EPA/600/R-04/109
September 2004
Demonstration of Low Cost, Low Burden, Exposure
Monitoring Strategies for Use in Longitudinal Cohort
Studies
Volume 1 - Final Report
by
Jerry D. Rench, RTI Task Order Leader, James H. Raymer, Lisa Thalji, Michelle Spruill,
Cynthia A. Salmons, Larry C. Michael, Monica J. Pecha, Elizabeth Dean, Gerry Akland
RTI International
3040 Cornwall! s Rd
POBox 12194
Research Triangle Park, NC 27709-2194
EPA Contract No. 68-D-99-012
Task Order 0014
RTI Project No. 07505.014
EPA Task Order Project Officer: Roy Fortmann
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and
Development funded and managed the research described here under Contract Number
68-D-99-012 to RTI International. It has been subjected to the Agency's peer and
administrative review and has been approved for publication as an EPA document
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
Abstract
The Federal Government is currently planning a large, prospective birth cohort
study known as the National Children's Study that will potentially involve 100,000
children and their families. The observation period will start as close to conception as
possible and will continue for 20 years after birth. Given the magnitude and expense of
such a large study, sample collection methods that are amenable to acquisition of samples
exclusively by the participants themselves followed by direct shipment to the analysis
laboratory present a cost-effective alternative to technician-based sampling procedures. In
this pilot study, the ability of participants in three age cohorts to collect environmental
and biological samples according to prescribed protocols was evaluated. The cohorts
consisted of parents and their children in the ages of 0-1 year, 3-5 years, and 6-8 years
old. Biological and environmental samples collected during the study included urine,
hair, saliva, breast milk, duplicate diet, tap water, vacuum cleaner dust, floor surface dust
wipes, air samples, cotton sock dosimeters, and humidity/temperature measurements.
Sample collection instructions and materials were prepared, subjected to evaluation and
modification using a test population, and shipped to participants over a 12-month period.
Participants were requested to collect the samples, complete questionnaires, and return
the samples to the laboratory within defined time periods. Upon receipt at the laboratory,
the condition of the samples was assessed by visual inspection and the details of the
receipt and evaluation were logged into a computer database; queries were subsequently
used to assess compliance. In some cases, chemical analysis was used to further evaluate
sample integrity.
The demonstration studies generated considerable information that favorably
supports sample collection by study participants and remote data collection via the web,
although the studies brought forward a number of issues that can impact a large-scale
study such as the NCS. Completion rates for the different on-line surveys were 73% or
better. For a relatively complex survey including pesticide use, the response rate was
92%. There was a 96% response rate for completion of a time/activity diary related to the
participant child's exposure to pesticides in the home. Participant compliance with
sample collection instructions appeared to be good for most sample types. One hundred
percent of the hair samples and vacuum cleaner dust samples, for example, were
considered to be acceptable. Acceptability rates were greater than 85% for breast milk,
urine and food. But the number of acceptable beverage (diet) and tap water samples was
lower due to return of leaking containers. The number of acceptable samples was
substantially lower for the more complex sampling methods. All of the VOC air sampling
badges were returned to the laboratory, but only 56% of the samples were determined to
be acceptable. Although an instructional video was included with the badge, the
participants found the sample collection method to be too complex. Samples were
generally returned in a timely manner, in compliance with the instructions. Results of this
project are very encouraging, indicating that remote data collection by study participants
is feasible. Results can be used to develop strategies that will maximize completeness of
sample collection while minimizing participant burden and study costs.
-------
Table of Contents
Section Page
ES Executive Summary ES-1
Key Findings for the 0-1 Cohort ES-2
Key Findings for the 3-5 Cohort ES-4
Key Findings for the 6-8 Cohort ES-4
1 Introduction 1-1
2 Study Methods 2-1
2.1 Overview of Study Design 2-1
2.2 Web-enabled Data Collection and Panel Description 2-5
2.2.1 Sampling Methodology for the Web-enabled Panel 2-5
2.2.2 Recruitment Procedures for the Web-enabled Panel 2-6
2.2.3 Survey Administration Procedures for the Web-enabled Panel 2-7
2.3 Selection and Recruitment of Study Subjects 2-8
2.3.1 Sample Selection of Study Subjects 2-8
2.3.2 Recruitment of Study Subjects 2-10
2.4 Study Methods: 0-1 Cohort - Breast Milk, Food, Beverage and Urine ..2-10
2.4.1 Questionnaires, Diaries, and Instructional Materials 2-10
2.4.2 Biological and Environmental Sample Collection 2-11
2.4.3 Compliance Monitoring and Sample Analysis 2-12
2.5 Study Methods: 3-5 Cohort - Urine/Socks, Hair/Vacuum Dust,
Tap Water 2-12
2.5.1 Urine and Socks 2-13
Questionnaires, Diaries, and Instructional Materials 2-13
Sample Collection 2-13
Compliance Monitoring and Sample Analysis 2-14
2.5.2 Hair and Vacuum Dust 2-14
Questionnaires and Instructional Materials 2-14
Sample Collection 2-15
Compliance Monitoring and Sample Analysis 2-15
2.5.3 Tap Water 2-16
Questionnaires and Instructional Materials 2-16
Sample Collection 2-16
Compliance Monitoring and Sample Analysis 2-17
2.6 Study Methods: 6-8 Cohort - Saliva/Dust, VOC/HOBO (Light Intensity)
HOBO (Temperature/Relative Humidity) 2-17
-------
Section Page
2.6.1 Saliva and Dust Wipe 2-17
Questionnaires and Instructional Materials 2-17
Sample Collection 2-18
Compliance Monitoring and Sample Analysis 2-18
2.6.2 VOC Badge and HOBO/Light Intensity 2-18
Questionnaires and Instructional Materials 2-19
Sample Collection 2-19
Compliance Monitoring and Sample Analysis 2-20
2.6.3 HOBO (Temperature) 2-21
Questionnaires and Instructional Materials 2-21
Sample Collection 2-21
Compliance Monitoring and Sample Analysis 2-22
2.7 Web Implementation of Questionnaires and Other Materials 2-22
2.8 Usability Testing 2-24
2.9 Protection of Study Subjects 2-26
2.10 Incentive Schedule 2-26
3 Results and Conclusions 3-1
3.1 Observations of the 0-1 Cohort 3-2
3.1.1 Recruitment and Retention 3-2
3.1.2 Web Data Collection: Metaquestionnaire, Food Diary and
Debriefing Questionnaire 3-4
Metaquestionnaire 3-5
Food Diary 3-5
Debriefing Questionnaire 3-7
3.1.3 Environmental Sample Collection: Breast Milk, Beverage, and
Food Samples 3-8
3.1.4 Biological Sample Collection: Urine 3-8
3.1.5 Indices of Compliance with Study Protocol 3-13
Acceptability 3-13
Timeliness 3-14
3.1.6 Summary and Conclusions: 0-1 Cohort 3-20
3.2 Observations of the 3-5 Cohort 3-22
3.2.1 Recruitment and Retention 3-22
3.2.2 Web Data Collection: Metaquestionnaire, Activity Diary, and
Debriefing Questionnaire 3-24
Urine Metaquestionnaire 3-24
Sock Activity Diary 3-24
Hair and Dust Metaquestionnaire 3-26
Tap Water Metaquestionnaire 3-26
Debriefing Questionnaire 3-27
3.2.3 Environmental Sample Collection: Cotton Socks, Vacuum Dust,
and Tap Water 3-30
3.2.4 Biological Sample Collection: Urine and Hair 3-30
3.2.5 Indices of Compliance with Study Protocol 3-30
Acceptability 3-30
-------
Section Page
Timeliness 3-37
3.2.6 Summary and Conclusions: 3-5 Cohort 3-43
3.3 Observations of the 6-8 Cohort 3-48
3.3.1 Recruitment and Retention 3-48
3.3.2 Web Data Collection: Metaquestionnaire, and
Debriefing Questionnaire 3-48
Saliva and Dust, VOC Badge and HOBO (Temperature)
Metaquestionnaire 3-49
Debriefing Questionnaire 3-50
3.3.3 Environmental Sample Collection: Dust Wipe, Volatile Organic
Compound (VOC) Badge, and HOBO (Temperature) 3-51
3.3.4 Biological Sample Collection: Saliva 3-51
3.3.5 Indices of Compliance with Study Protocol 3-51
Acceptability 3-51
Timeliness 3-58
3.3.6 Summary and Conclusions: 6-8 Cohort 3-63
4 Quality Assurance Procedures and Results 4-1
4.1 Documentation 4-1
4.2 Participant Follow-up 4-1
4.2.1 Visits 4-2
4.2.2 Telephone Calls 4-2
4.3 Laboratory Activities 4-2
4.4 QA Observations and Conclusions 4-2
5 Recommendations 5-1
5.1 Answers to Key Questions 5-1
5.2 Recommendations 5-2
Study Planning 5-2
Study Initiation 5-4
Study Implementation 5-4
-------
Exhibits
Number Page
ES-1 Observations on Level of Burden: 0-1 Cohort ES-3
ES-2 Observations on Level of Burden: 3-5 Cohort ES-5
ES-3 Observations on Level of Burden: 6-8 Cohort ES-7
2-1 Samples and Metaquestionnaire Content Collected for 0-1 Cohort 2-2
2-2 Samples and Metaquestionnaire Content Collected for 3-5 Cohort 2-3
2-3 Samples and Metaquestionnaire Content Collected for 6-8 Cohort 2-3
2-4 Longitudinal Collection of Metaquestionnaire Data and Samples
by Cohort and Month 2-4
2-5 Target Days for Sampling and Web Data Collection for a Sampling Event 2-5
2-6 Web-enabled Panel Recruitment Procedures 2-6
2-7 Sampling Domains for Sample Selection 2-9
2-8 Nine Strata Per Cohort 2-9
2-9 Example Practice Drop Down Menu 2-23
2-10 Example Activity Diary 2-24
2-11 Incentive Schedule Per Study Subj ect for Sample Collection Activities 2-27
3-1 Criteria for Defining Sample Acceptability 3-3
3-2 Recruitment and Retention of Study Subjects by Income
Characteristic: 0-1 Cohort 3-4
3-3 Results of Questionnaire and Food Diary Data Collection: 0-1 Cohort 3-5
3-4 Web Data Collection for Breast Milk 3-6
3-5 Web Food Diary 3-7
3-6 Collection Summary and Condition of Breast Milk Samples: 0-1 Cohort 3-9
3-7 Collection Summary and Condition of Duplicate Food Samples: 0-1 Cohort. 3-10
3-8 Collection Summary and Condition of Duplicate Beverage
Samples: 0-1 Cohort 3-11
3-9 Collection Summary and Condition of Urine Samples
Received: 0-1 Cohort 3-12
3-10 Timeliness of Breast Milk Sample Collection: 0-1 Cohort 3-15
3-11 Timeliness of Food Sample Collection: 0-1 Cohort 3-16
3-12 Timeliness of Beverage Sample Collection: 0-1 Cohort 3-17
3-13 Timeliness of Urine Sample Collection: 0-1 Cohort 3-18
3-14 Temporality of Sample Collection and Metaquestionnaire Completion
by Sample Type: 0-1 Cohort 3-19
3-15 Samples Collected by Day of Week: 0-1 Cohort 3-19
3-16 Summary of Data and Sample Collection: 0-1 Cohort 3-21
3-17 Observations on Level of Burden: 0-1 Cohort 3-23
3-18 Recruitment and Retention of Study Subjects by Income
Characteristic: 0-5 Cohort 3-24
3-19 Results of Questionnaire and Diary Collection: 3-5 Cohort 3-25
3-20 Pesticides Applied in Home 3-25
3-21 Still Image of Pesticide Label 3-26
3-22 Sock Activity Diary Screen 3-27
-------
Number Page
3-23 Room Vacuum Cleaner Was Used 3-28
3-24 Source of Drinking Water 3-29
3-25 Sample Collection - 3-5 Cohort: All Sample Types 3-31
3-26 Collection Summary and Condition of Sock Samples: 3-5 Cohort 3-32
3-27 Collection Summary and Condition of Vacuum Dust Samples: 3-5 Cohort.... 3-33
3-28 Collection, Summary and Condition of Tap Water Samples: 3-5 Cohort 3-34
3-29 Collection Summary and Condition of Urine Samples: 3-5 Cohort 3-35
3-30 Collection Summary and Condition of Hair Samples: 3-5 Cohort 3-36
3-31 Timeliness of Socks Sample Collection: 3-5 Cohort 3-38
3-32 Timeliness of Urine Sample Collection: 3-5 Cohort 3-39
3-33 Timeliness of Vacuum Dust Sample Collection: 3-5 Cohort 3-40
3-34 Timeliness of Hair Sample Collection: 3-5 Cohort 3-41
3-35 Timeliness of Tap Water Sample Collection: 3-5 Cohort 3-42
3-36 Temporality of Sample Collection and Metaquestionnaire Completion
by Sample Type: 3-5 Cohort 3-44
3-37 Samples Collected by Day of Week: 3-5 Cohort 3-44
3-38 Summary of Data and Sample Collection: 3-5 Cohort 3-45
3-39 Observations on Level of Burden: 3-5 Cohort 3-47
3-40 Recruitment and Retention of Study Subjects by Income
Characteristic: 6-8 Cohort 3-48
3-41 Results of Questionnaire Collection: 6-8 Cohort 3-49
3-42 Hours and Minutes in Various Locations 3-50
3-43 Sample Collection - 6-8 Cohort: All Sample Types 3-52
3-44 Collection Summary and Condition of Settled Dust Wipe
Samples: 6-8 Cohort 3-53
3-45 Collection Summary and Condition of VOC Badge/HOBO Sampling
Device: 6-8 Cohort 3-54
3-46 Collection Summary and Condition of HOBO (Temperature)
Sample: 6-8 Cohort 3-55
3-47 Collection Summary and Condition of Saliva Samples: 6-8 Cohort 3-56
3-48 Timeliness of Dust Wipe Sample Collection: 6-8 Cohort 3-59
3-49 Timeliness of Saliva Sample Collection: 6-8 Cohort 3-60
3-50 Timeliness of VOC Badge/HOBO Sample Collection: 6-8 Cohort 3-61
3-51 Timeliness of HOBO (Temperature) Sample Collection: 6-8 Cohort 3-62
3-52 Temporality of Sample Collection and Metaquestionnaire Completion
by Sample Type for 6-8 Cohort 3-64
3-53 Samples Collected by Day of Week: 6-8 Cohort 3-64
3-54 Summary of Data and Sample Collection: 6-8 Cohort 3-65
3-55 Observations on Level of Burden: 6-8 Cohort 3-68
4-1 Comments and Observations From Results of QA Visits
and Calls 0-1 Cohort 4-4
4-2 Comments and Observations From Results of QA Visits
and Calls 3-5 Cohort 4-6
4-3 Comments and Observations From Results of QA Visits
and Calls 6-8 Cohort 4-7
-------
Executive Summary
The federal government is planning a large prospective birth cohort study known as the
National Children's Study that will potentially involve 100,000 children and their families. The
observation period will start as close to conception as possible and follow-up will last for 20
years after birth. There are substantial challenges, both technical and logistical, associated with
successfully implementing a study of this magnitude. This demonstration project conducted for
U.S. EPA, which is working with NICHD, CDC, and NIEHS in planning the NCS, consisted of a
set of pilot studies that involved nine participants in three cohorts. The cohorts included parents
and their children in the ages of 0-1 years, 3-5 years, and 6-8 years old. These demonstration
studies were designed to address some of the problems that can be anticipated in carrying out the
NCS. Our aim was to evaluate the use of data collection methods that impose a minimal burden
on the study participants while maintaining high data quality. Broad study objectives included:
assessing the feasibility of recruiting and retaining study participants (children and
their caretakers) in a set of longitudinal exposure studies,
assessing the feasibility of employing readily available and easy to use sampling
methods, instruments, and/or techniques, and
demonstrating the feasibility of remote data collection methods with readily available,
easy to use, state-of-the-art methods, instruments, and techniques for assessing human
exposures to environmental contaminants.
Readily-available and commonly used methods, instruments, and techniques were tested
over a 12-month data collection period. Selected exposure data (meta data, environmental
samples, and biological samples) were collected periodically from participants who were
enrolled from an existing web-based panel. A primary objective of the study was to assess the
feasibility of remote (home) data collection using a web-based panel. The web is likely to be an
important avenue of communication and data collection for the NCS. We assessed the feasibility
of using the Internet for providing instructions to participants who collected their own samples
and completed questionnaires on-line. The selection of sample types was based on the desire to
capture and evaluate an array of typical methods and not to provide data for an actual exposure
ES-l
-------
study. Biological and environmental samples collected included:
Cohort: urine, breast milk or duplicate diet (food and beverage);
3-5 Cohort: urine, cotton socks, hair, vacuum dust, and tap water; and
6-8 Cohort: saliva, dust wipe, volatile organic compounds, and
humidity/temperature.
Key Findings for the 0-1 Cohort
We started data collection with 3 breast feeding and 6 non-breast feeding participants.
Approximately 45 % of the eligible families that were approached to participate agreed to enroll.
There were 6 of 9 participants still returning samples at end of the study. Participants dropping
out of the study came from the low and middle income groups. We observed a 73%+ rate for
completing metaquestionnaires, which included a food diary, and providing samples. The
majority of samples with the exception of the beverage were in good condition when they were
received in the lab. Acceptability rates were 85%+ for breast milk, urine, and food; only 69% of
the beverage samples were acceptable. Many were unacceptable because they were warm or
leaking or the date was not recorded. The study provided evidence that breast milk can be
collected without major difficulties. Participants were timely in collecting samples after kits
were sent (at least 80% of the samples were collected within 7 days of receiving the kits). An
important aspect of the demonstration studies was to determine the extent to which study
subjects complied with the instruction of dating samples and completing metaquestionnaires
after collecting the samples. Participants were remarkably compliant with these instructions. Of
all the samples collected, 94% were labeled with a collection date. Of those samples with
sample collection and questionnaire completion dates, 96% of the participants completed
metaquestionnaires within 1 day of collecting the sample. There were no significant problems
completing the metaquestionnaire and food diaries although additional analyses of the data may
be appropriate.
The data and samples returned, as well as the results of the quality assurance field visits
and calls, indicated that study subjects can successfully collect samples if carefully instructed. A
description of the data and sample collection activities is provided in Exhibit ES-1 While none
of the activities appeared to be overly burdensome, the retention rate of 66% which was observed
for a 12-month study period, might be an important consideration in designing and implementing
theNCS.
ES-2
-------
Exhibit ES-1. Observations on Level of Burden: 0-1 Cohort
Sample/
Data
Collection
Breast Milk
Food &
Beverage
Urine
MetaQx
and Food
Diary
Description of Activity
Breast pump provided.
One 2-oz sample per
month requested. Label
sample and store in
freezer. Disassemble
pump and wash. Package
sample for shipment.
Collect a second portion
of all food and beverage
consumed by infant in
provided containers for
each type of sample.
Label sample, store in
refrigerator through 12-
hour collection period,
then freeze. Package
sample for shipment.
Urine sample collected on
gauze pad (inserted into
diaper) that is worn the
evening after the day the
breast milk sample is
collected. Pad from
feces-free diaper is placed
in provided container,
which is stored in freezer
until shipped. Blue ice is
shipped with urine
sample.
Questionnaire and food
diary are to be completed
day after samples are
collected using WebTV.
Duration1
(min)
10
13
(duplicate
diet only)
303
combined
53
10
Level of
Difficulty2
3
1
2
4
Participant Comments
One participant indicated that the
pump would not work and it was
assembled incorrectly.
Three -fourths of the participants
thought it was "easy" to prepare the
duplicate food sample.
Most participants did not find the
urine pad collection to be
burdensome.
About one-half of participants
thought the food diary was
"difficult" to "very difficult" to
complete.
Average duration of time (in minutes) to complete sample activity as reported by sample participants during debriefing, unless
otherwise noted.
2RTI estimate of level of difficulty; Scale of 1 - 5, with 1 = easy and 5 = difficult.
3RTI estimate of duration
ES-3
-------
Key Findings for the 3-5 Cohort
The data collection period for the 3-5 year old cohort started with nine participants, but
the recruitment rate for this cohort was low. Of the eligible families contacted, only 30% agreed
to participate, but the retention rate was very good. Only one participant from the low income
strata dropped out, mid-way through the study. Response rates for completing the
metaquestionnaires for each of the monthly sampling events (e.g., urine and socks, vacuum dust
and hair, and tap water) were very good (-90% or better). Supplemental video instructions were
prepared for hair collection. Participants were also generally compliant with providing samples
(85%+). Parents of this cohort were successful in returning samples that were suitable for
analyses (88%+). Exceptional assistance was obtained in recording the date on the sampling
packages (97%+) and completing the metaquestionnaires shortly after the sample was collected
(94%+). For the 3-5 cohort, the timeliness for sample collection ranged from 77% for urine to
85% for socks and hair. We received feedback during the quality assurance interviews and the
debriefing survey about vacuum dust collection. One participant stated that it was a big problem
to collect the vacuum dust sample. Tap water was a relatively easy sample to collect, although
we found that several samples were leaking when received in the lab. The level of burden
associated with completing the metaquestionnaires and collection of the samples for the 3-5
cohort appears to have been acceptable (Exhibit ES-2). All of the metaquestionnaires could be
completed relatively quickly (10 minutes or less) and the samples generally took less than 10
minutes to collect, with the exception of the socks which were to be worn for 2 hours.
Key Findings for the 6-8 Cohort
Retention of participants in the 6-8 year old study was very good with eight of nine
families continuing to provide data and samples in the last study month. However, recruiting
study subjects was marginally successful in that only 31% of the eligible families agreed to
participate. The one study subject not completing the study was from the low income strata.
Response rates for completing the metaquestionnaires (94%+) and providing samples (94%+)
were very good. However, the samples collected by this cohort, notably saliva, VOCs, and
HOBO temperature data collection, were in many instances unacceptable for analyses in the lab.
Also, we noted that for many of the sample types, the samples were not collected in a timely
manner (29% to 69% collected within the specified time period). Parents and children were very
ES-4
-------
Exhibit ES-2. Observations on Level of Burden: 3-5 Cohort
Sample/
Data
Collection
Urine
Socks
Hair
Vacuum
Dust
Tap Water
MetaQx and
Sock
Activity
Diary
Description of Activity
Collect first morning urine
void in specimen cup. Label
sample with date and time, as
well as the time of last
urination. Store in freezer for
24 hours. Package sample for
shipment.
Wear socks indoors without
shoes for 2 hours, and then
place in sample container.
Freeze overnight. Package
sample for shipment.
Cut a 1-inch section of hair
from the scalp of the child
(instructional video provided).
Place in bag and record date
and time. Package sample for
shipment.
Remove vacuum cleaner bag
from vacuum and place in
plastic bag. Record date and
time on label. Package sample
for shipment.
Fill bottle to red line with tap
water from kitchen sink tap.
Test pH of water with pH
strip. Record the pH, date and
time on the label. Store in
refrigerator until shipped.
Package sample for shipment.
Questionnaire and sock diary
are to be completed day after
samples are collected using
WebTV.
Duration1
(min)
3
3
4
7
6
10
Level of
Difficulty2
2
2
4
2
1
3
Participant Comments
Participants indicated that it was
not a problem to collect the first
morning urine from their child.
Five of the eight study
participants or 63% indicated that
it was a small problem for them
to get their child to wear the
socks for a 2-hour period. The
remaining three indicated it was
not a problem. One child
indicated the socks were itchy.
Only one participant said that it
was a problem to collect the hair
sample; most said it was not a
problem.
One participant used a vacuum
cleaner with a water filtration
system. This participant was
individually instructed on how to
collect a sample of dust from the
vacuum.
No problems noted.
Participants were able to provide
information for each component
of the diary; item nonresponse
was negligible. Error messages
were mentioned by some
participants.
Average duration of time to complete (in minutes) sample activity as reported by sample participant during debriefing; unless
otherwise noted.
2RTI estimate of level of difficulty; Scale of 1 - 5, with 1 = easy and 5 = difficult.
3 RTI estimate of duration
ES-5
-------
willing to provide saliva samples (97% response rate); however, few of the samples collected
were acceptable for analysis (23%). These problems included no recorded collection date,
sample contained mouthwash, or an insufficient collection volume. Collection of dust wipe
samples did not present many problems but some were considered unacceptable for analyses
because participants did not record the collection date or because very little dust was present.
Even though the collection rate for VOC samples was 100%, use of the VOC sampling device
was a problem for participants. Only 56% of the samples were acceptable for conducting
analyses. We anticipated problems with this sample type and developed a video. Unfortunately,
the video was not as accessible via the web as intended. We suspect that one-half of the VOC
sampling badges were not worn for the entire 48-hour collection period. The 42-day HOBO
temperature collection was also not as successful as anticipated. This type of sample gave some
problems that appear to be instrument-related as well as compliance-related.
We noted a problem with the metaquestionnaires for this cohort regarding questions in
which respondents were required to account for time over a 24-hour period. There were
instances in which the participants had difficulties in accounting for and summing time over a
24-hour period. Cognitive testing and modifications in the web-based study instrument should
help to alleviate this kind of problem in future data collection efforts. The burden placed on the
study subjects was considered to be reasonable even though participants were asked to provide a
variety of samples and metadata (Exhibit ES-3). Based on the debriefing data and the quality
assurance visits and calls, we confirmed what we saw in the lab, that the VOC badge and saliva
sample collection may have imposed a level of burden and difficulty that was greater than that
for the other samples.
The demonstration studies generated considerable information, that favorably support
sample collection by study subjects and remote data collection via the web, although the studies
brought forward a number of issues that can impact a large-scale study such as the NCS and
deserve additional consideration. Many of these issues can be handled through, for example,
additional cognitive testing to address meta data collection issues, oversampling to compensate
for potential drop out rates and compliance issues, alterations in the manner in which samples are
shipped, and a recognition of physical limitations that could make sample collection troublesome
for some participants, e.g., reading colors on pH strips. Potential limitations of participants
should be taken into account. Other issues relate to the sample collection/monitoring methods
that can be realistically fielded, given that the participants themselves are to collect samples.
ES-6
-------
Exhibit ES-3. Observations on Level of Burden: 6-8 Cohort
Sample/ Data
Collection
Dust Wipe
Saliva
VOC Badge
HOBO worn
with VOC
HOBO (Temp)
MetaQx and
Activity Diary
Description of Activity
Using the template and wet
wipe provided, collect dust
wipe sample. Record date and
time on label. Store in freezer
for until shipment. Package
sample for shipment.
Rinse mouth with mouthwash.
Discard used mouthwash.
Collect saliva in cup and fill to
red line. Record date and time
on label. Store in freezer until
shipment. Package sample for
shipment.
Watch instructional video.
Remove VOC from can and
attach screen guard. Wear for
48 hrs. Record dates and
times. After 48 hrs, remove
screen guard and separate
sections of badge. Attach caps
to open ends. Return to can
for storage in freezer.
Package sample for shipment.
Attach HOBO to shirt near
VOC Badge. Wear for 48 hrs.
Record dates and times.
Return to plastic box for
storage. Package sample for
shipment.
Remove HOBO from package
and place on table or shelf.
Place thermometer near
HOBO. Twice per week
record the date, time, and
temperature for 6 weeks (42
days) on a data sheet. Package
materials for shipment.
Questionnaire and activity
diary are to be completed day
after samples are collected
using WebTV.
Duration1
(min)
7
10
12
(for
assembling
VOC badge)
53
2
304
7
Level of
Difficulty2
2
2
5
3
2
3
Participant Comments
Participants noted it was not a problem
to collect the dust sample or to place it
in the special shipping container.
Four of the participants did not have a
problem with collecting saliva from
their child, whereas two reported it was
a small problem.
One third said it was not a problem to
assemble the badge and have their child
wear it. One third said it was a small
problem. One third said it was very
difficult to assemble the VOC badges
and obtain the cooperation of their
child to wear it. There were problems
in making the videos available to
participants for viewing.
One child did not want to wear the
HOBO while at camp because there
were hobos nearby and was worried
about being teased.
Participants reported that it was not a
problem to setup the HOBO or record
the room and outdoor temperature.
It was difficult to make duration of
activities sum to 24 hours, potentially
making Qx more cumbersome.
Average duration of time to complete sample activity as reported by sample participant during debriefing; unless otherwise noted.
2RTI estimate of level of difficulty; Scale of 1 - 5, with 1 = easy and 5 = difficult.
3 RTI estimate of duration
4 RTI estimate of duration; participants may not have been taking into consideration the additional time to manually record data in the
debriefing questionnaire.
The development of simpler, participant-friendly measurement devices and methods are
anticipated in the future. These new technologies can be readily put to use in the NCS as they
become available.
ES-7
-------
Introduction
Interest continues to grow in environmental health studies of children. Not only do
children compared to adults have a higher physiological susceptibility to adverse exposure
effects as their body systems develop, their behaviors often increase the likelihood that they will
ingest or absorb various ambient environmental contaminants. To address the special risks that
young children encounter as part of their daily lives, President Clinton signed the Children's
Health Act of 2000 on October 17, 2000. This laid the groundwork for planning a major
prospective cohort study on the impact of environmental exposures on children's health known
as the National Children's Study (NCS). This report presents the results of a set of
demonstration studies that were conducted as a pilot study to support the planning and
implementation of the NCS.
Longitudinal cohort studies that obtain detailed exposure data on multiple chemicals for
young children over time have an opportunity to identify specific exposures that may cause
adverse health outcomes. Because the age at exposure and route of exposure may affect the
nature and magnitude of such adverse effects, measurements from different environmental media
repeated over time are necessary to thoroughly assess the impact of environmental exposures on
children.
However, the burden on study participants from prolonged and obtrusive exposure
measurement efforts would likely compromise a long-term study's success. For example, many
participants would probably drop out of the study, and those who remain in the study might
differ in important ways from those who leave. Even if participants remain in the study, their
degree of compliance with complicated instructions for obtaining samples might diminish over
time. Also, participants might alter their activities around obtrusive monitors (e.g., cigarette
smokers may consciously or unconsciously smoke away from an air monitor). Institutional
Review Boards may determine that the participant burden imposed on participating families is
inappropriate for studies of children and new mothers, two population subgroups that are given
special consideration by IRBs. Finally, if the complexity and cost of exposure monitoring
1-1
-------
become excessive, financial constraints may preclude a study's implementation or completion.
These demonstration studies were designed to address some of these problems through
the use of data collection methods that impose a minimal burden on the study participants while
maintaining high data quality. Broad study objectives included:
assessing the feasibility of recruiting and retaining study participants (children and
their caretakers) in a set of longitudinal exposure studies,
assessing the feasibility of employing readily available and easy to use sampling
methods, instruments, and/or techniques, and
demonstrating the feasibility of remote data collection methods with readily available,
easy to use, state-of-the-art methods, instruments, and techniques for assessing human
exposures to environmental contaminants.
We chose existing data collection methods that are most efficient for obtaining standard
questionnaire data, activity patterns data, and measurement data on pollutant concentrations. An
important feature of the study was the use of remote data collection through a Web-based panel.
The approach integrates an efficient, low-burden method for identifying study subjects and
collecting metadata via the web with the collection of biological and environmental samples
from the same subjects. The application of these technologies was aimed at reducing the need
for field investigators who have traditionally been an essential, but expensive, component of
environmental field data collection. Important questions that were addressed in the study include
those presented in the
accompanying box.
The demonstration studies
were conducted under a task order
contract with U.S. EPA. The
scope of the task order included
developing a quality systems
implementation plan (QSIP);
Key Questions Addressed
in the Demonstration Studies
* D Can study subjects be successfully recruited through a pre-existing
web-enabled panel?
> D Are the incentives used in the study appropriate for level of burden?
* D Is the Web a feasible way of collecting questionnaire, activity, and
food diary data?
D Can study participants coordinate the temporality requirements of
collecting questionnaire data shortly after biological and
environmental samples have been collected?
obtaining study approvals from
EPA and RTI Institutional Review
Boards (reported in Section 2.9);
conducting focus groups
meetings; and conducting the demonstration studies, which are described in this report.
* D Can study subjects follow instructions and successfully assemble
and/or use equipment for collecting samples of food and water,
volatile organic compounds, urine, hair, breast milk, and others?
* D Can study subjects successfully receive supplies and ship samples?
1-2
-------
Study Methods
2.1 Overview of Study Design
Planning for the demonstration studies
commenced in Summer 2001 and field data
collection started in February 2002. The
demonstration project consisted of three separate
studies involving parents and children of three birth
cohorts: 0 to 1 years old, 3 to 5 years, and 6 to 8
years. Selection of these age groups for the studies
was made by U.S. EPA investigators to provide a
broad distribution of children age groups and to test
a range of data collection methods and study burden
issues.
Study participants originated from a
nationally representative Web-enabled panel
developed by Knowledge Networks (KN). The
Web-enabled panel, as discussed further below, is
based on a probability-based random-digit dial
Selected Design Features of the
Demonstration Studies
An existing web-enabled panel was recruited
and received information about the study via
interactive television and the web.
Study communications and questionnaire data
were collected via innovative IT devices that
were already in study subject's homes.
Questionnaire data and biological and
environmental samples were collected for
members of each of three cohorts (0-1, 3-5, 6-
8 years old) that initially consisted of 9
subjects in each.
Environmental and biological sampling
equipment were shipped to study subject's
homes; study subjects collected samples. No
technicians were sent into the field.
Instructional materials were provided via the
web and with shipments.
Quality assurance included telephone and in-
home visits.
(RDD) sample drawn from all 10-digit telephone numbers in the U.S. Panelists utilize an
Internet appliance that connects to a television with Web access.
The demonstration study involved nine study participants (children-parent pairs) in each
of the three birth cohorts. Study subject candidates were randomly selected from across the
nation and came from homes from diverse income strata and urban and rural environments. All
households recruited into the demonstration studies provided consent during enrollment and
granted permission to obtain survey, physical, chemical, and biological data.
For each cohort, we collected metadata through an on-line questionnaire and biological
2-1
-------
and environmental samples. Exhibits 2-1, 2-2, and 2-3 present the sampling schema for each of
the three demonstration studies. Sample collection started in February 2002 for the 0-1 and 6-8
cohorts. Sampling for the 3-5 cohort started in March 2002. Sampling waves were generally
staggered so that in any given week, shipping preparation activities focused on a particular type
of sample for one of the cohorts. Exhibit 2-4 presents the sampling schedule established at the
beginning of the study.
A sampling wave started with an e-mail message sent by KN to the study participants
alerting them that sample collection materials were being sent early the following week.
Laboratory staff generally shipped materials on a Monday (unless it was a federal holiday).
Receipt of the shipment with the sample collection materials coincided with the release of the
web-based metaquestionnaire. Participants were instructed to complete the questionnaire only
after (and shortly after) collecting the biological and environmental samples. Participants were
asked to ship the samples back to the laboratory after they had been collected. Exhibit 2-5
shows the activities conducted by the laboratory and KN that needed to be coordinated for each
sampling wave. A measure of success for the demonstration studies was the extent to which
study subjects completed the questionnaire after collecting the samples and the amount of time
that they took to collect the samples and return them after they received the shipment. On the
third and seventh day after a questionnaire was released, KN sent reminder e-mails to encourage
study subjects to complete the survey if they had not done so. Telephone call reminders were
placed when study subjects did not respond after one month.
Exhibit 2-1. Samples and Metaquestionnaire Content Collected for 0-1 Cohort
Biological Sample
Collection
Urine one sample
collected overnight on
a specified day using
specially prepared
diapers
Environmental
Sample Collection
Breast milk1 one
sample collected on a
specified day
Once breast-feeding
stops, 1 day duplicate
food sample collected
Target No.
of Samples
per Year
per Home
12 milk, 12
urine
Monthly
Metadata Question
(Content)
Questions related to
mother's daily activities
with exposure questions
related to breast-feeding
and infant behaviors
Food diary
Indicators of
Compliance
Urine presence of
creatinine; evidence
that sample was
prepared and shipped
properly
Breast milk evidence
that the breast pump
was used; visual
inspection of sample
upon receipt
Breast milk sample considered to be an environmental sample (or source of exposure) with respect to the infant.
2-2
-------
Exhibit 2-2. Samples and Metaquestionnaire Content Collected for 3-5 Cohort
Biological Sample
Collection
Urine one sample
collected on a
specified day
Hair one sample
collected on a
specified day
None
Environmental
Sample Collection
Cotton socks worn
by children in the
home for 1-2 hours on
the day before the
urine sample was
collected
Vacuum cleaner
dust collected
immediately after hair
collection
Tap water one
sample collected on a
specified day
Target No.
of Samples
per Year
per Home
4 urine, 4 sets
of socks
Months 2, 5,
8,11
3 hair, 3 dust
Months 3, 6,
9
3 water
Months 4, 7,
9
Metadata Question
(Content)
Time and activity diary
data for day of wearing
the socks. Questions
related to types of sources
and times/activities in
specific locations of
exposure sources
Date of last change of
vacuum cleaner bag;
questions about
vacuuming frequency,
areas vacuumed and
potential sources
Date and time of
collection; questions
about water supply
system and potential
sources
Indicators of
Compliance
Urine presence of
creatinine; evidence
that the sample was
prepared and shipped
properly
Socks visual
inspection of sample
upon receipt
Hair evaluate
appearance of sample;
evidence that hair is
bundled properly and
sufficiently long
Dust visual inspection
Tap water compare
pH measured at time of
receipt to pH measured
by study participant.
Exhibit 2-3. Samples and Metaquestionnaire Content Collected for 6-8 Cohort
Biological Sample
Collection
Saliva one sample
collected on a
specified day
None
None
Environmental
Sample Collection
Settled dust collected
using a surface wipe
Personal VOC badge;
(with HOBO light
sensor). Collected for
2 days
Electronic
temperature/relative
humidity (HOBO)
logging data for 2
months with periodic
manual recordings of
temperature
Target No.
of Samples
per Year
per Home
4 dust, 4
saliva
Months 1, 4,
7, 10
2 samples
Months 5, 8
2 samples
Months 6, 9
Metadata Question
(Content)
Dates of collection.
Questions about potential
sources and activities
Questions about potential
sources and activities
Dates, times; window
open/closed; values
observed; concurrent
outdoor temperature and
source of information,
whether rained in past 24
hours
Indicators of
Compliance
Saliva sodium
analysis, visual
inspection.
Dust wipe visual
inspection
Measurement of VOC' s
in samples received;
visual inspection;
evaluation of light
sensor data
Comparison of the
temperature logged by
the device with those
recorded by study
participants
20
-3
-------
Exhibit 2-4. Longitudinal Collection of Metaquestionnaire Data and Samples by Cohort and Month
Age Measurement
Parameter
0 to 1 Urine (U)
Breast Milk or Diet (M)
Questionnaire (Q)
3 to 5 Urine (U)
Cotton Socks (S)
Questionnaire (Q)
Hair (H)
Vacuum Dust (V)
Questionnaire (Q)
Tap Water (W)
Questionnaire (Q)
6 to 8 Saliva (S)
Dust Wipe (D)
Questionnaire (Q)
VOC Badge (V)
Questionnaire (Q)
HOBO (H)
Questionnaire (Q)
2002
Feb
19 2
U
M
Q
S
D
Q
Mar
18
U
S
Q
Apr
2 15 2
U U
M M
Q Q
H
V
Q
May
13 20 2
U
M
Q
w
Q
S
D
Q
Jun
10 17 24
U
M
Q
u
S
Q
V
Q
Jul
8 15 22
U
M
Q
H
V
Q
H
Q
Aug
5 12 19
U
M
Q
w
Q
S
D
Q
Sep
3 9 16 30
U
M
Q
u
S
Q
H
V
Q
V
Q
Oct
7 14 28
U
M
Q
w
Q
H
Q
Nov
4 11 25
U
M
Q
u
S
Q
s
D
Q
Dec
9 2
U
M
Q
2003
Jan
6 20
U
M
Q
-------
Exhibit 2-5. Target Days for Sampling and Web Data Collection
for a Sampling Event
Day
Friday
Monday
Tuesday
Wednesday
Thursday
Friday
Web Data Collection
E-mail sent from KN alerting
participants that sampling materials
are being shipped next week.
Metaquestionnaire made available to
participants via web.
Metaquestionnaire completed and
received by KN1
Metaquestionnaire reviewed for
completeness1
Sample Collection
Sampling materials shipped by
laboratory staff
Sampling materials received by
participant
Sample collected by participant1
Sample and sampling equipment
shipped by participant to laboratory1
Sample examined for proper
collection, storage and shipment1
Target days for completion of activity
2.2 Web-enabled Data Collection and Panel Description
An important objective of the demonstration study was to assess the feasibility of using
remote data collection methods. Web data collection is one approach that may be appropriate for
collecting some data for the NCS. The existing panel assembled by Knowledge Networks
provided an opportunity to evaluate this approach for exposure monitoring.
2.2.1 Sampling Methodology for the Web-enabled Panel
Knowledge Network's Web-enabled panel is based on a nationally representative, list-
assisted, random-digit-dial (RDD) sample drawn from all 10-digit telephone numbers in the
United States. Only those banks of telephone numbers that have zero directory-listed telephone
numbers are excluded. Telephone numbers are selected from the 1+ banks with equal probability
of selection for each telephone number. Telephone exchanges with concentrations of Hispanics
and African-Americans are oversampled. Sampling is implemented without replacement to
ensure that telephone numbers already fielded are not fielded twice.
The sample is first screened for confirmed disconnected telephone numbers and for
businesses. Next, the sample is screened to exclude telephone numbers that are not in the
WebTV Internet Service Provider network. This process results in the exclusion of
approximately 6 percent to 8 percent of the U.S. population. Additionally, households that do
not have a telephone are not covered in the sample (approximately 6 percent of the U.S.
2-5
-------
households).
2.2.2 Recruitment Procedures for the Web-enabled Panel
Exhibit 2-6 illustrates the KN panel recruitment procedures. All telephone numbers that
pass the screening process are sent to a commercial reverse address-matching service. All
telephone numbers matched to an address receive an advance letter and $5 cash incentive. A
random 50 percent subsample of the unmatched telephone numbers is also included in the final
sample of telephone numbers sent for recruitment. All telephone numbers passing the screening
process are sent to a telephone interviewing organization for recruitment. Cases are dialed up to
90 days, with at least 15 dial attempts on cases in which no one answers the telephone and 25
dial attempts on telephone numbers known to be associated with households. Extensive refusal
conversion is also performed.
Exhibit 2-6. Web-enabled Panel Recruitment Procedures
ROD Sample
File
Reverse Address
Match
Telephone Recruitment
Introductory package, $5 bill
Installation and
service support
Household is survey ready
Background information collected
Experienced interviewers conduct all recruitment interviews. During the 10-minute
interview, interviewers ask to speak with an adult household member who is told that they have
been selected to join the research panel. Respondents are instructed that in exchange for
2-6
-------
participation in short weekly surveys over a 2- to 3-year period, the household is provided with
free hardware (an Internet appliance that connects to a television), free Internet access, free
password-protected e-mail accounts for each household member age 13 and older, ongoing
technical support, and an incentive program to encourage continued participation. All members
in the household are enumerated, and some initial demographic variables and background
information on prior computer and Internet usage are collected.
To ensure consistent delivery of survey content, each household is provided with
identical hardware, even if they currently own a computer or have Internet access. Microsoft's
WebTV is the hardware platform currently used. The device consists of a set-top box that
connects to a TV and the telephone. It also includes a remote keyboard and pointing device.
The WebTV device has a built-in 56K modem that provides the household with a connection to
the Internet. The base unit also has a small hard drive to accommodate large file downloads,
including video files. File downloads do not require any user intervention and usually occur at
night.
Prior to shipment, each unit is custom configured with individual e-mail accounts so that
it is ready for immediate use by the household. Most households are able to install the hardware
without additional assistance. Knowledge Networks maintains a telephone technical support line
and will, when needed, provide on-site installation. The call center also contacts household
members who do not respond to e-mail and attempts to restore contact and cooperation.
All new panel members are sent an initial survey (i.e., the Adult Profile Survey) to
confirm equipment installation and familiarize them with the WebTV unit. Demographics such
as gender, age, race, income, and education are collected for each participant to create a member
profile. This information can be used to determine eligibility for specific studies and does not
need to be gathered with each survey.
2.2.3 Survey Administration Procedures for the Web-enabled Panel
To initiate a survey, an e-mail message is sent to the panel members selected for the
survey. The e-mail invitation includes the key elements of an advance letter and informs the
recipient that a survey is waiting for him/her. The participant clicks on a button within the e-
mail to start the multimedia questionnaire. In general, nonrespondents to surveys are sent up to
two e-mail reminders to complete the survey. Telephone prompting to complete the survey via
2-7
-------
the Internet may also be employed. The Internet-connected family television set, rather than a
computer screen, serves as the monitor on which surveys are viewed and completed. All Web-
enabled panel surveys are self-administered, which allows respondents to complete the surveys
at their convenience in the comfort and privacy of home. Survey consistency across households
is assured because each household receives the same standardized hardware and high-speed
network connectivity. Therefore, each panel member receives the same stimulus. By controlling
the platform used by respondents, consistency in survey administration is achieved. From the
respondent's point of view, the inclusion of video, audio, and 3-D graphics in the questionnaire
makes the survey experience much more engaging and less burdensome than conventional
telephone interviews.
The survey administration procedures specific to the demonstration studies are described
under Study Methods, Sections 2.4, 2.5, 2.6.
2.3 Selection and Recruitment of Study Subjects
2.3.1 Sample Selection of Study Subjects
We selected parent-child study participant pairs from Knowledge Networks' nationally
representative web-enabled panel of U.S. households. We used pre-existing household data to
identify groups of households that have children in each of the three age cohorts: 0-1, 3-5, and
6-8 years of age. For each group, we then used the adult panel members' household information
as the basis for forming three sampling domains shown in Exhibit 2-7. Using these three
domains, we developed nine household strata per cohort by crossing the three categories of
income by region, and by rural area. The nine strata per cohort are reflected in Exhibit 2-8
below.
We imposed the requirement that participants with children in the 0-1 age cohort must be
breast feeding. To operationalize this requirement, we used household data to identify pregnant
women in their third trimester and/or women who had just given birth. We administered a brief
screener questionnaire over the Web via interactive TV to ascertain if the pregnant women, or
other women who had just given birth, were planning to breast feed. We recruited with certainty
any women who recently gave birth during the past three months, were already breast feeding
and planned to do so for at least three months. Additionally, we also sampled with certainty
2-8
-------
pregnant women who were giving birth within six weeks of the screener and planned to breast
feed. Because many households contained age-eligible children, we further imposed the
condition that a household may only be included in one cohort study. If a household fell into
more than one age eligible cohort, we randomly selected which cohort for inclusion.
Exhibit 2-7. Sampling Domains for Sample Selection
Domain
Definition
Household income
Low
Medium
High
Less than $24,999
Between $25,000 and $59,999
More than $60,000
Urbanicity
Urban/Suburban
Rural
Metropolitan areas
Non-metropolitan areas
Region*
East
West
New England, Middle Atlantic, South Atlantic,
East North Central, and East South Central
Pacific, Mountain, West North Central, and
West South Central
*U.S. Census categories
Exhibit 2-8. Nine Strata Per Cohort
Strata
Low Income Rural
Low Income East/Urban
Low Income West/Urban
Medium Income Rural
Medium Income East/Urban
Medium Income West/Urban
High Income Rural
High Income East/Urban
High Income West/Urban
2-9
-------
2.3.2 Recruitment of Study Subjects
We randomly ordered the households that met the sampling and eligibility criteria into
nine separate strata lists per cohort. Our initial approach was to recruit one household per list by
choosing the first household on the list, and contacting them through an e-mail message sent over
the Web via interactive TV. The recruitment e-mail (see Appendix A) described the purpose,
sponsor, and requirements of the study. It asked the panel member if he or she was willing to
participate in the demonstration study. As part of this e-mail survey, we also collected the age of
the child in order to confirm the information we already had on the household and ensure it
contained age appropriate children. Once a household agreed to participate in the study, we
mailed the informed consent form (See Section 2.9 and Appendix B) along with a cover letter
that summarized the requirements outlined in the e-mail. Households that did not return the
informed consent within one week were prompted by telephone.
When a household refused to participate, we contacted the subsequent household on the
list and re-initiated the process described above. However, about one month into participant
recruitment, we altered our approach to contact multiple households per cohort in parallel. The
elapsed time for the overall recruitment process during the busy holiday season (i.e., recruitment
e-mail, mailing of the informed consent, telephone follow-up, and receipt of the consent form)
was longer than anticipated. To meet the study schedule, we revised the recruitment e-mail to
indicate that the household may be selected to participate, and that we would re-contact them at
the end of the recruitment period. Once all nine cells across all cohorts were filled with at least
one or more eligible households, we selected one household per strata per cohort to participate in
the demonstration studies. We sent a confirmation/welcome e-mail to those households that
were selected to participate and a thank you e-mail to the households that were not selected.
2.4 Study Methods: 0-1 Cohort - Breast Milk, Food, Beverage and Urine
In this cohort, emphasis was placed on the collection of breast milk, beverage, or food
that the baby consumed (environmental) and on urine from the baby (biological).
2.4.1 Questionnaires, Diaries, and Instructional Materials
The metaquestionnaire used for this cohort was designed to obtain information that might
be linked to exposures to metals. Metals could be absorbed by the mother following exposures
2-10
-------
via dust, vapors, industrial emissions, work with metals, herbal supplements, and foods,
especially seafood (mercury, arsenic). Although kept purposefully short, the questionnaires
asked about possible exposures derived from activities, proximity to industry, and ingestion
exposures. A question was also included to probe an aspect of the child's development that
could be affected if exposures to metals were occurring. The metaquestionnaire for this cohort
can be found in Appendix C-l.
The diary for this cohort was a food diary developed to obtain information about the
types and quantities of foods/beverages consumed by the infant. Data were input as described in
Section 2.7 to capture each food item at each meal. The diary also sought to understand how the
food was handled so that any contamination by dust, from either surfaces or transfer to the food
via hands, might have contributed to ingestion exposure.
Instructional materials were prepared and accompanied each shipment. These multiple-
page instructions described and showed participants what was contained in the shipment, how it
was to be used, and how it was to be shipped back. A one-page version of the instructions was
also provided to the participants to serve as a quick reference guide. Instructions were also
available via the Web. A checklist was also provided to help participants remember to include
all of the items originally sent to them in the return shipment. Copies of instructional materials
are shown in Appendices D-l, E-l, and F-l.
A debriefing questionnaire was developed for the 0-1 cohort that was administered at the
end of the study (Appendix G-l). This study instrument requested information about the study
subject's impressions of the sampling equipment and procedures, ease of completing the
metaquestionnaire, and appropriateness of the incentives.
2.4.2 Biological and Environmental Sample Collection
Prior to use, all sample collection materials (breast pump, collection bottles, etc.) were
verified to be clean with regard to the potential target metals. The collection bottles were soaked
in concentrated nitric acid at least overnight. They were rinsed with de-ionized water before
being tested using the following procedure. The breast pump/sample bottles were
rinsed/equilibrated with dilute acid. The recovered leachate was analyzed by ICP/MS to
determine the elements present. The extracts were shown not to contain metals that would
contaminate the sample during collection and storage. Sample collection materials were
2-11
-------
prepared and packaged as described in Appendix H-l and shipped to the participant along with a
pre-paid FedEx return-shipping label as described in Appendix D-l The breast pump was
considered to be a personal item; the same pump was used for each participant each time. It was
returned to RTI for cleaning in order to be sure that all pumps were subject to the same cleaning
procedure.
2.4.3 Compliance Monitoring and Sample Analysis
When samples were received at RTI from participants, they were inspected and logged
into an Access database with a Visual Basic interface. This interface prompted the person
logging in samples to answer a series of questions that would later be used to assess compliance
and the condition of samples upon receipt. The questions used for the 0-1 cohort are shown in
Appendix 1-1.
As a further measure of compliance for urine collection, samples collected in Months 1,
6, and 8 (February, July, and September) were analyzed for creatinine. The pads containing the
desired urine samples were thawed, and the urine was expressed into a urine container provided
by Quest Diagnostics, the vendor that conducted the creatinine analyses. Samples were analyzed
for creatinine using a spectrophotometric assay as described in the QSIP for this project. A total
of four blanks (deioninzed water) were sent for analysis to serve as laboratory blanks.
The timeliness with which samples were returned was also evaluated. A target of sample
collection within seven days of participant receipt of sample collection materials was established
for all sample types in this cohort. We considered the last day of sample collection to be the
collection date for all samples collected in the demonstration studies.
2.5 Study Methods: 3-5 Cohort - Urine/Socks, Hair/Vacuum Dust, Tap Water
In this cohort, three different types of samples were used. The first set of samples consisted
of socks worn by the child to collect dust and this was linked to urine. Such a pairing would be
amenable to the study of metals or pesticides, depending on how the samples were analyzed.
The second set of samples combined vacuum dust and hair, a scenario that could be used to
evaluate chronic exposures to metals in the environment. Finally, tap water was collected.
These combinations were chosen to cover a range of sample types that would potentially be
utilized in a longitudinal study.
2-12
-------
2.5.1 Urine and Socks
In this set of samples, pre-cleaned, white cotton socks were provided and were to be
worn, without shoes, by the child for a 2-hour period while indoors. A urine sample was
requested for the next morning. This mimics a link between dermal exposure to pesticides via
floor dust with pesticides/metabolites measured in urine the following day. This sample type
was collected quarterly over the course of the year.
Questionnaires, Diaries, and Instructional Materials. The metaquestionnaire
used for this group focused on pesticide use in and around the home as well as at any daycare
centers that the child might attend. In each case, information on the pesticides used, the form in
which they were applied or used, and the places of application were gathered. This included
both indoor and outdoor uses as well as the types of pests for which the pesticide was applied. In
addition, questions were asked about the types of foods the child ate within the last 24 hours,
whether or not the child's hands were washed before meals, and fractions of time the child spent
in various microenvironments both inside and outside of the home. The metaquestionnaire for
this set of samples, including the activity diary for the socks, can be found in Appendix C-2
Instructional materials were prepared and accompanied each shipment. These
instructions described and showed participants what was contained in the shipment, how it was
to be used, and how it was to be shipped back. A one-page version of the instructions was also
provided to the participants to serve as a quick reference guide. Instructions could also be
viewed via the Web in the event that hardcopy instructions were misplaced. A checklist was
provided to help participants remember to include all of the items originally sent to them in the
return shipment. These instructions are shown in Appendices D-2, E-2, and F-2.
A debriefing questionnaire was developed for the 3-5 cohort that was administered at the
end of the study (Appendix G-2) This questionnaire requested information about the study
subject's impressions of the sampling equipment and procedures for the urine and socks (and
other sampling procedures for the 3-5 cohort).
Sample Collection. Prior to use, all sample collection materials (socks, collection
bottles, etc.) were cleaned. Sample collection materials were prepared and packaged as
described in Appendix H-2 and shipped to the participant along with a pre-paid FedEx return-
shipping label as described in Appendix D-2
2-13
-------
Compliance Monitoring and Sample Analysis. When samples were received at
RTI from participants, they were inspected and logged into an Access database with a Visual
Basic interface. This interface prompted the person logging in samples to answer a series of
questions that would later be used to assess compliance and the condition of samples upon
receipt. The questions used for the 3-5 cohort are shown in Appendix 1-2.
As a further measure of compliance for urine collection, samples collected in Months 2,
8, and 11 (March, September, and December) were analyzed to determine if creatinine was
present. The measurement of creatinine indicates the presence of urine. Urine samples in the
specimen cups were allowed to thaw and an aliquot was transferred into a urine container
provided by Quest Diagnostics. Samples were analyzed using a spectrophotometric assay. A
total of four samples were sent for analysis to serve as laboratory blanks. In addition, aliquots of
two of the urine samples were split and sent for analysis in two different batches. A comparison
of the measured creatinine values in these split samples served as an indicator of inter-day
precision.
The timeliness with which samples were returned was also evaluated. A target of sample
collection within seven days of participant receipt of sample collection materials was established
for this set of samples.
2.5.2 Hair and Vacuum Dust
In this subset of samples, a sample of hair was collected along with vacuum dust from the
home vacuum cleaner. This environmental/biological sample pair was chosen to mimic the
evaluation of the longer-term exposure to metals via dust to their accumulation in hair.
Questions about hair and vacuum dust collection were included on the debriefing questionnaire.
Questionnaires and Instructional Materials. To minimize the burden to the
participant, the metaquestionnaire (Appendix C-2) asked many of the same questions as the
sock/urine survey. Although this sample pair was more appropriate to metals, this demonstration
study was not designed to be a full-scale exposure study but rather a test of the ability of
participants to handle questionnaires and various sample collection approaches. An additional
question was added that defined whether or not the child's hair had been washed that day. That
2-14
-------
would be important to know in order to differentiate between metals or other compounds in the
hair vs. on the hair.
Instructional materials were prepared and accompanied each shipment. These
instructions described and showed participants what was contained in the shipment, how it was
to be used, and how it was to be shipped back. A one-page version of the instructions was also
provided to the participants to serve as a quick reference guide. A checklist was provided to help
participants remember to include all of the items originally sent to them in the return shipment.
These instructions are shown in Appendices D-2, E-2, and F-2. It is important to point out that
this mode of dust collection assumed that participants used vacuum cleaning systems in which
dust was collected in the dry state. One of the participants owned and used a Rainbow vacuum
cleaner that collects dust into water. This problem was discussed and the participant was
instructed to collect dust by disassembling the vacuum cleaner and scraping dust out of the
tubing that leads to the water. The collected dust was placed into a small, zip-loc bag.
Given the possible difficulty in conveying to the participant the appropriate method to
collect the hair sample, a short (approximately 1 minute) video was prepared. This video was
available via the Web for participants to view. It was thought this might provide for the most
consistent sample collections. In addition, demonstration of the use of video instructions and an
evaluation of participant's use of such a medium would help to define how future studies could
convey visual information. A copy of the video is supplied separately with this report and the
script can be found in Appendix D-2.
Sample Collection. Prior to use, all sample collection materials (scissors, comb) were
cleaned. Sample collection materials were packaged as described in Appendix H-2 and shipped
to the participant along with a pre-paid FedEx return-shipping label as described in Appendix D-
2.
Compliance Monitoring and Sample Analysis. When samples were received at
RTI from participants, they were inspected and logged into an Access database with a Visual
Basic interface. This interface prompted the person logging in samples to answer a series of
questions that would be used later to assess compliance and the condition of samples upon
receipt. The questions used for the 3-5 cohort are shown in Appendix 1-2.
The timeliness with which samples were returned was also evaluated. A target of sample
2-15
-------
collection within seven days of participant receipt of sample collection materials was established
for this set of samples.
2.5.3 Tap Water
In this subset of samples, tap water was the focus of collection efforts. There was no
biological sample collected concurrently with the collection of tap water. Tap water collected in
this study could be used to evaluate drinking water as a source of exposures to pesticides, metals,
and other toxic compounds. (An increased use of bottled water for drinking purposes can affect
the impact of tap water on total exposures, but studies can be designed to take this into account.)
No preservative, such as sodium thiosulfate, was added to the water at the time of collection. It
was never anticipated that these samples would be analyzed. A preservative would be necessary
to prevent further action of water disinfectants (residual chlorine, chloramines) on chemicals in
the water during transport.
Questionnaires and Instructional Materials. The questions asked of this cohort
(Section 2.7) consisted of items related to the source of their dinking water, whether or not they
used water treatment, what water the child drank, how long the water was run before a glass for
drinking was taken for consumption, what other beverages were mixed using the tap water, and
how much tap water was directly consumed by the child. The questionnaire for this set of
samples can be found in Appendix C-2. Questions about tap water collection were included on
the debriefing questionnaire.
Instructional materials were prepared and accompanied each shipment (see below).
These instructions described and showed participants what was contained in the shipment, how it
was to be used, and how it was to be shipped back. A one-page version of the instructions was
also provided to the participants to serve as a quick reference guide. Instructions were also
available via the Web. A checklist was also provided to help participants remember to include
all of the items originally sent to them in the return shipment. These instructions are shown in
Appendices D-3, E-3, andF-3.
Sample Collection. Sample collection materials were prepared and shipped as
described in Appendix H-2.
2-16
-------
Compliance Monitoring and Sample Analysis. When samples were received at
RTI from participants, they were inspected and logged into an Access database with a Visual
Basic interface. This interface prompted the person logging in samples to answer a series of
questions that would later be used to assess compliance and the condition of samples upon
receipt. The questions used for the 3-5 cohort are shown in Appendix 1-2.
The timeliness with which samples were returned was also evaluated. A target of sample
collection within seven days of participant receipt of sample collection materials was established
for this sample type.
2.6 Study Methods: 6-8 Cohort - Saliva/Dust, VOC/HOBO (Light Intensity),
HOBO (Temperature/Relative Humidity)
In this cohort, three different types of samples were used. The first set of samples,
collected quarterly, consisted of saliva and settled dust collected using a dust wipe. The second
set of samples was badges designed to collect volatile organic compounds (VOCs) over a 48-
hour period (2 sampling periods). VOC collection was linked with measurement of light
intensity via HOBO. Connecting light intensity measurements with the VOC badge provided an
indication of whether the badge had been worn. Finally, a small monitor (HOBO) was deployed
that would track temperature and relative humidity for 42 days.
2.6.1 Saliva and Dust Wipe
Questionnaires and Instructional Materials. The metaquestionnaire developed
for this set of samples (Appendix C-3) focused on whether or not there were smokers in the
home, use of materials that would release VOCs, whether or not windows were open/closed, and
riding in motor vehicles. In addition, there were questions that asked about time spent in various
microenvironments.
Instructional materials were also sent to participants with each shipment. These
instructions described and showed participants what was contained in the shipment, how it was
to be used, and how it was to be shipped back. A one-page version of the instructions was also
provided to the participants to serve as a quick reference guide. Instructions were also available
over the Web in case the hardcopy version was misplaced. A checklist was also provided to help
participants remember to include in the return shipment all of the items originally sent to them.
These instructions are shown in Appendices D-3, E-3, and F-3.
2-17
-------
A debriefing questionnaire was developed for the 6-8 cohort that was administered at the
end of the study (Appendix G-3) This questionnaire requested information about the study
subject's impressions of the sampling equipment and procedures for the saliva and dust wipes
(and other sampling procedures for the 6-8 cohort).
Sample Collection. Sample collection materials were prepared and packaged as
described in Appendix H-3 and shipped to the participant along with a pre-paid FedEx return-
shipping label as described in Appendix D-3
Compliance Monitoring and Sample Analysis. When samples were received at
RTI from participants, they were inspected and logged into an Access database with a Visual
Basic interface. This interface prompted the person logging in samples to answer a series of
questions that would later be used to assess compliance and the condition of the samples. These
questions are shown in Appendix 1-3.
The timeliness with which samples were returned was also evaluated. A target of sample
collection within seven days of participant receipt of sample collection materials was established
for this set of samples.
2.6.2 VOC Badge and HOBO/Light Intensity
In this subset of samples, a 3M 3520 Organic Vapor Monitor badge was used in
conjunction with a continuous light monitor/data logger (HOBO). Exposures to VOCs are
common and the 3M badge is a convenient way to obtain a time-weighted average of exposure to
volatile chemicals. The VOC badge is a passive device approximately 43 mm in diameter and
23 cm thick worn by participants by attaching the clip to their clothing. Prior to use, it must be
removed from the sealed can and assembled properly (see instructional materials section below).
Use of this device is complicated and its use was designed to see if participants could
successfully collect a sample. The HOBO (Anset Computer Corporation) is a small monitoring
device (58 x 44 x 17 mm) that is capable of recording relative humidity, temperature, and light
intensity. It can be pre-programmed to start and stop data logging prior to shipment. Data from
the logger are downloaded and the device is programmed through the use of a PC and
manufacturer-supplied software. For this application, the HOBO was pre-programmed to stop
2-18
-------
light intensity data acquisition (see compliance monitoring below) after 16 days, long past the
time that the VOC badge sampling should have been completed (7 days); the participant did not
need to do anything with the HOBO, except to wear it along with the VOC badge during the
monitoring period. No biological samples were collected in conjunction with VOC sampling.
Questionnaires and Instructional Materials. The metaquestionnaire used for this
group (Appendix C-3) was the same as for the saliva/dust collection and focused on whether or
not there were smokers in the home, use of materials that would release VOCs, whether or not
windows were open/closed, and riding in motor vehicles. In addition, there were questions that
asked about time spent in various microenvironments. Questions about VOC badge/HOBO
sample collection were included on the debriefing questionnaire.
Instructional materials were prepared and accompanied each shipment. These
instructions described and showed participants what was contained in the shipment, how it was
to be used, and how it was to be shipped back. A one-page version of the instructions was
provided to the participants to serve as a quick reference guide. Instructions were also available
via the Web. A checklist was provided to help participants remember to include all of the items
originally sent to them in the return shipment. These instructions are shown in Appendices D-3,
E-3, and F-3.
Given the possible difficulty in conveying to the participant exactly the right way to
collect the VOC sample, a short (approximately 1 minute) video was prepared. This video was
available via the Web for participants to view. It was thought this might provide for the most
consistent sample collections. Proper assembly, use, termination of sampling and preparation for
shipment of the badges was the most difficult thing asked of any of the participants.
Demonstration of the use of video instructions along with an evaluation of participant's use of
such a medium would help to define how future studies could convey visual information. A
copy of the video is supplied separately with this report and the script can be found in Appendix
D-3
Sample Collection. Sample collection materials were prepared and packaged as
described in Appendix H-3 and shipped to the participant along with a pre-paid FedEx return-
2-19
-------
shipping label as described in Appendix D-3 The HOBO was pre-programmed to begin and end
sample collection automatically.
Compliance Monitoring and Sample Analysis. When samples were received at
RTI from participants, they were inspected and logged into an Access database with a Visual
Basic interface. This interface prompted the person logging in samples to answer a series of
questions that would later be used to assess compliance and the condition of samples upon
receipt. The questions used for the 6-8 cohort are shown in Appendix 1-3 Also, data from the
HOBOs were downloaded and examined to determine if it appeared that the VOC badge and
HOBO had been worn for 48 hours. It was assumed that major light variations would be
monitored during the course of the sampling period. Relatively "flat" light signals would be
anticipated only during periods when participants' were asleep or if the device was placed on a
shelf and not worn.
An additional test of compliance was to actually extract and analyze the badges for VOCs
to determine if the badge was really used along with the HOBO (VOCs measured) or if it was
assembled properly (reasonable concentrations of VOCs measured). If unusually high
concentrations were measured, one potential cause would be that the white, permeation barrier
was not on the device for extended periods of time (i.e., not assembled properly or allowed to sit
open during preparation for return shipment). Another possibility is that the badge fell open
during return shipment and became contaminated. It was anticipated that the nine badges from
the first collection would be analyzed. However, participants had trouble with this method and it
was necessary to combine both sets in order to obtain nine samples that appeared to be
acceptable for VOC analysis.
The first stage of each badge was extracted with carbon-disulfide/acetone and analyzed
for typical indoor air VOCs by gas chromatography/mass spectrometry as described in the QSIP.
Two laboratory blanks (unexposed badges) were extracted and analyzed as a measure of
background. Two method controls were prepared by exposing each badge for a known time to a
stream of VOC-containing nitrogen generated using a permeation system. In addition, one
extract was analyzed in duplicate to assess instrumental analysis precision. Although not ideal,
badges used for the blank and control samples were not taken from the same lot as those shipped
to the field.
2-20
-------
The timeliness with which samples were returned was also evaluated. A target of sample
collection within seven days of participant receipt of sample collection materials was established
for this sample type.
2.6.3 HOBO (Temperature)
In this subset of samples, temperature and relative humidity were the focus. There was
no biological sample collected concurrently. This sample was used to determine how well an
electronic device could be deployed in the home for extended periods. This was used as a model
of various real-time monitors that might be anticipated to be used in future, longitudinal studies.
The HOBO, however, is a very rugged device and might not be typical of all real-time monitors
that might be proposed in the NCS. For this study, the HOBOs were pre-programmed to collect
data for 44 days for the first shipment (July). Given that some participants delayed starting
sample collection, the recording duration for the second (October shipment) was extended to 110
days.
Questionnaires and Instructional Materials. The metaquestionnaire used for this
group (Appendix C-3) was the same as for the saliva/dust and the VOC/HOBO collections and
focused on whether or not there were smokers in the home, use of materials that would release
VOCs, whether or not windows were open/closed, and riding in motor vehicles. In addition,
there were questions that asked about time spent in various microenvironments. Questions about
HOBO (temperature) data collection were included on the debriefing questionnaire.
Instructional materials were prepared and accompanied each shipment (see below).
These instructions described and showed participants what was contained in the shipment, how it
was to be used, and how it was to be shipped back. A one-page version of the instructions was
also provided to the participants to serve as a quick reference guide. Instructions were also
available via the Web. A checklist was also provided to help participants remember to include
all of the items originally sent to them in the return shipment. These instructions are shown in
Appendices D-3, E-3 and F-3.
Sample Collection. Materials were prepared and shipped as described in Appendix H-
3. Following deployment in the home, participants were asked to record the temperature read
2-21
-------
from a thermometer that was to be placed next to the HOBO for the duration of the monitoring
period. The HOBO was pre-programmed to start the day after shipment and was to record data
within 44 or 110 days, a period of time sufficient to cover the desired 42-day monitoring period.
Compliance Monitoring and Sample Analysis. When samples were received at
RTI from participants, they were inspected and logged into an Access database with a Visual
Basic interface. This interface prompted the person logging in samples to answer a series of
questions that would later be used to assess compliance and the condition of samples upon
receipt. The questions used for the 6-8 cohort are shown in Appendix 1-3 Also, data from the
HOBOs were downloaded and examined to determine if it appeared that the device had not been
moved around during the monitoring period.
The timeliness with which samples were returned was also evaluated. A target of
completion of sample collection within 49 days of participant receipt of the HOBO was
established for temperature measurement in this cohort.
2.7 Web Implementation of Questionnaires and Other Materials
The metaquestionnaires, food diary, and socks activity diary (Appendix C) were
programmed for self-administration over the web via interactive TV. While the substantive
content of each questionnaire differed, each survey contained similar questions and web screen
layout. Screens were designed using web survey features such as drop down menus for diaries,
radio buttons for checking all that apply, and the use of still images and graphics. Several
screens in each web questionnaire collected dates for the various biological or environmental
specimen collections. We programmed soft checks for each date and soft checks between dates
such that various prompts would appear if the subject did not record an answer, or recorded an
inconsistent answer. For example, "Please provide answer" would appear if the study participant
attempted to skip over date information. For inconsistent dates, the prompt, "There is a problem
with the date entered. Please re-enter date" would appear.
Because these surveys were self-administered, we also programmed some practice
questions and provided an example of what a completed activity diary should look like. See
Exhibits 2-9 and 2-10 below.
2-22
-------
Exhibit 2-9. Example Practice Drop Down Menu
3 Knowledge Networks - Microsoft Internet Explorer
File Edit View Favorites Tools Help
f>Go
Hi 33 I '^Search [^Favorites ^History
Address ]^T] http://nq9.intersurvey.conn/cgi-bin/nqcgi/nqint.cgi
As described in Section 2.5, participants learned how to properly collect and ship the
biological and environmental samples in three ways, two of which exploited the benefits of web
technology. Subjects were able to access a copy of the printed instructions included with each
sample kit over the web on Knowledge Networks website. Instructions on accessing the online
instructions were included with the prenotification e-mail that subjects received two days prior to
the sample kit arriving at their homes. Another mechanism for providing instructions over the
web was through the use of short video clips on how to collect the more complicated samples of
hair and VOC badge. These 6075 second videos were developed by RTI and made available to
subjects via interactive TV one week prior to the sample collection. Due to technological
difficulties solely related to the Web TV platform and corresponding modem speed, anywhere
between 40 75% of the videos were successfully watched prior to sample collection. Future
studies using video clips over the web are likely to avoid these specific difficulties with faster
modem connections and a different Internet platform.
2-23
-------
Exhibit 2-10. Example Activity Diary
File Edit View Favorites lools Help
lSearcri [^Favorites ^History
Address #] http: ,i'/nq9, intersurvey. com/cgi-bin/nqcgi/nqint .cgi
In the screens that follow you will be able to record details
about each activity. Begin on the first screen by entering
information about the first activity your child engaged in
and end on the last screen with the last activity your child
engaged in. The information you enter for each activity
over the 2 hour period will allow the researchers to create a
log of your child's activities that looks much like the
following table.
ExamDle Activity Lot
[[Activity [[Location [ISurface |lBeginTime [EndTime [Socks
Eating (kitchen vinyl 1:15pm 1:30pm lyes
Note that the total time you report in your child's activity
SSjDone
£ Internet
2.8 Usability Testing
Prior to the start of data collection, we conducted usability testing on each of the five
programmed web questionnaires and diaries, and the instructional materials that accompanied
each specimen kit. Five participants (one in the 0-1 cohort, three in the 3-5 cohort, and one in
the 6-8 cohort) were selected from a convenience sample of non-technical RTI staff with
children in the corresponding age groups. The usability testing was conducted at RTFs usability
lab and used the interactive TV platform.
For each session, the participant was asked to review the sample specimen kit, related
equipment and read the instructional materials. The survey methodologist probed for areas of
ambiguity or confusion. She also encouraged the participant to volunteer any feedback on how
to improve the overall clarity, presentation and understanding of the instructional materials.
Some of the results included the identification of minor inconsistencies between labeling of the
kits and instructions, lack of clarity on using Federal Express for return shipment, and some
2-24
-------
confusion over the packing materials. Findings from this aspect of the usability testing were
incorporated and used to refine the instructional materials, contents and packaging of the
specimen kits.
Following testing of the specimen kits, each participant was asked to self-administer the
questionnaire, thinking aloud while they completed the interview. They were encouraged to
volunteer any criticisms of the instrument and to explain any confusion they had while answering
questions. During the interview, the survey methodologist made note of any data entry errors the
participant made and any problems they tried to fix. For example, the interactive TV
questionnaires require a respondent to press a "Submit" button after entering their answer. The
first question in each questionnaire asked the respondent about their relationship to the child in
question, and listed about 10 possible relationships. The "Submit" button was displayed below
this list, and since the list was quite long, it was not visible on screen. Without fail, each time a
participant was answering his or her first questionnaire, the participant would select the answer
then wait to be taken to the next screen. After a few seconds, participants would either ask what
to do next, or start pressing buttons to try to get to another question. The survey methodologist
made note of this each time. The interviews continued in this manner. After completing a
questionnaire, the survey methodologist administered several debriefing questions to obtain the
participant's overall impression of the questionnaire.
Where possible, refinements were made to each web questionnaire and/or diary. For
example, one participant was asked to record the number of hours her child spent in specific
rooms within the homes. She was given a list of rooms in this order: Living Room»Dining
Room^Family Room»Bedroom»Study/Library»Bathroom»Kitchen>Other. However,
during the testing it became obvious that the participant wanted to enter the answers in order of
the rooms her child spent the most hours in. She started with the bedroom, then bathroom,
dining room, and living room. Since the rooms categories were not listed in this order, she
arrowed up and down the list repeatedly to enter each new answer. As a result of this
observation, the listed order was changed to: Bedroom>Bathroom>Kitchen>Dining
room>Living room>Family room>Study/library>Other.
2-25
-------
2.9 Protection of Study Subjects
Protection of the study subjects throughout the duration of the demonstration studies was
a high priority. The 1KB approval process commenced with the development of the study
protocol (Appendix J) in June 2001. The protocol served as a basis for obtaining approvals from
RTI and U.S. EPA. Questionnaires for all three birth cohorts were developed and submitted with
the RTI and EPA IRB review package. The package was submitted to the RTI and EPA IRBs in
August and October, respectively. Approvals were received from both RTI and EPA in
September and November, respectively. The RTI IRB requested several changes in the consent
forms, none of which changed the objectives or direction of the three demonstrations studies.
Several changes and additions to questionnaires and instructions were submitted for
review throughout the startup and the beginning of data collection. These changes included the
addition of an instructional video to be shown for collecting hair samples and packing checklists
to assist the participants in packaging the collection materials and specimens to be returned to
RTI. Other changes consisted of the addition of photographs to the full length instructions and
minor changes due to usability test results to both the instructions and questionnaires. A quality
control checklist which was used by QA staff for conducting the telephone and in-person audits
was developed, submitted, and approved by the IRB. All revisions submitted throughout the
study were approved.
IRB renewal for RTI was submitted for full review in August 2002. The IRB reviewed
and approved the continuation through an expedited review. This approval sustained the
demonstration activities for the remainder of the data collection period. This project resulted in
no incidents or adverse reactions.
2.10 Incentive Schedule
The type and duration of data collection activities assumed by the study subjects made it
appropriate to consider incentives for the demonstration studies. The schedule of payments was
agreed to with U.S. EPA at the outset of the project for budgeting purposes. Incentives were
distributed by KN according to the schedule in Exhibit 2-11. There was no negative feedback
from study subjects during the course of the 12-month study period; however, some dissatisfaction
was express by the 0-1 cohort study subjects in the debriefing questionnaire. Questions and
comments from participants were phoned in and arose during phone and in-person audits. For
example, one participant phoned to ask if she would be receiving a check before Christmas.
2-26
-------
Exhibit 2-11. Incentive Schedule Per Study Subject
for Sample Collection Activities
Cohort
0-1
3-5
6-8
Sample
Types
Urine/
Breast Milk
Urine/Sock
Hair/Dust
Tap Water
Saliva/Dust
Volatile
Organic
Chemical
Temp/
Relative
Humidity
Cash
Payment
$210
$135
$70
$60
$135
$40
$40
Gift
Certificate
for Child
$0
$25
$25
Total
Payment
Value
$210
$160
$70
$60
$160
$40
$40
Cash Payment Distribution
by Quarter
1234
$30 45 60 75
$25 30 35 45
$15 25 30
$15 20 25
$25 30 35 45
$15 25
$15 25
2-27
-------
Results and Conclusions
These demonstration studies involved the collection of considerable data ranging from
the recruitment and retention of the study subjects to the content of the metaquestionnaires and
the condition of the environmental and biological samples. Key objectives were to determine
whether remote data collection is feasible and to assess the ability of study subjects to collect
environmental and biological samples. To that end, it was not necessary nor the original
intention of the task order to analyze all the data that were collected. Therefore for this report,
we selected key variables for analysis that essentially examined the degree to which this
approach to exposure monitoring is feasible and the extent to which these data collection
technologies could be successfully implemented with study subjects. Therefore we present in
this report:
Measures of success in recruiting and retaining study subjects
Response rates for completion of the study questionnaires and diaries
Assessments of the success and difficulties in completing the web questionnaires
Study subjects' willingness to provide a wide range of biological and environmental
sample types
Condition and acceptability of the samples for laboratory analyses
Timeliness of sample collection.
We requested study subjects to start collecting samples shortly after they received the
sampling materials. If the instructions were followed to the letter, most samples would have
been returned approximately 7 days from the time that sampling materials were shipped. We
recognize that this schedule is somewhat arbitrary and that successful exposure monitoring can
allow sample collection schedules that are longer than 1 week. However, in a large scale study
in which the sampling materials are to be cleaned and reused, there will be additional project
expenses if there is a significant inventory of sampling materials in the field at any given time.
As an index of timeliness for this report, we used 7 days or less as the threshold for defining a
3-1
-------
sample that was returned in a timely manner.
The condition of the sample upon receipt helps to determine whether or not a sample
could be properly identified and analyzed to yield a result that would not be suspect, for
whatever reason. These evaluations are discussed below. Exhibit 3-1 summarizes the
acceptance criteria used in this study. Although other criteria could be proposed for some of the
sample types, those shown in the exhibit were expected to be of greatest concern from the
perspective of the desired state of the sample for shipping; the need to link samples to
participants; and their importance in a longitudinal exposure study, e.g., sample collection
dates/times. Notes were kept for each sample shipment received. Observations included those
about samples along with comments about packaging and missing items. General text comments
for each shipment are included in Appendices K-l, K-2, andK-3 for the 0-1, 3-5, and 6-8
cohorts, respectively. Data tables generated from the master database used to develop summary
statistics for the 0-1, 3-5, and 6-8 cohort are shown in Appendices L-l, L-2, andL-3 ,
respectively.
In accordance with agreements made at the outset of the task order, we have purposely
reported the findings for each cohort separately. Additional analyses of the data collected for
these cohorts may be beneficial and appropriate as hypotheses for the NCS are finalized. For
each cohort we present selected findings that are directed at responding to the study objectives.
3.1 Observations of the 0-1 Cohort
3.1.1 Recruitment and Retention
We started the demonstration study with nine study subjects. As discussed in Section
2.3, our intent was to include a broad representation of participants in each cohort with regard to
household income status, region of country and urban-rural location of home {Exhibits 2- 7 and
2-8). We were successful in recruiting a study subject for each of the nine sampling strata per
cohort with two exceptions. We were unable to recruit a study subject for the rural, low income
strata and selected a subject from rural medium income group. We were also unable to recruit a
rural, high income participant and substituted with a study subject from the west/urban, high
income group. The need for these substitutions was driven by two factors: the availability of
study subjects in KN's panel and more importantly, the screening criteria we imposed of
selecting women who were currently breastfeeding or planning to after birth.
3-2
-------
Exhibit 3-1. Criteria for Defining Sample Acceptability
COHORT
0-1
3-5
6-8
SAMPLE MATRIX
Breast milk
Food
Beverage
Urine
Socks
Urine
Vacuum dust
Hair
Tap water
Dust wipe
Saliva
VOC badge
HOBO (Temp)
SAMPLE ACCEPTABILITY
Sample received cold
Container not leaking
Collection date recorded
Sample received cold
Container not leaking
Collection date recorded
Sample received cold
Container not leaking
Collection date recorded
Sample observed to be urine
Sample received cold
Collection date recorded
Sample in amber jar
Collection date recorded
Sample received cold
Collection date recorded
Collection date recorded
Sample in plastic bag
Collection date recorded
Sample received cold
Sample not frozen
Collection date recorded
Sample contained only dust
Collection date recorded
Sample received cold
Container filled to the mark
No evidence of mouthwash
Collection date recorded
Cap properly affixed
Badge secured with Teflon tape
Badge packaged correctly
Collection date recorded
HOBO data downloadable
HOBO matched temp, data sheet
Collection dates recorded
Exhibit 3-2 presents selected statistics regarding the recruitment and retention of study
subjects for the 0-1 cohort. Important observations and outcomes for this cohort include the
following:
3-3
-------
Of the nine study subjects that started the study in February 2002, seven were still
completing questionnaires at the end of the study and six were still providing samples.
One study subject (mother) dropped out of the study in Month 2 after completing one
questionnaire and without providing any samples; sampling shipments were terminated
after Month 2. However, KN continued to post the questionnaire. The study subject
unexpectedly "re-entered" the study in Months 10, 11, and 12 when metaquestionnaires
were completed. A decision to not participate in the demonstration study for several
months may have related in part to moving to another home and failing to "reconnect"
with KN. Participation was encouraged through a number of telephone calls by KN staff
during Month 2 without success.
Two other study subjects stopped providing samples in Months 7 and 9 of the 12-month
study period.
Although our intent was to start the demonstration study with nine breastfeeding mothers,
there were only three that were breastfeeding when data collection commenced in late
February 2002. When study subjects were identified and signed consent forms in late
2002, all participants were breastfeeding. The important lesson learned from this aspect
of the study is that breastfeeding status can change relatively quickly.
There were 44 eligible families with children 0 to 1 who were asked to participate.
Slightly more than half (54%) agreed to enroll in the study. (Only nine were accepted.)
Exhibit 3-2 presents response rates by household income status, but the small numbers
limit the conclusions that can be drawn. The lowest response rate was observed for the
high income group.
One of two low income study subjects was not providing samples at the end of the study
and two of four middle income participants dropped out of the study. All three high
income study subjects finished the study.
Exhibit 3-2. Recruitment and Retention of Study Subjects by Income
Characteristic: 0-1 Cohort
Income
Characteristic
Low*
Medium
High
Total
Recruitment
No.
Contacted
3
20
21
44
No. (%)
Agreeing
to
Participate
2 (67%)
15 (75%)
7 (33%)
24 (54%)
No. in
Study
2
4
3
9
Retention
No. (%)
Completing
MetaQx in
Month 12
2
2
O
7 (78%)
No. (%)
Providing
Samples in
Month 12
1
2
3
6 (67%)
*Number contacted does not include participant substituted from medium income group.
MetaQx = metaquestionnaire
3.1.2 Web Data Collection: Metaquestionnaire, Food Diary and Debriefing
Questionnaire
The questionnaire completion results for the 0-1 cohort are shown in Exhibit 3-3.
Overall, web survey completion for both the metaquestionnaire and food diary was fairly
3-4
-------
consistent with an average of seven of eight subjects completing each month. As noted in
Section 3.1.1., one subject dropped out of the demonstration study after the first month of data
collection. We employed the use of two reminder e-mails and telephone prompting for
nonrespondents. Some demonstration participants voiced complaints during the in-person and
telephone QA checks that the level of e-mail prompting, while only twice, was excessive. While
best practices for self-administered web surveys prescribe a two e-mail approach whereby the
first e-mail is a thank-you/reminder prompt, and the second is a prompt for nonrespondents, there
may be additional considerations for a longitudinal study with a monthly periodicity that should
be investigated for future studies.
Exhibit 3-3. Results of Questionnaire and Food Dian
Questionnaire
Metaquestionnaires
Debriefing (end of study) Questionnaires
1 Data Collection:
0-1 Cohort
Completed (%)
81 (87/108)
100 (8/8)
Food Diary
Food Diary
81 (87/108)
Metaquestionnaire. The first set of web screens (Exhibit 3-4) for the
metaquestionnaire collected dates for breast milk and urine collection. As described in Section
2.7, soft range and consistency checks were programmed to improve data quality. Overall, study
participants were able to record date information for both types of specimen collection.
The remainder of the metaquestionnaire was administered to women who were still breast
feeding. All of these questions were straightforward and answered by a subset of study
participants. These questions gathered information about specific activities conducted in the past
month or week by anyone in the household or the mother. We also collected data on whether the
mother was taking vitamin supplements or had eaten fish in the past 24 hours. At the end of this
section, we collected information related to the child's development.
Food Diary. The food diary was administered to all study participants whose child was
consuming solid foods. The diary recorded the food intake for one day and collected the
following eight items per food item:
Meal (e.g. breakfast, lunch, dinner, or snack)
Name of child's food item
3-5
-------
Quantity of food item
Weight or volume of food item before handling
Weight or volume of food item actually consumed by child
Weight or volume of the leftover
Contact with utensil
Contact with preparers' hands
Exhibit 3-4. Web Data Collection for Breast Milk
31 Knowledge Networks - Microsoft Internet Explorer
File Edit View Favorites lools Help
jfrSQ I
J-'Back r
@! $ | '^Search Q|j Favorites ^History
Address j^] http://nq9,intersijrvey,com/cgi-bin/nqcgi/nqint.cgi
Study participants recorded eight pieces of information for one food item per web screen.
Subjects were prompted to record additional items, and routed to subsequent food diary screens
appropriately. See Exhibit 3-5. Overall, this design appears to have worked well. Study
participants were able to provide estimates for weight of food consumed and leftover. The
validity of these estimates was not verified; follow-up work may be appropriate. The rate of
item nonresponse was negligible, and was the result of one participant not answering one of the
subitems within one month.
3-6
-------
Exhibit 3-5. Web Food Diary
3 Knowledge Networks - Microsoft Internet Explorer
mmm
File Edit View Favorites lools Help
4^ Back < -f - Q gj fS | | Search [^Favorites QfHistary
Address j^J http://nq9,intersurvey,conn/cgi-bin/nqcgi/nqint.cgi
Internet
Debriefing Questionnaire. Of the subjects who provided breast milk samples, one
indicated that the pump provided in the sampling kit was difficult to use. Additionally, it took
about 10 minutes on average for these same subjects to pump the milk into the container provider
and store in the freezer.
Study participants did not think it was a problem to collect urine with the special diaper
and gauze pad insert, or insert the gauze pad into the container, or store it in their freezer until
shipping. Subjects were split equally on whether it was a problem to complete the food diary
with about half saying it was "very easy" to "easy" to complete, and others indicating it was
"difficult" to "very difficult" to complete. On average, it took 10 minutes to complete the diary.
Finally, about 75% of the 0-1 participants indicated it was easy to prepare the duplicate food
sample with the remaining 25% noting it was very difficult. About half the subjects indicated
that the total incentive amount of $210 was appropriate, and about half indicated it was not.
3-7
-------
3.1.3 Environmental Sample Collection: Breast Milk, Beverage, and Food
Samples
Exhibit 3-6 shows the results for those participants from whom we were able to collect
breast milk, and whether or not samples were provided at each time point. As mentioned
previously, all but three participants had stopped breastfeeding by the time data collection
commenced. Such a situation would not be anticipated if women were recruited before delivery.
Of those three, one participant stopped breastfeeding by the fourth month (June). Of the possible
samples, 89% were collected.
An overview of the results for the food and beverage sample collection is shown in
Exhibits 3-7 and 3-8. There were a possible 108 participant-months of data collection for the 0-
1 cohort, because the research protocol called for 12 months of sampling for the nine study
subjects. One participant dropped out of the study during Month 2. Therefore, we used a
denominator of 98 participant-months, defined as those months when participants received
sample collection materials, for reporting response rates. A total of 83% and 73% of the
beverage and food samples, respectively, were collected.
Appendix M-l shows the masses of food and beverage collected. The data are organized
by participant by month. In general, the masses of solid food and beverages consumed were
higher for babies that were not breast feeding than those who were. No other analyses have been
performed on these data.
3.1.4 Biological Sample Collection: Urine
Taking into account the study subject that essentially withdrew during Month 2, we
attempted to collect 98 urine samples (108 possible collections less 10 collections from
withdrawn participant). Of these, there were 86 samples (88%) submitted by the participants
(Exhibit 3-9}. The urine samples that were not collected were attributed to the two study
subjects that dropped out of the study in Months 7 and 9 when they stopped providing any
biological or environmental samples.
3-8
-------
Exhibit 3-6. Collection Summary and Condition of Breast Milk Samples: 0-1 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample received cold, container was not leaking, and
collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as sample received either warm, container was leaking, or collection
date not recorded.
Participant not breast feeding.
No sample collected
Sample Collection Summary
Mothers breast feeding at
outset of study
Participant - months of breast
milk sample collection1
Breast milk samples collected
Acceptable sample rate
3
28
89% (25 samples/28 sampling events)
88% (22 acceptable samples/25 samples)
Defined as the number of months when the participant was breast feeding
3-9
-------
Exhibit 3-7. Collection Summary and Condition of Duplicate Food Samples: 0-1 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample received cold, container was not
leaking, and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as sample received either warm, container was leaking
or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of food
collection1
Food samples collected
Acceptable food samples
collected
98
73% (72/98)
86% (62/72)
Defined as the number of months when participants received materials to collect a food sample.
3-10
-------
Exhibit 3-8. Collection Summary and Condition of Duplicate Beverage Samples: 0-1 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample received cold, container was not
leaking, and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as sample received either warm, container was leaking,
or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of beverage
collection1
Beverage samples collected
Acceptable beverage samples
collected
98
83% (8 1/98)
69% (56/81)
Defined as the number of months when participants received materials to collect a beverage sample.
3-11
-------
Exhibit 3-9. Collection Summary and Condition of Urine Samples Received: 0-1 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample received cold, contained only urine,
and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as sample received either warm, container other than
urine, or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of urine sample
collection1
Urine samples collected
Acceptable urine samples
Urine samples with detectable levels
of creatinine2
98
88% (86/98)
87% (75/86)
100% (22/22)
Defined as the number of months when participants received materials to collect a urine sample.
2 Defined as the number of urine samples with detectable creatinine per number of samples analyzed.
3-12
-------
3.1.5 Indices of Compliance with Study Protocol
Acceptability. Of the breast milk samples collected, 88% were in acceptable condition
(i.e., cold, not leaking, and collection date recorded), when they reached the laboratory, as shown
in Exhibit 3-6.
For the food and beverage samples, 86% of the food and 69% of the beverage samples
were received cold in a non-leaking container with the collection date recorded. These data are
shown graphically in Exhibits 3-7 and 3-8. We noted that only one of the study subjects was
able to provide duplicate beverage samples in acceptable condition for all 12 months.
Consistency was better for the duplicate food samples. Most of the beverage samples were
unacceptable because of leakage. This indicates that the samples were not properly frozen
before return shipment, the cooling capacity of the blue ice packs was insufficient or they were
not properly frozen, or the sample sat outside in the heat and sun prior to FedEx pickup. These
problems can be addressed, assuming full compliance with protocols, by more detailed protocols
(e.g., freeze samples for at least 24 hours before return shipment) and increased cooling capacity
(more ice packs).
A total of 86 urine samples were collected and 75 of these were acceptable (Exhibit 3-9)
for an acceptability rate of 87%. An acceptable sample was one that was received free of feces,
cold (if not frozen), in the proper container, and labeled with the collection date. Participants
were very good about following instructions with regard to collecting only those samples free of
feces. Multiple diapers were provided to be sure that this could be done for each monitoring
event. Since contamination of the gauze pads by feces was not a problem, this question,
although asked during sample log-in, was not probed specifically in the statistical analysis.
The results of the creatinine analyses are summarized in Exhibit 3-9. Complete data are
shown in Appendix M-2. In all cases, creatinine was detected in the urine samples. Two of the
measurements made on samples from Month 8 indicated very low values. This suggests very
dilute urine, either natural or because of added water. However, it is clear that all of the samples
contained urine. Data quality for the creatinine analyses was assessed through the use of lab
blanks. Water samples were shipped to Quest Diagnostics to serve as method blanks. Two
samples were shipped with Batch 1, and one each with Batches 2 and 3. No creatinine was
detected in any of the blank samples.
3-13
-------
Timeliness. We looked at the amount of time it took the parents of the 0-1 cohort to
return breast milk, food, beverage and urine samples as an index of timeliness for completing the
sampling protocol. The timeliness of breast milk sample collection (specified to be within one
week of kit receipt) is shown in Exhibit 3-10. A total of 92% of the samples were collected in a
timely manner. It is important to recognize that these samples were being provided by three
individuals, two of which breastfed the duration of the study. For food and beverage samples,
approximately 80% were collected within seven days (Exhibits 3-11 and 12). Three of nine
participants consistently collected their samples in a timely fashion. A total of 81% of the urine
samples were collected in a timely manner (Exhibit 3-13).
An important aspect of the demonstration studies was determining the extent to which
study subjects complied with the instruction of dating samples and completing
metaquestionnaires after collecting the samples. We found that study subjects were remarkably
compliant these instructions for all categories of the samples collected from the 0-1 cohort. Of
all the samples collected, 94% were labeled with a collection date. Of those samples with
sample collection and questionnaire completion dates, 96% of the participants completed
metaquestionnaires within 1 day of collecting the sample (Exhibit 3-14).
Based on comments we received during the quality assurance visits and calls, we
expected to find that most participants in the 0-1 cohort were collecting samples on the weekend
after they received the kit. Using only those samples in which a date was recorded on the sample
received, Exhibit 3-15 indicates that many participants did find the weekend to be a convenient
time to collect the urine, food, and beverage samples. However, we were pleased to see that in
many instances samples were collected in the middle of the week, presumably soon after the
sampling materials were received and in accordance with the instructions. Collectively, slightly
more samples were being collected on weekdays and than weekends.
3-14
-------
Exhibit 3-10. Timeliness of Breast Milk Sample Collection: 0-1 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collection date not provided, but sample was collected
Participant not breast feeding
No samples collected
Sample Collection Summary
Timeliness rate
92% (23 samples collected within 7 days/25 samples)
3-15
-------
Exhibit 3-11. Timeliness of Food Sample Collection: 0-1 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
Sample collection date not provided, but sample was collected
No samples collected
Sample Collection Summary
Timeliness rate
81% (58 samples collected
within 7 days/72 samples)
3-16
-------
Exhibit 3-12. Timeliness of Beverage Sample Collection: 0-1 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
Sample collection date not provided, but sample was collected
No samples collected
Sample Collection Summary
Timeliness rate
79% (64 samples collected within 7 days/81
samples)
3-17
-------
Exhibit 3-13. Timeliness of Urine Sample Collection: 0-1 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
Sample collection date not provided, but sample was collected
No samples collected
Sample Collection Summary
Timeliness rate
81% (70 samples collected
within 7 days/86 samples)
3-18
-------
Exhibit 3-14. Temporality of Sample Collection and Metaquestionnaire
Completion by Sample Type: 0-1 Cohort
Type of Sample
Breast Milk
Food
Beverage
Urine
Total
Total No. of
Samples
Collected
25
72
81
86
264
Samples Collected
with Date
Recorded (%)
92 (23/25)
94 (68/72)
93 (75/81)
95 (82/86)
94 (248/264)
Samples Meeting
Temporality
Requirement* (%)
100 (23/23)
94 (64/68)
96 (72/75)
96 (79/82)
96 (238/248)
Temporality requirement defined as sample was collected same day or day before metaquestionnaire was completed.
Exhibit 3-15. Samples Collected by Day of Week: 0-1 Cohort
Sunday
Monday
Tuesday
Wednesday
Day of the Week
Thursday
Friday
Saturday
3-19
-------
3.1.6 Summary and Conclusions: 0-1 Cohort
We started data collection with 3 breast feeding and 6 non-breast feeding participants
who were asked to provide metaquestionnaire data, urine, breast milk samples or and duplicate
food and beverage samples in each of 12 sampling months. Approximately 45 % of the eligible
families that were approached to participate agreed to enroll. Retention over the 12-month study
period (6 of 9 still returning samples at end of the study) was not as favorable as hoped for, but
this outcome might have been expected given the demands that are placed on mothers with
newborns. Participants dropping out of the study came from the low and middle income groups.
Study participants in the 0-1 cohort were generally compliant with what was asked. We
observed an 80%+ rate for completing metaquestionnaires and providing samples (Exhibit 3-16).
The one exception was the duplicate food sample in which 73% of the samples were provided
perhaps indicating a level of burden that was different than the other collected samples. This
finding may not be surprising when considering that the samples were collected during a busy
time of the day and that diet samples, unlike the others, posed an economic burden. However,
three-quarters of the participants said in the debriefing survey that the sample was easy to
prepare. The remainder said it was very difficult. The debriefing questionnaire also indicated
that participants were split equally on the difficulty in completing the accompanying food diary;
about one-half said it was easy to very easy.
The majority of samples with the exception of the beverage were in good condition when
they were received in the lab. Acceptability rates were 85%+ for breast milk, urine, food; but
only 69% of the beverage samples were acceptable. Many were unacceptable because they were
warm or leaking or the date was not recorded. Recommendations for additional cooling and
clarification of instructions will be necessary to improve the success rates for collecting these
types of samples.
The study provided evidence that breast milk can be collected without major difficulties.
Special attention needs to be paid to keeping samples cold during return shipment, especially in
hot summer months. Stressing the importance of allowing sufficient time for the milk and cold
packs to freeze prior to return shipment is important. Recognizing the small number of women
in the demonstration study, we were impressed with the overall willingness to provide samples
and the condition of the samples when they were received.
3-20
-------
Exhibit 3-16. Summary of Data and Sample Collection: 0-1 Cohort
Sample/Data
Collection
Percent
Compl.1
Percent
Accept.2
Percent
with
Acceptable
Timeliness3
Percent
with Date
Recorded4
Percent with
Acceptable
Temporality5
Major
Problems/Comments
Breast Milk
89
92
92
100
One participant (of three)
thought that the pump
was difficult to use.
Food
73
86
81
94
94
No significant problems.
Beverage
83
69
79
93
96
One-third of beverage
samples were in
unacceptable condition
(e.g., leakage) upon
receipt.
Urine
MetaQx6
Food Diary
No significant problems.
No significant problems.
Two email prompts to
complete the Qx were
considered excessive by
some participants.
Percent Completed (Number of samples or questionnaires received -=- total number of possible samples).
2 Percent Acceptable (Number of samples considered to be acceptable for analyses -^ total number of samples received). See exhibits and
text for definitions of acceptability.
3 Percent with Acceptable Timeliness (Number of samples received within 7 days of kit shipment -^ total number of samples received).
Percent with Date Recorded (Number of samples with date properly recorded -=- total number of samples received).
Percent with Acceptable Temporality (Number of samples collected same day or day before meta questionnaire completed -=- total
number of samples collected with date recorded).
6 MetaQx (Metaquestionnaire)
Generally, the participants were timely in collecting samples after kits were sent (at least
80% of the collected samples were obtained within 7 days of receiving the kits). As an
acceptance criterion, timeliness might not be as important as other aspects of sample collection.
However, delays in receiving and recycling sampling equipment is a consideration that may
increase equipment and labor costs during field operations.
Parents were found to be compliant with dating the samples (90%+) and completing the
metaquestionnaires (90%+) on the same day (or within 1 day) after collecting the sample, as
requested. Even with what seems to be a high level of compliance, anything less than 100% will
increase the costs of a longitudinal study given the over-sampling and additional materials that
would be required to compensate.
A key aspect of this demonstration study was to evaluate the extent to which study
subjects, rather than field technicians, could successfully and reliably collect samples if a
reduced rate of acceptable samples is acceptable. The data and samples returned, as well as the
results of the field visits and calls, indicated that study subjects can successfully collect samples
3-21
-------
if carefully instructed. Also, we observed no signs of sample tampering. Diaper insert samples
were analyzed for creatinine to confirm that the samples did indeed contain urine. In all
analyzed samples, we detected creatinine. With infants, we can anticipate problems with
obtaining a "good" sample from diapers, if they become soiled overnight. This might result in a
reduced sample expectation for this type of sample.
For the number of samples collected over a 12-month period, the level of burden placed
on the 0-1 cohort parents did not appear to be excessive by our own estimates and the
information provided by the participants (Exhibit 3-17). The fact that samples were generally
returned in a condition acceptable for analyses and collected in a timely fashion provides some
evidence that the overall level of burden was satisfactory, but as pointed out earlier, we lost one-
third of the study subjects in this 12-month study. However, the extent to which participants had
difficulties with receiving shipments and returning them was not explored as thoroughly as we
would have liked. We know from the quality assurance contacts, that participants occasionally
had difficulties with shipping, but we did not quantify these types of problems. A study protocol
that places considerable long-term responsibility on the study subjects to monitor receipt of
shipments and their return is an important consideration and if not handled properly could affect
the overall level of burden and satisfaction.
It is important to note that we received a few negative comments about burden in the
debriefing survey (e.g., food diary was difficult to complete and some saying that the duplicate
food sample was difficult to prepare) and the quality assurance visits/calls (e.g., "the hardest
thing has been the time it takes to track everything"). Also, participants were evenly split on
whether the $210 incentive that was offered was adequate; about one-half said that it was not
enough.
3.2 Observations of the 3-5 Cohort
3.2.1 Recruitment and Retention
Nine children-parent study subjects were successfully enrolled in the 3-5 cohort. We
were able to identify and recruit a study subject for each of the sampling strata discussed
previously in Exhibit 2-8. Selected recruitment and retention characteristics for this
demonstration study are presented in Exhibit 3-18 and summarized below.
3-22
-------
There were 79 eligible families that were contacted and asked to participate. Of these, 23
(29%) agreed to enroll and nine were selected. The lowest response rate was observed in
the high income group (20%).
Retention rates for this cohort were encouraging. Eight of nine study subjects (89%) were
still completing questionnaires and providing samples at the end of the study.
The one study subject that withdrew stopped providing samples and data midway through
the 10-month study period that started in March 2002 (Exhibit 2-4).
One of three low income study subjects was not providing samples or completing
metaquestionnaires at the end of the study. All three of the high income and all three of
the middle income study subjects finished the study.
Exhibit 3-17. Observations on Level of Burden: 0-1 Cohort
Sample/
Data
Collection
Breast Milk
Food&
Beverage
Urine
MetaQx and
Food Diary
Description of Activity
Breast pump provided. One
2-oz sample per month
requested. Label sample
and store in freezer.
Disassemble pump and
wash. Package sample for
shipment.
Collect a second portion of
all food and beverage
consumed by infant in
provided containers for each
type of sample. Label
sample, store in refrigerator
through 12-hour collection
period, then freeze. Package
sample for shipment.
Urine sample collected on
gauze pad (inserted into
diaper) that is worn the
evening after the day the
breast milk sample is
collected. Pad from feces-
free diaper is placed in
provided container, which is
stored in freezer until
shipped. Blue ice is shipped
with urine sample.
Questionnaire and food
diary are to be completed
day after samples are
collected using WebTV.
Duration1
(min)
10
13 (duplicate
diet only)
303
combined
53
10
Level of
Difficulty2
3
1
2
4
Participant Comments
One participant indicated that the pump
would not work and it was assembled
incorrectly.
Three-fourths of the participants
thought it was "easy" to prepare the
duplicate food sample.
Most participants did not find the urine
pad collection to be burdensome.
About one-half of participants thought
the food diary was "difficult" to "very
difficult" to complete.
Average duration of time (in minutes) to complete sample activity as reported by sample participants during debriefing, unless
otherwise noted.
2 RTI estimate of level of difficulty; Scale of 1 - 5, with 1 = easy and 5 = difficult.
3 RTI estimate of duration
3-23
-------
Exhibit 3-18. Recruitment and Retention of Study Subjects by Income
Characteristic: 3-5 Cohort
Income
Characteristic
Low
Medium
High
Total
Recruitment
No.
Contacted
27
22
30
79
No. (%)
Agreeing
to
Participate
8 (30%)
9 (41%)
6 (20%)
23 (29%)
No. in
Study
3
3
3
9
Retention
No. (%)
Completing
MetaQx in
Month 10*
2
3
3
8 (89%)
No. (%)
Providing
Samples in
Month 10*
2
3
3
8 (89%)
*Month 10 = Last month of data collection for this cohort (See Exhibit 2-4).
MetaQx = metaquestionnaire
3.2.2 Web Data Collection: Metaquestionnaire, Activity Diary, and Debriefing
Questionnaire
The questionnaire and diary completion results for the 3-5 cohort are shown in Exhibit 3-
19 below. Overall, web survey completion for the metaquestionnaires and diary was high,
ranging from 89 to 96% complete each survey period across the various instruments.
Urine Metaquestionnaire. The urine questionnaire was administered four times
through out the 12-month demonstration period. It captured information on seven questions in
addition to date and time for urine and sock sample collection. Overall, the level of item
nonresponse per question was negligible. To facilitate reporting which pesticides were applied
in the home, we used still images to help the subject find the EPA regulation number on the back
of a pesticide product. As shown in Exhibit 3-20, subjects were able to view the location of an
EPA regulation number on a bottle, and enter the numbers into the web survey. They were also
able to click and expand the still image, as shown in Exhibit 3-21.
Sock Activity Diary. The sock activity diary (Exhibit 3-22) was scheduled for four,
but inadvertently administered only three times (months 2, 5 and 8), throughout the 12-month
study demonstration field period. KN did not field the sock diary on the final sampling month.
This event points out a difficulty that is a concern in complex surveys but can be overcome with
additional oversight processes. The sock activity diary asked about the activities the child
participated in while wearing the socks during a 2-hour period on the same day in which the
urine sample was collected. Overall study participants were able to provide information for each
component of the diary; item nonresponse was negligible.
3-24
-------
Exhibit 3-19. Results of Questionnaire and Diary Collection: 3-5 Cohort
Questionnaire
Urine
Hair and Dust
Tap Water
Debriefing Questionnaire
Diary
Sock Activity Diary
Completed (%)
92 (33/36)
89 (24/27)
89 (24/27)
89 (8/9)
96 (26/27)
Exhibit 3-20. Pesticides Applied in Home
File Edit View Favorites Tools Help
^Search gjFavorites -3HistorV
Address j^S] http;//nq9,intersurvev,com/cgi-bin/nqcgi/nqint.cgi
If pesticides were applied in or around your
home within the past 2 weeks, what was
applied? Give the name and U.S. EPA
number (if known) of the product. The U.S.
EPA registration number is located on the Fmnt / Qack
label Of the product. Select to view example
Pesticide #1:
Pesticide #2:
Pesticide #3:
Pesticide #4:
Pesticide #5:
U.S. EPA Reg. No:
U.S. EPA Reg. No:
U.S. EPA Reg. No:
U.S. EPA Reg. No:
Next Question
Internet
3-25
-------
Exhibit 3-21. Still Image of Pesticide Label
3 Knowledge Networks - Microsoft Internet Explorer
File Edit View Favorites Tools Help
3 fflS I C| Search j*jFavorites ^History 1^- g
Address |^] http://nq9.intersurvey.conn/nqproj/002/4548/pest,html
I
.,*. >
- R5J004
Hair and Dust Metaquestionnaire. The hair and dust metaquestionnaire was
administered three times throughout the 12-month study demonstration field period. A total of
six questions were administered. Overall item nonresponse for this web survey was negligible.
In one question, see Exhibit 3-23 below, subjects were asked to indicate the rooms of the house
in which they last vacuumed. All subjects who vacuumed in the past 24 hours were asked this
question about which room they last vacuumed.
Tap Water Metaquestionnaire. The tap water metaquestionnaire was administered
three times throughout the 12-month study demonstration field period. A total of four questions
were administered. Overall item nonresponse for this web survey was negligible. In one
question, see Exhibit 3-24 below, subjects were asked to indicate the source of tap water in their
home. All but one subject was able to identify the source of their water supply.
3-26
-------
Exhibit 3-22. Sock Activity Diary Screen
3 Knowledge Networks - Microsoft Internet Explorer
mmm
File Edit View Favorites lools Help
4^ Back < -f - Q gj fS | | Search [^Favorites QfHistary
Address j^J http://nq9,intersurvey,conn/cgi-bin/nqcgi/nqint.cgi
Internet
Debriefing Questionnaire. Based on the debriefing survey, seven of eight study
participants indicated that it was not a problem to collect the first morning urine from their child.
On average it took three minutes to prepare the urine specimen for the freezer.
Five of the eight study participants or 63% indicated that it was a small problem for them
to get their child to wear the socks for a 2-hour period. The remaining three indicated it was not
a problem. Our QA checks indicated that one child reported the socks were itchy. On average it
took four minutes to prepare and package the cotton socks.
Overall, study participants reported that it was not a problem to collect the vacuum
cleaner dust, although one subject indicated it was a significant problem. On average it took
seven minutes to collect the dust sample. Similarly, seven subjects also reported that it was not a
problem to collect the hair sample or bundle the strands of hair together. Only one participant
reported difficulty with both. On average it took four minutes to collect and bundle the hair
sample. No problems were reported in collecting or testing the water sample in the debriefing
survey.
3-27
-------
Exhibit 3-23. Room Vacuum Cleaner Was Used
3 Knowledge Networks - Microsoft Internet Explorer
File Edit View Favorites Tools Help
4= Back - * -
-------
Exhibit 3-24. Source of Drinking Water
3 Knowledge Networks - Microsoft Internet Explorer
File Edit View Favorites Tools Help
<>>Go
^Search n£| Favorites ^History
Address ]^£] http://nq9.intersurvey.conn/cgi-bin/nqcgi/nqint.cgi
zl
Internet
3-29
-------
3.2.3 Environmental Sample Collection: Cotton Socks, Vacuum Dust, and Tap
Water
Exhibit 3-25 shows the number of kits shipped and the number of kits returned. One
study subject dropped out halfway through the study. Out of a possible 90 sample sets, 85 were
returned for a response rate of 94%. Collection results for socks, vacuum dust, and tap water are
shown in Exhibits 3-26, 3-27, 3-28 respectively. The collection rate for socks was 92%; the
collection rate for vacuum dust was 96%; and the collection rate for tap water was 93%.
The pH values measured by the participants and by RTI staff upon receipt of the tap
water sample are shown in Appendix M-3. In almost all cases, the pH value was lower that that
recorded by the participant. Small changes, e.g., half a pH unit, could be attributed to shifts
during shipment, especially as carbon dioxide dissolved into the water because of the head space
in the bottle. The pH values were recorded after the samples had warmed to room temperature.
The impact of sample storage on pH was not investigated using controlled laboratory
experiments. Large pH differences could be due to the type of water collected (different
amounts of organic matter, etc.) or difficulty experienced by the participant in reading the color
of the pH test strips. Each strip had three indicating regions and the color of each segment was
to be compared to a standard chart provided with the sample. An inability to distinguish colors
could have hampered this test.
3.2.4 Biological Sample Collection: Urine and Hair
Collection response rates for urine and hair (Exhibit 3-29 and 3-30) were 86% and 96%,
respectively. The lower rate of collection for urine is consistent with some of the participants
indicating, in the QA visits and calls, that some children did not cooperate and failed to provide
urine samples.
3.2.5 Indices of Compliance with Study Protocol
Acceptability. The acceptability of the samples collected was also evaluated.
For socks, Exhibit 3-26 shows that of the samples collected, 97% were considered to be
acceptable (i.e., received in proper condition and collection date recorded). For the
corresponding urine samples (Exhibit 3-29), the acceptability (i.e., received cold and the
collection date recorded) rate was 90%. Additional cooling capacity and a clarification of pre-
shipment freezing durations to study subjects would help improve this situation.
3-30
-------
Exhibit 3-25. Sample Collection - 3-5 Cohort: All Sample Types
Part
010
011
012
013
014
015
016
017
018
Mar1
Apr2
May3
Jun1
Jul2 | Aug3
^
Sep1
Oct2
Nov3
Dec1
Socks and Urine Kit
2Vacuum Dust and Hair Kit (October kits sent September 20; November kits sent October 28; December kits
November 25; See Exhibit 2-4.)
3Tap Water Kit
Participant returned kit.
No kit returned.
sent
Sample Collection Summary
Total number of kits shipped
Total number of kits returned
Return sample rate
90
85
94% (85
returned kits/90 shipped kits)
3-31
-------
Exhibit 3-26. Collection Summary and Condition of Sock Samples: 3-5 Cohort
Part
010
011
012
013
014
015
Mar
016
017
018
Jun
Sep
Dec
Sample in acceptable condition for analysis. Acceptable sample defined as sample received in proper container and
collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as either sock sample not received in the proper
container or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of sock collection1
Sock samples collected
Acceptable sock samples collected
36
92% (33/36)
97% (32/33)
Defined as the number of months when participants received materials to collect a urine and sock sample.
3-32
-------
Exhibit 3-27. Collection Summary and Condition of
Vacuum Dust Samples: 3-5 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample collection date recorded.
No sample collected
Sample Collection Summary
Participant-months of vacuum dust
collection1
Vacuum dust samples collected
Acceptable vacuum dust samples
collected
27
96% (26/27)
100% (26/26)
Defined as the number of months when participants received materials to collect samples.
3-33
-------
Exhibit 3-28. Collection Summary and Condition of Tap Water Samples: 3-5 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample received cold, not frozen,
container not leaking, and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as either water sample not received cold, not
received frozen, container was leaking, or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of tap water
collection1
Tap water samples collected
Acceptable tap water samples collected
27
93% (25/27)
88% (22/25)
Defined as the number of months when participants received materials to collect samples.
3-34
-------
Exhibit 3-29. Collection Summary and Condition of Urine Samples: 3-5
Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as urine sample received cold and
collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as either urine sample not received cold or
collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of urine collection1
Urine samples collected
Acceptable urine samples collected
Urine samples with detectable levels of
creatinine2
36
86% (3 1/36)
90% (28/31)
100% (23/23)
Defined as the number of months when participants received materials to collect a urine sample.
2 Defined as the number of urine samples with detectable creatinine per number of samples analyzed.
3-35
-------
Exhibit 3-30. Collection Summary and Condition of Hair Samples: 3-5 Cohort
Part
Apr
010
011
012
013
014
015
016
017
018
Jul
Oct
Sample in acceptable condition for analysis. Acceptable sample defined as sample received in proper
container and collection date recorded.
No sample collected
Sample Collection Summary
Participant-months of hair collection1
Hair samples collected
Acceptable hair samples collected
27
96% (26/27)
100% (26/26)
Defined as the number of months when participants received materials to collect samples.
3-36
-------
Creatinine (Exhibit 3-29) was measured in all of the urine samples. The complete data
set is shown in Appendix M-4. Water samples were sent to Quest Diagnostics to serve as
method blanks. Two samples were shipped with Batch 1, and one each with Batches 2 and 3.
No creatinine was detected in any of the blank samples. In order to assess repeatability, two
urine samples from this cohort were split and samples differed by less than 4% of the first
measurement obtained.
Acceptability of the vacuum dust samples was 100%, as shown in Exhibit 3-27. Since
these samples did not need to be cooled and were solid to begin with, this matrix presented few
problems. Comparable acceptability rates (100%) were seen for hair (Exhibit 3-30).
The acceptability of tap water samples, shown in Exhibit 3-28, was 88%. Three of 25
samples were leaking when received. However, all samples were cold when they reached the
laboratory. The larger volume, and increased heat capacity, of the water samples likely
contributed to their remaining cold. This supports the need for additional cooling capacity in
shipments containing urine samples.
Timeliness. Data for socks and urine are shown in Exhibits 3-31 and3-32,
respectively. It is interesting to note that a timeliness rate of 85% was achieved for socks, but a
77% rate was achieved for urine samples. The variation in time between the collection of socks
and the collection of urine was not investigated. Such a time difference could be reflective of the
cooperation of this age group to provide urine samples. In addition, the needs for first morning
void could easily be forgotten by children aged 3-5 years; they could easily get up and go to the
bathroom before the parent can intervene. Delays between the collection of an environmental
sample and the corresponding biological sample can make studies difficult if the temporal
linkage of samples is critical to success.
The timeliness rates for vacuum dust and hair were 81% and 85%, respectively as shown
in Exhibit 3-33 and 3-34. It was evident that at least one study subject in April did not comply
with the instructions that the dust and hair samples be collected on the same day. Whether or not
this was true for sets where both types were collected outside the 7 day time frame was not
evaluated. The timeliness data for the collection tap water is shown in Exhibit 3-35. For these
samples, 84% of the samples were collected on time. It is noted that two study subjects tended to
be late across all sample types.
3-37
-------
Exhibit 3-31. Timeliness of Socks Sample Collection: 3-5 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
Sample collection date not provided, but sample was collected
No samples collected
Sample Collection Summary
Timeliness rate 85% (28 samples collected within 7 days/33 samples)
3-38
-------
Exhibit 3-32. Timeliness of Urine Sample Collection: 3-5 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
No samples collected
Sample Collection Summary
Timeliness rate
77% (24 samples collected within 7 days/3 1 samples)
3-39
-------
Exhibit 3-33. Timeliness of Vacuum Dust Sample Collection: 3-5 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
No samples collected
Sample Collection Summary
Timeliness rate
81% (21 samples
collected within 7 days/26 samples)
3-40
-------
Exhibit 3-34. Timeliness of Hair Sample Collection: 3-5 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
No samples collected
Sample Collection Summary
Timeliness rate
85% (22 samples collected within 7 days/26 samples)
3-41
-------
Exhibit 3-35. Timeliness of Tap Water Sample Collection: 3-5 Cohort
Part May Aug
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
No samples collected
Sample Collection Summary
Timeliness rate
84% (21 samples collected
within 7 days/25 samples)
3-42
-------
For the 3-5 cohort, nearly all (140 of 141 or 99%) of the collected samples across all
sample types were dated (Exhibit 3-36). Temporality requirements of completing the
metaquestionnaire within 1 day of collecting the sample also revealed surprisingly favorable
results with 138 of 140 samples (99%) being compliant with the instructions.
Participants in the 3-5 cohort preferred to collect samples on weekdays, shortly after the
kits arrived (Exhibit 3-37). There was considerably more weekday sample collection than
weekend.
3.2.6 Summary and Conclusions: 3-5 Cohort
The data collection period for the 3-5 year old cohort started with nine participants, but
the recruitment rate for this cohort was low. Of the eligible families contacted, only 30% agreed
to participate. On the other hand, the retention rate was very good. Only one participant from
the low income strata dropped out, mid-way through the study. The data and sample collection
demands were significant because the participants were asked to collect a variety of samples that
cycled over a 10-month period {Exhibit 2-4). Response rates for completing the
metaquestionnaires for each of the monthly sampling events (e.g., urine and socks, vacuum dust
and hair, and tap water) were very good (-90% or better).
Participants were also generally compliant with providing samples (90%+), although
urine samples were returned 86% of the time, which was impressive but slightly less than the
excellent response rate for the other sample types (Exhibit 3-38). This finding may be indicating
that first morning void samples, which require cooperation from the child, may be more difficult
to collect than some of the other samples. However, most of the parents indicated in the
debriefing survey that collection of the urine sample was not a problem. Creatinine was detected
in all of the tested samples indicating that participants were providing urine.
Interestingly, we saw two data collection sets (i.e., dust and hair; tap water) in which the
response rate was higher for returning the samples (>90%) than for completing the questionnaire
(<90%), possibly indicating that there were a few instances in which study subjects forgot or
were unable to complete the questionnaire after collecting the samples. Additional follow-up
and analyses of the data are appropriate to better understand these study dynamics and the
unusual finding for the urine and socks sampling in which there was a higher response rate for
metaquestionnaire completion (96%) than for the submission of samples (92%).
3-43
-------
Exhibit 3-36. Temporality of Sample Collection and Metaquestionnaire
Completion by Sample Type: 3-5 Cohort
Type of Sample
Socks
Urine
Vacuum Dust
Hair
Tap Water
Total
Total No. of
Samples
Collected
33
31
26
26
25
141
Samples Collected
with Date
Recorded (%)
97 (32/33)
100(31/31)
100 (26/26)
100 (26/26)
100 (25/25)
99(140/141)
Samples Meeting
Temporality
Requirement* (%)
100 (32/32)
94(29/31)
100 (26/26)
100 (26/26)
100 (25/25)
99(138/140)
Temporality requirement defined as sample was collected same day or day before metaquestionnaire was
completed.
Exhibit 3-37. Samples Collected by Day of Week: 3-5 Cohort
Sunday Monday Tuesday Wednesday Thursday
Day of the Week
Friday
Saturday
3-44
-------
Exhibit 3-38. Summary of Data and Sample Collection: 3-5 Cohort
Sample/Data
Collection
Urine
Socks
Vacuum
Dust
Hair
Tap Water
MetaQx:
Urine
MetaQx:
Hair & Dust
MetaQx: Tap
Water
Sock
Activity
Diary
Percent
Compl.1
86
92
96
96
93
92
89
89
96
Percent
Accept.2
90
97
100
100
88
Percent
with
Acceptable
Timeliness3
77
85
81
85
84
Percent
with Date
Recorded4
100
97
100
100
100
Percent with
Acceptable
Temporality5
94
100
100
100
100
Major
Problems/Comments
No significant problems.
Some resistance from
several children to wear
socks for full 2-hour
period.
Only one parent
reported difficulty
collecting sample.
Only one parent
reported difficulty
collecting sample.
Leaking containers.
No significant problems.
Percent Completed (Number of samples or questionnaires received -=- total number of possible samples).
Percent Acceptable (Number of samples considered to be acceptable for analyses -=- total number of samples received). See
exhibits and text for definitions of acceptability.
3 Percent with Acceptable Timeliness (Number of samples received within 7 days of kit shipment -^ total number of samples
received).
Percent with Date Recorded (Number of samples with date properly recorded -=- total number of samples received).
Percent with Acceptable Temporality (Number of samples collected same day or day before meta questionnaire completed -=-
total number of samples collected with date recorded).
6 MetaQx (Metaquestionnaire)
Parents of the 3-5 cohort were also very successful in returning samples that were
suitable for analyses (88%+). Exceptional assistance was obtained in recording the date on the
sampling packages (97%+) and completing the metaquestionnaires shortly after the sample was
collected (94%+). For the 3-5 cohort, the timeliness for sample collection (i.e., collecting the
sample within 7 days of when the kit was sent) ranged from 77% for urine to 85% for socks and
hair. This difference may have been due in part to our selection of 7 days as a cut-off point for
defining timeliness. However, additional time might be needed to permit the most flexible
scheduling of the collections by the participants. The difficulty in collecting the urine on time
might be suggesting that either the sock collections should be repeated if the urine sample is lost
or that the temporal linkage should be known and dealt with in data analysis.
3-45
-------
We received feedback during the quality assurance interviews and the debriefing survey
about vacuum dust collection. One participant stated that it was a big problem to collect the
vacuum dust sample. Additional follow-up after the debriefing survey was not possible;
however, we are aware that some study subjects had vacuum cleaning systems in their homes
that did not involve a dry dust collection. The collection scheme for vacuum dust conveyed to
participants will need to be broadened to encompass a range of vacuum cleaner types, including
those that use water or a bag-less canister. This does not present an obstacle to sample
collection. It could have an impact on limits of detection if only small samples can be collected
from water-based collections. In addition, the mode of scrapping can impact the samples. The
exact method proposed will depend on the target analytes for that sample. We made adjustments
in our procedures and instructions for collecting samples for water-based and bag-less cleaning
systems and believe that the demonstration studies indicated that vacuum dust sample collection
is feasible in a full scale study.
Tap water was a relatively easy sample to collect, although we found that several samples
were leaking when received in the lab. We believe this problem is correctable with
modifications in the instructions. An additional factor identified for this sample type is that it
was assumed that all participants could differentiate color equally well; the ability to match
colors on the pH test strips was critical to this test. In retrospect, there is no basis for this
assumption. Any tests incorporated into the NCS should be free of error that is the result of
differences in color perceptions among study participants.
The level of burden associated with completing the metaquestionnaires and collection of
the samples for the 3-5 cohort appears to have been acceptable (Exhibit 3-39). All of the
metaquestionnaires could be completed relatively quickly (10 minutes or less) and the samples
generally took less than 10 minutes to collect, with the exception of the socks which were to be
worn for 2 hours. Parents noted in the debriefing survey and the quality assurance visits that
some children did not like wearing the socks (e.g., found them to be "itchy"). One parent
thought that collecting the hair sample was difficult, but most parents did not think it was a
problem to collect this type of sample. We received no complaints about the incentive which
was $265 for the 3-5 cohort members, if all samples were collected. The instructions provided
3-46
-------
Exhibit 3-39. Observations on Level of Burden: 3-5 Cohort
Sample/
Data
Collection
Urine
Socks
Hair
Vacuum
Dust
Tap Water
MetaQx and
Sock
Activity
Diary
Description of Activity
Collect first morning urine
void in specimen cup. Label
sample with date and time, as
well as the time of last
urination. Store in freezer for
24 hours. Package sample for
shipment.
Wear socks indoors without
shoes for 2 hours, and then
place in sample container.
Freeze overnight. Package
sample for shipment.
Cut a 1-inch section of hair
from the scalp of the child
(instructional video provided).
Place in bag and record date
and time. Package sample for
shipment.
Remove vacuum cleaner bag
from vacuum and place in
plastic bag. Record date and
time on label. Package sample
for shipment.
Fill bottle to red line with tap
water from kitchen sink tap.
Test pH of water with pH
strip. Record the pH, date and
time on the label. Store in
refrigerator until shipped.
Package sample for shipment.
Questionnaire and sock diary
are to be completed day after
samples are collected using
WebTV.
Duration1
(min)
3
3
4
7
6
10
Level of
Difficulty2
2
2
4
2
1
3
Participant Comments
Participants indicated that it was
not a problem to collect the first
morning urine from their child.
Five of the eight study
participants or 63% indicated that
it was a small problem for them
to get their child to wear the
socks for a 2-hour period. The
remaining three indicated it was
not a problem. One child
indicated the socks were itchy.
Only one participant said that it
was a problem to collect the hair
sample; most said it was not a
problem.
One participant used a vacuum
cleaner with a water filtration
system. This participant was
individually instructed on how to
collect a sample of dust from the
vacuum.
No problems noted.
Participants were able to provide
information for each component
of the diary; item nonresponse
was negligible. Error messages
were mentioned by some
participants.
Average duration of time to complete (in minutes) sample activity as reported by sample participant during debriefing; unless
otherwise noted.
2RTI estimate of level of difficulty; Scale of 1 - 5, with 1 = easy and 5 = difficult.
3 RTI estimate of duration
3-47
-------
with the sampling kits and via WebTV, including the video for collecting the hair sample,
appeared to have been read (or viewed) and understood.
3.3 Observations of the 6-8 Cohort
3.3.1 Recruitment and Retention
For the 6-8 cohort demonstration study, we were able to recruit a study subject for each
of the sampling strata identified in Exhibit 2-8. Recruitment and retention rates for the 6-8
cohort are presented in Exhibit 3-40. Notable features were the following.
We approached 87 families about participating in the study; 27 (31%) agreed to
participate. Response rates were comparable for all three income strata.
Eight of nine (89%) study subjects were completing questionnaires and providing
samples at the end of the study.
The one study subject that withdrew stopped providing samples and metadata three-
fourths of the way through the study.
One of three low income study subjects was not providing samples or completing
metaquestionnaires at the end of the study. All three of the high income and all three of
the middle income study subjects finished the study.
3.3.2 Web Data Collection: Metaquestionnaire and Debriefing Questionnaire
The results for completing the questionnaires for the 6-8 cohort are shown in Exhibit 3-
41. Overall, web survey completion for the metaquestionnaires was very high, ranging from 94
to 97% completed each survey period across the various instruments.
Exhibit 3-40. Recruitment and Retention of Study Subjects
by Income Characteristic: 6-8 Cohort
Income
Characteristic
Low
Medium
High
Total
Recruitment
No.
Contacted
32
31
24
87
No. (%)
Agreeing
to
Participate
10(31%)
10 (32%)
7 (29%)
27(31%)
No. in
Study
3
3
3
9
Retention
No. (%)
Completing
MetaQx in
Month 10*
2
3
3
8 (89%)
No. (%)
Providing
Samples in
Month 10*
2
3
3
8 (89%)
*Month 10 = Last month of data collection for this cohort (See Exhibit 2-4).
MetaQx = metaquestionnaire
3-48
-------
Exhibit 3-41. Results of Questionnaire Collection: 6-8 Cohort
Questionnaire Results
Saliva and Dust
VOC and HOBO
HOBO
Debriefing Questionnaires
Completed (%)
97 (35/36)
94(17/18)
94(17/18)
75 (6/8)
Saliva and Dust, VOC Badge and HOBO (Temperature) Metaquestionnaire.
The periodicity for this questionnaire varied with the collection of the biological or
environmental sample collection. Saliva and dust was administered four times; VOC and HOBO
were administered twice; and HOBO alone was also administered twice during the 12-month
demonstration study field period. This 8-item questionnaire collected information about
exposures inside and outside the home. One series of questions in particular sought to collect the
number of hours the child spent in various locations during a 24-hour period (Exhibit 3-42). The
collection of hours and minutes per location for the child in a 24-hour period was obtained over
several questions and hence web screens. The screen presented in Exhibit 3-42 is an example of
the question asked in this series and is the first or initial question asked about hours and minutes
in locations. If the subject indicated any number of hours or minutes in this screen, subsequent
questions specific to that location were asked.
The item nonresponse for this particular item was higher than other web surveys, ranging
from 20 to 30%. It was clear that the subjects were having difficulty with the logic checks
programmed to ensure the hours reported summed to 24, and that the hours by location were
fully accounted. It also appears that they didn't provide an answer in hours or minutes when
they found themselves not summing to the hours for a given location.
Asking subjects to account for a full 24-hour period and then further subdividing that
time by hours and minutes per location was a cognitively challenging task. This was further
complicated by programming logic which produced error messages when the number of hours
and minutes did not add up to the 24-hour period. In the future, this data element would benefit
greatly from cognitive testing to get a better sense of how to capture these data from subjects.
3-49
-------
Exhibit 3-42. Hours and Minutes in Various Locations
3 Knowledge Networks - Microsoft Internet Explorer
File Edit View Favorites Tools Help
~B i»G° I
I -4-iBack
J±] fjJ ^Search Q£jFavorites {^History
Address ]fT| http://nq9.intersurvey.corn/cgi-bin/nqcgi/nqint.cgi
Debriefing Questionnaire. Data from the debriefing questionnaire indicated that four
study participants did not have a problem with collecting saliva from their child, whereas two
reported it was a small problem. On average it took 10 minutes to collect and prepare the saliva
specimen. Study participants also noted that it was not a problem to collect the dust sample or to
place it in the special shipping container.
Information in the debriefing questionnaire regarding the VOC badge was inconsistent
with what we observed in the lab, namely that it was very difficult for study participants to
assemble the VOC badges and obtain the cooperation of their child to wear it. The debriefing
data were evenly split across three categories not a problem, a small problem, and a big
problem to assemble and have their child wear it.
Finally, study participants reported that it was not a problem to setup the HOBO or record
the room and outdoor temperature.
3-50
-------
3.3.3 Environmental Sample Collection: Dust Wipe, Volatile Organic Compound
(VOC) Badge, and HOBO (Temperature)
Exhibit 3-43 shows the number of kits shipped and the number of kits returned. One
study subject stopped providing samples in the fourth quarter of the study. Out of a possible 72
sample sets, 70 were returned for a response rate of 97%.
Collection results for settled dust, VOC badges, and HOBOs are shown in Exhibits 3-44,
3-45, and 3-46, respectively. The collection rate for dust wipes was 97%; the collection rate for
VOC badges was 100%; and the collection rate for HOBO samples was 94%.
3.3.4 Biological Sample Collection: Saliva
Collection results for saliva showed a 97% rate of return (Exhibit 3-47). This high
response rate is consistent with the response rates seen for other samples in which there is direct
parental involvement. Also, relative to urine, the collection of saliva is more difficult to forget
than is a first morning void.
3.3.5 Indices of Compliance with Study Protocol
Acceptability. For dust wipes, Exhibit 3-44 shows that of the samples collected, 89%
were acceptable (i.e., appeared to contain dust and collection date recorded). For the
corresponding saliva samples (Exhibit 3-47), the acceptability rate (i.e., received cold, filled to
the red line, contained no mouthwash, and the collection date recorded) was 23%. Saliva was
the most problematic sample type, despite its seemingly easy collection. The unacceptable
samples were received warm, did not contain the requested volume, or contained the mouthwash
that was supposed to be used only to rinse out the mouth prior to saliva collection. Some
samples were received warm; additional cooling capacity and a clarification of pre-shipment
freezing durations would help this situation. The volume requirements should be defined to be
larger than what is actually needed to ensure that sufficient sample is collected. The presence of
mouthwash in the sample suggests that the instructional materials, although clear in our minds,
were not followed. This would need to be presented in stronger terms and some supplemental
instructions, such as a video, might be needed. Discarding the first few saliva "deliveries" prior
to the collection of the sample to be analyzed could also be a beneficial approach.
3-51
-------
Exhibit 3-43. Sample Collection - 6-8 Cohort: All Sample Types
Part
019
020
021
022
023
024
025
026
027
Feb1
May1
Jun2
Jul3
Aug1
Sep2
Oct3
Nov1
Saliva and Dust Wipe Kit
2 VOC Badge and HOBO Kit
3 HOBO 42-Day Kit
Participant returned kit.
No kit returned.
Sample Collection Summary
Total number of kits shipped
Total number of kits returned
Return sample rate
72
70
97% (70 returned kits/72
shipped kits)
3-52
-------
Exhibit 3-44. Collection Summary and Condition
Of Settled Dust Wipe Samples: 6-8 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as dust wipe appeared to contain dust
and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as either dust wipe did not appear to contain
dust or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of dust wipe
collection1
Dust wipe samples collected
Acceptable dust wipe samples collected
36
97% (35/36)
89% (3 1/3 5)
'Defined as the number of months when participants received materials to collect samples.
3-53
-------
Exhibit 3-45. Collection Summary and Condition of
VOC Badge/HOBO Sampling Devices: 6-8 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as cap was found on the VOC badge,
secured with tape, badge packed correctly and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined either cap was not found on the VOC badge, not
secured with tape, badge not packed correctly or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of VOC
badge collection1
VOC badge samples collected
Acceptable VOC badge
samples collected
18
100% (18/18)
56% (10/18)
1 Defined as the number of months when participants received materials to collect samples.
3-54
-------
Exhibit 3-46. Collection Summary and Condition of
HOBO (Temperature) Sample: 6-8 Cohort
Part
019
020
021
022
023
024
025
026
027
Jul
Oct
Sample in acceptable condition for analysis. Acceptable sample defined as HOBO could be downloaded
upon receipt, temperature was recorded correctly, samples collected for the entire 42-day period, and
collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as either HOBO could not be
downloaded upon receipt, temperature was not recorded correctly, samples not collected for the 42-day
period, or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of HOBO collection1
HOBO samples collected
Acceptable HOBO samples collected
18
94% (17/18)
35% (6/17)
Defined as the number of months when participants received materials to collect samples.
3-55
-------
Exhibit 3-47. Collection Summary and Condition of Saliva Samples: 6-8 Cohort
Sample in acceptable condition for analysis. Acceptable sample defined as sample received cold, filled to the red
line, contained no mouthwash, and collection date recorded.
Sample in unacceptable condition. Unacceptable sample defined as either sample not received cold, not filled to
the red line, contained mouthwash, or collection date not recorded.
No sample collected
Sample Collection Summary
Participant-months of saliva collection1
Saliva sample collected
Acceptable saliva samples collected
36
97% (35/36)
23% (8/35)
Defined as the number of months when participants received materials to collect samples.
3-56
-------
Acceptability of the VOC badge/HOBO samples was 56%, as shown in Exhibit 3-45.
This sampling procedure was problematic for several study subjects. Three of the participants
never provided an acceptable sample, and two of the participants collected the sample properly
the first time but not the second.
Selected VOC badges were extracted and analyzed. The VOC analysis data are shown in
Appendix M-5. The samples contained many of the target analytes except for o-
dichlorobenzene. All of the sampled badges appeared to have been exposed for sample
collection. There was variability among the concentrations measured and none of the samples
indicated suspiciously-high VOC concentrations; this suggests that those samples that appeared
to be collected properly were in fact collected properly. The blank samples showed small
amounts of background for toluene and/>-dichlorobenzene. Reproducibility of injections was
very good, with percent differences between measurements ranging from 1 to 5%. However, the
method recoveries were generally low, with most of the analytes showing approximately 40-50%
recovery, with the exception of p-xylene, which showed a recovery of 88-99%. The reason for
this is not clear but it is important to note that this does not in any way affect the conclusions of
the study. The VOC analyses confirmed that samples considered acceptable based on visual
inspection were probably handled correctly by the study participants.
The light intensities measured by the HOBOs with each VOC badge showed some
compliance problems. As anticipated, light variations were substantial during the day while the
badge was being worn and fairly small variations, if any, during the night time hours. From the
June 18, 2002 shipment, the data suggested that four of the nine participants did not wear the
badge for the entire 48 hours period, i.e., data looked as though the device was sitting on a shelf
for prolonged periods when activity, and thus, light intensity variations, would have been
anticipated. For the September 9, 2002 shipment, five of the nine participants did not wear the
HOBO/badge combination for the required 48 hours. The HOBO's were pre-programmed to
stop recording on September 26, 2002. Participants #21, 23, and 24 did not start the VOC data
collection until after the HOBO had stopped recording data, thus no light data were available.
Finally, the acceptability of the 42-day HOBO data (i.e., data could be downloaded,
temperatures recorded by the participant were in line with those recorded by the HOBO, data
collected for full 42-day period, and the collection date was recorded) are shown in Exhibit 3-46.
Only 35% of the collected samples provided acceptable data. Acceptable samples were defined
as those in which the data could be downloaded, temperature was manually recorded properly,
3-57
-------
the sampling duration was 42 days, and the collection date was recorded. There were no
participants that provided acceptable data for the July sampling; however six of eight that
collected data in October provided acceptable data. A possible explanation for the unacceptable
data collected during July was the timeliness with which study participants initiated sample
collection. The impending need for the HOBOs to be shipped out for VOC sampling resulted in
a request from KN to the participants that they return the HOBOs. If participants waited for
several days before initiation of collection, they would have had to return the devices short of the
requested collection duration of 42 days.
Timeliness. Data for dust wipes and saliva are shown in Exhibit 3-48 and 3-49,
respectively. We observed the same timeliness rates for dust wipes and saliva (66%). However,
using a 7-day threshold for defining timeliness does not mean they were collected at the same
time or on the same day. The variation in time between the collection of wipes and the
collection of saliva was not investigated. Delays between the collection of an environmental
sample and the corresponding biological sample can make studies difficult if the temporal
linkage of samples is critical to success.
The data for timeliness of collection for the VOC badges are shown in Exhibit 3-50.
Only 50% of the samples were collected on time. However, given that this sample needed to be
collected for a total of 48 hours, there was less flexibility in our threshold of 7 days for defining a
timeliness collection compared to some of the other matrices. Any delay in starting collection
following receipt of the package would make the deadline on the back end tighter.
Finally, the timeliness data for the HOBO monitors alone is shown in Exhibit 3-51. The
number of samples promptly returned after 42 days was only 29%. Greater tardiness was
observed in the second collection relative to the first collection. This could suggest fatigue with
the study or that a sampling device that is left for a long period of time gets forgotten.
3-58
-------
Exhibit 3-48. Timeliness of Dust Wipe Sample Collection: 6-8 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
Sample collection date not provided, but sample was collected
No sample collected
Sample Collection Summary
Timeliness rate
66% (23 samples collected within 7 days/35 samples)
3-59
-------
Exhibit 3-49. Timeliness of Saliva Sample Collection: 6-8 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
Sample collection date not provided, but sample was collected
No samples collected
Sample
Timeliness
Collection
rate
Summary
66% (23
samples collected within 7
days/35
samples)
3-60
-------
Exhibit 3-50. Timeliness of VOC Badge/HOBO Sample Collection: 6-8 Cohort
Sample collected within 7 days of receipt of collection materials
Sample collected after 7 days
No sample collected
Sample
Timeliness
Collection
rate
Summary
50% (9
samples
collected within 7 days/18
samples)
3-61
-------
Exhibit 3-51. Timeliness of HOBO (Temperature) Sample Collection: 6-8 Cohort
H
Sample collected and returned within 7 days of the 42-day collection period
Sample collected and returned after 7 days of the 42-day collection period
Sample collected and returned before the 42-day collection period was over. Date and temperature were not
recorded on sampling log by study subject for the entire 42-day period
No sample collected
Sample Collection Summary
Timeliness rate
29% (5 samples collected within 49 days/17 samples)
3-62
-------
Compliance with instructions to date sample containers and complete metaquestionnaires
after sampling was notable for the 6-8 cohort (Exhibit 3-52). For all samples collected across all
sample types, 100 (95%) had sample identification, and the collection date had been recorded.
Of the 100 samples, non-compliance was most evident for the 42-day HOBO collection and was
not surprising given the length of time associated with collecting this sample compared to other
samples which typically involved a sample collection period of less than 48 hours.
For the 6-8 cohort, we saw that participants were not predisposed to collecting samples
on any one day of the week (Exhibit 3-53), but generally there were more samples collected on
weekdays than weekends. One notable observation was the VOC badge sampling, in which it
appeared that many participants preferred to collect this 48-hour sample over the weekend.
3.3.6 Summary and Conclusions: 6-8 Cohort
Retention of participants in the 6-8 year old study was very good with eight of nine
families continuing to provide data and samples in the last study month. However, recruiting
study subjects was marginally successful in that only 31% of the eligible families agreed to
participate. The one study subject not completing the study was from the low income strata.
Response rates for completing the metaquestionnaires (94%+) and providing samples
(94%+) were very good (Exhibit 3-54). However, the samples collected by this cohort, notably
saliva, VOCs, and HOBO temperature data collection, were in many instances unacceptable for
analyses in the lab. Also, we noted that for many of the sample types, the samples were not
collected in a timely manner (29% to 69% collected within the specified time period). Parents
and children were very willing to provide saliva samples (97% response rate); however, few of
the samples collected were acceptable for analysis (23%). These problems included no recorded
collection date, sample contained mouthwash, or an insufficient collection volume. Alternative
methods could be considered such as use of the salivette (pad that is chewed and absorbs saliva)
or the approach of delaying the start of collection after mouthwash use. Selection of methods for
future studies will depend upon the nature of the analysis, e.g., DNA from buccal cell,
metabolites of organic compounds. If a certain volume is needed, investigators will need to
request a larger volume than what is needed.
3-63
-------
Exhibit 3-52. Temporality of
Completion bj
Type of Sample
Dust Wipe
Saliva
VOC/HOBO
HOBO (Temp)
Total
Total No. of
Samples
Collected
35
35
18
17
105
Sample Collection and Metaquestionnaire
f Sample Type: 6-8 Cohort
Samples Collected
with Date
Recorded (%)
91 (32/35)
94 (33/35)
100(18/18)
100(17/17)
95 (100/105)
Samples Meeting
Temporality
Requirement* (%)
91 (29/32)
94(31/33)
89(16/18)
82 (14/17)
90 (90/100)
Temporality requirement defined as sample was collected same day or day before metaquestionnaire was completed.
Exhibit 3-53. Samples Collected by Day of Week: 6-8 Cohort
Sunday
Monday
Tuesday
Wednesday
Day of the Week
Thursday
Friday
Saturday
3-64
-------
Exhibit 3-54. Summary of Data and Sample Collection: 6-8 Cohort
Sample/Data
Collection
Saliva
Dust Wipe
VOC Badge
with HOBO
HOBO
(42-day)
MetaQx:
Saliva &
Dust
MetaQx:
VOC&
HOBO
MetaQx:
HOBO
Percent
Compl.1
97
97
100
94
97
94
94
Percent
Accept.2
23
89
56
35
|
Percent
with
Acceptable
Timeliness3
66
66
50
29
Percent
with Date
Recorded4
94
91
100
100
Percent with
Acceptable
Temporality5
94
91
89
82
Major
Problems/Comments
Large number of
samples were
unacceptable for
analyses. Samples were
received warm, had low
volume, or contained
mouthwash.
No significant problem.
Participants found VOC
badge difficult to
assemble. Many badges
unacceptable for
analyses. Some children
did not like wearing the
badge; concerns about
peer acceptance. There
were difficulties in
downloading
instructional video via
WebTV.
Low number of HOBO
sampling events with
acceptable data.
Participants had
problems with a
question about amount
of time spent in various
locations over 24-hour
period.
Percent Completed (Number of samples or questionnaires received -^ total number of possible samples).
2 Percent Acceptable (Number of samples considered to be acceptable for analyses -^ total number of samples received). See exhibits
and text for definitions of acceptability.
Percent with Acceptable Timeliness (Number of samples received within 7 days of kit shipment -=- total number of samples received).
Percent with Date Recorded (Number of samples with date properly recorded -=- total number of samples received).
Percent with Acceptable Temporality (Number of samples collected same day or day before meta questionnaire completed -=- total
number of samples collected with date recorded).
6 MetaQx (Metaquestionnaire)
3-65
-------
Collection of dust wipe samples did not present many problems but some were
considered unacceptable for analyses because of not noting the collection date or because very
little dust was present. The use of an alternate method for settled dust, such as a dust plate that
can integrate over a longer period of time could be helpful if the experimental design would
allow for an integrated sample.
Even though the collection rate for VOC samples was 100%, use of the VOC sampling
device was a problem for participants. Only 56% of the samples were acceptable for conducting
analyses. The focus group participants also advised us that these samples would be difficult to
collect because of assembling problems and acceptance by the child. We anticipated these
problems and developed a video. Unfortunately, the video was not as accessible as we had
intended. Even with limited access to the video and detailed instructions that were accessible to
all via hardcopy and web, many participants could not effectively handle this sample type. The
light intensity data suggested that that compliance was an issue. The data showed that one-half
of the samples were not worn for the entire 48-hour period, and thus any conclusions based on
the data would be suspect. Additionally, three of the participants were delinquent in starting the
sampling activities during the second monitoring event.
The 42-day HOBO temperature collection was also not as successful as anticipated. This
type of sample gave some problems that appear to be instrument-related as well as compliance-
related. In one instance, the data could not be downloaded from the HOBO; return of the device
to the manufacturer resulted in a successful download. Although only 35% of the collected
samples were considered to be acceptable, the very low compliance rate for July was likely the
result of a need to re-cycle the HOBOs for the VOC badge sampling effort. There was
sufficient time for participants to complete the 42-day collection activity but not if they did not
start for several days. As the time for shipment of the VOC badges approached, KN sent an e-
mail to participants to request that they return the HOBOs. This could have caused folks not to
sample for the requested duration. Some of the participants did not record the temperature
manually using the schedule requested. A lack of manual recording would not likely impact data
quality but served here as a measure of how well participants could comply with study demands.
Parents said in the debriefing that the sample was not difficult to collect.
We also noted a problem with the metaquestionnaires for this cohort regarding questions
in which respondents were required to account for time over a 24-hour period. There were
instances in which the participants had difficulties in accounting for and summing time over a
3-66
-------
24-hour period. Cognitive testing and modifications in the web-based study instrument should
help to alleviate this kind of problem in future data collection efforts.
The burden placed on the study subjects was considered to be reasonable even though
participants were asked to provide a variety of samples and metadata (Exhibit 3-55). Based on
the debriefing data and the quality assurance visits and calls, we confirmed what we saw in the
lab, that the VOC badge and saliva sample collection may have imposed a level of burden that
was greater than that for the other samples.
3-67
-------
Exhibit 3-55. Observations on Level of Burden: 6-8 Cohort
Sample/ Data
Collection
Dust Wipe
Saliva
VOC Badge
HOBO worn
with VOC
HOBO (Temp)
MetaQx and
Activity Diary
Description of Activity
Using the template and wet
wipe provided, collect dust
wipe sample. Record date and
time on label. Store in freezer
for until shipment. Package
sample for shipment.
Rinse mouth with mouthwash.
Discard used mouthwash.
Collect saliva in cup and fill to
red line. Record date and time
on label. Store in freezer until
shipment. Package sample for
shipment.
Watch instructional video.
Remove VOC from can and
attach screen guard. Wear for
48 hrs. Record dates and
times. After 48 hrs, remove
screen guard and separate
sections of badge. Attach caps
to open ends. Return to can
for storage in freezer.
Package sample for shipment.
Attach HOBO to shirt near
VOC Badge. Wear for 48 hrs.
Record dates and times.
Return to plastic box for
storage. Package sample for
shipment.
Remove HOBO from package
and place on table or shelf.
Place thermometer near
HOBO. Twice per week
record the date, time, and
temperature for 6 weeks (42
days) on a data sheet. Package
materials for shipment.
Questionnaire and activity
diary are to be completed day
after samples are collected
using WebTV.
Duration1
(min)
7
10
12
(for
assembling
VOC badge)
53
2
304
7
Level of
Difficulty2
2
2
5
3
2
3
Participant Comments
Participants noted it was not a problem
to collect the dust sample or to place it
in the special shipping container.
Four of the participants did not have a
problem with collecting saliva from
their child, whereas two reported it was
a small problem.
One third said it was not a problem to
assemble the badge and have their child
wear it. One third said it was a small
problem. One third said it was very
difficult to assemble the VOC badges
and obtain the cooperation of their
child to wear it. There were problems
in making the videos available to
participants for viewing.
One child did not want to wear the
HOBO while at camp because there
were hobos nearby and was worried
about being teased.
Participants reported that it was not a
problem to setup the HOBO or record
the room and outdoor temperature.
It was difficult to make duration of
activities sum to 24 hours, potentially
making Qx more cumbersome.
Average duration of time to complete sample activity as reported by sample participant during debriefing; unless
noted.
2RTI estimate of level of difficulty; Scale of 1 - 5, with 1 = easy and 5 = difficult.
3 RTI estimate of duration
4 RTI estimate of duration; participants may not have been taking into consideration the additional time to manually
the debriefing questionnaire.
otherwise
record data in
3-68
-------
Quality Assurance Procedures
and Results
Quality assurance is a typical and integral part of exposure monitoring studies conducted
and sponsored by US EPA. For these demonstration studies, a variety of QA activities were
performed which included:
developing and implementing a Quality Systems and Implementation Plan (QSIP)
which included standard operating procedures (SOPs) (the QSIP and SOPs were
submitted as a separate deliverable under the task order; the QSIP with its
modifications received final approval on October 21, 2002);
making calls and visits to participants to assess their compliance with procedures and
discuss any problems that they had;
communicating with RTI laboratory and project staff; and
observing RTI activities.
4.1 Documentation
A QSIP was prepared at the outset of the task order, which detailed the quality assurance
and quality control procedures for the study. The QSIP was revised when the laboratory
analyses were added through a task order modification. SOPs describing RTFs activities were
written and revised as needed. QA staff assisted with preparation of these documents and
reviewed the final documents. QA staff also reviewed monthly progress reports to keep abreast
of the overall project status.
4.2 Participant Follow-up
To follow up with participants, QA staff prepared a QA Audit Procedures guidance
document that included questionnaire checklists and can be found in Appendix N for each cohort
and received Institutional Review Board (IRB) approval before proceeding with contacting
participants in late April 2002. The checklists covered the main points of the procedures that
4-1
-------
participants were following. Calls and visits were generally scheduled approximately one week
after sampling supplies were sent to participants. In addition to asking specific questions about
the procedures, participants were asked to provide any comments that they had on the project and
the procedures.
4.2.1 Visits
RTI staff visited two 3-5 cohort households in Ohio during July 2002, one 6-8 cohort
household in North Carolina during May 2002, and one 0-1 cohort household in Illinois in
December 2002. The same checklists (Appendix N) were used for both the visits and the calls,
with a few additional items included for the visits.
4.2.2 Telephone Calls
RTI staff called to follow up with participants for 5% of the sampling events, for a total
of 14 calls. The calls were spread over the period of the project and over the 3 cohorts - 5 for 0-
1 cohort, 5 for 3-5 cohort, and 4 for 6-8 cohort. All sample types were covered with the
telephone calls. The results of the calls and visits are summarized in Exhibits 4-1, 4-2 and 4-3.
4.3 Laboratory Activities
QA staff observed RTI staffs shipping and receiving procedures during the second
month of field data collection (March 2002) and confirmed that detailed records were maintained
on outgoing and incoming shipments. QA staff also routinely reviewed weekly summaries and
communicated with laboratory staff about any problems encountered. They also participated in
some calls that laboratory staff made to follow up on problems with incoming shipments.
QA staff observed extraction of the VOC badges on January 6, 2003. Notes were made
on problems found at the start of the extraction step, such as missing parts and loose lids. A
written standard operating procedure (SOP) was followed for the extractions, and a separate SOP
is available for analysis of the extracts by GC/MS.
4.4 QA Observations and Conclusions
The study approach minimized the level of human contact with participants, compared
with what would normally be associated with this type of study when field technicians travel to
the participants' homes. For the QA follow-up calls and visits, the participants were very
4-2
-------
accommodating and seemed genuinely interested in performing the procedures properly. The
calls and visits provided insight into the participants' experiences.
Even though the participants had RTI contact information, almost all had concerns and
observations that had not motivated them to call us, but were brought to our attention during the
QA interview. Many of the participants asked if they would be given the results of the research.
As Exhibits 4-1, 4-2, and 4-3 indicate, a variety of topics were noted during the calls and
visits. These may be broadly categorized as follows:
Comments about logistics, such as shipping and Federal Express coordination and
being prompted to complete the metaquestionnaire before they had time to collect the
samples.
Comments regarding procedures, convincing children to wear the monitors, accessing
video instructions, completing surveys, assembling VOC badges, and using
appropriate sock and diaper sizes.
Generally, the calls generated similar comments to the visits to the participants' home,
but imposed less of a burden on the participants and were more cost effective. For a larger study,
both calls and visits should be considered, but with the emphasis on calls.
4-3
-------
Exhibit 4-1: Comments and Observations From Results of QA Visits
and Calls: 0-1 Cohort
Participant No./
Date/Type of
Sample
Comments and Observations
Part. 006
5/7/02 (C)
Breast milk and urine
Participant asked about the point of the study and the payments for participating
Part. 009
6/7/02 (C)
Duplicate diet and
urine
Participant is at work when the packages are picked up.
Participant tries to complete the questionnaire within 24 hours, but was a little behind this time.
She had 2 sick children, but completed data collection within 48 hours.
The sample kits usually come on Wednesday. The participant works outside the home and so
generally collects the samples on the weekend. KN starts sending reminders too early, in her
opinion, before she has had time to collect the samples.
For the on-line survey question about when the diaper is put on and taken off, she consistently gets
an error message about the times she enters being incorrect. She puts the diaper on in the evening
(around 8 p.m.) and takes it off the next morning (around 6 a.m.).
Participant tries to collect the same amount of food that the child has eaten, but it is sometimes
hard to judge.
Part. 004
7/1/02 (C)
Duplicate diet and
urine
FedEx office in town so she drops the packages off. There was confusion one time with an old
label. FedEx called thinking that they were shipping the package to her, not from her-so she now
removes all old labels. The confusion delayed shipment when it occurred.
This time, she was unable to get the metaquestionnaire to come up and e-mailed KN about the
problem.
Participant works nights several nights a week so she often has the samples in the freezer for more
than one night.
Part. 005
8/26/02 (C)
Duplicate diet and
urine
Participant did not try to send the same volume of milk that her son drank, but did provide a
sample of the whole milk that he was drinking.
Part. 008
12/17/02 (V)
Duplicate diet and
urine
The study subject reported that the sample collection has gone well and that she liked participating
in the study. The hardest thing has been the time it takes to track everything, but that is not a
problem. Nothing went wrong with the latest mailing and sample collection. She collected the
duplicate diet sample on Saturday and the baby urine Sunday morning. She is not breast feeding.
The samples were stored in her freezer. She had not mail them because she thought that the RTI
QA specialist might want to see them. She was planning to mail them that week.
For one of the sampling waves, only one diaper was sent, but she said that was not a problem. She
found it easier to take the samples to the Fed Ex office near her office than to wait for pickup. Fed
Ex was late twice in picking up. She said dropping the shipment off was not a problem. With
regard to the WebTV, she thought they send too many reminders. Because she works, she can only
do the sample collection on the weekend. She puts the diaper on the baby at night and takes the
urine sample the next morning, she has had a small problem with the WebTV program, which
assumes that the diaper is put on and the sample taken the same day. The program will tell her
that the sample was collected before the diaper was put on. But all she has to do is hit ENTER.
4-4
-------
Exhibit 4-1: Comments and Observations From Results of QA Visits
and Calls: 0-1 Cohort (cont.)
Participant No./
Date/Type of
Sample
Part. 001
12/28/02 (C)
Duplicate diet and
urine
Comments and Observations
She is no longer breast feeding. The participant said that she has sometimes had problems with
error messages when she enters the times for the diaper on the questionnaire.
4-5
-------
Exhibit 4-2: Comments and Observations From Results of QA Visits
and Calls: 3-5 Cohort
Participant No./
Date/Type of
Sample
Comments and Observations
Part. 015
4/22/02 (C)
Socks and urine
Vacuum cleaner is water based and does not have a bag. Participant needed smaller socks for her
child, and thought it would help to know the metaquestionnaire questions ahead of time.
Part. 017
4/23/02 (C)
Vacuum bag and hair
Participant has a central vacuum system, which does not have a bag. She asked if notification about
upcoming vacations is needed to coordinate schedules.
Part. 013
5/22/02 (C)
Tap water
Participant turned off the water conditioner before collecting the sample. Participant did not
observe a red line on the sample bottle that he used.
Part. Oil
7/8/02 (V)
Socks and urine
Participant has had trouble with the questionnaire at times.
The child found the socks "itchy."
The participant could not access the video instructions for the hair sample collection.
Part. 014
7/8/02 (V)
Socks and urine
Participant had trouble with the questionnaire one time.
The family replaced hardwood floors during one of the samples so there was probably extra dust.
Participant wants to know if they will see the results of the study.
Participant was unable to access the video instructions for the hair sample.
Part. 018
9/9/02 (C)
Socks and urine
Participant noted that it is difficult to include the tape in the shipment back if she needs the tape to
prepare the box for shipping. She pulls some strips of tape off to use to close the box before she
puts the tape in the box.
The socks were a little tight and not that stretchy.
Part. 012
12/3/02 (C)
Tap water
Participant had received the box for socks and urine samples, but had not collected the samples yet.
Participant does not check webTV electronic mail frequently.
Child is stubborn about providing the urine sample.
Participant asked about seeing the research findings.
4-6
-------
Exhibit 4-3: Comments and Observations From Results of QA Visits
and Calls: 6-8 Cohort
Participant No./
Date/Type of
Sample
Comments and Observations
Part. 026
5/28/02 (V)
Saliva and dust
Participant was not sure when to expect shipments.
The box that the participant opened during the visit included a note asking that the participant return
the dust template from last time, but she thinks she threw it away. The instructions did not clearly
indicate that the template should be returned, she thought. She suggested adding the dust template
to the checklist.
There was an extra piece of paper or wipe in the plastic bag with the mouthwash and the sampling
container this time. Was this supposed to be there? (It is to absorb any spills.)
According to the instructions, the participant should e-mail KN to let them know that she is sending
the package. The participant did not know an e-mail address for KN and thought that more
specifics should be included about this.
The participant was not sure how quickly the samples should be collected after she receives the
sampling kit. (The instructions say that the samples should be collected within a week.)
Since the last samples were collected, the road leading to the participant's home and the
participant's driveway have been paved. They were gravel before. The house is less dusty now.
Part. 020
6/24/02 (C)
HOBO and VOC
badge
Earlier in the project, Federal Express left the packages at her house for her without a signature, but
later Federal Express did not leave the packages if no one was home; she has now signed for them
to always leave them; she takes the packages to work to ship back.
Her son was concerned about being teased, especially with the name HOBO showing on the
monitor-there are hobos near his day camp that the campers have been discussing.
For the VOC badge, she was confused about which numbers to write where. Perhaps the
instructions could be clarified. Some of the numbers were already written on the lid.
The O-rings were black, not white as stated in the instructions.
She had assumed that the samples would be the same as for previous months and did not open the
box until Sunday. Her son was wearing the monitors the day of the call. They were beside his bed
while he slept at night.
Part. 027
7/22/02 (C)
HOBO and VOC
badge
Participant had just returned from vacation and opened box, but had not started the 42-day HOBO
sampling yet.
Her child was not happy about wearing the monitors. She found wearing the monitors
inconvenient, particularly during activities away from home. Her husband set up and packaged the
monitors. They watched the video instructions.
Part. 025
10/16/02 (C)
HOBO
The TV is currently out for repair and is expected back in about 2 weeks. Unable to access WebTV
during this time.
Participant asked about seeing the results of the study.
The VOC badge was the most difficult of the measurements. She viewed the video instructions
before she used it the first time.
Participant had a period of trouble with viewing videos.
4-7
-------
Exhibit 4-3: Comments and Observations From Results of QA Visits
and Calls: 6-8 Cohort (cont.)
Participant No./
Date/Type of
Sample
Comments and Observations
Part. 019
11/11/02 (C)
HOBO
Participant had received the box for saliva and dust samples, but had not collected the samples yet.
She asked if another check will be sent before Christmas.
Federal Express leaves packages at her door if she is not home; she prefers that they leave them
with a neighbor.
The questionnaire is sometimes frustrating, because it will not accept certain numeric answers and,
thus, she cannot always provide accurate information.
She had viewed the video instructions for the VOC badge.
4-8
-------
Recommendations
At the outset of the demonstration studies we posed several questions (Section 1.0) that
we hoped would be answered by this project. In Section 5.1 below, we provide short answers to
these questions. Section 5.2 summarizes our recommendations and suggestions for additional
research.
5.1 Answers to Key Questions
Can study subjects be successfully recruited through a pre-existing web-enabled panel?
Yes. The web can be used successfully to recruit subjects. A very detailed
recruiting e-mail/survey was critical in ensuring that potential subjects understood
what was expected of them and why. Data from the debriefing surveys indicate
that subjects felt informed and knowledgeable about the study requirements prior
to hard copy informed consent.
Are the incentives used in the study appropriate for level of burden?
Possibly. We were able to ask this question of the 0-1 participants in the
debriefing survey. Samples collected by this cohort included breast milk,
duplicate diet and urine samples. Those data show mixed results, some said the
incentive amounts for the 12-month study were appropriate, others said they were
not.
Is the web a feasible way of collecting questionnaire, activity and food diary data?
Yes. The consistently high completion rates for web surveys, and activity and
food diaries demonstrate that when survey burden is kept to a minimum, subjects
will comply.
Can study participants coordinate the temporality requirements of collecting data shortly after
biological and environmental samples have been collected?
Yes. Study participants were instructed to complete metaquestionnaires shortly
after collecting the samples. Compliance rates were well over 90% across all
sample types. In a few instances questionnaires appeared to have been completed
before the sample was collected.
5-1
-------
Can study subjects follow instructions and successfully assemble and/or use equipment for
collecting samples of food and water, volatile organic compounds, urine, hair, breast milk and
others?
Mostly, but some samples were easier to collect than others. Overall, we had
relatively high collection rates for the biological and environmental samples that
were collected. Some samples, such as hair, dust, and water were relatively easy
to collect. Others, such as the VOC and saliva samples were difficult to collect
correctly. Sample condition and acceptability was a problem with some sample
types.
Can study subjects successfully receive supplies and ship samples?
Overall, yes. Study subjects were generally successful in receiving and shipping
supplies. However, placing the burden of shipping on the study subject has its
own set of logistical and technical issues. These include, for example, a greater
number of communications between project staff and participants to discuss
packaging problems and missing items; a potentially greater number of samples
that have not been adequately cooled or packaged; or shipments that arrive at the
laboratory on weekends and holidays even though instructions requested
otherwise.
5.2 Recommendations
We made a number of observations while conducting the three demonstration studies. In
general, the study demonstrated that a web-based approach to collecting data and samples from
study subjects is feasible. We offer below several recommendations for additional research and
suggestions for the NCS.
Study Planning
The web is a feasible approach for collecting information that does not have to be
collected in a clinical environment and for providing instructional and other information
that includes graphics and images. This form of data collection, which should be
considered for the NCS, is dependent upon the study subject having a television (and
telephone in the case of the panel we used) or computer equipment. The web approach
allows distribution of instructional materials in two formselectronic and hardcopy,
which may be beneficial to some families. Problems that we encountered with study
subject access to videos can be addressed with new technologies. The QA interviews
5-2
-------
indicated that in some instances the study subjects had computer-related difficulties with
a particular question. These types of problems are not unusual and can be successfully
addressed in a large-scale study.
Cognitive interviewing and usability testing of the study instruments and instructional
materials are needed for a large-scale study. Instances in which we asked participants to
account for their activities over a 24-hour period and to sum time spent in various
environments posed problems using the web-based approach.
Consideration should be given to asking study participants, while they are completing
questionnaires, about the level of burden being imposed on them by participating in the
study. The information can be used to re-design data collection approaches, if they
become necessary, in a long-term longitudinal study such as the NCS.
To ensure an appropriate sample size of women who breast feed, it is important to
consider over sampling this subgroup to maintain the requisite number of cases at the end
of the study. It's also important to confirm breast feeding status at least two weeks prior
to the start of the survey because it is difficult to predict loss of subjects who have ceased
breast feeding.
Planning activities need to ensure that activities requested of study participants are not
affected by physical limitations, such as color blindness, etc. An example can be found
in the pH data from tap water generated during this study. Although all possible causes
for large differences between the pH values measured by participants and those measured
at RTI have not been evaluated, an inability of a participant to distinguish shades of color
on the pH test strips could have been a cause. It is recommended that proposed
participant activities be evaluated with physical limitations in mind.
There is a need to consider the effect of socioeconomic status (e.g., income level) on
potential biases in data acquired to address specific hypotheses. Although the number of
participants in this study was relatively small, there were indications that lower income
participants were less compliant and had higher drop-out rates than those in higher
income strata. Should a particular hypothesis need a study population from lower income
strata, over-sampling will need to be considered to avoid biases in the resulting data.
5-3
-------
Study Initiation
Recruitment response rates for each of the three demonstration studies were relatively
low. Each of the cohorts originated from a pre-existing panel that presumably had a
favorable predisposition toward participating in surveys. Unlike phone and in-person
surveys that give control to the interviewer to contact and recruit subjects, web
recruitment delegates this control to the potential subject. Based on our experience, the
recruitment period should allow at least four weeks per subject for the process to be
completed. This process includes time for the potential subject to check their e-mail,
respond to the recruitment invitation, receive via regular mail the hardcopy informed
consent form, and mail it back.
Study Implementation
Pre-notification and e-mail reminders proved to facilitate compliance and timely
completion of the web surveys, and the activity and food diary in a low cost and efficient
manner. Subjects were expected to complete the survey requirements (web survey or
diary and sample collection) within one week of receipt of the sampling materials. As a
result, e-mail reminders were sent after Day 3 for all subjects and after Day 7 for non-
respondents. Future longitudinal surveys that use the web may want to tailor the
frequency of the e-mail reminders to the needs and habits of the household. Some study
subjects voiced complaints about the frequency of the e-mail reminders and noted that
they did not think it was necessary.
We demonstrated that timely shipments to participants are possible. However, return
shipments were more problematic. There were numerous instances in each of the cohort
studies in which supplies, like the breast pump, were not returned and staff needed to
follow up. The expense associated with staff communications with study participants
needs to be factored into future studies of a similar nature. Another important
consideration is the fact that participants sometimes ignored requests to not ship samples
for Saturday or holiday delivery. Careful attention needs to be paid to this aspect of
study implementation so that samples are in good condition when received.
Procrastination in the collection of and return of samples was also a problem in some
cases. The consequences on shipping equipment for subsequent sample collection events
5-4
-------
can be problematic, especially in large studies with limited equipment. It might be worth
considering a decrease in the incentive for very late samples.
Additional testing of samples that are collected in a liquid state but are required to be
shipped frozen or cool is warranted. We observed instances in the 0-1 cohort, for
example, in which samples were leaking or warm when they were received. Given the
variability that may exist in participants' homes for preparing, storing, and shipping
samples, increased cooling capacity during shipping and longer freezing times are among
the items needing to be investigated to ensure that samples are suitable for analyses when
they are received.
Tap water was relatively easy to collect in the 6-8 cohort study. However, we used a
straight forward and simple protocol that might not be adequate for the full range of
analyses needed for the NCS. Should preservatives be needed, e.g., thiosulfate or acid,
additional thought needs to be given to how this can be accomplished without
compromising participant safety. For the collection of water for the analysis of organic
compounds of low volatility, thiosulfate and a solid buffer could be added at the time of
collection, or even be placed in the bottle itself beforehand to optimize conditions for the
target analyte. Extra care will be needed to explain the safety concerns to participants. In
some cases, the risk might be deemed unacceptable. Alternatively, water could be
acidified upon receipt at the lab if stability over a 2-day shipping period is not a major
consideration. The sampling approach will obviously depend upon the purpose of the
sample. Shipping sample collection containers with additives should not be a concern
given the low toxicities. Leakage of sample containers should be addressed.
Use of the VOC sampling device was difficult for participants. Improved instructional
materials might improve the acceptability of samples arriving at the lab. However, the
situation observed in the 6-8 cohort might indicate that such devices cannot be used when
a high rate of acceptability is needed. This might also argue for a much more passive
approach such as monitoring based on sensors or some other means that does not require
the participant to do much. Additional research into an appropriate method for measuring
VOCs is warranted.
The feasibility of using less rugged instruments/devices should be evaluated as
appropriate. For example, an aerosol nephelometer is a device used to measure real-time
particle concentrations and is a candidate for use should particles be the focus of a
5-5
-------
specific hypothesis. This instrument is somewhat fragile and requires set-up. In such
cases participants would need to set up the device and take special care in packaging the
device for return shipment. Given that participants in the current demonstration study
had difficulty following assembly and packaging instructions, more complex and fragile
devices could be problematic.
Even though the purpose of these demonstration studies was to evaluate the feasibility of
having study participants collect biological and environmental samples, we need to
consider the use of technicians to aid in the collection of important, yet difficult, samples
to maximize the number of such samples that are acceptable for laboratory analysis. Less
expensive collection approaches for important samples is of no value if the resulting data
are unreliable.
We should also evaluate further the ability of study participants to comply with
temporality of environmental and biological samples. This factor will be very important
if a specific, time-dependent linkage is requested to evaluate a biological marker
following an environmental exposure. That is, temporal linkages can be critical,
depending on the hypothesis being tested, and it is important to understand anticipated
compliance and its effect of the value of the resulting data.
Each of the demonstration studies collected more information than could be analyzed for
this task order. Additional analyses of the information collected in the metaquestionnaires, such
as the extent to which study subjects in the 3-5 cohort successfully responded to the questions
about pesticide use, may be worthwhile. The chemical analyses samples that we performed were
conducted primarily for purposes of assessing compliance and to confirm that study subjects had
collected the samples as requested (e.g., creatinine in urine). Many of the samples were not
analyzed and further research may be appropriate including looking at associations between
demographic factors or responses on the metaquestionnaires and the levels of toxic substances in
the samples. Using the web in the NCS is a worthwhile consideration and may be appropriate
for data collection and study communications with the public, health care community, and study
participants. For these demonstration studies, we used an existing panel as a source of study
subjects. The NCS participants could be considered a special panel that would require
establishing links and processes for transmitting information. Additional research may be
appropriate to investigate the logistical aspects of implementing web-based data collection in the
NCS.
5-6
------- |