1
SF­
83
SUPPORTING
STATEMENT
ENVIRONMENTAL
PROTECTION
AGENCY
1.
IDENTIFICATION
OF
THE
INFORMATION
COLLECTION
1(
a)
TITLE
Detroit
Children's
Health
Study:
Health
Effects
of
Environmental
Exposures
among
Children
Living
in
the
Detroit,
Michigan
Area
1(
b)
ABSTRACT
This
epidemiologic
study
will
be
conducted,
and
the
information
collected,
by
the
Epidemiology
and
Biomarkers
Branch,
Human
Studies
Division,
National
Health
and
Environmental
Effects
Research
Laboratory,
Office
of
Research
and
Development,
U.
S.
Environmental
Protection
Agency
(
EPA).
The
participation
of
parents
and
children
in
this
collection
of
information
is
strictly
voluntary.
This
study
will
examine
of
the
role
of
long­
term
environmental
exposures,
particularly
local
sources
of
ambient
air
pollution,
in
children's
respiratory
health
conditions
such
as
allergies
and
asthma.
Children's
respiratory
health
will
be
assessed
by
a
parent­
completed
questionnaire
and
by
simple
measures
of
the
child's
lung
function
and
exhaled
breath.
The
twenty­
page
questionnaire
will
be
distributed
to
the
parents
of
children
enrolled
in
the
fourth
and
fifth
grades
at
selected
public
schools
in
the
Detroit
and
Dearborn.
The
questionnaire
includes
a
request
for
permission
for
their
child
to
participate
in
measurements
of
lung
parameters.
Later,
at
a
subset
of
these
schools,
those
children
with
parental
consent
will
be
asked
to
perform
simple
measurements
of
air
flow
rates,
lung
volumes,
and
exhaled
nitric
oxide.
The
schools
will
be
selected
based
on
school
location
with
respect
to
major
roadways
and
major
emission
point
sources.
Children's
long­
term
environmental
exposures
will
be
assessed
based
on
a
detailed
lifetime
residential
history
from
the
questionnaire,
supplemental
air
quality
measurements
collected
at
a
selected
group
of
schools
distributed
throughout
the
school
district,
and
an
air
quality
model
incorporating
geographic
information
on
local
sources
of
air
pollutants.
Central
site
measurements
of
ambient
air
pollutants
will
be
collected
at
the
time
of
the
lung
function
examinations.
The
collected
information
will
be
used
to
estimate
the
epidemiologic
associations
between
respiratory
health
conditions
and
long­
term
exposures
to
ambient
air
pollutants
in
the
study
community.
The
epidemiologic
analysis
will
examine
the
association
of
air
quality
parameters
with
a
higher
prevalence
of
respiratory
conditions,
with
lower
lung
function,
or
with
higher
levels
of
exhaled
nitric
oxide
after
appropriate
control
for
other
determinants
of
respiratory
health.
The
estimated
respondent
burden
for
this
study
is
8,250
hours
and
$
113,468.
This
study
requires
no
maintenance
of
records
by
the
respondents.

2.
NEED
FOR
AND
USE
OF
THE
INFORMATION
COLLECTION
2(
a)
NEED
/
AUTHORITY
FOR
THE
INFORMATION
COLLECTION
This
information
collection
is
part
of
a
research
program
consistent
with
the
strategic
plan
for
EPA's
Office
of
Research
and
Development
(
ORD)
which
has
identified
research
on
airborne
particulate
matter
as
a
highpriority
research
area
with
particular
emphasis
on
"
studying
the
heightened
sensitivity
/
susceptibility
of
certain
subpopulations
(
e.
g.,
children)".
EPA
has
broad
legislative
authority
to
establish
air
quality
criteria
and
to
conduct
research
to
support
these
criteria.
This
data
collection
is
for
an
epidemiological
study
to
identify
and
evaluate
exposure
to
and
effects
of
air
pollutants
on
human
health
as
part
of
EPA's
research
2
program
on
the
short­
term
and
long­
term
effects
of
air
pollutants
on
human
health
pursuant
to
42
U.
S.
C.
§
7403
(
d)
(
1)
(
A).
These
studies
will
provide
information
on
currently
regulated
air
pollutants,
including
nitrogen
dioxide,
fine
particulate
matter
(
PM2.5),
and
several
urban
air
toxics
(
volatile
organic
compounds).
These
studies
will
also
provide
information
on
the
epidemiologic
association
of
these
air
pollutants
with
respiratory
symptoms
and
illness
in
children.
EPA
has
identified
the
role
of
ambient
air
pollutants
in
the
initiation
and
exacerbation
of
asthma
as
an
issue
of
scientific
uncertainty
underlying
current
EPA
regulations
(
Asthma
Research
Strategy.
U.
S.
Environmental
Protection
Agency,
Office
of
Research
and
Development,
Washington,
DC,
EPA
600/
R­
01/
061,
2002).
EPA
has
recognized
children
with
asthma
as
a
susceptible
subpopulation
(
Human
Health
Research
Strategy,
U.
S.
Environmental
Protection
Agency,
Office
of
Research
and
Development,
Washington,
DC,
EPA
600/
R­
02/
050,
2003).
This
information
collection
is
part
of
ORD's
current
multiyear
plans
for
particulate
matter
(
www.
epa.
gov/
osp/
myp/
pm.
pdf,
pp.
46­
47)
and
human
health
 
susceptible
subpopulations
(
www.
epa.
gov/
osp/
myp/
hh.
pdf,
pp.
38).

2(
b)
PRACTICAL
UTILITY
/
USERS
OF
THE
DATA
The
information
will
primarily
be
used
by
epidemiologists
in
the
EPA's
Epidemiology
and
Biomarkers
Branch
to
prepare
scientific
papers
for
peer­
reviewed
publications
and
presentations
at
scientific
conferences.
These
peer­
reviewed
scientific
publications
will
be
further
reviewed
by
EPA's
National
Center
for
Environmental
Assessment
for
inclusion
in
Air
Quality
Criteria
Documents
and
may
be
used
by
EPA's
Office
of
Air
Quality
Planning
and
Standards
to
set
air
quality
standards
and
to
evaluate
the
benefit
of
the
current
National
Ambient
Air
Quality
Standards.
EPA
investigators
also
will
prepare
a
report
with
school­
specific
summary
information
from
this
study
for
use
by
the
local
community,
local
school
and
public
health
officials,
and
EPA
regional
scientists.
Information
from
the
survey
may
be
useful
to
local
school
and
public
health
officials
in
the
design
and
implementation
of
public
health
education
and
intervention
programs.

3.
NONDUPLICATION,
CONSULTATIONS,
AND
OTHER
COLLECTION
CRITERIA.
3(
a)
NONDUPLICATION
The
children's
health
questionnaires
and
lung
measurements
do
not
duplicate
any
existing
information.
While
this
study
closely
follows
the
design
and
conduct
of
several
previous
studies,
no
previous
study
has
been
conducted
of
the
health
effects
of
children's
exposures
to
motor
vehicle
emissions
in
the
Detroit
metropolitan
area.
Supplemental
air
quality
measurements
will
complement,
but
do
not
duplicate,
the
local
air
pollution
measurements
data
routinely
collected
and
reported
by
the
City
of
Detroit,
Wayne
County,
and
the
Michigan
Department
of
Environmental
Quality.
The
existing
five,
compliance­
oriented
monitors
are
located
at
sites
that
are
not
representative
of
the
full
range
of
population
exposures
in
the
Detroit
metropolitan
area.
To
account
for
the
spatial
variability
in
air
pollution
levels,
supplemental
air
quality
measurements
will
be
collected
at
a
selected
group
of
schools
distributed
throughout
the
school
district.
3(
b)
FEDERAL
REGISTER
NOTICE
The
announcement
of
a
public
comment
period
for
this
ICR
was
published
in
the
Federal
Register
on
December
7,
2004
(
69
FR
70680).
3(
c)
CONSULTATIONS
The
following
people
have
been
consulted
in
the
Detroit
metropolitan
area:
3
Dr.
Noble
Masseree,
Director,
Detroit
Department
of
Health
and
Wellness
Promotion
Dr.
Joyce
Hargrove,
Detroit
Department
of
Health
and
Wellness
Promotion
Dr.
Ann
Chevalier,
Air
Quality
Evaluation
Section,
Michigan
Dept.
of
Environmental
Quality
Dr.
Darrell
Rodgers,
Exec.
Director
of
Environmental
Health
and
Safety,
Detroit
Public
Schools
Dr.
Linda
Leddick,
Exec.
Director
of
Research,
Evaluation,
and
Assessment,
Detroit
Public
Schools
Dr.
John
Artis,
Superintendent,
Dearborn,
MI
City
School
District
Mr.
Jerry
Uchtorff,
Director
of
District
Operations,
Dearborn,
MI
City
School
District
Dr.
Adnan
Hammad,
Arab
Community
Center
for
Economic
and
Social
Services
Dr.
Kathryn
Savoie,
Arab
Community
Center
for
Economic
and
Social
Services
Ms.
Kathy
Edgren,
Project
Manager,
Community
Action
Against
Asthma
Dr.
Jerry
Keeler,
Michigan
Center
for
the
Environment
and
Children's
Health
Dr.
Christine
Johnson,
Henry
Ford
Health
System,
Dept.
of
Biostatistics
and
Research
Epidemiology
Dr.
Christine
Joseph,
Henry
Ford
Health
System,
Dept.
of
Biostatistics
and
Research
Epidemiology
The
following
scientists
have
provided
favorable
extramural
peer­
reviews
of
the
study
protocol:
Dr.
Joachim
Heinrich,
Institute
of
Epidemiology,
GSF
National
Research
Center
for
Environment
and
Health,
Munich,
Germany
Dr.
Michael
Brauer,
University
of
British
Columbia,
Vancouver,
Canada
This
information
collection
has
been
publicly
discussed
at
the
following
meetings:
EPA
Region
5
Briefing,
Chicago,
March,
2004
EPA
Scientist­
to­
Scientist
Meeting
on
Detroit
Research
Activities,
June,
2004
DEARS
Community
Kick­
off
Meeting,
Detroit,
June,
2004
EPA
Scientist­
to­
Scientist
Meeting
on
Near
Roadway
Studies,
January,
2005
EPA
Board
of
Scientific
Councilors
review
of
Human
Health
Program,
February,
2005
EPA
Board
of
Scientific
Councilors
review
of
Particulate
Matter
Program,
March,
2005
The
response
from
these
organizations
and
at
these
meetings
has
been
favorable.
3(
d)
EFFECTS
OF
LESS
FREQUENT
COLLECTION
The
questionnaire
information
and
the
lung
function
measurements
will
be
collected
only
once
during
the
study.
This
factor,
therefore,
does
not
apply
to
this
study.
3(
e)
GENERAL
GUIDELINES
We
have
adhered
to
all
of
OMB's
general
guidelines.
3(
f)
CONFIDENTIALITY
We
will
adhere
to
all
standard
steps
to
maintain
confidentiality.
These
include
the
exclusion
of
personal
identifiers
from
the
statistical
database
and
the
storage
of
hard
copies
in
locked
files.
Contractor
involvement
will
be
monitored
by
requiring
hard
copies
to
be
stored
at
EPA
after
the
contractor
has
completed
data
processing.
The
contractor
has
agreed
to
observe
the
rules
of
confidentiality
regarding
the
health
information
provided
by
individuals
and
their
families.
We
will
control
access
to
the
data
in
accordance
with
Federal
privacy
regulations
and
OMB
Circular
A­
110.
3(
g)
SENSITIVE
QUESTIONS
No
questions
of
a
sensitive
nature
will
be
required.
4
4.
THE
RESPONDENTS
AND
THE
INFORMATION
REQUESTED
4(
a)
RESPONDENTS
/
SIC
CODES
This
study
will
be
conducted
among
schoolchildren
who
are
enrolled
in
the
fourth
and
fifth
grades
of
selected
elementary
schools
in
the
Detroit
and
the
Dearborn
public
school
systems.
The
questionnaire
will
be
completed
by
the
parents
of
these
children.
At
a
subset
of
schools,
children
with
written
parental
permission
may
participate
in
measurements
of
lung
parameters.
Schoolchildren
were
selected
for
this
study
because
they
may
constitute
a
susceptible
population
due
to
the
rapid
growth
and
development
of
the
respiratory
system
in
early
childhood.
Elementary
schoolchildren
generally
do
not
smoke,
have
less
spatial
mobility
than
most
adults,
and
have
a
more
limited
lifetime
residential
history
than
adults.
4(
b)
INFORMATION
REQUESTED
(
i)
Data
Items:
The
questionnaire
is
attached,
and
a
detailed
discussion
of
the
questionnaire
design
is
presented
in
section
B.
2(
d)
below.
The
requested
information
includes
questions
specific
to
the
child
including
residential
history,
general
demographic
information,
childhood
respiratory
illness
and
history
of
asthma,
current
respiratory
health
conditions,
and
general
household
characteristics.
The
lung
measurements
will
include
the
child's
peak
expiratory
flow
rate,
forced
expiratory
volume
at
one
second,
forced
vital
capacity,
and
exhaled
nitric
oxide.
The
child's
height
and
weight
will
also
be
measured
as
part
of
the
pulmonary
examination.
Measurements
of
fine
particles,
nitrogen
dioxide,
and
selected
volatile
organic
compounds
will
be
collected
outside
a
subset
of
elementary
schools
along
with
publicly­
available
geographic
information
on
the
locations
of
schools,
major
roadways,
and
major
point­
sources
of
air
pollutants.
No
maintenance
of
records
by
respondents
is
required
for
this
study.
(
ii)
Respondent
Activities
The
parents
of
each
child
will
be
asked
to
complete
a
twenty­
page
questionnaire
that
will
be
available
in
English,
Spanish,
and
Arabic.
The
parents
will
also
be
asked
to
decide
whether
their
child
should
participate
in
the
lung
examination
by
responding
to
the
informed
consent
form.
Finally,
the
parents
will
be
requested
to
seal
the
completed
questionnaire
into
the
provided
envelope
and
return
the
package
as
directed
to
the
child's
school.
In
a
subset
of
schools,
children
with
parental
permission
will
be
asked
to
participate
in
standard
measurements
of
breathing
ability
and
exhaled
breath.
The
breathing
tests
involve
blowing
three
to
eight
times
into
a
tube
(
changed
for
each
child).
The
breathing
test
is
no
more
stressful
than
blowing
out
the
candles
on
a
birthday
cake.
Usually
four
to
five
forced
breathing
maneuvers
are
sufficient
to
capture
the
flow
rates
in
children
from
this
age
group.

5.
THE
INFORMATION
COLLECTED
­­
AGENCY
ACTIVITIES,
COLLECTION
METHODOLOGY,
AND
INFORMATION
MANAGEMENT
5(
a)
AGENCY
ACTIVITIES
EPA's
Epidemiology
and
Biomarkers
Branch
will
be
responsible
for
administering
the
questionnaire,
conducting
the
examinations
of
lung
function
and
exhaled
breath,
answering
respondent
questions,
reviewing
data
quality,
converting
the
data
to
electronic
form,
developing
and
maintaining
the
database,
epidemiologic
analysis,
and
preparation
of
reports
and
peer­
reviewed
publications.
Some
of
these
activities
may
be
performed
by
an
EPA
contractor.
5
5(
b)
COLLECTION
METHODOLOGY
AND
MANAGEMENT
The
respiratory
symptom
and
child
characteristics
information
will
be
collected
using
a
scannable
questionnaire.
The
lung
function
measurements
will
be
collected
by
automated
spirometers
and
nitric
oxide
analyzers
directly
into
an
electronic
format.
The
air
quality
measurements
will
be
conducted
by
an
EPA
contractor
using
Federal
reference
methods
and
by
the
existing
air
quality
monitoring
system.
No
additional
burden
will
be
placed
on
state
and
local
air
pollution
agencies
in
the
Detroit,
Michigan
metropolitan
area.
All
data
will
be
reviewed
for
unusual
or
unacceptable
values.
The
scanned
questionnaire
data
will
be
verified
through
comparisons
with
the
original
records.
Statistical
data
will
be
maintained
in
electronic
format
using
the
Statistical
Analysis
System
(
SAS).
We
will
control
access
to
the
data
in
accordance
with
Federal
privacy
regulations
and
OMB
Circular
A­
110.
5(
c)
IMPACTS
ON
SMALL
BUSINESSES
OR
OTHER
SMALL
ENTITIES
The
procedures
for
questionnaire
distribution
and
collection,
and
for
lung
function
measurements,
have
been
designed
to
minimize
the
burden
to
the
teachers
and
administrators
of
the
Detroit
and
Dearborn
Public
Schools.
We
have
taken
special
steps
to
brief
the
school
superintendents
regarding
this
study
and
we
will
conduct
informational
meetings
with
the
involved
teachers.

5(
d)
COLLECTION
SCHEDULE
The
study
will
be
conducted
over
the
2005­
2006
school
year
and
will
be
coordinated
with
the
academic
calendar.
The
study
time
line
is
(
a)
prepare
the
questionnaire
packets
by
mid­
September,
(
b)
hold
short
meetings
at
each
school
to
inform
principals,
teachers,
and
school
nurses
between
early­
October
and
mid­
November,
(
c)
distribute
the
questionnaire
packets
to
students
following
each
school
meeting,
(
d)
collect
the
returned
questionnaires
prior
to
mid­
December,
(
e)
prepare
an
annotated
list
of
children
with
parental
permission
for
the
lung
function
test
by
February,
2006,
and
(
f)
conduct
the
lung
function
tests
of
the
students
with
parental
permission
during
March
and
April.
Data
sets
will
be
delivered
by
February
for
the
questionnaires,
and
by
September
for
the
lung
examinations.
These
preliminary
data
sets
will
be
extensively
reviewed
for
quality
assurance
prior
to
analysis.

6.
ESTIMATING
THE
BURDEN
AND
COST
OF
THE
COLLECTION
6(
a)
ESTIMATING
RESPONDENT
BURDEN
Questionnaire:
Based
on
a
similar
information
collection
in
El
Paso,
Texas
(
ICR
1940.02),
we
estimate
that
each
respondent
parent
will
spend
an
average
of
25
minutes
completing
the
questionnaire.
This
includes
time
for
reviewing
the
informed
consent
form
and
instructions,
filling
out
the
questionnaire,
and
returning
the
questionnaire.
No
record
keeping
is
required.
Lung
Function
Tests:
Based
on
our
field
experience,
we
estimate
that
each
of
the
3,500
selected
children
will
spend
an
average
of
30
minutes
performing
the
lung
function
examination
and
that
2,000
of
these
children
will
spend
an
additional
15
minutes
performing
the
exhaled
nitric
oxide
measure.
This
includes
time
spent
giving
assent
to
participate,
measuring
height
and
weight,
receiving
instruction,
and
performing
repeated
breathing
maneuvers.
These
children
are
a
subset
of
the
children
with
completed
parental
questionnaires.
6
TABLE
1.
ESTIMATED
RESPONDENT
BURDEN
Type
of
Respondent
Respondent
Activities
Estimated
Number
of
Respondents
Burden
Hours
Frequency
Annual
Reporting
Bu
Annual
Cost
Parent
Complete
Questionnaire
15,000
0.40
hr.
1
6000
hr
$
101,880a
Child
Lung
Function
Examination
3,500
0.50
hr.
1
1750
hr
$
9,013
Child
Exhaled
Nitric
Oxide
Measure
2,000
0.25
hr.
1
500
hr
$
2,575
Totals
20,500
8,250
hr
$
113,468
a
$
16.98/
hour
(
average
hourly
wage
for
private
industry,
July
2003)
b
$
5.15/
hour
(
minimum
wage)
ANNUAL
REPORTING
BURDEN:
8,250
hours
ANNUAL
RESPONDENT
COST:
$
113,468
NO
ANNUAL
RECORD
KEEPING
BURDEN
7
6(
b)
ESTIMATING
RESPONDENT
COSTS
Questionnaire:
The
cost
estimate
for
the
parental
respondent
is
based
on
the
average
hourly
earnings
for
private
industry
workers
reported
by
the
US
Bureau
of
Labor
Statistics'
National
Compensation
Survey
(
July
2003)
at
$
16.98
per
hour.
Given
the
0.40
hour
burden,
the
respondent
cost
for
each
child's
parent
is
$
6.79.
Lung
Function
and
Exhaled
Breath
Tests:
The
wage
estimate
for
the
child
was
the
minimum
hourly
wage
($
5.15
per
hour).
Given
the
0.75
hour
burden,
the
respondent
cost
for
each
child
is
$
3.86.
8
TABLE
2.
AGENCY
COST
BURDEN
Burden
Hours/
Unit
Cost
($)

Agency
Activities
Contractor
EPA
Units
Total
Hours
Contractor
$
108/
hour
EPA
$
54/
hour
Prepare
and
format
questionnaires
411
240.0
1
651
$
44,388
$
12,960
Prepare
questionnaire
packages
0.015
0.0
15,000
225
$
24,300
$
0
Plan
and
conduct
school
meetings
7.25
2.0
80
740
$
62,640
$
8,640
Distribute/
collect
questionnaires
14.00
2.0
80
1,280
$
120,960
$
8,640
Process
collected
questionnaires
0.05
0.05
15,000
1,500
$
81,000
$
40,500
Prepare
statistical
data
sets
and
reports
360
0.0
1
360
$
38,880
$
0
Perform
lung
function
examinations
1.50
0.2
3,500
5,950
$
567,000
$
37,800
Perform
exhaled
nitric
oxide
measures
0.75
0.2
2,000
1,900
$
162,000
$
21,600
Project
management
/
data
analysis
360
960.0
1
1,320
$
38,880
$
51,840
TOTAL
1
13,926
$
1,140,048
$
181,980
AGENCY
TOTAL
ANNUAL
BURDEN:
13,926
hours
AGENCY
TOTAL
ANNUAL
COST:
$
1,322,028
9
6(
c)
ESTIMATING
AGENCY
BURDEN
AND
COST
These
contractor
and
agency
burden
estimates
were
based
on
our
prior
experience
in
developing
and
gathering
information
for
research
purposes.
The
agency
costs
have
been
based
on
a
GS­
13(
6).
This
ICR
is
a
one­
time
collection
of
information
with
analysis
extending
into
future
years.
10
6(
d)
ESTIMATING
THE
RESPONDENT
UNIVERSE
AND
TOTAL
BURDEN
AND
COSTS
Based
on
sample
size
requirements,
the
respondent
universe
is
15,000
parents
for
the
questionnaire,
3,500
children
for
the
lung
function
and
2,000
children
for
the
exhaled
nitric
oxide
measures.
Thus,
the
total
estimated
respondent
burden
for
this
study
is
8,250
hours
and
$
113,468.
6(
e)
BOTTOM
LINE
BURDEN
HOURS
AND
COSTS
/
BURDEN
TABLES
The
estimated
respondent
burden
for
this
study
is
8,250
hours
and
$
113,468.
The
estimated
agency
cost
for
conducting
this
study
is
13,926
hours
and
$
1,322,028.
6(
f)
BURDEN
STATEMENTS
Questionnaire:
Parental
reporting
burden
for
this
collection
of
information
is
estimated
to
average
25
minutes
and
will
require
no
record
keeping.
This
includes
the
time
for
reviewing
the
informed
consent
form,
reading
the
instructions,
filling
out
the
questionnaire,
and
returning
the
questionnaire.
Lung
Function
and
Exhaled
Breath
Tests:
The
children's
reporting
burden
for
this
collection
of
information
is
estimated
to
average
45
minutes
per
child.
This
includes
times
for
receiving
instruction,
giving
verbal
assent
to
participate,
and
performing
repeated
breathing
maneuvers.

An
agency
may
not
conduct
or
sponsor,
and
a
person
is
not
required
to
respond
to,
a
collection
of
information
unless
it
displays
a
currently
valid
OMB
control
number.
The
OMB
control
numbers
for
EPA's
regulations
are
listed
in
40
CFR
part
9
and
48
CFR
chapter
15.

To
comment
on
the
Agency's
need
for
this
information,
the
accuracy
of
the
provided
burden
estimates,
and
any
suggested
methods
for
minimizing
respondent
burden,
including
the
use
of
automated
collection
techniques,
EPA
has
established
a
public
docket
for
this
ICR
under
Docket
ID
No.
ORD­
2004­
0019,
which
is
available
for
public
viewing
at
the
Office
of
Environmental
Information
(
OEI)
Docket
in
the
EPA
Docket
Center
(
EPA/
DC),
EPA
West,
Room
B102,
1301
Constitution
Ave.,
NW,
Washington,
DC.
The
EPA
Docket
Center
Public
Reading
Room
is
open
from
8:
30
a.
m.
to
4:
30
p.
m.,
Monday
through
Friday,
excluding
legal
holidays.
The
telephone
number
for
the
Reading
Room
is
(
202)
566­
1744,
and
the
telephone
number
for
the
OEI
Docket
is
(
202)
566­
1752.
An
electronic
version
of
the
public
docket
is
available
through
EPA
Dockets
(
EDOCKET)
at
http://
www.
epa.
gov/
edocket.
Use
EDOCKET
to
submit
or
view
public
comments,
access
the
index
listing
of
the
contents
of
the
public
docket,
and
to
access
those
documents
in
the
public
docket
that
are
available
electronically.
Once
in
the
system,
select
"
search,"
then
key
in
the
docket
ID
number
identified
above.
Also,
you
can
send
comments
to
the
Office
of
Information
and
Regulatory
Affairs,
Office
of
Management
and
Budget,
725
17th
Street,
NW,
Washington,
DC
20503,
Attention:
Desk
Officer
for
EPA.
Please
include
the
EPA
Docket
ID
No.
ORD­
2004­
0019
in
any
correspondence.
11
B.
STATISTICAL
APPROACH
1.
SURVEY
OBJECTIVES,
KEY
VARIABLES,
AND
OTHER
PRELIMINARIES
1(
a)
SURVEY
OBJECTIVES
This
study
will
examine
of
the
role
of
long­
term
environmental
exposures,
particularly
local
sources
of
ambient
air
pollution,
in
children's
respiratory
health
conditions
such
as
allergies
and
asthma.
The
study
objective
is
to
estimate
the
epidemiologic
associations
between
long­
term
environmental
exposures
and
prevalences
of
respiratory
conditions
among
children
living
in
a
major
U.
S.
urban
area.
EPA's
National
Center
for
Environmental
Assessment
will
consider
the
findings
of
this
study
in
conjunction
with
similar
epidemiologic
studies
and
with
animal
and
in
vitro
toxicologic
studies.
EPA
has
previously
held
that
epidemiologic
associations
found
in
one
study
population
may
be
generalized
to
similar
populations
exposed
to
similar
levels
and
composition
of
ambient
air
pollution.
The
current
information
collection
will
fully
characterize
the
study
population
with
respect
to
the
observed
prevalence
of
respiratory
conditions
and
environmental
exposures.
EPA
does
not
wish
to
estimate
the
overall
prevalence
of
any
respiratory
condition
or
environmental
exposure
for
the
general
U.
S.
population
on
the
basis
of
this
information
collection.
No
attempt
will
be
made
to
generalize
the
observed
prevalence
of
any
respiratory
condition
or
environmental
exposure
to
any
other
community
or
to
the
nation
as
a
whole.
1(
b)
KEY
VARIABLES
The
key
respiratory
condition
is
the
child's
lifetime
history
of
asthma.
Other
asthma­
related
respiratory
symptoms
and
conditions
of
interest
include
hay
fever
or
nasal
allergies,
wheeze,
chronic
morning
cough,
productive
cough
apart
from
colds,
and
bronchitis.
The
key
lung
measures
are
the
child's
peak
expiratory
flow
rate,
forced
expiratory
volume
at
one
second,
forced
vital
capacity,
and
exhaled
nitric
oxide.
The
child's
height
and
weight
will
also
be
measured
as
part
of
the
pulmonary
examination.
The
key
environmental
exposures
are
indicated
by
the
child's
lifetime
residential
history.
Ambient
measurements
of
fine
particles,
nitrogen
dioxide,
and
selected
volatile
organic
compounds
will
be
collected
at
a
subset
of
elementary
schools.
These
ambient
measurements
will
be
modeled
with
publicly­
available
geographic
information
on
local
sources
of
air
pollutants.
This
exposure
assessment
model
will
be
combined
with
each
child's
residential
history
and
used
to
estimate
each
child's
lifetime
exposure
to
the
modeled
air
pollutants.
The
key
alternative
determinants
of
respiratory
health
(
potential
confounders
of
the
epidemiologic
association)
are
age,
sex,
race,
ethnicity,
severe
chest
injuries
or
operations,
parental
education
(
socioeconomic
status),
parental
history
of
respiratory
illness,
household
characteristics,
exposure
to
environmental
tobacco
smoke
in
the
home,
and
indoor
combustion
sources.
1(
c)
STATISTICAL
APPROACH
The
statistical
approach
proposed
for
this
information
collection
is
a
standard
method
employed
by
many
academic
investigators
in
epidemiologic
studies
of
ambient
air
pollution,
including
the
Harvard
Six
Cities
Study,
the
Harvard
/
Health
Canada
Twenty­
four
Cities
Study,
the
Kanawha
Valley
Health
Study,
the
Bitterfeld
Study
(
Germany),
and
the
Southern
California
Children's
Health
Study.
This
statistical
approach
was
also
used
in
EPA's
El
Paso
Children's
Health
Study
(
ICR
1940.02).
Epidemiologic
studies
do
not
require
the
selection
of
a
random
sample
of
subjects
from
a
larger
target
population
for
internal
validity,
provided
that
the
selection
process
is
not
related
to
both
the
underlying
12
distribution
of
health
conditions
in
the
study
area
and
the
exposure
of
interest.
The
cohort
of
fourth
and
fifth
grade
children
at
selected
public
elementary
schools
in
Detroit
and
Dearborn
will
provide
a
reasonable
study
population
that
will
fully
meet
the
objectives
of
this
information
collection.
The
study
will
fully
characterize
the
study
cohort
so
that
the
epidemiologic
associations
found
in
this
study
might
be
compared
with
associations
reported
for
other
studies.
As
in
other
epidemiologic
studies,
the
statistical
analyses
will
account
for
other
major
determinants
of
the
respiratory
conditions
(
potential
confounders)
through
stratification;
either
by
modeling
or
by
restriction.
For
rare
genetic
or
accidental
conditions
that
are
major
determinants
of
respiratory
conditions,
such
as
cystic
fibrosis,
chest
injuries
or
chest
operations,
confounding
will
be
prevented
through
restriction
of
the
study
cohort
to
exclude
subjects
with
these
conditions.
The
information
on
respiratory
conditions
will
then
be
analyzed
by
logistic
regression
(
symptom
prevalence)
or
by
linear
regression
(
lung
measurements)
using
epidemiologic
models
that
will
adjust
for
other
known
determinants
of
these
respiratory
conditions.
1(
d)
FEASIBILITY
Obstacles:
Based
on
previous
experience
with
health
diaries,
the
major
problem
has
been
non­
response.
This
is
usually
improved
by
providing
an
envelope
to
return
the
questionnaire
and
non­
monetary
incentives
for
participation.
Sufficient
funds:
This
study
has
been
budgeted
with
sufficient
funds
from
EPA's
Science
and
Technology
budget
authority.
The
funding
already
secured
in
fiscal
year
2005
is
adequate
for
the
entire
information
collection.
Timeline:
The
information
collection
will
be
completed
by
October
2006.
A
detailed
schedule
is
proved
in
section
A.
5(
d)
above.
This
specific
information
collection
will
not
be
repeated.

2.
SURVEY
DESIGN
2(
a)
Target
Population
and
Coverage
Epidemiologic
studies
do
not
require
the
selection
of
a
random
sample
of
subjects
from
a
larger
target
population
for
internal
validity.
Thus,
the
target
population
and
the
sampling
frame
are
essentially
identical
for
this
study.
The
study
will
fully
characterize
the
students
actually
enrolled
in
the
study
cohort
so
that
the
epidemiologic
associations
found
in
this
study
might
be
compared
with
associations
reported
for
other
studies.
In
keeping
with
standard
methods
for
risk
assessment,
EPA
has
previously
held
that
epidemiologic
associations
found
in
one
study
population
may
be
generalized
to
similar
populations
exposed
to
similar
levels
and
composition
of
ambient
air
pollution.
In
keeping
with
EPA's
multi­
year
research
plans
and
the
study
objective,
this
information
collection
will
examine
the
epidemiologic
associations
between
long­
term
environmental
exposures
and
prevalences
of
respiratory
conditions
among
children
living
in
a
major
U.
S.
urban
area.
The
Detroit
metropolitan
area
was
selected
because
of
the
presence
of
local
sources
of
air
pollutants,
including
mobile
sources
and
major
emission
point­
sources;
but
not
with
any
prior
knowledge
regarding
the
overall
prevalence
of
respiratory
conditions
or
the
distribution
of
respiratory
conditions
across
the
metropolitan
area.
The
choice
of
elementary
schoolchildren
as
a
target
population
is
responsive
to
a
Congressional
mandate
to
evaluate
the
effects
of
environmental
pollutants
on
sensitive
sub­
populations.
2(
b)
Sample
Design
(
i)
Sampling
frame.
13
The
sampling
frame
will
be
students
enrolled
in
the
fourth
and
fifth
grades
of
selected
public
elementary
schools
in
Detroit
and
Dearborn
during
the
2005­
2006
academic
year.
Children
below
the
fourth
grade
are
often
not
capable
of
performing
consistently
in
lung
function
examinations,
while
children
enrolled
in
middle
schools
may
have
begun
smoking
cigarettes.
Public
schools
will
be
selected
on
the
basis
of
proximity
to
major
local
sources
of
air
pollutants
to
ensure
an
exposure
gradient.
(
ii)
Sample
size.
The
proposed
sample
size
is
up
to
15,000
questionnaires,
up
to
3,500
lung
function
measurements,
and
2,000
measurements
of
exhaled
nitric
oxide.
The
Detroit
and
Dearborn
public
schools
have
sufficient
schoolchildren
enrolled
in
the
fourth
and
fifth
grades
to
provide
the
necessary
sample
size.
A
complete
discussion
of
the
statistical
power
and
the
precision
targets
for
the
study
is
in
Appendix
A.
(
iii)
Stratification
variables.
To
ensure
proper
epidemiologic
control
of
race
and
socioeconomic
status
as
confounders,
the
selection
of
public
schools
will
be
stratified
by
city
(
Detroit
and
Dearborn)
and
by
the
five
Census
enumeration
districts
within
Detroit,
for
a
total
of
six
strata.
Each
of
these
strata
is
more
homogeneous
with
respect
to
race
and
socioeconomic
status
than
the
metropolitan
area
as
a
whole.
(
iv)
Sampling
method.
For
each
of
these
six
strata,
public
elementary
schools
will
be
selected
to
provide
an
exposure
gradient
across
each
stratum.
The
selection
will
be
based
on
the
proximity
to
major
local
sources
of
air
pollutants
and
will
be
performed
by
exposure
assessment
specialists
who
will
be
blinded
to
the
distribution
of
health
status
across
the
study
area.
A
subset
of
these
elementary
schools
will
be
selected
for
the
measurements
of
lung
function
and
exhaled
breath.
Among
the
selected
public
elementary
schools
in
Detroit
and
Dearborn,
all
fourth
and
fifth
grade
students
will
be
eligible
for
the
questionnaire
study.
At
the
subset
of
schools
selected
for
the
lung
examinations,
all
students
with
parental
permission
will
be
eligible
for
the
measurements
of
lung
parameters.
Thus
the
sampling
is
on
the
level
of
the
school
and
not
among
individual
students
within
a
selected
school.
Schoolspecific
effects
will
be
handled
analytically
through
a
random­
effect
model.
(
v)
Multi­
stage
sampling.
Not
applicable.
2(
c)
PRECISION
REQUIREMENTS
(
i)
Precision
targets.
The
survey
is
designed
to
show,
with
95%
confidence,
the
odds
ratio
associated
specifically
with
a
10
ppb
increment
of
nitrogen
dioxide
provided
that
risk
is
21%
or
greater.
A
more
complete
discussion
of
the
precision
targets
is
provided
in
Appendix
A.
(
ii)
Non­
sampling
error.
Exposure
misclassification
will
be
minimized
by
the
collection
of
the
child's
residential
history,
collecting
air
pollution
measurements
at
selected
locations
throughout
the
study
area,
and
by
the
development
of
a
GIS­
based
exposure
assessment
model.
Differential
reporting
of
respiratory
conditions
will
be
minimized
by
the
use
of
a
standard
questionnaire
and
by
a
standardized
protocol
for
the
collection
of
information
across
the
study
area.
Non­
differential
under­
reporting
of
respiratory
conditions
will
reduce
the
statistical
power
of
the
study,
but
will
not
create
a
bias
away
from
the
null
hypothesis
of
no
air
pollution
effect.
Confounding
will
be
minimized
by
restriction
and
analytically
by
an
epidemiologic
model.
Children
with
rare
respiratory
conditions
such
as
cystic
fibrosis
will
be
excluded
from
the
analysis.
School­
level
clustering
effects
will
be
handled
analytically
through
a
random­
effect
model.
Epidemiologic
models
will
14
make
adjustments
for
other
major
determinants
of
the
health
outcome
including
age,
sex,
race,
ethnicity,
socioeconomic
status
(
as
indexed
by
parental
education),
and
household
characteristics.
The
epidemiologic
model
for
lung
function
will
also
adjust
for
height
and
weight.
A
substantial
portion
of
the
questionnaire
is
devoted
to
the
determination
of
information
on
potential
confounders.
2(
d)
QUESTIONNAIRE
DESIGN
The
questionnaire
is
divided
into
several
sections
of
related
questions:
personal
identifying
information,
child's
personal
data,
child's
medical
history,
child's
current
health
(
common
respiratory
symptoms,
chest
illnesses,
gastrointestinal
illness,
and
asthma),
child's
nutrition,
child's
activities,
home
characteristics,
family
history,
and
current
household
members.
The
structures
of
the
specific
questions
in
each
section
are
detailed
below.
The
numbers
in
parentheses
refer
to
the
question
number
on
the
questionnaire.
A
copy
of
the
questionnaire
is
attached.
The
personal
identifying
information
on
the
detachable
cover
page
of
the
questionnaire
(
child's
name,
address,
phone
number,
school,
grade,
teacher,
and
parent's
name)
are
required
to
correctly
identify
the
child.
The
permission
signature
and
date
are
required
to
document
informed
consent
by
a
parent
or
guardian
for
subsequent
testing
of
the
child's
breathing
ability
and
exhaled
breath.
The
information
on
residential
history
(
child's
home
address
at
birth,
the
address
of
the
child's
biological
mother
when
the
child
was
born,
the
child's
address
when
he/
she
entered
the
first
grade,
and
the
last
previous
addresses
for
children
who
moved
in
the
prior
two
years)
are
necessary
for
exposure
assessment.
These
cover
pages
will
be
detached
and
stored
physically
separate
from
the
statistical
information
on
the
remainder
of
the
questionnaire.
The
personal
identifying
information
will
not
become
part
of
the
statistical
data
base.
The
child's
personal
data
are
required
to
determine
the
child's
age
(
1,
2,
3),
weight
(
4),
height
(
5),
sex
(
6),
Hispanic
status
(
7)
and
race
(
8).
These
variables
are
major
determinants
of
respiratory
health
and
potential
confounders
of
the
association
of
interest.
The
current
date
and
the
child's
birth
date
will
allow
us
to
check
the
direct
parental
report
of
the
child's
age.
The
self­
identification
of
ethnicity
is
especially
relevant
for
the
evaluation
of
the
subsequent
lung
function
measures
which
show
considerable
ethnic
variation
in
anthropomorphic
relationships.
The
child's
medical
history
section
includes
three
distinct
categories
of
questions:
restriction
variables
(
16­
21),
outcome
measures
(
22­
26,
28­
38),
and
potential
effect
modifiers
(
9­
15,
27).
Children
with
rare
conditions
that
are
major
determinants
respiratory
health
(
16­
21)
must
be
restricted
from
the
sample
to
prevent
confounding
of
the
epidemiologic
association.
Severe
chest
illnesses
(
22­
23),
especially
before
the
age
of
2
years,
are
an
important
outcome
of
interest
and
may
also
increase
the
child's
susceptibility
to
environmental
insults.
Hay
fever
and
other
allergies
characteristic
of
atopy
(
24­
26)
are
an
important
outcome
of
interest
and
a
potential
indicator
of
increased
susceptibility
to
environmental
insults.
Children
with
low
birth
weight
or
born
prematurely,
especially
those
kept
in
intensive
care
or
given
supplemental
oxygen,
may
be
more
sensitive
to
inhalation
exposures
in
later
life
(
9­
15).
Diabetics
(
27)
have
recently
been
identified
as
a
sensitive
subpopulation.
The
child's
lifetime
history
of
asthma
is
a
key
outcome
variable
for
this
study.
The
preceding
questions
on
a
physician's
diagnosis
of
asthma
related
to
severe
chest
illness
(
22­
23)
may
not
be
sufficient
to
identify
children
with
mild
or
subclinical
asthma.
As
a
further
screening
for
asthma,
this
section
of
the
questionnaire
is
modeled
on
the
elements
of
a
careful
medical
history
for
wheeze­
related
symptoms
(
28­
33),
medication
(
34),
emergency
visits
(
35),
and
hospitalization
(
36).
Asthma
may
be
indicated
by
wheezing
with
colds
(
28­
29),
wheezing
apart
from
colds
(
30),
shortness
of
breath
with
wheezing
(
31),
night­
time
wheezing
(
33),
and
exercise
induced
wheezing
(
32).
We
also
ask
about
the
life­
time
history
15
of
wheeze
(
28a,
33a),
the
seasonal
pattern
of
wheezing
(
28b),
and
the
severity
of
the
wheezing
(
30a­
b).
We
ask
about
the
child's
history
of
wheeze
medication
(
34)
and
severity
of
symptoms
(
35­
36).
This
section
concludes
with
questions
regarding
the
clinical
diagnosis
of
asthma
(
37),
the
timing
of
this
diagnosis
(
37a),
and
the
current
level
of
asthma
(
38).
By
providing
parents
of
wheezing
children
an
opportunity
to
report
on
their
children's
symptoms
(
28­
36),
we
hope
to
focus
questions
37
and
38
solely
on
clinical
asthma.
The
child's
current
health
section
(
39­
50)
relates
to
health
outcomes
during
the
12
months
prior
to
the
questionnaire.
While
this
section
is
lengthy,
the
information
solicited
by
these
questions
is
important
and
directly
related
to
the
purpose
of
this
study.
The
parents
of
healthy
children
will
rapidly
pass
through
these
questions,
while
the
parents
of
children
with
respiratory
conditions
are
given
the
opportunity
to
fully
and
accurately
describe
these
conditions.
Common
respiratory
conditions
including
cough
(
39,
40),
phlegm
(
41,
42),
hay
fever
(
43),
runny
nose
(
44),
ear
infections
(
45)
and
general
chest
illnesses
(
46)
are
covered
before
proceeding
to
other
respiratory
illnesses:
nasal
allergies
(
47),
asthma
(
48),
pneumonia
(
49),
and
bronchitis
(
50).
We
want
to
provide
parents
of
children
with
relatively
mild
respiratory
symptoms
an
opportunity
to
report
these
symptoms
in
order
to
limit
the
over­
reporting
of
the
more
severe
respiratory
illnesses.
Chronic
cough
(
39a,
40a),
especially
chronic
morning
cough
(
39a),
and
chronic
productive
cough
(
41­
42),
especially
apart
from
colds
(
42a),
are
indicators
of
chronic
respiratory
conditions
such
as
asthma
and
bronchitis,
but
these
symptoms
may
be
present
in
children
with
subclinical
disease.
The
hay
fever
question
is
repeated
with
reference
to
the
last
12
months
and
treatment
by
a
physician
for
this
condition
provides
an
indicator
of
the
condition's
severity
(
43).
Similarly,
persistent
runny
nose
and
ear
infections
are
an
indicator
of
upper
respiratory
infections
(
44­
45).
The
questions
on
chest
illness
questions
focus
on
restriction
of
activity
(
46a),
a
physician's
visit
related
to
the
illness
(
46b),
and
the
physician's
diagnosis
(
46c).
The
questions
on
specific
respiratory
illnesses
(
47­
50)
address
the
study's
principal
outcomes
and
focus
on
differentiating
between
nasal
allergies,
asthma,
pneumonia,
and
bronchitis;
on
the
severity
of
each
condition;
and
on
seasonal
patterns
of
disease.
In
each
of
the
four
subsections,
the
first
question
refocuses
the
parent
on
illnesses
during
the
last
12
months
and
asks
about
the
frequency
and
duration
of
such
illnesses
(
47,
47a;
48,
48a;
49,
49a;
50,
50a).
The
questions
on
visits
to
doctor
or
health
practitioner
provide
an
indication
of
the
severity
of
the
illness
(
47b,
48b,
49b,
50b).
The
questions
on
medications
provide
a
check
on
the
differential
diagnosis
of
these
four
conditions,
for
example,
steroids
would
be
inappropriate
for
pneumonia
and
antibiotics
are
not
generally
prescribed
for
asthma
(
47d­
e,
48d­
e,
49d­
e,
50d­
e).
Since
air
pollution
and
respiratory
disease
follow
different
seasonal
patterns,
we
ask
about
illness
seasonality
(
47c,
48c,
49c,
50c).
These
questions
on
health
outcomes
are
placed
at
the
beginning
of
the
questionnaire
to
ensure
that
the
parental
respondent
is
relatively
fresh
and
that
the
health
focus
of
the
questionnaire
is
clear.
The
child's
nutrition
section
(
51­
55)
uses
a
common
question
format
to
assess
the
frequency
of
specific
foods
in
the
child's
diet
(
51,
52)
with
particular
attention
to
dietary
sources
of
anti­
oxidants.
Dietary
anti­
oxidants
may
play
a
role
in
modifying
the
sensitivity
of
children
and
adults
to
air
pollutants.
The
questions
on
breast
feeding
during
infancy
(
53­
55)
have
shown
a
relationship
to
respiratory
allergies
and
asthma
in
previous
studies.
The
child's
activities
section
(
56­
77)
collects
information
on
potential
environmentally­
related
behavioral
problems
(
56­
68,
70­
74),
on
exposures
to
tap
water
(
69),
on
participation
in
sport
teams
related
to
time
spent
outdoors
(
75),
and
on
school­
related
exposures
to
vehicle
exhaust
(
76­
77).
16
The
home
characteristics
section
includes
three
distinct
categories
of
questions:
exposure­
related
variables
(
78­
83),
potential
confounders
(
87­
98,
101­
104,
106­
107),
and
potential
effect
modifiers
(
84­
86,
90,
99­
100,
105).
The
exposure
assessment
based
on
residential
history
may
be
more
accurate
for
children
who
have
lived
a
greater
portion
of
their
lives
in
the
city
(
78),
near
their
current
residence
(
79),
in
their
current
residence
(
80,
82­
83),
and
for
children
who
spend
more
time
in
the
city
during
the
past
year
(
81).
Time
spent
outdoors
by
children
(
84)
as
well
as
exposures
to
other
children
in
pre­
school
settings
(
85­
86)
may
be
related
to
increased
sensitivity
to
air
pollutants.
Socioeconomic
status
as
a
potential
confounder
may
be
indicated
by
the
type
of
residence
(
87),
age
of
residence
(
88),
size
of
residence
(
89,
90)
and
ownership
status
(
91).
Other
potential
confounders
include
indoor
exposures
to
furry
pets
and
birds
(
92),
pests
(
93),
pesticides
(
94),
gas
stoves
(
95),
especially
those
with
pilot
lights
(
95a)
or
those
used
as
supplemental
heaters
(
95b),
and
unvented
heaters
(
96­
98).
Indoor
mold
exposures
are
of
special
interest
as
potential
confounders
and
are
characterized
by
water
damage
(
101),
visible
molds
(
102),
dehumidifiers
(
103),
or
humidifiers
(
104).
Air
conditioners
(
99),
air
cleaners
(
100),
and
open
windows
(
105)
may
act
as
effect
modifiers
by
altering
the
air
exchange
rate
and
by
cleaning
the
indoor
air.
A
final
potential
confounder
is
exposure
to
traffic­
related
noise
(
106)
and
outside
noise
disturbance
of
the
child's
nighttime
sleep
pattern
(
107).
The
family
history
section
(
108­
118)
is
focused
on
genetic
and
early
life
factors
as
potential
confounders.
A
genetic
predisposition
of
the
child
to
respiratory
illness
may
be
indicated
by
a
history
of
respiratory
illness
in
either
the
biological
father
(
108­
110)
or
the
biological
mother
(
111­
113).
Early
exposures
to
environmental
tobacco
smoke
are
assessed
by
the
child's
exposure
during
pregnancy
(
114),
during
the
first
two
years
of
life
(
115­
116),
and
after
two
years
of
age
and
prior
to
first
grade
(
117­
118).
Some
studies
have
found
that
maternal
smoking
(
114­
115,
117)
may
be
more
important
than
smoking
by
others
in
the
home
(
116,
118).
The
child's
family
section
(
119­
136)
is
largely
focused
on
current
family
characteristics
and
behaviors
as
potential
confounders.
Socioeconomic
status
as
a
potential
confounder
is
best
indicated
by
the
educational
levels
of
the
respondent
(
129)
and
the
other
primary
adult
(
131a);
and
also
may
be
indicated
by
family
size
and
household
relationships
(
119­
122),
access
to
educational
resources
(
123)
and
parental
attention
to
the
child
(
124).
We
have
omitted
sensitive
questions
regarding
family
income
which
have
been
used
by
other
investigators.
The
number
of
persons
aged
14
years
and
over
(
119)
may
be
related
to
environmental
tobacco
smoke,
while
the
number
of
children
under
14
(
120)
may
be
related
to
the
frequency
of
respiratory
infections.
The
respondent's
sex
(
126),
relationship
to
the
child
(
127),
and
language
proficiency
(
128)
may
be
related
to
the
accuracy
of
the
questionnaire.
The
child's
current
exposures
to
environmental
tobacco
smoke
in
the
home
are
determined
by
questions
on
the
smoking
status
(
130,132,133),
number
of
cigarettes
(
130b,
132b,
134),
and
smoking
of
pipe
or
cigars
(
130a
,
133)
for
the
respondent
(
130),
for
the
other
primary
adult
(
132),
and
for
any
other
smokers
in
the
home
(
133).
Finally,
awareness
of
a
local
Air
Quality
Index
(
135)
and
behavior
modification
appropriate
to
the
daily
index
(
136)
will
be
modifiers
for
health
outcomes.
Style
of
questions
While
most
questions
have
forced
choice
between
multiple
responses,
we
also
provide
opportunities
for
the
respondent
to
provide
a
free­
form
response
to
some
questions.
Free­
form
responses
are
an
easy
way
of
asking
about
continuous
variables
(
1,
2,
3,
4,
5,
10).
A
combination
of
multiple
choice
and
free­
form
responses
is
used
to
avoid
forcing
a
response
into
an
inappropriate
multiple
choice
(
7,
8,
16­
18,
65),
when
the
respondent
may
know
a
medication's
name
but
not
its
general
classification
(
47e,
48e,
49e,
50e),
to
17
obtain
more
information
on
a
positive
response
in
order
to
classify
an
illness
(
22a,
23a,
46c),
or
when
a
set
of
multiple
choices
may
not
be
exhaustive
(
24a,
75a,
87,
92­
93,
96).
The
possible
responses
to
most
questions
provide
the
respondent
with
an
opportunity
to
indicate
that
the
answer
is
unknown.
Only
the
questions
on
nutrition
(
51­
52),
school
performance
(
56­
59),
and
school
transportation
require
a
forced
choice
without
an
opportunity
to
indicate
a
lack
of
knowledge.
Purpose
of
Questions
Outcomes
of
interest
include
questions
on
common
respiratory
conditions
(
22­
26a)
and
chest
illnesses
including
asthma,
pneumonia,
bronchitis,
and
other
chest
illnesses
(
23a).
The
questionnaire
solicits
information
on
subclinical
asthma
evidenced
by
wheezing
(
28)
and
a
further
characterization
of
clinical
asthma
(
37).
The
quality
of
parental
reporting
may
depend
on
the
respondent's
sex
(
126),
relationship
to
the
child
(
127),
and
language
proficiency
(
128).
Potential
confounders
include
the
child's
age
(
1,
2,
3),
sex
(
6),
Hispanic
status
(
7)
and
race
(
8);
severe
preexisting
health
conditions
of
the
child
(
12­
21);
a
history
of
respiratory
illness
in
the
biological
parents
as
an
indicator
of
genetic
predisposition
(
108­
113);
and
socioeconomic
status
as
indicated
by
the
child's
residence
(
78­
91),
family
size
(
119­
120)
or
the
parent's
education
level
(
129,
131a).
We
also
examine,
as
potential
confounders,
various
indoor
exposures
(
92­
98),
especially
mold
exposures
(
102­
102a),
and
environmental
tobacco
smoke
exposures
(
130­
130b,
132­
134).
Potential
effect
modifiers
of
the
association
of
interest
include
the
child's
age
(
1,
2,
3),
sex
(
6),
and
residential
history
(
78­
83).
Low
birth
weight
(
9­
10),
prematurity
(
11),
neonatal
care
(
12­
15),
an
early
history
of
severe
chest
illness
(
22,
22a),
or
atopy
(
24,
24a)
may
also
increase
the
child's
susceptibility
to
environmental
insults.
Air
conditioners
(
99­
99b)
and
air
cleaners
(
100)
may
act
as
effect
modifiers
by
altering
air
exchange,
and
by
removing
gasses
and
particles
of
outdoor
origin
from
the
indoor
air.
3.
PRETESTS
AND
PILOT
TESTS
This
questionnaire
follows
the
standard
questions
for
respiratory
symptoms
in
children
established
in
1978
by
the
Epidemiology
Standardization
Project
of
the
American
Thoracic
Society
and
the
National
Institutes
of
Health's
Division
of
Lung
Diseases.
A
similarly
worded
questionnaire
was
successfully
used
in
the
Harvard
Six
Cities
Studies
and
in
the
Harvard
24­
Cities
Study
of
over
13,000
children
in
24
communities
in
the
United
States
and
Canada.
A
similar
questionnaire
was
used
in
the
El
Paso
Children's
Health
Study
(
OMB
Control
Number
2080­
0065).
These
earlier
successful
studies
constitute
the
pretest
of
this
questionnaire.
The
current
layout
of
the
questionnaire
has
been
improved
to
address
issues
regarding
questionnaire
completion
observed
in
the
El
Paso
study.
In
particular,
an
issue
regarding
the
pattern
of
skipping
questions
has
been
addressed
by
increasing
the
visibility
of
question
numbers
and
by
blocking
questions
into
subsections.
The
current
questionnaire
has
been
reviewed
as
part
of
the
consultation
process.
4.
COLLECTION
METHODS
AND
FOLLOW­
UP
4(
a)
COLLECTION
METHODS
Informational
meetings
were
held
in
2004
with
EPA
regional
scientist(
s),
representatives
from
the
Detroit
and
Dearborn
Public
School
Systems,
local
public
health
officers,
state
and
local
environmental
agencies,
and
other
persons
as
part
of
the
consultation
process.
A
more
formal
application
and
hearing
will
be
conducted
to
secure
approval
for
the
study
from
each
school
district.
After
this
approval,
a
single­
page
letter
will
be
requested
from
each
school
district
superintendent
providing
a
general
explanation
and
endorsement
of
the
study.
These
letters
will
be
reviewed
and
approved
by
the
Institutional
Review
Board
and
provided
to
the
parents
with
the
questionnaire.
18
Once
the
study
is
approved
by
the
school
districts,
the
school
administrations
will
be
asked
for
readily
available
information
on
the
number
of
fourth
and
fifth
grade
classes
in
the
school,
a
list
of
the
fourth
and
fifth
grade
teachers
responsible
for
each
class,
and
the
number
of
students
in
each
class.
These
items
are
necessary
to
ensure
that
sufficient
questionnaires
are
provided
to
each
school
and
to
track
the
participation
rates
for
each
class.
Individual
packets
will
be
prepared
for
each
of
the
fourth
and
fifth
grade
students
in
the
selected
elementary
schools.
Each
packet
will
consist
of
a
copy
of
the
appropriate
superintendent's
letter,
a
questionnaire,
and
a
sharpened
#
2
lead
pencil
(
with
a
toll­
free
EPA
telephone
number)
enclosed
within
a
plain
envelope
with
a
gummed
flap.
Each
questionnaire
will
be
have
a
unique
serial
number
(
lithography
code)
printed
on
each
page.
The
individual
packets
will
be
assembled
into
bundles
with
the
appropriate
number
for
each
class,
the
bundle
will
be
labeled
for
a
particular
class,
and
a
record
will
be
made
of
the
range
of
serial
numbers
assigned
to
each
class.
To
accommodate
Hispanic
or
Arabic
parents
who
may
have
difficulty
with
English,
packets
with
Spanishlanguage
or
Arabic­
language
translations
of
the
questionnaire
will
be
available
upon
request
at
every
elementary
school.
If
the
school
district
indicates
that
a
school
has
a
high
proportion
of
parents
who
are
Hispanic
or
Arabic,
questionnaires
in
both
or
all
languages
will
be
provided
in
all
packets
for
that
school.
English­
language
questionnaires
will
be
provided
to
all
parents.
The
questionnaire
distribution
process
will
begin
with
short
meetings
at
each
school
for
the
involved
teachers,
the
school
nurse,
the
school
principal,
and
any
other
school
personnel
with
an
interest
in
the
study.
These
meetings
will
explain
the
logistics
for
questionnaire
distribution
and
collection,
provide
sample
questionnaires,
and
review
the
study
schedule
for
each
class.
Each
teacher
will
receive
a
bundle
of
questionnaire
packets
for
his/
her
students.
The
questionnaire
packets
will
be
distributed
by
the
teacher
to
the
students
at
the
end
of
the
school
day.
The
students
will
carry
the
packets
home
for
their
parents
to
complete
the
questionnaire,
seal
the
completed
questionnaire
in
the
packet
envelope,
and
return
the
packet
over
the
next
few
days.
The
parents
will
keep
the
pencil
with
the
toll­
free
EPA
telephone
number.
As
the
questionnaires
are
returned
by
the
students,
the
teachers
will
collect
the
sealed
packets
and
hold
them
in
a
secure
place.
Within
one
week,
the
questionnaires
will
be
collected
by
an
EPA
contractor
from
the
teachers.
At
a
secure
facility,
the
EPA
contractor
will
open
the
questionnaire
packet
and
detach
the
consent
form
with
personal
identifying
information
from
the
questionnaire.
The
only
identifier
which
links
questionnaire
to
the
child's
name
and
other
personal
information
is
the
pre­
printed
serial
number
on
the
questionnaire
and
the
consent
form.
The
consent
forms
will
be
stored
in
a
locked
filing
cabinet
at
a
secure
facility,
separate
from
the
statistical
information
on
the
remainder
of
the
questionnaires.
Later
in
the
school
year
at
a
subset
of
elementary
schools,
those
children
with
written
permission
from
their
parent
or
guardian
will
be
scheduled
for
measurements
of
lung
function
and
exhaled
breath.
The
lung
function
examinations
will
be
conducted
in
the
child's
elementary
school
during
normal
school
hours,
no
child
will
be
left
alone
at
any
time
in
the
custody
of
a
single
field
technician,
and
all
personnel
will
adhere
to
school
policies
and
procedures.
A
school­
nurse
will
be
present
at
the
elementary
school
during
the
examinations.
Prior
to
examination,
a
trained
field
technician
will
individually
explain
the
testing
process
to
each
child,
ask
for
his/
her
assent
to
participate
in
the
exam,
and
return
unwilling
children
to
their
classroom.
The
field
technician
will
then
ask
each
child
whether
they
have
ever
smoked
more
than
five
cigarettes
in
their
lifetime
and
whether
they
have
had
a
recent
respiratory
infection.
Any
child
who
answers
positive
to
either
question
will
be
returned
to
their
classroom.
In
any
of
these
cases,
the
reason
for
the
child's
non­
19
participation
will
be
held
confidential.
All
children
who
report
to
the
examination
area
will
receive
a
small,
non­
monetary
incentive
regardless
of
their
participation.
In
the
exam,
the
field
technician
will
record
each
child's
height
and
weight,
explain
in
more
detail
the
forced
expiratory
and
exhaled
breath
maneuvers,
and
coach
the
child
to
perform
a
series
of
forced
expiratory
maneuvers
and
exhaled
breath
procedures.
The
lung
function
examination
and
exhaled
nitric
oxide
procedure
will
be
conducted
according
to
standard
guidelines
developed
by
the
American
Thoracic
Society.
A
sterile,
disposable
mouthpiece
will
be
used
for
each
child.
4(
b)
SURVEY
RESPONSE
AND
FOLLOW­
UP
In
the
El
Paso
Children's
Health
Study
(
ICR
1940.02)
9,797
questionnaires
were
distributed,
1,564
questionnaires
were
not
returned,
and
461
packets
were
returned
with
a
refusal
to
participate.
Thus,
the
initial
response
rate
was
79
percent
and
most
classrooms
achieved
an
85%
response
rate.
A
further
674
questionnaires
had
substantial
missing
responses
and
1,171
were
missing
selected
covariates.
In
El
Paso,
93%
of
the
children
were
able
to
perform
reproducible
spirometry.
For
this
study,
the
estimates
of
statistical
power
anticipate
a
similar
attrition
from
the
initial
sampling
frames
for
the
questionnaire
(
15,000
parents)
and
the
lung
function
examination
(
3,500
children).
5.
ANALYZING
AND
REPORTING
SURVEY
RESULTS
5(
a)
DATA
PREPARATION
The
completed
questionnaires
will
be
optically
scanned
into
an
electronic
database.
A
sample
of
questionnaires
will
be
visually
compared
with
the
scanned
data
to
ensure
that
the
scanning
software
assigns
the
correct
values
to
each
variable.
The
resulting
statistical
data
file
will
be
held
as
the
primary
or
canonical
data
set.
The
statistical
data
will
then
undergo
an
extensive
quality
assurance
review
by
EPA
investigators
to
identify
erroneous
and
inconsistent
responses.
This
review
will
involve
both
the
cross­
tabulation
of
results
and
a
thorough
review
of
individual
questionnaires.
As
each
issue
is
identified
and
resolved,
a
single,
evolving
computer
program
will
be
prepared
to
generate
and
document
all
revisions
of
the
primary
data
set
in
the
preparation
of
an
analytic
data
set,
as
well
as
the
generation
of
all
secondary
variables.
The
quality
assurance
documentation
will
include
the
edited
values,
the
date,
and
the
initials
of
the
investigator
in
the
body
of
the
computer
program.
5(
b)
ANALYSIS
This
multilevel
analytic
study
(
also
called
semi­
ecological
or
semi­
individual)
will
use
both
group­
level
or
ecological
level
information
on
environmental
exposures
and
individual­
level
information
on
health
status
and
potential
confounders.
This
design
has
particular
utility
whenever
a
common
environmental
exposure
is
shared
by
individuals
clustered
within
a
geographic
area.
Environmental
exposures
will
be
assessed
on
the
basis
of
life­
long
(
up
to
five
residences
and
school
location)
proximity
to
major
sources
of
air
pollutants
through
a
multi­
stage
approach
that
was
successfully
used
in
the
El
Paso
Children's
Health
Study.
 
The
first
stage
will
involve
the
acquisition
of
geographic
information
on
stationary,
mobile,
and
area
sources
of
air
pollutant
emissions
in
Wayne
County
and
the
integration
of
this
information
into
a
geographic
information
system
(
GIS).
 
The
second
stage
will
be
a
campaign
of
air
quality
monitoring
using
standard
passive
sampling
devices
to
measure
ambient
concentrations
of
selected
air
pollutants,
including
nitrogen
dioxide,
during
periods
of
summertime
air
stagnation
at
25
elementary
schools.
20
 
The
third
stage
will
be
the
development
of
a
predictive
model
for
long­
term
average
air
pollution
concentrations
from
the
air
quality
measurements
and
the
GIS
information
system.
 
The
fourth
stage
will
involve
the
questionnaire­
based
collection
of
a
detailed
residential
history
at
four
stages
of
the
child's
life
and
the
geocoding
of
these
addresses.
 
The
fifth
stage
will
be
the
application
of
the
predictive
air
quality
model
to
the
geocoded
address
history
of
each
study
participant
and
the
estimation
of
long­
term
exposures.
The
epidemiologic
model
for
respiratory
symptoms
and
other
binary
indicators
will
adjust
for
random
school
effects;
fixed
design
effects
for
city
(
Detroit
or
Dearborn),
enumeration
district
within
Detroit
(
five
districts),
and
questionnaire
language;
and
fixed
covariate
effects
for
sex,
age,
race
(
four
categories),
Hispanic
ethnicity,
parental
education
(
four
categories),
parental
allergies,
gas
cooking
stove
with
pilot
light,
and
environmental
tobacco
smoke
in
the
home.
The
potential
confounding
effect
of
these
covariates
will
be
assessed
in
comparison
with
a
simplified
model
that
included
only
the
design
factors.
The
lung
function
levels
will
be
modeled
as
the
natural
logarithm
of
lung
function
adjusted
for
random
spirometer/
technician
effects;
fixed
design
effects
for
city
(
Detroit
or
Dearborn)
and
enumeration
district
within
Detroit
(
five
districts);
and
fixed
covariate
effects
for
sex,
age,
race
(
four
categories),
Hispanic
ethnicity,
parental
education
(
four
categories),
environmental
tobacco
smoke
in
the
home,
the
natural
logarithms
of
height
and
weight,
an
interaction
between
sex
and
height,
and
interactions
between
Hispanic
ethnicity
and
height
and
weight.
For
ease
of
interpretation,
the
regression
coefficients
will
be
converted
to
the
percent
change
in
the
level
of
lung
function.
Exhaled
nitric
oxide
levels
will
be
modeled
adjusting
for
random
analyzer/
technician
effects;
fixed
design
effects
for
city
(
Detroit
or
Dearborn)
and
enumeration
district
within
Detroit
(
five
districts);
and
fixed
covariate
effects
for
sex,
age,
race
(
four
categories),
Hispanic
ethnicity,
parental
education
(
four
categories),
and
environmental
tobacco
smoke
in
the
home.
An
alternative
model
using
the
natural
logarithm
of
exhaled
nitric
oxide
will
be
considered.
5(
c)
REPORTING
RESULTS
Publications
related
to
exposure
assessment
in
the
DCHS
will
include
field
method
evaluations
of
passive
samplers
for
air
toxics
and
development
of
a
land
use
regression
model
to
estimate
interurban
air
pollution
for
the
Detroit/
Dearborn
airshed.
Additional
presentations
related
to
geographical
analysis
for
mapping
urban
air
pollution
exposure
and
assessing
spatial
differences
of
source
impacts
(
in
conjunction
with
source
apportionment
models)
may
be
considered.
It
is
anticipated
that
the
application
and
development
of
these
exposure
assessment
techniques
within
a
geographical
analysis
framework
will
help
facilitate
improved
exposure
estimations
for
epidemiological
studies
linking
respiratory
health
effects
to
urban
air
pollution
sources.
The
proposed
study
will
be
the
first
large
field
data
collection
of
both
exhaled
nitric
oxide
and
lung
function
measurements
among
schoolchildren.
We
anticipate
one
or
more
papers
related
to
an
intercomparison
of
exhaled
nitric
oxide
with
questionnaire
indicators
of
allergies
and
asthma,
exhaled
nitric
oxide
as
an
indicator
of
lung
inflammation,
and
standard
spirometric
measures
of
pulmonary
obstruction.
The
main
health
paper
will
be
an
epidemiologic
assessment
of
whether
higher
long­
term
exposures
to
air
pollutants,
as
determined
by
residential
history,
are
associated
with
the
higher
prevalence
of
allergies
and
asthma
and
lower
levels
of
lung
function
among
schoolchildren.
In
addition
to
the
main
paper,
we
envision
a
series
of
additional
papers
examining
questionnaire­
based
indicators
of
environmental
21
exposures,
including
environmental
tobacco
smoke,
gas
cooking
stoves
with
pilot
lights,
and
unvented
gas
heaters.
22
APPENDIX
A:
Statistical
Power
and
Precision
The
statistical
power
and
precision
of
this
study
has
been
evaluated
using
the
observed
results
from
EPA's
El
Paso
Children's
Health
Study
as
a
reference.
This
approach
is
most
suitable
for
epidemiologic
studies
where
the
crude
associations
are
adjusted
for
multiple
covariates.
On
the
basis
of
this
procedure,
the
Detroit
Children's
Health
Study
will
have
an
86
percent
probability
(
statistical
power)
to
detect
an
association
with
current
asthma
as
strong
as
those
found
in
the
El
Paso
Children's
Health
Study
(
odds
ratio
(
OR)
=
1.65
for
a
10
ppb
increment
in
GIS­
modeled
nitrogen
dioxide).
The
Detroit
Children's
Health
Study
will
have
an
98
percent
probability
(
statistical
power)
to
detect
an
association
with
forced
vital
capacity
less
than
85
percent
of
predicted
at
least
as
strong
as
those
found
in
the
El
Paso
Children's
Health
Study
(
OR
=
3.10
for
a
10
ppb
increment
in
GIS­
modeled
nitrogen
dioxide).
Expected
exposure
gradient
(
exposure
variance)
The
local
point,
area,
and
line
sources
of
air
pollutants
in
Detroit
and
Dearborn,
Michigan,
are
more
numerous
and
extensive
than
in
El
Paso,
Texas.
El
Paso
County
and
surrounding
Texas
Counties
had
fewer
than
five
significant
industrial
point
sources,
while
Wayne
County
has
31
distinct
point
sources.
The
annual
average
daily
traffic
(
AADT)
volumes
on
road
segments
in
Wayne
County
are
greater
than
in
El
Paso
County
and
the
road
grid
is
more
interconnected.
Our
indicators
of
area
sources,
the
ranges
of
population
and
housing
density
in
Wayne
County,
are
at
least
as
great
as
those
in
El
Paso
County.
As
shown
in
several
studies,
the
contributions
of
air
pollutants
from
local
sources
tends
to
show
a
sharp
decay
in
ambient
concentrations
as
one
moves
away
from
the
local
source.
For
major
roadways,
some
pollutants
decay
to
background
levels
within
150
meters
of
the
roadway.
Thus,
we
expect
that
the
exposure
gradients
in
the
Detroit
Children's
Health
Study
will
be
similar
to
those
found
in
the
El
Paso
Children's
Health
Study.
Expected
outcome
prevalences
and
measurements
(
outcome
variance)
Symptoms
and
conditions
In
the
El
Paso
Children's
Health
Study,
the
school­
specific
prevalence
of
respiratory
symptoms
and
conditions
varied
considerably
across
the
community,
even
after
adjustment
for
a
variety
of
factors
including
ethnicity,
questionnaire
language,
domestic
exposures,
and
socioeconomic
status.
For
the
lifetime
history
of
illness,
allergy
prevalence
varied
from
9
to
33
percent
and
asthma
varied
from
3
to
15
percent.
For
respiratory
symptoms
and
conditions
in
the
prior
year,
asthma
prevalence
varied
from
1
to
9
percent
and
bronchitis
prevalence
varied
from
less
than
1
percent
to
nearly
8
percent.
Since
children
of
Hispanic
ancestry,
especially
those
children
whose
parents
used
the
Spanish
language
questionnaire,
had
a
much
lower
prevalence
of
allergies
and
asthma
than
the
non­
Hispanic
children
in
El
Paso
and
in
other
studies,
we
believe
that
the
expected
prevalences
in
these
symptoms
and
conditions
will
be
slightly
higher
in
the
Detroit
Children's
Health
Study
because
of
Detroit's
proportionally
smaller
Hispanic
population.
Spirometry
In
the
El
Paso
Children's
Health
Study,
the
overall
prevalence
of
a
forced
vital
capacity
less
than
85
percent
of
predicted
was
7.5
percent.
Higher
prevalences
were
observed
for
forced
expiratory
volume
at
one
second
less
than
85
percent
of
predicted
(
8.8
percent)
and
peak
expiratory
flow
less
than
80
percent
of
predicted
(
11.1
percent).
We
believe
that
the
expected
prevalences
of
low
lung
function
levels
below
these
cut­
points
will
be
similar
in
the
Detroit
Children's
Health
Study.
Sample
size
based
on
budgetary
and
logistic
constraints
Questionnaires
23
In
the
El
Paso
Children's
Health
Study,
we
distributed
9,797
questionnaires
in
54
elementary
schools
and
received
7,772
completed
questionnaires,
461
refusals,
and
1,564
questionnaires
not
returned
(
16
percent).
After
a
stringent
quality
assurance
review
of
the
completed
questionnaires,
we
excluded
674
children
missing
multiple
covariates
and
274
children
with
severe
conditions
or
chest
injuries.
We
were
successful
in
geocoding
7,529
residences
(
97
percent).
For
the
analyses
of
respiratory
symptoms
and
conditions,
we
excluded
a
further
457
children
with
missing
confounder
data
and
714
children
with
missing
symptom
data.
Of
the
remaining
5,654
children,
4,231
children
had
resided
in
their
current
residence
for
at
least
one
year
prior
to
the
questionnaire.
Thus,
the
original
number
of
questionnaires
dropped
from
9,797
to
4,231
(
43
percent).
We
think
that
similar
levels
of
participation
and
exclusion
may
be
expected
in
the
Detroit
Children's
Health
Study.
We
are
preparing
to
distribute
up
to
15,000
questionnaires
and
expect
to
be
able
to
use
43
percent
in
our
analyses
or
about
6,500
children.
As
a
conservative
estimate,
we
expect
that
this
50
percent
increase
in
sample
size
will
lead
to
a
22
percent
increase
( 
1.5
=
1.22)
in
the
precision
of
this
study,
which
will
produce
a
decrease
in
the
expected
confidence
interval
and
an
improvement
in
statistical
power.
Lung
function
examinations
In
the
El
Paso
Children's
Health
Study,
we
attempted
to
perform
lung
function
examinations
of
2,687
children
with
written
parental
consent
and
the
child's
verbal
assent.
Almost
all
of
these
children
(
93
percent)
were
successfully
able
to
perform
reproducible
spirometry.
Similar
exclusions
to
those
for
the
questionnaire
produced
a
set
of
1,871
children
who
resided
in
El
Paso
for
at
least
one
year
prior
to
the
questionnaire
(
70
percent).
We
think
that
similar
levels
of
participation,
ability
to
perform
spirometry,
and
exclusion
may
be
expected
in
the
Detroit
Children's
Health
Study.
We
are
preparing
to
attempt
up
to
3,500
examinations
and
expect
to
be
able
to
use
70
percent
in
our
analyses
or
about
2,500
children.
As
a
conservative
estimate,
we
expect
that
this
30
percent
increase
in
sample
size
will
lead
to
a
14
percent
increase
( 
1.3
=
1.14)
in
the
precision
of
this
study,
which
will
produce
a
decrease
in
the
expected
confidence
interval
and
an
improvement
in
statistical
power.
Expected
statistical
power
Symptoms
and
conditions
In
the
El
Paso
Children's
Health
Study,
a
10
ppb
increment
in
GIS­
modeled
nitrogen
dioxide
was
associated
with
current
asthma
(
OR
=
1.65,
95%
confidence
interval
(
CI):
1.08,
2.50;
t­
statistic
2.50)
after
adjustment
for
demographic,
socioeconomic,
parental,
and
random
school
effects.
Assuming
a
similar
range
of
exposures,
a
slightly
higher
prevalence
of
asthma,
and
a
50
percent
increase
in
sample
size,
we
expect
that
the
Detroit
Children's
Health
Study
will
have
86
percent
probability
(
statistical
power)
to
detect
an
association
at
least
as
strong
as
that
found
in
El
Paso.
Spirometry
In
the
El
Paso
Children's
Health
Study,
a
10
ppb
increment
in
GIS­
modeled
nitrogen
dioxide
was
associated
with
a
2.4
percent
decrement
in
forced
vital
capacity
(
95%
CI:
­
4.0%,
­
0.7%;
t­
statistic
­
2.82)
after
adjustment
for
demographic,
anthropomorphic,
socioeconomic,
and
spirometer/
technician
effects.
Assuming
a
similar
range
of
exposures,
a
similar
variability
in
lung
function,
and
a
50
percent
increase
in
sample
size,
we
expect
that
the
Detroit
Children's
Health
Study
will
have
89
percent
probability
(
statistical
power)
to
detect
an
association
at
least
as
strong
as
that
found
in
El
Paso.
For
indicators
of
lung
function
levels
less
than
85
percent
of
predicted,
a
10
ppb
increment
in
GISmodeled
nitrogen
dioxide
was
associated
with
low
forced
vital
capacity
(
OR
=
3.10,
95%
CI:
1.65,
5.82;
t­
statistic
3.51).
Assuming
a
similar
range
of
exposures,
a
similar
variability
in
lung
function,
and
a
50
24
percent
increase
in
sample
size,
we
expect
that
the
Detroit
Children's
Health
Study
will
have
98
percent
probability
(
statistical
power)
to
detect
an
association
at
least
as
strong
as
that
found
in
El
Paso.
