Journal of Exposure Analysis and Environmental Epidemiology (2000) 10, 159 ± 167
# 2000 Nature America, Inc. All rights reserved 1053-4245/00/$15.00
www.nature.com/jea
Pesticide storage and use patterns in Minnesota households with children
JOHN L. ADGATE,a ANNE KUKOWSKI,b CHUCK STROEBEL,b PAMELA J. SHUBAT,b SHANA MORRELL,b
JAMES J. QUACKENBOSS,c ROY W. WHITMOREd AND KEN SEXTONa
a
School of Public Health, University of Minnesota, Box 807 UMHC, Minneapolis, Minnesota 55455
Minnesota Department of Health, 121 E 7th Place, St. Paul, Minnesota 55101
c
U.S. Environmental Protection Agency, P.O. Box 93478, Las Vegas, Nevada 89193
d
Research Triangle Institute, P.O. Box 12194, Research Triangle Park, North Carolina 27709
b
As part of the National Human Exposure Assessment Survey ( NHEXAS ) , residential pesticide storage and use patterns were evaluated in a population - based
sample of Minnesota households with children aged 3 ± 13. In - home interviews and inventories were conducted to identify pesticide products stored and used
in and around 308 households. This statistically based sample represents more than 49,000 urban and rural households in the census tracts sampled. More than
850 unique products were identified using Environmental Protection Agency ( EPA ) registration numbers. Pesticide products were found in 97% and reported
used in 88% of study households. Population - weighted mean values for pesticide storage and use were 6.0 and 3.1 products per household, respectively. The
most common active ingredients found were diethyl toluamide ( DEET ) and related compounds, piperonyl butoxide, pyrethrins, dimethylamine 2 - [ 2 methyl - 4 - chlorophenoxy ] propionate ( MCPA ) and chlorpyrifos. Household socio - demographic characteristics explained little of the variability in pesticide
storage and use patterns, and there were no significant differences in residential storage and use patterns between households located in urban versus non - urban
census tracts. Although the prevalence of households with pesticide products was similar to recent national surveys, observed storage and use rates were almost
twice those obtained in recent national studies, reflecting improved inventory techniques used by this study and / or increased rates of pesticide presence and use
in study households. Journal of Exposure Analysis and Environmental Epidemiology ( 2000 ) 10, 159 ± 167.
Keywords: children, National Human Exposure Assessment Survey ( NHEXAS ) , pesticide active ingredients, residential pesticide use.
Introduction
Although pesticides are widely used in the United States,
relatively few population - based studies have quantified
non -occupational exposure to these compounds. Most past
studies have attempted to evaluate potential residential
exposures through surveys of pesticide storage and use
( Finklea et al., 1969; Savage et al., 1981; Kamble et al.,
1982; Bennett et al., 1983; Davis et al., 1992; Whitmore et
al., 1992 ). One study, the Non -Occupational Pesticide
Exposure Study ( NOPES ) , augmented survey data with
measures of pesticides in environmental media inside
residences (Whitmore et al., 1994 ) . Given the limitations
of available data and increasing concerns about children's
non -dietary exposures to pesticides (National Research
Council, 1993 ), there is an acute need to better understand
residential pesticide exposures.
1. Address all correspondence to: Dr. John L. Adgate, School of Public
Health, University of Minnesota, Box 807 UMHC, Minneapolis, MN
55455. Tel.: ( 612 ) 624 - 2601. Fax: ( 612 ) 626 - 0650.
E-mail: [email protected]
Received 6 April 1999; accepted 16 November 1999.
This paper examines household pesticide inventory data
collected as part of the Minnesota Children's Pesticide
Exposure Study ( MNCPES ), which is described in
Quackenboss et al. (2000 ). The MNCPES is a purposeful
Phase III special study that was part of the National Human
Exposure Assessment Survey ( NHEXAS ) (Sexton et al.,
1995 ). MNCPES was designed, in part, to examine the
relationship between relatively simple means of exposure
estimation, such as surveys, and more complex and time consuming physical measurements of pesticide concentrations, such as measurements in environmental media and
human tissues. This manuscript describes analysis of
pesticide inventory (survey ) data; analyses of measurements in environmental media and human tissues will be
described in subsequent manuscripts by MNCPES collaborators. The statistically based survey was designed to
identify and preferentially select a higher proportion of
households in which children were likely to experience
exposures to atrazine, chlorpyrifos, diazinon, and malathion. It was postulated that children living in these
homes and participating in the MNPES were more likely to
have measurable concentrations of pesticides or pesticide
metabolites in their urine than children living in homes
with fewer pesticides available. As described in Quack-
Adgate et al.
enboss et al. (2000 ), household -level sample weights
were developed to adjust for the process of oversampling
some households. The weights were incorporated into this
analysis so inferences could be drawn about the prevalence
and use of pesticide products in the census tracts sampled.
The subsequent discussion describes our analysis of
these data, which allowed us to: 1) characterize the
presence and reported use of pesticides in this sample of
Minnesota households with children; 2 ) determine prevalence of product types, chemical classes, and active
ingredients; and 3 ) examine relationships between demographic characteristics and pesticide storage and use
patterns. These findings allow us to make inferences about
current pesticide storage and use patterns in the census tracts
sampled, and provide insight into the strengths and
limitations of this approach for use in exposure analysis.
Methods
This study used a telephone survey and in - home interviews
to collect data on pesticide storage and use patterns in
selected urban and non -urban households. The telephone
survey confirmed the eligibility of residences and participants for the study and identified households reporting
frequent or regular pesticide use, as described in Quackenboss et al. (2000 ). A subsample of these respondents
was selected for in - home visits, which included completion of a questionnaire on household characteristics and
occupant demographics, a technician inventory of pesticide
products present in and around the residence, and subject
recollection of products used in the year previous to the
interview. These data were used to select participants for
intensive environmental and biological sampling in the
MNCPES.
Sample Selection
The following briefly summarizes the important features of
the study design; a detailed description of the sample
selection process is found in Quackenboss et al. (2000 ).
The study population was restricted to ``urban'' residences
in the cities of Minneapolis and St. Paul, MN, and to nonurban areas of Goodhue and Rice Counties, MN, located
just south of the metropolitan area. A list of phone numbers
for households predicted to have children between 3 and 12
years of age was obtained from a commercial vendor. Innercity census tracts were oversampled to reduce underrepresentation of inner-city neighborhoods, which could
result from initiating selection from a list including only
households with phone numbers. Non - urban households
were defined as those located in non- urban census tracts in
Rice or Goodhue Counties. Initially, only those households
obtaining drinking water from private wells were eligible
160
Pesticides in Minnesota households
for inclusion in the non- urban cohorts; however, this
requirement was later dropped for households in Goodhue
County.
In- Home Data Collection
Participant interviews and pesticide inventories were
conducted between May and August 1997 by field
personnel from the Minnesota Department of Health and
the Community Health Service of Goodhue and Wabasha
Counties. Respondents were parents or guardians of the
resident child. If repeated restatement of a question did not
result in a response, non- leading probes were employed and
``Don't Know'' was offered as an acceptable response.
Participants were asked: ``Do you currently have any
pesticide products used to control for insects or weeds in or
around your home?'' Simultaneous with this question, the
interviewer handed the respondent a card listing types of
pesticide products (e.g., pet collar, insect baits, repellents)
and types of pests. Several techniques were used to ensure a
comprehensive inventory of pesticide products, including
interviewer training, use of ``prompt cards'' to help
respondents recall pesticide product types, and repeated
prompting to identify all storage areas inside and outside the
home. Disinfectants were specifically excluded from the
inventory due to concerns that inclusion might increase the
length of the interview and thus discourage participation in
later phases of the MNCPES. Agricultural pesticides were
also excluded from the inventory, unless they had been used
in or around the home.
Technicians recorded the name and Environmental
Protection Agency ( EPA ) registration numbers of all
pesticide products found in and around the residence.
Duplicate products within a household and product volumes
were not recorded. As each product was recorded,
respondents were asked whether the product had been used
during the past year.
Data Analysis
Pesticide products were entered into an electronic database,
facilitated by a specially designed program that matched
EPA registration numbers of inventoried products to
pesticide products registered for sale in Minnesota at any
time since 1983 (Minnesota Department of Agriculture,
1996 ). In a few instances, the EPA registration number
could not be identified and an exact of product name was
used to match products found to those in the MDA database.
Product types (i.e., insecticide, insect repellent, herbicide, etc. ) were determined based on the primary use on the
label and ancillary information from the MDA database.
Active ingredients for inventoried products were identified
by matching EPA registration numbers to a pesticide
database maintained by the California Environmental
Protection Agency (California Environmental Protection
Agency, 1997 ). Active ingredients were also assigned to
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Pesticides in Minnesota households
one of 50 chemical classes ( e.g., pyrethrins and pyrethroids,
diethyl toluamide [ DEET ] and related compounds, synergists, organophosphates, etc. ) based on chemical structure
or a common mechanism of toxicity, as described by
Morgan ( 1989 ).
Household level weights were developed using a process
described in Quackenboss et al. (2000 ). These weights
were used to adjust the data so that the participating
households represented the overall population of the census
tracts sampled. Tabulations of chemical classes and active
ingredients were restricted to categories found in at least 50
households because the weighted percentages become
unreliable below this level, i.e., the error associated with
the measurement becomes a significant proportion of the
weighted percentage.
The questionnaire and inventory data files were converted to SAS files and checks were made for consistency,
completeness, and coding errors. Unweighted data summary statistics, weightings of product types and active
ingredients, and multivariate analyses of unweighted data
were calculated using Statistical Analysis Systems (1989 ).
Since data were skewed, multivariate analyses were
conducted using log and square root transformations to
comply with standard regression assumptions. Weighted
pesticide storage and use values, standards errors, and 95th
percent confidence intervals (95th% CIs) were computed
using SUDAAN (Shah et al., 1997 ).
Results
Study Population
Participant response rates and demographic characteristics
are shown in Table 1. Sample households were selected
from a commercial phone list of 2303 households expected
to have children aged 3 to 12 years, although the final
sample included children aged 3 ±13 years ( Quackenboss
et al., 2000 ) . Using the statistical weightings developed as
part of MNCPES, the 308 participating households
represent 49,091 households with children in the census
tracts sampled. Inventories were completed for 70% of
selected households. The oversampling of inner- city
census tracts was effective in selecting non -white and
Hispanic participants in the study population at rates
equivalent to their prevalence in the overall Minnesota
population, as projected from the 1990 census. Selected
households that did not participate in the inventory were
not appreciably different from the inventoried households
with regard to household size or urban /non -urban status. A
large proportion of the study households was single -family
residences, however, and only 16 ( 5.2% ) households
reported being renters. One third of the non -urban
households indicated that they resided in working farms.
A majority of the non - urban households were located in
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Adgate et al.
Table 1. Response rates and demographic characteristics of the surveyed
population.
Households or
residents ( N )
Percentage of
respondentsa
Response rates
Home inventory
appointments scheduled
335
73b
Home inventory
appointments completed
308
92c
1209
87
Race d
White
Black / African American
Hmong
61
55
4.4
3.9
Other Asian or
Pacific Islander
15
1.1
8
0.6
50
3.6
Native American
Other
Ethnicity d
Non - Hispanic
1360
Hispanic
38
97
2.7
Household location
Minneapolis or St. Paul
Goodhue or Rice County
not on working farms
Goodhue or Rice County
households on working farms
House type
Single family homes
Apartments or mobile homes
225
55
73
18
28
9
302
98
6
1.9
Reported household income ( US$ )
3
1.0
10,000 ± 19,999
20,000 ± 29,999
< 10,000
10
29
3.2
9.4
30,000 ± 39,999
31
10
40,000 ± 49,999
53
17
50,000 ± 74,999
112
36
75,000 ± 99,999
37
12
> 100,000
25
8.1
Don't know
2
0.7
Refusal
6
1.9
a
Totals within each descriptive category do not always add to 100% due to
rounding.
b
Response rate calculated based on a denominator n = 477, i.e., households
meeting the geographic and age - eligible child criteria and were invited to
participate, as described in Quackenboss et al. ( 2000 ) .
c
As a percentage of scheduled visits.
d
Number of residents in each category.
small communities, or were on small parcels of land that
were not themselves part of a farm. The reported median
household income was more than US$56,000, substantially
161
Adgate et al.
Pesticides in Minnesota households
greater than the average of median income for Minnesota in
1995 and 1996, i.e., approximately US$40,000 (U.S.
Census Bureau, 1998 ).
Weighted Survey Sample Results
A total of 2058 individual pesticides, corresponding to more
than 850 unique pesticide products as determined by EPA
registration number and product name, were inventoried in
the participating households ( Table 2 ). Products used
within the past year represented at least 482 different brand
names and 392 separate chemical formulations. An EPA
registration number was identified for all but 39 products
( 1.9% ) . Although no registration number was available for
these 39 products, they were included in tabulations of
products found, products reported used, product types and
active ingredients, if this information was available from
their label. Nearly all households had at least one pesticide
product found and reported used, only nine households had
no pesticide products, five of which were non -urban.
Respondents indicated that more than half the products
found in residences had been used within the past year. The
distribution of the total number of products found and used
per household was positively skewed, with a maximum of
45 products found in one home, although the maximum
number of products reported used in the previous year was
16 ( Figure 1 ).
Although the proportion of households with products
found was slightly higher for urban than non -urban
households, there were no statistically significant differences between them for prevalence or for the total number
of products found or reported used during the preceding
year ( Table 3) .
Product Types and Chemical Constituents
Product Types Table 4 presents a tabulation of product
types, e.g., insecticides, herbicides, and repellents. More
than half of the pesticide products found and nearly half of
the products reported used were insecticides, while the next
most common product types were insect repellents and
herbicides. The top three categories comprised nearly 95%
of the products reported used in the previous year. Of the
pesticide products found in study households, insect
repellents were the most likely to have been used in the
past year (Table 4) .
Chemical Classes Pyrethrins and pyrethroids were the most
commonly found class of chemicals, but DEET and related
Table 2. Pesticide storage and use values in study households.
Variable
Weighted
percentage
or mean
Standard
error
Lower bound
299
97.6%
1.02%
95.6%
Number of products found ( all households )
2058
5.99
0.32
5.36
6.62
Number of products found ( among households with products )
2058
6.14
0.32
5.51
6.77
Households reporting use in previous year
278
88.4%
83.6%
93.2%
Number of products used ( all households )
1083
3.12
0.18
2.77
3.47
Number of products used
( among households using products )
1083
3.53
0.18
3.18
3.88
Urban households with pesticide products
Number of products found ( all households )
221
1451
98.0%
5.98
Number of products found ( among households with products )
1451
6.10
Urban households reporting use
203
88.3%
Number of products used ( all households )
767
3.12
0.19
2.75
3.49
Number of products used ( among households using products )
767
3.53
0.19
3.16
3.90
Non - Urban households with pesticide products
Number of products found ( all households )
78
607
93.1%
6.16
86.5%
5.00
99.7%
7.32
Number of products found ( among households with products )
607
6.62
5.42
7.82
80.5%
97.0%
Households with pesticide products
Non - Urban households reporting use
Unweighted
count
( n = 308 )
2.45%
1.06%
0.34
0.34
2.61%
3.36%
0.59
0.61
95.9%
5.31
Upper bound
99.6%
100%
6.65
5.43
6.77
83.2%
93.5%
75
88.7%
Number of products used ( all households )
316
3.11
0.32
2.48
3.74
Number of products used ( among households using products )
316
3.50
0.34
2.83
4.17
162
4.20%
95% Confidence interval
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Pesticides in Minnesota households
Adgate et al.
Figure 1. Frequency distribution of pesticide products found and reported used in Minnesota households with children.
compounds were the most commonly used class of
chemicals. DEET and related compounds, which are insect
repellents, were reported used in almost half of study
households. They were found in 637 ( 31% ) individual
products and reported used in 445 ( 22% ) of the individual
products tabulated. In descending order, the next most
common classes of compounds used were pyrethrins and
pyrethroids, organophosphates, synergists, chlorophenoxy
herbicides, and carbamates, all of which were reported used
in more than 20% of study households.
Active Ingredients The top 18 active ingredients found and
reported used are shown in Table 5. A total of 166 active
ingredients were identified in the 2058 inventoried products,
and 107 of these ingredients were reported used in the year
preceding the inventory. DEET was found in more than half
of study households. DEET and isomers of DEET were
reported used most frequently within the past year. Other
common active ingredients include the synergist, piperonyl
butoxide, and plant - derived pyrethrins, which were found
in more than 40% and reported used in 25% of study
households. Fourteen of the top 18 active ingredients were
found in at least 20% of sampled households, and 15 of
the top 18 common active ingredients were reported used
in at least 10% of study households. Active ingredients
could not be identified for 49 ( 2.4% ) of the products
found.
Banned and Restricted Chemicals Banned pesticides were
found in 28 (9% ) households in the inventory. A sodium
arsenate - containing ant control product banned in 1989 was
found in 23 (7.5% ) and reported used in 20 ( 6.5% ) of
Table 3. Urban versus non - urban homes: differences in pesticide storage and use patterns.
Mean difference: urban
minus non - urban
Standard
error
t - Statistics
p - value
Households with pesticide products
4.88%
3.52%
1.39
0.167
Number of products found ( all homes )
0.18
0.68
0.27
0.789
Number of products found ( among homes with products )
0.52
0.70
0.74
0.459
Households reporting use in previous year
0.38%
4.95%
0.08
0.938
Number of products used in previous year ( among all homes )
0.01
0.37
0.02
0.980
Number of products used in previous year ( among homes using products )
0.03
0.39
0.07
0.947
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
163
Adgate et al.
Pesticides in Minnesota households
Table 4. Summary of pesticides found and reported used by product type.
Product type
N
Insecticides
Useda
Found
Weighted
percentage
of totalb
N
Weighted
percentage
of total
1052
52
539
49
479
369
21
22
306
165
27
18
Rodenticides
59
1
37
2
Fungicides
51
2
19
1
Otherc
48
3
17
2
Totals
2058
Insect repellents
Herbicides
1083
a
Reported used in the previous year.
Totals do not always add to 100% due to rounding.
c
Includes molluskicides, growth regulators, vertebrate repellents, and other
special use product types.
b
participating households. Despite being banned from sale
for more than 7 years, this particular product was the fifth
most commonly used individual pesticide formulation.
Three households had products containing silvex, and one
household had products containing DDT, atrazine, and
2,4,5 -T. Only one of these six pesticide products, which had
silvex as its active ingredient, was reported used in the year
preceding the inventory.
Socio -Demographic Factors and Pesticide Storage and
Use
Multivariate regression analysis was used to examine
whether demographic characteristics were predictive of
the dependent variables, household pesticide storage and
use values. Independent variables included the demographic
characteristics listed in Table 1, as well as parental level of
education, household size, home type, and home ownership
data obtained from the household screening questionnaire
( Quackenboss et al., 2000 ). The numbers of pesticides
found and used in and around the home were only weakly
and inconsistently related to socio - demographic factors.
Thus, for this population, the demographic characteristics
did not predict potential presence or use of pesticides by
households.
Discussion
In order to increase the likelihood of obtaining detectable
concentrations of chlorpyrifos and other MNCPES target
pesticides in environmental media and biological samples,
this study oversampled frequent pesticide users living in
households with children. Since the oversampling was
done intentionally, and the probabilities were assessed at
each level of selection, it was possible to weight the sample
results and thereby estimate the characteristics of the
164
49,091 households represented by this survey. These
results represent storage and use patterns in a relatively
large population of households that have children between
the ages of 3 and 13.
While our population mirrored the racial makeup
projected from the 1990 census, reported income levels
were greater than the average of median income for
Minnesota by nearly 40%. This may be due to use of a
commercial phone list, selection of frequent pesticide users,
or by biased reporting of household income by study
subjects. Despite these potential limitations, our findings are
in substantial agreement with previous population -based
studies reporting prevalence of households with pesticide
products. Most previous studies of pesticides in the home
have not focused exclusively on households with children.
Some regional studies (Kamble et al., 1982; Bennett et
al., 1983; Davis et al., 1992 ) have used telephone surveys,
while others ( Finklea et al., 1969; Savage et al., 1981;
Whitmore et al., 1992 ) conducted in - home inventories
similar to the one used in this study. Reported use rates in
these studies range from 73% to 98%, although in some
cases, this includes yard and garden uses, as well as indoor
use. In a population -based survey of more than 8000
households conducted in 1976 ±1977, Savage et al. ( 1981 )
Table 5. Household storage and use patterns for common active
ingredients.
Active ingredient
Households found
Households used
N
Weighted
percentage
of total
N
Weighted
percentage
of total
DEET
196
58
162
47
Piperonyl butoxide
152
45
91
25
Pyrethrins
147
43
88
25
DEET- related isomers
123
35
108
30
MCPA
107
35
55
17
Permethrin, mixed cis, trans
93
24
65
15
Chlorpyrifos
89
29
55
17
Propoxur
Octyl bicycloheptene
dicarboximide
84
83
25
25
53
43
17
12
Allethrin
81
24
49
13
2,4 - D (Dichlorophenoxyacetic acid)
Borax ( sodium tetraborate
decahydrate )
74
23
37
11
73
19
56
14
2,4 - D, dimethylamine salt
69
23
36
12
Diazinon
65
18
37
11
Glyphosate, isopropylamine
salt
Tetramethrin
62
18
37
12
62
18
32
8.5
Resmethrin
60
20
24
8.1
Carbaryl
50
14
24
5.4
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Pesticides in Minnesota households
reported that 91% of households used pesticides in their
house, garden, or yard. In EPA Region V, which includes
Minnesota, 79% of households reported using pesticides in
their households in the previous year; households averaged
1.7 (range 1.3± 2.3 ) containers of pesticides in their
households. A more recent national probability - based in home survey of 2447 households conducted in 1988 found
an average of 3.8‹ 0.5 pesticide products (95th% CI: 3.34 ±
4.34 ) ( Whitmore et al., 1992 ) . That survey, however,
included disinfectants in its inventory of pesticide products.
Our weighted results differ significantly from previous
studies: homes in our sample had approximately two times
higher average numbers of pesticides present and reported
used than the two largest previous studies. A number of
factors relevant to study design and analysis, however,
complicate attempts to compare our results to previous
studies. First, relatively few studies have been performed:
three have been national in scope, while the others have
been regional or local. Variability in climate, pests, and
cultural attitudes on application of chemicals and tolerance
of pests confounds comparisons across studies. Second,
temporal trends may also confound comparison. Although
household pesticide expenditures have remained relatively
constant in dollar terms over the period 1979 ± 1995, the
number of people paying for outdoor professional treatments has been growing, especially for relatively wealthy
households with large yards (Templeton et al., 1998 ).
Third, both sample selection and study design focused on
other goals may affect inventory results in unknown ways.
In MNCPES and NOPES ( Whitmore et al., 1994 ),
participants were selected to increase the probability of
finding measurable pesticide concentrations. The comparison of these studies with inventories conducted as part of
case ±control epidemiological studies of childhood cancer
may be instructive for comparing common product types or
active ingredients or to try to assess changes in trends
between geographic regions, but cannot be used for more
rigorous comparisons. Fourth, studies have used different
methods for naming and quantifying active ingredient use.
Fifth, for purposes of tabulation, some studies appear to
combine related active ingredients, ( e.g., dichlorophenoxyacetic acid [ 2,4- D ] with its dimethylamine salt )
although these are usually listed as separate ingredients on
labels. Therefore, the variation between our results and
previous studies may reflect our sampling techniques,
regional variation in use, the definition of ``pesticide
product'' used in the various surveys, actual changes in
pesticide storage and use over time, or some combination of
these factors.
Strengths and Limitations
This study is the largest population - based sample of
pesticide storage and use in households with children.
Although statistically robust, a number of limitations are
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Adgate et al.
inherent in the sample selection, data collection, and
analysis methods used in this study. First, potential
participants in the study were identified from a telephone
list purchased from a commercial vendor. Given that the
median reported income was considerably greater than
median income for Minnesota in 1995 and 1996, it appears
that higher -income families may be more likely to engage
in activities that result in their inclusion in commercial
mailing lists. In addition, lower income families are less
likely to have telephones, and thus may have been excluded
from the selection process at the outset. While oversampling
``inner city'' census tracts to include non -Whites and
Hispanics at rates approximately equal to their proportions
in the 1990 census was successful, the ultimate study
population had disproportionately low numbers of families
who lived in non -single family detached homes or who
rented rather than owned.
The study population was divided between ``urban'' and
``non -urban'' households in order to achieve an underlying
objective of MNCPES: to examine the relative contribution
of various sources, such as drinking water, in geographical
areas where supplies are expected to offer different potential
exposures. In this study, ``non -urban'' status was determined by the 1990 census data. Initially, only households
with private wells were included in the non- urban cohorts;
eventually, however, this requirement was dropped for
Goodhue County due to an insufficient number of private
wells. Given the relatively small number of working farms
in this survey, our results are not representative of farm
children's potential non -dietary exposures from working on
or living near agricultural production. These results are
representative, nevertheless, for potential residential exposures for the population in the census tracts sampled.
A number of potential forms of bias are inherent in
survey study designs. Interviewers relied on respondents to
show them pesticide products stored in and around the
residence and to provide information on product use.
Therefore, results may be subject to recall bias introduced
by the timing of the interview and information bias arising
from imperfect knowledge, memory, or unwillingness to
reveal all products or use rates. This concern was partly
alleviated through use of the pesticide product prompting
card presented to participants during the survey. Knowledge
bias may have occurred because respondents might not have
associated certain pest control products with the term
``pesticide'' or known of use by other household members.
Products that are not packaged similarly to other pesticide
products, are removed from their packaging for use, and /or
are not stored with other pesticide products (e.g., flea
collars, shampoos, and moth balls) seem most likely to have
been inadvertently omitted. Prevarication bias may have
occurred because respondents did not want to inform the
interviewer of all pesticide products or product use in the
house due to fear of disapproval, adverse consequences, or a
165
Adgate et al.
desire to terminate the interview quickly. Lastly, because
pesticide use is seasonal in the upper Midwest, subject
recollections may have been more accurate for respondents
contacted later in the survey period ( May ± August ). Since
most previous studies have used a similar time frame and
sampling techniques, these biases are likely relatively
consistent across studies.
The process used for identification of product active
ingredients in this study was generally successful: <7%
were unidentifiable. The primary reason for failure to
identify active ingredients was lack of a valid EPA
registration number due to a missing, torn or obscured
label, often as a result of product age. As a result, products
for which active ingredients could not be identified may be
more likely to contain ingredients subsequently banned or
more restrictively regulated. An additional limitation was
the lack of information requested on how much, where, and
how recently each product had been used. These data were
not collected because of insufficient time to complete a
more in -depth questionnaire for each household. Thus, we
have product -specific information only for the presence and
use of a product during the past year. Some non -product specific information about the frequency and location of
insect control inside and outside the residence, or in the
lawn and garden was collected as part of a Household
Screening Questionnaire, which will be reported in a
subsequent paper.
As new studies of residential pesticide storage and use are
conducted, our data will be useful for making comparisons
across geographic regions and in assessing changes in
residential pesticide storage and use over time. In addition,
these data also will useful to help evaluate relationships
between household pesticide storage and use, pesticide
residues in environmental media, and body burden. Results
from the intensive environmental monitoring phase of
MNCPES, which was conducted in a subset of 102
households participating in this survey, emphasize measurements of pesticide residues in air, food, water, soil, house
dust, and children's urine (Quackenboss et al., 2000 ).
Conclusions
o In this statistically based sample of Minnesota households with children, pesticide products were found in 97%
and reported used in 88% of residences in the previous year.
These prevalence rates are similar to most previous
population -based surveys.
o The population -weighted mean number of products
stored (5.99; 95th% CI: 5.36, 6.62 ) and used (3.12; 95th%
CI: 2.77, 3.47 ) were almost twice most recent national
surveys. These differences may reflect actual temporal shifts
or geographic variation in pesticide storage and use patterns,
or differences in survey techniques.
166
Pesticides in Minnesota households
o There were no significant differences in the number of
pesticide products stored or used between urban and nonurban households. Since less than half the non -urban
households were on working farms, this inventory provides
limited information about residential pesticide storage and
use rates for farm families.
o Identification of active ingredients in a wide variety of
products was easily obtained via a database available on the
Internet, which allowed for identification of commonly used
active ingredients: DEET and related compounds, piperonyl
butoxide, pyrethrins, dimethylamine 2- [ 2- methyl -4 chlorophenoxy ] propionate (MCPA ) , permethrin, and
chlorpyrifos.
o With the exception of one sodium arsenate - containing
product, banned and restricted pesticides were found in
fewer than 2% of participating households. Despite being
banned from sale for more than 7 years, this particular brand
was the fifth most commonly used individual pesticide
formulation.
o Socio -demographic factors (i.e., race, income, etc. ) did
not explain variations in pesticide storage or use patterns in
this statistically based sample of Minnesota households with
children.
Acknowledgments
We thank the families who participated in the survey for
their cooperation, and Pam Holst and Lori Anderson of the
Community Health Service of Goodhue and Wabasha
Counties for their assistance in the field, and the Minnesota
Department of Agriculture for providing technical assistance on use of its electronic registry of pesticide products.
This study was funded, in part, through the NHEXAS
Pesticides and PAH Module Cooperative Agreement
R821902 between the U.S. EPA and the consortium of
Research Triangle Institute / Environmental and Occupational Health Sciences Institute (RTI /EOHSI ), by U.S.
EPA STAR grant R825283 to the University of Minnesota,
and by a grant from the Legislative Commission on
Minnesota Resources.
References
Bennett G.W., Runstrom E.S., and Wieland J.A. Pesticide use in homes.
Bull. ESA 1983: 64: 31 ± 38.
California Environmental Protection Agency ( Cal. EPA ) . Department of
Pesticide Regulation. ``USEPA / OPP Pesticide - Related Database
Queries.'' USEPA / OPP Pesticide Products Database ( Oct. 1997 ) .
Online: http: / / www.cdpr.ca.gov / docs / epa / epamenu.htm.
Davis J.R., Brownson R.C., and Garcia R. Family pesticide use in the
home, garden, orchard, and yard. Arch. Environ. Contam. Toxicol.
1992: 22: 260 ± 266.
Finklea J.F., Keil J.E., Sandifer S.H., and Gadsden R.H. Pesticides and
pesticide hazards in urban households. J. South Carolina Med. Assoc.
1969: 65: 31 ± 33.
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Pesticides in Minnesota households
Kamble S.T., Gold R.E., and Parkhurst A.M. Nebraska residential
pesticide use survey ( 1979 and 1980 ) . The Agricultural Experiment
Station, Institute of Agriculture and Natural Resources, University of
Nebraska, 1982.
Minnesota Department of Agriculture ( MDA ) . Minnesota Pesticide
Information Retrieval System ( MNPIRS ) . State of Minnesota, St.
Paul, MN, 1996.
Morgan D.P. Recognition and Management of Pesticide Poisonings, 4th
edn. U.S. Environmental Protection Agency Washington, DC. EPA 540 / 9 - 88 - 001 1989.
National Research Council ( NRC ) . Pesticides in the Diets of Infants and
Children. National Academy Press, Washington, DC, 1993.
Quackenboss J.J., Pellizzari E.D., Shubat P.J., Whitmore R.W., Adgate
J.L., Thomas K.W., Freeman N.C.G., Stroebel C., Lioy P.J., Clayton
C.A., and Sexton K. Design strategy for a multipathway pesticide
exposure study in children. J. Expos. Anal. Environ. Epidemiol. 2000:
10: 146 ± 159.
Savage E.P., Keefe T.J., Wheeler H.W., Mounce L., Helwic L., Applehans
F., Goes E., Goes T., Mihlan G., Rench J., and Taylor D.K. Household
pesticide usage in the United States. Arch. Environ. Health 1981: 36:
304 ± 309.
Sexton K., Kleffman D.E., and Callahan M.A. An introduction to the
Journal of Exposure Analysis and Environmental Epidemiology (2000) 10(2)
Adgate et al.
national human exposure assessment survey ( NHEXAS ) and related
phase I field studies. J. Expos. Anal. Environ. Epidemiol. 1995: 5 ( 3 ) :
229 ± 232.
Shah B.V., Barnwell B.G., and Bieler G.S. SUDAAN User's Manual.
Release 7.5. Research Triangle Institute, Research Triangle Park, NC,
1997.
Statistical Analysis Systems ( SAS ) . Version 6.12. 1989 ± 1996.
Templeton S.R., Zilberman D., and Yoo S.J. An economic perspective on
outdoor residential pesticide use. Environ. Sci. Technol. 1998: 32 ( 17 ) :
416A ± 423A.
U.S. Census Bureau. U.S. Census Bureau, the Official Statistics. ``1996
Income'' ( April 2, 1998, last update ) . Online: http: / / www.census.gov / hhes / income / income96 / in96med1.html.
Whitmore R.W., Kelly J.E., and Reading P.L. National Home and Garden
Pesticide Use Survey. Final Report, Vol. I: Executive Summary,
Results, and Recommendations. Prepared by Research Triangle
Institute for U.S. EPA Office of Pesticides and Toxic Substances.
NTIS PB92 - 174747INZ, 1992.
Whitmore R.W., Immerman F.W., Camann D.E., Bond A.E., Lewis R.G.,
and Schuam J.L. Non - occupational exposures to pesticides for
resident of two U.S. cities. Arch. Environ. Contam. Toxicol. 1994:
26: 47 ± 59.
167