
Person Glasses
Thank you for your participation in the study.
-------
APPENDIX B
MARKETING REPORT
49
-------
INTRODUCTION
This is a summary report compiled for the manufacturers
who participated in the Yale GAG point-of-use filter study.
This report is comprised of two parts. The first is the results
of questions sent to potential study participants. These questions
mainly pertain to the subjects perception of their water quality.
The second part of this report concerns the results of questions
asked of study participants about their filter.
PART 1
The initial pre-study questionnaires included questions
that would help us evaluate people's opinions regarding the
general present quality of their drinking water before entering
the study. The total number persons included in the evaluation
of the pre-study questionnaires was 343.
We asked their opinion on the following:
1) how the water tastes
2) the amount of odor in their water and
3) the occurrence of color and/or turbidity in their
water
With respect to taste27.7% found their water acceptable all
the time and 57.4% found it acceptable most of the time. Much
smaller proportions of the population found the water unacceptable
most of the time (8.5%) and unacceptable all of the time (5.5%).
Only 0.9% of the polled population had no opinion at all regarding
the tasteof their drinking water.
The majority of respondents (45.8%) found that the water
never had an odor, while 42.9% found that it occasionally had an odor.
50
-------
Again/ smaller percentage comprised the remaining categories of
those who felt their water had a constant odor (5.2%) and those who
had no opinion about drinking water odor (6.1%).
Only 9.3% of the population were of the opinion that their
water had color and/or turbidity constantly. A small percentage
(3.5%) had no opinion regarding this issue, while 50.7% felt they
occasionally experienced color and/or turbidity in their drinking
water. The remaining 36.5% felt that their water never had color
or turbidity.
We also inquired about the use of water softeners/ faucet
filters and iron filters in the subjects' homes. Of the three
devices the most prevalent one was a faucet filter (8.2%). This
was followed by a water softener (1.2%) and an iron filter (0.6%).
An overwhelming majority answered no to whether or not a
treatment device was being used in their home; faucet filter (81%)
an iron removal filter (81.9%) and a water softener (85.7%). A
small percentage (13.1%) did not know whether or not their house
had a water softener. Approximately ten percent (10.2%) did not
know if there was a faucet filter being used and another 17% were
not aware of whether or not an iron removal filter had been installed
in their house. This is not surprising since the participants do
not own their homes and are not responsible for its maintenance.
For convenience the results have been summarized below in
table form.
51
-------
Table 1. Results of pre-study questionnaire.
Question: Which of the following best describes how /our water
TASTE (N=343)
27.7% Acceptable all the time
57.4% Acceptable most of the time
8.5% Unacceptable most of the time
5.5% Unacceptable all the time
0.9% NO opinion
ODOR (N=343)
45.8% Never had an odor
42.9% Occasionally has an odor
5.2% Has a constant odor
6.1% No opinion
COLOR (N=343)
36.5% Never has color
50.7% Occasionally has color
9.3% Has a constant color problem
3.5% No opinion
Question: Are there any of the following water treatment devices
being used in your home?
WATER SOFTENER (N=343)
1.2% Yes
85.7% No
13.1% Don't know
FAUCET FILTER (N=343)
8.2% Yes
81.0% NO
10.8% Don't know
IRON FILTER (N=343)
0.6% Yes
81.9% No
17.6% Don't know
52
-------
PART II
With regard to actual use of POLLENEX faucet filters, statistics
were compiled on subject responses to marketing questions placed
at the bottom of each health diary. The questions were provided
toy Associated Mills, the filter manufacturer. After some modifica-
tions by the Yale research team for use in the diaries/ the
questions were approved by the Human Investigations Committee
at Yale University. Each month for 6 months/ the families were
given a set of questions.
The first set of questions asked the following: DID THE ODOR/
TASTE/ OR COLOR CHANGE IN THE LAST MONTH? A yes/no response was
provided with the exception of the first month (of filter use) which
asked to clarify the change as to better or worse.
The results in Table 2 reflect percents of subject response
and is broken down on a monthly basis. We've delineated the two
filter types as well as the response with respect to ODOR/ TASTE/
COLOR/ and NO ANSWER.
Those with carbons seemed to respond to odor and color with
no change/ 70.8% and 67.4%, respectively. As you can see 55.6% of
those with blanks thought the change in odor was for the better/
while a larger majority s aw no change in color (77.7%). It seems
that more of the participants noticed a change for the better with
regards to taste: 65.6% for carbons/ while those with blanks were
nearly evenly split between better and no change in taste: 52.9%
to 47.0%.
There is a large discrepency between filter types in the
53
-------
percentage of people who did not respond: 18.3% for carbons/
45.7% for blanks. And yet when you look at the next two months/
the blanks have a zero/ non-respondent^percentage.
Month 2 yields a generally higher no change response for all
three elements in both groups. Those with blanks may have caught
on because their response to change in odor plummeted to 0% in the
yes category.
Month 3 holds similar percentages for those with carbons
with the exception that the ratio of yes to no is closer to 1:4 than
1:3. The blanks responded to both odor and taste with 100% no change
and did not answer the color portion.
The second set of questions were aimed at evaluating two things:
1) DID THE FILTER PERFORM AS EXPECTED?
2) WOULD THE SUBJECTS HAVE BOUGHT A FILTER PRIOR TO
THE STUDY?
Before looking at the statistics, it's important to note that
generally/ for months 1-3/ people had the same filter type. Month
4-6/ the filter type should have been switched. By month 4, the
time of the switch/ people had an "expectation" because they had
become familiar with the filter and its previous performance. This
may be reflected (Table 3) in the increase in the positive response
category for the carbons for months 4-6. The positive responses
for the blank filters also increased for months 4-6.
Those with blanks tended to have less of an expectation. This
is reflected in the higher percentages in their "no expectations"
responses.
Evaluation of the second question is not an easy task. In
54
-------
Odor
No
Answer
Table 2.
FAUCET FILTER AESTHETICS
% Of Subject Responses to Marketing Questions
First Set: Did the Odor/ Taste/ or Color Change in the Last Month?
Month 1
Carbon „, Blank
No No
Better Worse Change Better Change
28.1 1.1 70.8 55.6 9.0
65.6 1.1 33.3 52.9 47.0
31.5 1.1 67.4 22.2 77.7
Month 2 Month 3
Carbon Blank Carbon Blank
Yes No Yes No Yes No Yes No
23.3 76.7 0 100% 22 78% 0 100%
35.2 64.8 33.3 66.7 24.4 75.6 0 100%
24.7 75.3 33.3 66.7 17.1 82.9 0 0
18.3%
45.7%
20.6%
0%
14.6%
0%
m
U1
-------
Month 1&2
Table 3.
Subject Responses to Marketing Question
Second Set
Month 3 Month 4 Month 5
Month 6
Carbon
Yes 50%
Did the water
filter perform
aa expected?
No 0
No Expectations
Yes 20%
Would you
have bought
a filter be-
fore the
study? No 30
No Expec-
tations -
Blank
15
27
7
23
15
11
.4%
.0
.7
.1
.4
.5
Carbon Blank
46.
11.
9.
9.
53.
9.
5 42.9
6 28.6
3 28.6
3 28.6
5
3 28.6
Carbon
64.
8.
13.
9.
44.
16.
5
3
2
9
6
5
Blank Carbon Blank
44.1 82
23.5 2
1
8
41.2 55
8.8 16
. 8 36 . 6
.8 20.0
.1 13.3 .
.3 10.0
.6 40.0
.7 20.0
Carbon
54
5
11
22
37
11
.3
.7
.4
.9
.1
.4
Blank
50.0
27.3
13.6
9.1
50.0
31.8
No Answer
50% 46.2% 30.2% 14.3
21.5
41.2 19.4 30.0
28.6
9.1
o\
-------
month to month collections there were people who were "true?
converts" and loved their filters and then there were those who
found the "inconvenience" almost too demanding. These preferences
may be reflected in the evaluation of whether people would have
purchased a water filter prior to the study.
It appears that more people would not have bought a filter
previously. For both groups the response is variable across the
6 months/ but generally we can see an increase in those persons
who would not have purchased a filter. Perhaps prolonged use of an
"inconvenient" water device has lead some subjects to vote "no".
The final question was "How could the water filter be improved?"
There were 42 individual suggestions given into three categories.
The first and most common suggestion concerned the space the filter
occupied. Many suggested making it smaller/ attaching it someplace
else/ or designing it so it wasn't so "bulky". The second category
of suggestions concerned durability. Many people said it leaked at
the connection to the sink/ the switches leaked and that the filter
was not durable enough to attach the dishwasher to it. The third
category concerned the actual mechanics of the filter. That is
to say people felt they should be able to run not water through it
and that the filter was too slow. A few subjects did comment that
they were satisfied with the filter and had no suggestions.
At the end of the faucet filter study/ the majority of subjects
kept their POLLENEX filter.
57
-------
PART II
With regard to actual use of the Everpure by-pass filters,
statistics were compiled on subject responses to marketing questions
placed at the bottom of each health diary. The questions were
provided by Everpure/ Inc., the filter manufacturer. After some
modifications by the Yale research team for use in the diaries,
the questions were approved by the Human Investigations Committee
at Yale University. Each month for 6 months, the families were
given a set of questions. The results are given by question.
1. During the past month, has there been any change in odor,
taste or color of your drinking water?
ODOR (N=60)
67% Did not detect a change in odor
33% Did notice a change in odor
For those that detected a change in odor, 95% indicated
that the odor of the water was better. Only 5% (one person) thought
that the odor was worse.
TASTE (N=60)
40% Did not detect a change in taste
58% Did detect a change in taste
2% Did not answer the question
For the families that detected a change in taste 97% thought
that the water tasted better. Only 3% indicated that the water
tasted worse.
53
-------
COLOR (N=60)
72% Did not detect a change in color
27% Did detect a change in color
2% Did not answer the question
For those that "detected a change in color, 88% indicated
that the color of the water had improved. Only 13% thought that
the color was worse.
QUESTION 2
Do you feel that your water filter has performed as you
expected?
(N=62)
84% Yes
5% No
11% No expectations
QUESTION 3
Would you have bought a water filter before the study?
(N=62)
19% Yes
53% No
27% Don't know
QUESTION 4
How could the water filter you have now be improved?
*Note: Some people supplied more than one answer. There were 21
responses.
57% Faster water flow/more pressure
14% longer neck on faucet
10% connect it to the main faucet
10% filter hot water
5% keep sprayer
5% colder water through filter
5% make it taste as" good as norelco clean water machine
5% faucet switch i-s rough on bottom; should not be sharp
59
-------
QUESTION 5
What single feature/benefit do you like about your under-
counter drinking water system?
*Note: Some people supplied more than one answer. There were
60 answers.
42% improved taste
13% convenient
8% takes up little space/not bulky
7% filter hidden out of sight
7% on/off valve easy to use
7% pure water
3% use of portable dishwasher without disconnecting filter
3% clearer water
2% double filter system
2% cold water
2% better water
2% easy to keep clean
2% separate spigot
2% chemical removal
2% drink more water
2% better odor
QUESTION 6
Is there something you do not like about it?
(N=60)
22% Yes
78% No
If yes/ what is it? (N=13)
*Note: Some people supplied more than one answer.
46% slow water flow
31% loss of sprayer hose
15% faucet small
8% no hot filter water
60
-------
8% not attached to main faucet
8% bland taste
QUESTION 7
How could the water filter system be improved?
*Note: Some people supplied more than one answer. There were
26 answers.
54% increase water flow/pressure
19% increase height and extension of faucet
12% improve taste
8% keep sprayer hose
4% attach to main faucet
4% prevent leaking of cartridges
4% colder water through filter
4% filter hot water
QUESTION B
Please rank the benefits the system provides in order of
importance to you and your family. (l=most important; 6=least
important}
Taste and odor removal
——^Improved clarity and color of water
Health benefits of cyst and asbestos removal
Health benefits of organic chemical removal
Appliance protection provided by lime-scale inhibition
Chlorine removal
61
-------
Not all respondents answered this question in the same
manner. Foe example/ some ranked the benefits from most important
to least important/ while others ranked the importance of each
benefit.
The responses were analyzed in three different ways:
1) Numerical means for each benefit; 2) the frequency that each
benefit was ranked 1 or 2; and 3} the frequency that each benefit
was ranked 5 or 6.
#1 NUMERICAL MEANS
Numerical Means
Taste and odor removal 2.0
Organic chemical removal 2.0
Cyst and asbestos removal 2.3
Chlorine removal 3.2
Improved clarity and color 3.7
Appliance protection 4.4
The lower the mean/ the more important the benefit.
t2 Percentage that benefit was ranked 1 or 2 - Most important
Organic chemical removal 72%
Taste and odor removal 67%
Cyst and asbestos removal 35%
Improved clarity and color 20%
§3 Percentage that benefit was ranked 5 or 6 - Least important
Taste and odor removal 5%
Organic chemical removal 6%
Cyst and asbestos removal 17%
Chlorine removal 23%
Improved clarity and color 44%
Appliance protection 61%
62
-------
The three different analyses indicate that taste and odor
removal/ organic chemical removal and cyst and asbestos removal
are of equal importance to the filter users. Chlorine removal is
i
less important followed by improved clarity and color, and finally,
appliance protection.
QUESTION 9
Did you formerly buy bottled drinking water?
(N=56)
14% Yes
86% No
4% No answer
If no, did you ever consider doing so?
54% Yes
43% No
2% No answer
QUESTION 10
Are you aware of the difference between a water softener
and a drinking water system? (N=56)
52% Yes
46% No
2% No answer
If Yes, in your opinion, what is the major benefit each
provides?
Water softener
Improves cleaning
Whiter laundry; better for skin
Healthier bathing and washing
Less soap and detergent
Clean clothes
Not sure
Better for skin
Use less soap
Provides elements and mineral to soften water
Adds salts; less soap; no mineral build up
63
-------
Removes hardness
Removes minerals for better cleaning/ laundry
Softer water
Reduced mineral content
Better for washing
Washing
Hard water to soft/ removing lime deposits
Conserves soap used/ prevents scum
Better for washing
Filters out impurities and water is not so hard on skin/
clothes/ etc.
Drinking water system
Removes odor; improves taste
tastes and feels better
better taste; cleaner water
tastes and looks better
better drinking water
healthier
clean pipes; good drinking water
convenience
filters out chemicals and parasites
better water
healthier water; better taste
removes elements to purify
removes organic and inorganic impurities/ improves taste and
smell
removes impurities
purifies water; removes chlorine
less minerals
remove minerals/ contaminants/ particles; cleaner, better
taste and smell
cleaner and healthier water
improves taste; removes impurities
taste
taste/ purity
purity for drinking
health benefits from chemical removal
better tasting; purer water
improved taste
------- |