<?xml version="1.0" encoding="UTF-8"?>
<itemContainer xmlns="http://omeka.org/schemas/omeka-xml/v5" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://omeka.org/schemas/omeka-xml/v5 http://omeka.org/schemas/omeka-xml/v5/omeka-xml-5-0.xsd" uri="https://www.nal.usda.gov/exhibits/speccoll/items/browse?advanced%5B0%5D%5Belement_id%5D=39&amp;advanced%5B0%5D%5Btype%5D=is+exactly&amp;advanced%5B0%5D%5Bterms%5D=Spivey%2C+Gary+H.&amp;output=omeka-xml" accessDate="2026-11-07T20:47:15+00:00">
  <miscellaneousContainer>
    <pagination>
      <pageNumber>1</pageNumber>
      <perPage>15</perPage>
      <totalResults>2</totalResults>
    </pagination>
  </miscellaneousContainer>
  <item itemId="2713" public="1" featured="0">
    <fileContainer>
      <file fileId="1497">
        <src>https://www.nal.usda.gov/exhibits/speccoll/files/original/106ca43931f33c9eaa68ba14b5e16658.pdf</src>
        <authentication>4c2bdb89daf0bd16b250e5a5c53f995a</authentication>
        <elementSetContainer>
          <elementSet elementSetId="4">
            <name>PDF Text</name>
            <description/>
            <elementContainer>
              <element elementId="60">
                <name>Text</name>
                <description/>
                <elementTextContainer>
                  <elementText elementTextId="63446">
                    <text>Item ID Number

°1686

Author

Spivey, Gary H.

Corporate Author
Report/Article Title Protocol for Epidemiologic Studies of Agent Orange

Journal/Book Title
Year

1982

Month/Day

January 22

n
Number of Imaoes

138

Descplpton Notes

Wednesday, June 06, 2001

Page 1687 of 1688

�PROTOCOL

for
Epidemiologic Studies of Agent Orange

Submitted January 22, 1982

Gary H. Spivey, MD, MPH
Principal Investigator
Roger Detels, MD, MS
Co-Principal Investigator
Division of Epidemiology
School of Public Health
University of California
Los Angeles, California

Veterans Administration contract:
Contract amount:
Type of award:
Official requesting consultation

Contractor:

V101 (93) P-842
$133,951
Competitive
Barclay Shepard, MD
Department of Medicine
and Surgery
Veterans Administration
(202) 389-5411

Regents, University of California
405 Hilgard Avenue
Los Angeles, California 90024

�LIST OF INVESTIGATORS AND CONSULTANTS

I nvesti gate r5
Arthur K, Cho, PhD

Professpr of Pharmacology
School of Medicine
University of California, Los Angeles
Virginia Clark, PhD
Professor of Biostatistics and Biomathem,atics
Division of Biostatistics
School of Public Health
University of California, Los Angeles
Anne Hersey Coulson
Research Epidemiologist
Division of Epidemiology
School of Public Health
University of California, Los Angeles
William G. Cumberland, PhD
Assistant Professor
Division of Biostatistics
School of Public Health
University of California, Los Angeles
Roger Detels, MD, MS
Dean
School of Public Health
University of California, Los Angeles
Wendell W. Kilgore, PhD
Professor and Former Chairman of Environmental Toxicology
University of California, Davis
Jess F. Kraus, MPH, PhD

Professor
Division of Epidemiology
School of Public Health
University of California, Los Angeles

�Investigators (con*d)
Gary H. Spivey, MD, MPH
Associate Professor
Division of Epidemiology
School of Public Health
University of California, Los Angeles
Julian L. Van Lancker, MD
Chairman of Pathology
School of Medicine
University of California, Los Angeles

Consultants
Irvin Emanuel, MD
Director of Child Development and Mental Retardation Center
Professor Epidemiology and Pediatrics
University of Washington
Seattle, Washington
Dennis Cope, MD
Associate Professor of Medicine
School of Medicine
University of California, Los Angeles
Carol Aneshensel, PhD
Senior Statistician
Division of Epidemiology
School of Public Health
University of California, Los Angeles
Eve P. Fielder
Director, Survey Research Center
Institute for Social Science Research
University of California, Los Angeles
Francis J. Pirozzolo, PhD
Chief of Neuropsychologic Service
Assistant Professor
Department of Neurology
Baylor College of Medicine, Texas Medical Center, Houston
Associate Professor
Department of Psychology
University of Houston, Houston

�TABLE OF CONTENTS

I. Introduction *•
II. Background -•*•&lt;—&lt;--—

--—

....-..._._— 1
*-_—..._...

——.............»..
—....,.*..........

A. Environmental Toxicology ~

.
—

B. Animal Studies —-—-•—

-*—-

.——

5

—--—*—

C. Human Health Effects

*

6

:»--

D. Popular Press

5

——

7

*-

9

E. Studies of Agent Orange by'Other Investigators

11

III. Proposed Protocol
A. Introduction
B. Methods
1.
2.
3.
4.
5.

6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.

—

-—

*

;-18

-—

Summary of study designRationale for historical cohort study-—
Development of Exposure Likelihood Index
Establishing cohorts
Examination procedures
—
a. introduction —-•—-—
•
•
h. suggested questionnaire
data collection form
-c. physical examination
data collection form
•
d. laboratory tests
e. psychologic tests
psychological assessment
f. neuropsychologic tests
neuropsychological assessment
Possible confounding factors
Tracing of men selected for study
Subject contact procedures
Quality control
Data management
Data analysis
Sample size
Organization
Notification of subjects
Pilot testing of protocol
Timetable

17. Future follow-up of subjects

20
*

20
20
23
-37
42
•
42
.....4
.....5
•
lb-203b
49
lc-3c
51
56
le-9e
56
If-lOf
61
•
63
72
77
89
99
107
113
rl!8
—
120
124
125

�Appendix A - Glossary
Appendix B - Epidemiologic Methods
Appendix C - Review of Environmental Behavior of Agent Orange
Appendix D - Review of Literature of Effects of Agent Orange in Animals
Appendix E - Review of Literature on Reproductive Effects of Agent Orange
Appendix F - Review of Literature on Human Health Effects from Agent Orange
Appendix G - Review of Popular Press Articles on Agent Orange
Appendix H - Proposed Agent Orange Troop Exposure and Non-Exposure
Cohort Selection Concept Paper
Appendix I - Potential Confounders
Appendix J - Interviewer Training Manual, UCLA, SRC
Appendix K - Introduction to RAMIS II Users Manual

�I.

Introduction
M a n y technical military

and

scientific

abbreviations are used throughout this text.
glossary

and

Please see the

(Appendix A) for definitions, if necessary.
The major portion of the Vietnam

troops was between the years 1965-1972.

war

for

From 1965 the number of U.S.

American

American troops had

been in Vietnam prior to 1965» although their
small.

terras

numbers

were

troops grew rapidly to

a peak in 1968. During that year there

were

estimated

to

have been over a million men present in Vietnam.
the
officially

Vietnam
dates

era, by

from

Presidential

August 5, 1961, to May 7» 1975. In

this time period an estimated 9 million men
U.S.

Armed

Forces.

million served
combat;

in

309.000

Proclamation,

Of

these,

Vietnam;

1.6

served

in

the

approximately 2.2 to 2.8
million

were wounded in battle;

saw

action

in

47,000 died as a

result of acts of war.
Defoliants were used in South Vietnam from
1962,

to

cover.

February,

1971, mainly

January,

to deprive the enemy of

they were also used, to a much

lesser

degree,

for

crop destruction in areas of heavy Viet Cong activity and to
maintain

a

Herbicides
the

clear
Green,

estimated

perimeter
Pink,
368

Tetrachlorodibenzo-p-dioxin

around

some

base

camps.

and Purple contained about 40£ of
pounds
(TCDD)

of
dropped

2,3,7,8
in

Vietnam.

However, the use of these herbicides is not considered to be

�a

relevant

concern

because

they

were sprayed in limited

q u a n t i t i e s on less than 90,000 acres primarily from 1962

to

1964,

In

a tine when few U.S.

servicemen were in Vietnam.

contrast, 90/t of Agent Orange was used on over
acres

2.9

during a time when large numbers of U.S.

in Vietnam,.early 1965 to April 1970.
approximate

50:50

mixture

dichlorophenoxy

acetic

trichlorophenoxy

acetic

of

the

acid
acid

Agent

million

troops were

Orange

is

herbicides

(2,4-D)

and

(2,4,5-T).

an
2,4,

2,4,5

This

latter

substance was contaminated with varying amounts of TCDD.
brief

summary

Appendix A.
summaries

of

the use of the defoliants is included in

The interested reader is referred to

excellent

presented in the 1974 National Academy of Science

report (1), the Air Force document prepared by
Young (2),

Major

Alvin

and the new VA sponsored literature review

The work represented in this protocol was
under

A

contract

to

the

Veterans

conducted

Administration.

charged with developing the protocol

for

an

(3).

We were

epidemiologic

study of ground troops exposed to Agent Orange while serving
in

South

Vietnam.

extensive discussions

To

complete

this

materials;

animal effects;

documentaries,

held

reviewed

investigated

Army

reviewed literature on Agent Orange,

its constituents, its environmental behavior and
and

we:

with Army records personnel;

selected Army and Marine Corps records;
historical

charge

its

human

reviewed the popular press, television

congressional

testimony

and

other

�non-scientific

sources

of information; consulted with the
)
Air Force and Australian investigators about their studies;
and followed-up many leads on other possible sources of data
and information relevant to this study.
The
studies

initial

rather

draft,

than

which

emphasized

preliminary

details of the recommended historical

cohort study, has been reviewed by three review groups.
the

basis

of

detailed

revised the protocol.

On

comments from these groups we have

The original

draft

section

III

on

epidemiologic studies has been moved to Appendix B.
This protocol draft is organized as follows:
In section II (Background) we provide a brief review
of

information

relevant to the design of this epidemiologic

study including:
of

Agent

the environmental fate of the constituents

Orange,

the

animal

literature, a review of
summary

of

several

the

and

popular

human

health

press,

and

a

brief

current approaches to studies of Agent

Orange, both epidemiologic and

non-epidemiologic.

of

are

the

effects

background

section

presented

in

Details
separate

appendices.
In section III we present a
the

historical

necessary
conduct

study

which

provides

for an epidemiologic study team
a

incorporates
the

cohort

detailed

first

pilot

study

of

the

protocol.

many of the suggestions made
draft

protocol.

to

by

protocol
the

for

detail

organize

and

This protocol
reviewers

of

We believe a pilot study to be

�mandatory to adequately test the feasibility
the

full

sjbudy,

of

conductinc

to fully develop all study procedures and

data collection forms and to estimate study costs.
We originally proposed a series of studies utilizing
death
this

certificates.
draft.

We

proportionate

We have not included these studies in

recommend
mortality

that

the

analysis,

VA
and

complete
a

a

frequency

distribution of all complaints in the Agent Orange Registry,
as

rapidly

incorporated

as

possible

into

the

so
final

instruments for this study.

that

the

design

information can be
of

data

collection

�II.

Background
A.

Environmental Toxicology

A detailed l i t e r a t u r e review and discussions of this
topic is included as A p p e n d i x C and summarized here.
Herbicides were used in South Vietnam
and

1971

with

the

majority

of

use,afte/ 1965.

different compounds were used including 2,4
acetic

acid

(2,4-D),

2,4,5

name

given

and 2,4,5-T.
2,3.7,8
0.02

to an approximately

The 2,4,5-T component

1962

Several

Dichlorophenoxy

Trichlorophenoxy

(2,4,5-T), picloram and cacodylic acid.
the

'between

acetic acid

Agent

Orange

was

50:50 mixture of 2,4-D
was

Tetrachlorodibenzo-p-dioxin

contaminated

(TCDD),

with

at a level of

to 47 (average about 2) parts per million.
The

2,4,5-T)

two

have

herbicides

Agent

Orange

(2,4-D

and

relatively short half-lives in soil (up to 3

weeks) with little soil
apparently

in

even

penetration.

faster

in

tropical

The

degradation

soils.

is

TCDD may be

persistent in soil but the persistence seems to

be

limited

to areas where massive application has occurred.
TCDD when present in soil appears to be stable
exceedingly

low

levels

present

present in water

systems,

sediments.

2,4-D

The

contaminate runoff water.
compounds
Vietnam,

is
TCDD

and

TCDD

in

water

appears

2,4,5-T

runoff.

to

are

photodegradation

be

more

Photodegradation

well documented.

with

of

When

bound

to

likely

to

all

three

Under conditions present in
is

measured

in

hours.

�Degradation

of 2.H-D, 2,4,5-T and, to a lesser degree, TCDD

is enhanced by cicrobial action.

TCDD

appears

not

to

be

The environmental fate of the three constituents

of

volatilized.

Agent

Orange

are

not fully understood.

degradation appears most likely,
devise

exposure

assumptions.
degradation
another.

indices

for

it
this

may

be

one

index

and

necessary

to

assume

rapid

long tern persistence for

The use of several assumptions would increase

likelihood

of

to

study based on several

At a minimum these would be
for

Even though rapid

the

detecting an Agent Orange effect if there is

one.
Piclorazn and cacodylic acid
sufficiently
be possible

do

not

appear

toxic to be of major concern.
to

incorporate

some

to

be

However, it may

investigation

of

these

agents into the study as well.
B.

Animal Studies

The components of Agent
extensively

in

a

wide

variety

laboratory and in the field.

Orange
of

have

been

studied

animal species in the

A review of selected

articles

is presented in Appendix D and is summarized here.
The dioxin contaminant (2,3,7,8-TCDD)
considerable

attention

least

one

Additionally

received

in the literature and is thought to

be the most toxic synthetic chemical known.
at

has

It is toxic

in

organ system in every animal species tested.

every organ system

has

been

affected

in

at

�least

one

animal

anirials are:
liver

species.

high fatality

toxicity,

Among
(low

the

LD50)

major
in

effects in

cany

animals,

effects on the lymphatic system and iccune

response and teratogen^Lcity.

Carcinogenic effects have been

reported

Thus far, no evidence has been

in

some

species.

reported for transraittal of genetic effects from an
male

to

its

offspring

extensively studied.
used

nassive

doses,

although

this

area

exposed

has not been

The teratogenic studies have typically
unlikely

to

.be encountered in hunan

exposures, and the teratogenic effect has not been found
sone

spec'ies.

in

(See Appendix E for a summary/discussion of

the literature on reproductive effects.)
While toxic
different

species,

effects
the

have

severity

been

reported

in

many

of the effects, the dose

required to produce the effect, the LD 50 and the metabolism
and

storage of the dioxin vary widely.

Thus, extrapolation

to humans should be done with extreme caution.
The other two components of Agent Orange
found

to

effects

be
have

abnormalities
conclusive and

only

moderately

included
and
the

have

toxic in animals.

teratogenesis,

been

Reported

nervous

system

carcinogenesis but the literature is not
required

doses

were

generally

quite

large.
C.

Human Health Effects

Information on the human effects of the constituents
of

Agent

Orange

(2,4-D;

2,4,5-T

and

its

contaminant

�2,3»7,8-TCDC)

has

come

exposures,

occupational

population

exposures.

exposures

in

fron

small

studies

accidents,
Many

populations,
studies.

review

is

and

discussion

occupational

poisonings and general

involve

conducted epidemiologic

of

fairly

and
A

few

high
are

selective

presented

in

dose

properly

literature

Appendix F

and

summarized here.
i

The least hazardous
symptoms

Common

of

acute

following

include

headache,

Central

nervous

weakness

the

substances
exposure

and

to

is

2,^-D.

high doses

gastrointestinal

upset.

system dysfunction, peripheral neuropathy,

nephropathy

and

asthenia

Apparently,

2,i*-D

is

not

have

also

been

reported.

contaminated with 2,3,7,8-TCDD.

The potential exposures to 2,^-D among the ground troops

in

Vietnam were not likely to be high enough to be of concern.
Since 2,4,5-T
these

substances

is

have

contaminated

with

2,3,7i8-TCDD,

been largely studied together.

The

2,3,7,8-TCDD contaminant appears to be the more toxic of the
two

(2,4,5-T

appears

to

have

only moderate toxicity) in

animals, but the toxicity in man is uncertain.
the

only

exposure.
however,

Chloracne is

established health outcome associated with dioxin
It is not an adequate marker of dioxin
since

it

is

exposure,

difficult to diagnose, occurs after

exposures to other chemicals and does not always appear even
after

heavy

dioxin

exposure.

A

number

of other health

effects have been reported affecting every organ

system

of

�the body.

These inqlucje porphyria cutanea tarda, other skin

disorders, liver damage, disorders of lipid and carbohydrate
metabolise,

peripheral

dysfynction,
disorders.

neuropathy, • central nervous system

teratogenesis,

cancer

and

psychiatric

The evidence is not clear for any of these.

The literature on reproductive effects (see Appendix
E)

does

not

support

exposure were having
expected

to

most

a

effect

at

teratogenic

this

effect,

likely

An effect through the'

to

be

a

mutational

study

it

If the

could

difficult.

male,

effect.

mutations occur spontaneously and increase
making

time.

be

produce a common syndrome as is characteristic

of other teratogens.
is

any

with

however,
Dominant
age,

thus

Recessive mutations would not be

demonstrable in the offspring of the exposed male, and would
only

be

demonstrable

in

later

generations with enormous

numbers of study subjects.
While the scientific literature does not
unequivocal

focus

for

a

study

be

examined.

a

of effects of exposure to

Agent Orange, it does suggest many types
should

provide

of

effects

which

These suggested effects are addressed

by the protocol.
D.

Popular Press

Because this study has been mandated
highly

to

address

emotional

as

well as to

and

political

situation

answer an epideraiologic question, nonscientific as
scientific issues must be considered.

well

a

as

The popular press has

�10

been selectively reviewed to gain an

understanding

of

the

nonscientific issues important in designing a study which is
a c c e p t a b l e to veterans and the public and is
sound.

A

detailed review is presented in A p p e n d i x G and a

summary is presented here.

.

Much information on
Vietnam

has

been

the

.

use

of

Agent

public

on

this

subject.

of

are

engineers.

who

sprayers;

service
About

door

helicopter
50

and

gunners

units;

symptoms/diseases

may

Operation Ranch

Ranch Hand ground support personnel;

backpack

helicopters;

details.

veterans

those

have had direct contact with the herbicide:

handlers;

in

Veterans have been informed

that those at highest risk of exposure

Hand personnel;

Orange

presented, but with conflicting

Confusion is likely to exist in the minds
the

scientifically

drum

for

spray

and
which

combat
veterans

attribute to Agent Orange exposure have been reported.
The nonscientific
factors

which

might

media

complicate

have

highlighted

a study of health effects

among Agent Orange exposed ground troops.
general
the

public

following

possible

issues

sources

the

addressed

in

our

other

design:

exposures

which

with outcomes similar to those veterans

are attributing to Agent Orange (e.g.,
the

study

of 2,4,5-T and dioxin exposure other than

be associated

experience",

Veterans and

are likely to be aware of and expect to see

the use of Agent Orange in Vietnam;
might

several

use

the

total

"Vietnam

of other herbicides, the spraying of

�,

,

insecticides in V i e t n a m , wide use
U.S.

troops);

Orange

suspected

related

VA

problems;

11

of

illicit

bias

drugs

among

acainst finding Agent

difficulty

of

using

existing

records

to determine the exposure of ground troops to Agent

Orange;

and the proposed

delayed

effects

of

"time-bomb" mechanism

by

which

Agent Orange are experienced following

weight loss.
It will be difficult to control
sources of exposure to dioxins.
may

be

possible

stratification
region

of

on

(confounding

dealt

with

possibility
evaluated

in

of

selected

for

study.

However,
the

U.S.

such

and

of

as

geographic

other

Vietnam

and the suspected VA bias

protocol

term

improbable mechanism since
service

problems

"time-bomb"

long

techniques

residence

factors)

the

the

through

analytic

The

possible

Some control of the problem

urban/rural

residence.

exposures
are

through

for

(section
mechanism

follow-up

of

III).

The

can

only

be

the

veterans

biologically this seems an
men

during

their

Vietnam

would be likely to have been at their leanest.

major portion of their stored fat is

most

likely

to

The
have

been added since the time of exposure, and major weight loss
is most likely to be a sign of disease rather than a cause.
E.

Studies of Agent Orange by Other Investigators
The focus on Agent Orange in the

years

has

resulted

past

five

in a number of different studies among

Vietnam veterans, some completed and some ongoing.

Many

of

�12

these

we

consist

have
of

classified

the

volunteers.

uncontrolled

They

information.

as non-epidemiologic since they

may,

We

non-epideraiologic

collection

of

nonetheless,

summarize
studies

below

and

the

data

from

provide
some

current

useful
of

the

epidemiologic

studies which have come to our attention.
1.

Non-epidemiologic

Several studies have
groups

been

undertaken

and by the Veterans Administration

the health effects that could be
exposure.

Participants

volunteers.
biased

in

have,

Thus any results
the

direction

related
for

obtained

of

Agent

most
are

control

to

be

over-reporting of conditions

(All of these studies could

series.)

In addition,

for

be

considered

Despite the problems, these studies may have

value in identifying types
looked

been

group (i.e., veterans not exposed} is available

for comparison.
case

Orange

part,

likely

linked to a presumed exposure to Agent Orange.
no

veterans

to gather data on
to

the

by

in

a

of

conditions

that

should

be

controlled study of the effects of Agent

Orange.
The first of these ad hoc studies was done
by

Maude

de

Victor

a

VA

claims worker in Chicago.

processed a claim, made by a veteran who was dying
cancer,
Vietnam.
who

that

his

illness

in

of

1977
She
lung

was due to "those chemicals" in

She identified 35 other Vietnam

veteran

patients

she felt could be suffering from dioxin poisoning.

Her

�13

findings were publicized through a Chicago TV
and Ensign,

1980, pp.

Citizen

Soldier,

a

G.l.

service

to

Orange exposure.

veterans

who

had

de

Victor's

veterans

offered

a

rights

toll

self-administered

medical

free

concerns about Agent

Callers to this "hot line" numbered

3.000 by the end of the summer in 1978.
page

from

and

organization based in New York City,
phone

(Uhl

193-195).

Following the interest developed
results,

station

about

Each was sent a six

questionnaire

that

had

detailed questions in the areas of military-service history,
perceived herbicide exposure, personal health history,
medical

history,

and

family

stillbirths, miscarriages,
November,
coded.

1,000

history, with an emphasis on

and congenital birth defects.

questionnaires

Reported cases of

past

By

were returned and 536 were

cancer

and .birth

defects

verified with local physicians or medical records.

were

From the

results of this select group of responses 35 cases of cancer
were

reported

(included

3

cases

of

kidney cancer, 3 of

testicular cancer and a number of lymphatic system
77

children

were

cancers);

reported to have been born with defects;

and large numbers of respondents complained

of

skin

nervous

color,

difficulties

sensitivity

to

light,

and

changes

in

system

(Uhl and Ensign, 1980, pp 197-209).

Other "hot line" programs have also been established
and

have resulted in reports on the health effects of Agent

Orange.

For

example

in

September,

1980,

the

Vietnam

�14

Veterans

of America announced the development of a hot line

to provide vets with information.
medical

questionnaire

that

The callers were

of

Public

1981,

not

found

analysis.
called

others

have

Health.

(Muller,

any report of this

Another veterans group in St.

CAVEAT

conducted

We

a

was to be evaluated by faculty

members at Columbia's School
pp.32-33).

sent

Louis,

Missouri,

(Concerned American Veterans Against Toxins)

extensive

phone

interviews

with

veterans and

(eg., highway crew members) concerned about exposure

to herbicides and provided a listing
obtained

including

of

numbness,

reported symptoms

nervous

disorders,

psychological effects, skin rashes, alterations of

the

sex

drive, cancer, and birth defects (Furst, 1981, pp.39).
The Veterans Administration has also
development

health

effects

herbicides in Vietnam.
a

questionnaire,

laboratory tests.
been

the

of a data base called the Agent Orange Registry

to record information on veterans who
possible

initiated

are

resulting

concerned
from

about

exposure

to

Information obtained is derived from

physical

examination

and

a

set

As of October, 1981, 51»658 veterans

examined and their data computerized.

of
had

A Data Analysis

Task Force has been established to evaluate the

information

obtained*
2) Epidemiologic
The Air Force has designed
early

stages

of

and

is

now

in

implementation of a study of the aircraft

�15

crews who were involved in Operation Ranch Hand.
!

the

This

was

«.

;

name given to the defoliant spraying operation in South

Vietnam.

The Air Force study is a historical cohort

design

(see Appendix B for a discussion of this designjin which the
study group consists of all C123 crew
the

Ranch

Hand

Operation.

involved

Air

Force

were not involved in the spraying program.

personnel
Each of the

study and control members'will be examined according
extensive

protocol

involving

procedures,

laboratory

and

calls

physical

to

an

examination

questionnaire.

The

protocol

for long-term follow-up and regular re-examination of

both groups.
should

in

A control group of larger size

will be constructed utilizing similar
who

members

be

Preliminary results of the

Ranch

Hand

study

available in time to assist in early phases of a

ground troop study.
The Ranch Hand study presents a situation distinctly
different

from that confronted in a study of ground troops.

The Ranch Hand cohort is a

well-defined

group

of

limited

size, with known very heavy exposure to Agent Orange.
the conduct of a historical cohort study in
relatively

straightforward.

sample size, (approximately

small, the criticism

is

group

is

The study has been criticized

on the basis of small sample size.
the

this

Thus,

only

While it

is

true

that

1200 exposed) is relatively
partially

warranted.

When

faced with a limited population of this type, in which there
is no

possibility

of

an

increase

in

sample

size,

the

�16

epidemiologist

must

make

limited population.
limitations

Care

the

best

must

be

possible
taken

of

use

of the

the

obvious

in the results concerning rare disease, but the

population

should

be

fully

information

which

can

utilized

for

the

important

be obtained on more common diseases

from a group with heavy exposure.
The
planning

Australian

stages

of

government

a

study

of

is

currently

in

the

Australian ground troops

potentially exposed to Agent Orange in a fashion similar
the

American troops.

Currently the focus is on development

of a historical cohort study in which the study group

would

be defined on probability of exposure to Agent Orange.
are considering a control
The
test

Australian
of

the

considering

group

of

non-Vietnam

They

veterans.

group is currently working on a feasibility
exposure

a

variety

gradient
of

construction

preliminary

and

studies

is
of

case-control nature (see Appendix B for definition and
of

this

to

design).

a

uses

Results from the Australian preliminary

and pilot studies should be available in .time to

assist

planning

Since

the

exposures,

the

of

the

U.S.

ground

troop ' study.

Australians are also studying ground
U.S.

and Australian

troop

studies should be closely coordinated.

The Center for Disease Control is now
case-control

study

of

birth

birth defects drawn

from

the

defects

The

cases

registry.

in

defects.

conducting

a

The cases will be

metropolitan

Atlanta

birth

will be matched with normal

�17

births in the same area.

The case and control fathers

be compared as to Vietnam service experience.

will

Early results

of this study may be available in time to assist the, ground
troop study.
Several state health departments are either planning
/

or

conducting

studies of their own veterans.

Michigan particularly has conducted a
,survey
\

mailed

The State of
questionnaire

of its veterans and is pursuing some additional data
:

gathering.

The results of this work should be

further design of the ground troop study.

of

help

in

�18

III.

Proposed Protocol
A.

Introduction

In the following sections we detail the protocol for
the proposed historical cohort study.
a full

questionnaire,

likelihood

index

and

details

of

These details include
the

proposed

exposure

the complete physical exam protocol.

We feel that these materials

should

scientific

Knowledge by veterans of their

review

groups.

be

presumptive exposure (according to the

released

study)

only

and

to

of

the

details of the examination instruments would probably make a
valid epidemiologic study virtually impossible.
The planning of an epidemiologic study is usually
lengthy

process.

The process in the current study is much

more complex than usual.
population

a

involved

The large size

of

the

potential

(2.5 million men estimated), the large

and varied geographic region in which military activity took
place,

the

complexity

of

the

military

operations,

the

variety of potentially confounding exposures (such as combat
experience

itself,

possible

exposure

defoliants, chloroquine, Dapsone,
control

agents)

along

with

illicit

the

sorting

out

the

specific

drugs

problem

effects

In addition, in

the

to

records

and

confounding factors which must

understand
be

taken

of

of

exposure much more complex.
military

and

riot

lack of detailed data on

exposure to Agent Orange all make the
and

to malathion, other

Agent
order
the

into

defining
Orange
to

use

variety of
account

to

�19

construct

a

study

and

control

group

requires

contractor become relatively expert in military
and

operations.

study imposes a
planning

of

that the

terminology

The extreme interest qf the public in this
greater

than

usual

need

for

additional

safeguards against bias which are beyond those

normally required in an epidemiologic study.
The Air Force has spent more than 3 years developing
the protocol for their study which is a more straightforward
research problem than that of a
Australian

working

proposing

preliminary
completing

troop

study.

The

group planning their study of Australian Vietnam

veterans has been
currently

ground

studies
a

design

Thus it is clear that

for

almost

cohort
to

more

a

years

construction

gather

for

two

additional

and

testing
relevant

is
and
data

full historical cohort study.
than

required to complete this study.

several

months

will

be

�20

B.

Methods
1.

Summary of study design

The recommended
cohort

design

comparing

exposed cohorts

for

study

presumed

health

uses
highly

outcome.

the

an

historical

and

minimally

study

cohorts

should be define/d through the use of Army and perhaps Marine
Corps records for the period 1965-1971.

The

most

feasible

plan is to select high and low likelihood of exposure groups
which

have

maximal

probability.

Battle

possible

separation

casualties

of

exposure

should not be included in

the cohorts, nor should re-enlistees, officers or other with
multiple

Vietnam

tours, unless an easy method of selecting

comparable cohorts of officers and
personnel

can

be

devised.

multiple

Existing

active duty and veteran periods
each member of the cohorts.

should

tour

of

duty

records of both the
be

abstracted

for

All members should be traced as

necessary to determine current vital status, and be examined
by a standard protocol including an extensive questionnaire,
physical examination and laboratory testing with input
current

and past physicians..

Data analysis should address

the quality of the data collected and the
the

cohorts

as

well

as

from

comparability

of

the association between exposure

status and health status.
2.
In
possible

health

Rationale for historical cohort study
many
effects

epidemiologic
of

studies

environmental

of

exposure,

the
the

�anticipated outcomes associated with
exposure

nay

be

very

poorly

hi£h

defined,-

otherwise difficult to predict
These

even

in

advance

of

of

diffuse

very

levels

or

the

study.

problems of defining the possible outcomes may be due.

to 1.) an incomplete understanding (if any) of the effects of
the

chemical

constituents

of the exposure on hunans, 2) a

possible effect on multiple systems which can
number

result

in

a

of widely differing observed outcomes among affected

persons

and

themselves

3)
to

the

tendency

have

been

a

wide

persons

exposed

particular health problems to
investigator

of

to

who

attribute all their

the

exposure,

of

complaints

variety

believe

offering
of

the

exposure

effects with which to deal.
For Agent Orange, a great deal of
been

done

on

exposure.

the

These

consistent.

subject
results

of
of

animal

2,4-D,
animal

work

2,4,5-T,
studies

and TCDD
are

of

not

Effects have been observed in all organ systems

but with much variation from one species to another
range

has

observed

effects

and

in

the

in

the

dose required to

produce an effect.
The knowledge of human health effects is
complete

than

contradictory.
from

the

that

of

animal

We have compiled a

studies
list

(see

much

and

less

is

more

Appendix

F)

literature of more than 100 suspected symptoms or

diseases

potentially

Orange.

Thus,

there

associated
is

as

yet

with

exposure

to

Agent

no firm disease outcome

�22

established in any hunan population which could be used
developing

a

for

case-control study of t^he,effects of exposure

to Agent Orange in Vietnam veterans.
Because the outcome is difficult if

npt

impossible

to define, the case-control approach, dependent on the clear
identification of an outcome cannot be used.
of

a

well

required.

defined

a

cohort

study

Since exposures took place 10-15

since . a , rapid
effects

outcome,

from

historical

answer
Agent

form

to

the

Orange

of

In the absence
design

years

ago

is
and

question of adverse health

exposure

is

desired,

the

the cohort study, in which persons are

identified from past records in terms of their exposure
followed for outcome status over time, is required.

and

In this

case, it may be possible to identify exposures occurring ten
to

fifteen

years

ago, and thus to start the post exposure

follow-up at the point of
allow

10-15

years

probable

cancer,

For certain outcomes,

would

such

it could be necessary to follow the cohort into

the future, 25-35 years post exposure, because of
induction

This

of observation of health and disease in

exposed and unexposed veterans.
as

exposure.

or

latent

period

for

the

both

the

various cancers and

because the exposed veterans will, on the average, only then
have reached the age of most frequent diagnosis of cancer.

�23

3.

Development of Exposure Likelihood Index

Limitations
The most critical step in establishing an historical
cohort

study of Agent Orange is to estimate the exposure to

Agent Orange with sufficient detail and
the

identification

accuracy

to

allow

of two or more groups with a meaningful

difference in exposure.

Unfortunately we have identified no

mechanism

which would allow precise documentation of actual

exposure.

We recommend, therefore, the

exposure

likelihood

index.

This

development
index,

of

if properly

constructed, should allow separation of individuals along
gradient

of

their

probability

Agent Orange in Vietnam.

of

an

a

having been exposed to

The development of

this

exposure

likelihood index is beyond the scope of our current contract
because of 1) the volume of records which must be
2)

the

disorganized

state

of

the

records

reviewed,

and

3)

necessity for hand review and abstraction of all records

the
by

persons with appropriate security clearances.
Unfortunately,
review

of

spraying

Vietnam conflict.
types

of

forms

the

need

for

Thus, the records are
and

systematic

records was not anticipated during the

in

varying

found

degrees

without any organized listing of data
we

eventual

in

varying

of completeness

sources.

Therefore,

can at this time only outline below the general approach

which should be taken to develop this index from the records

�24

currently known to be available.

Summary of Procedure
The developnent of

likelihood

index

Document Agent Orange use (including

Ranch

should follow six steps.
Step 1:
Hand,

helicopter

the

exposure

These steps are:,

spraying,

surface

spraying,

aborted

missions, military occupational exposures and accidents)
i

by

i

location, date and quantity used per unit area.
Step 2:

Identify company headquarters in

high

and

use

and

low use areas.
Step 3:

Integrate data

on

Agent

Orange

troop location to develop likelihood of exposure indices for
individual companies for short tine intervals (day-by-day if
possible).
based

This

primarily

consideration

likelihood
on

should

of

frequency
be

given

exposure
of

index should be

exposures

although

to differentiating between

exposure to aerial spraying which is probably lower dose and
other

exposures such as hand spraying and being under abort

mission drops which were probably higher dose exposures.

�'

25
1

Step J»:

Identify soldiers

in

companies

present

dates

exposure

daily

$
Step 5:

indices
in

by

for

step

4

Using

likelihood

of

the companies in which the soldiers identified
served,

likelihood

the

of

develop

exposure

for

a

cueraulative

each

soldier

individual

by sunning the

number of exposures he encountered.
Step
likelihood
low.

6:
of

Order

by

individual

cohorts

fron

the

high

and

low

the

extremes of the spectrum.

exact cutoff levels, whether in percentiles
of

exposure

cumulative

exposure levels into a spectrun froa high to

Select members of

exposure

troops

will

likelihood

or

should

The

frequencies

have to be set when the number of troops

and breadth of the spectrum of exposure are known.
offs

of

The

cut

be set to maximize the differences in exposure

between the two cohorts.
We

have

proposal,

received

"Proposed

a

Agent

copy

of

Orange

the
Troop

December
Exposure

lion-Exposure Cohort Selection concept Paper" by Dr.

1981
and

Bricker

of the Department of Defense (DOD) (included as Appendix H).
The approach outlined by Dr.
with

our own.

Bricker is in basic

agreement

The major difference is that we propose that

the presumed exposure likelihood level be verified for
individual

soldier

dates of service.
B

and

R

by

establishing

the

company(ies) and

This procedure would allow for

leaves,

hospitalizations

and

each

turnover,

temporary

duty

�26

assignments all of which can introduce serious
cisciassification
The

individual

problecs

of

into an index based only on cocpany data.
index

is

also

necessary

to

verify

non-exposure for the low likelihood cohort.
A flow chart illustrating these steps

is

given

in

Figure 1 and details are discussed further below.
%
1) Documentation of Agent Orange Use
The first step in the development
likelihood

index

has,

of

Orange

exposure

in

South

Vietnac.

course, been done for the Air Force spraying

missions which are recorded on
A-gent

the

should be the collection of all available

data on the actual use of Agent
This

Of

the

HERBS

tape.

However,

Orange was also used at times in helicopter missions,

riverine

operations,

basecacip

perimeter

roadside

spraying

operations

and

spraying, and probably in other ways as

well.

In addition, Secretary

Schweiker

on

September

1961,

announced that there were records of 92 aborted Ranch

Hand missions in which the load of Agent Orange was
dumped

for

aircraft.
aborted

safety

reasons

to

lighten

the

There are now known to have been
missions.

There

appear

to

men.

180

such

of

over

There have also been reports of several major

accidents involving leakage of
facilities.

on the

have been four bases

where aborted missions potentially exposed a total
25,000

rapidly

load

over

23.

Agent

Orange

from

storage

Any of these situations could potentially have

�26i

Fig.1

FLOW CHART OF DEVELOPMENT OF EXPOSURE
LIKELIHOOD INDEX
Review records for
. evidence of Agent
Orange use

heavy

STEP 1

use

little or

no use

medium
identify
battalions
In area

Select battalions
which appear to

Search
records

STEP 2

Identify
battalions
in area

Set aside
A

Search
•army
records

except for AO
exposure

1
"

Identify

Identify
• individual
companies

Inrliwirlnnl

companies

STEP 3

Do
company &gt;t

STEP 4

STEP 5

Develop company
likelihood of
exposure index

Identify individual
soldiers by dates
in company

Create individual
exposure likelihood
index

Set aside
B

heavy
exposure
o

Do
company
ecords verify
little or no
exposure
7
Develop company
likelihood of
exposure index

Identify individual
soldiers by dates
in company

no

Create individual
exposure likelihood
index

yes

yes

I

STEP 6

no

Do not
conduct
study

Rank all individuals on
exposure likelihood index
Can a
sufficient
pool with clear
separation of
exposure
be developed

yes
Continue to
develop index

Return to records
, A and B above and
evaluate for possible
inclusion

no

Is
the pool
large
enough

Select top and
bottom groups for
study cohorts

no
i
- x ' ..
Set aside
B

�exposed

groundtroops

records

of

such

to

Agent

Orange.

Unfortunately,

activities were not kept in any organized

fashion and may be difficult to verify.
A complete search of all Array records would be quite
costly.

We

recommend

as

a

niniinua

Assistance Command Vietnam (MACV)
detail

for

evidence

of

that

records

the

be

Military

reviewed

in

Agent Orange use and to establish

geographic areas and time periods of most frequent and least
frequent
comnand

use
level

postponed

(hopefully
records

zero
for

until' tentative

use).

Agent

The review of lower

Orange

selection

of

The

could

be

likely candidate

battalions is Bade so that review could be
these.

use

concentrated

on

battalions which were operating in areas known

to be heavily sprayed through Ranch Hand records and/or

who

have been associated with possible heavy exposure frou other
types

of

records

should

be

selected

for

screening

of

companies and platoons for which exposure likelihoods should
be developed.
largely

free

Likewise,
of

these

battalions
possible

which

appear

likelihoods

be developed.
battalions

have

for

which

(which are expected to be low) should

Concentrating efforts on

would

be

exposures should also be

selected for screening of companies and platoons
exposure

to

these

2

types

of

the best probability of success and

keep the cost of the record review down.

�In all cases of identified use of Agent Orange,
locations

must

be recorded as precisely as possible in the

UTM (Universal Transverse Mercator) system by calendar
and

tine.

In

the

date

addition, if possible, the magnitude of use

should also be recorded.

2 ) Tr opp Hovement s
In order to develop the exposure likelihood index
thorough

review

of troop locations which can be correlated

to areas of exposure must
should

a

be

made.

The

troop

locations

be identified at the company level, or platoon level

if possible.

While the location of the company headquarters

is certainly not as precise an estimate of the location of a
given soldier as that soldier's actual location on

a

given

day, the company headquarters is the most precise locational
information which is consistently available.
consistent

records

individual

soldiers

individual

in

the

which

can

identify

reliably.
proximity

The
of

There

the

battalion headquarters.

would

locate

company

an

the

location

According to DOD personnel
units,

at

headquarters in a relatively

stable position over a period of several weeks.
would

of

his company headquarters

with whoa we have discussed this problem, combat
least,

of

placement

provides better precision than placing him at
of

no

location

the

are

The

troops

then operate in a relatively confined area around the

headquarters location.

It should be possible to locate

the

�conpany

headquarters

through

(operations officer) records.
for

greater

accuracy

records inducing:
and

the

special

by

use of the battalion S3

These should be

the

use

of

After-Action
operations

Reports,

both

activities;

the

Lessons

of

which

Brigade

Learned
relate to
Situation

the company Horning Reports and

the Organizational Histories.
potentially

supplemented

a variety of other

the Operational Report

Reports and Daily Journals;

other

the

useful

In addition to conbat

groups,

units, cheuical detachments and

such

as special forces

engineering

also be thoroughly investigated.
other units nay require the review

units,

units,

should

The investigation of these
of

different

types

of

records.

3) Company Likelihood of Exposure
The third step in the development
likelihood

index

would

be

likelihood

exposure

of

exposure

This index should be constructed for as snail a tine

unit as possible, preferably on a day-by-day
are

the

to relate steps one and two in

order to construct a company level
index.

of

basis.

There

a variety of potential ways in which this conpany level

exposure

likelihood

possibility

would

index
be

to

could

be

constructed.

divide the area of South Vietnac

into grids of a standard arbitrarily .defined
squares

of

10

kia

on

a

One

side.

size

such

as

Every use of Agent Orange

within a grid for a given time period, such

as

one

month,

�30

could be tabulated and the number of such uses assuned as an
arbitrary level of exposure.
within

that

grid

in

the

The

placement

same

of

a

company

time period could then be

assumed to represent an exposure of all individuals
company

at

that

in

time to that level of Agent Orange.

the
This

system, while perhaps relatively easy and unambiguous in its
construction;

1)

would assume that Agent Orange persisted

in the environment, 2) would assume
grid

that

location

in

prior to actual spraying would constitute an exposure,

3) would assume that the average exposure over the
was

the

that

interval

of
and

the

highest

H) would

exposure

tabulated during that

lead

to

a

depending

misclassification,

interval

certain

on

the

vagaries

location and troop locations in relation

to

the

level
of

of
spray

arbitrary

division of the grids.
A second method of relating
troop

locations

would

Agent

Orange

use

be to utilize a computer program to

compare the day-by-day locations of Agent Orange use to
day-by-day

locations

arbitrary

set

proximity

as

of

criteria

for

both

ti&amp;e

and

distance

For instance, for
»

the fixed wing spraying, the proximity definition could
location

of

the

of company headquarters, utilizing an

the definition of a "hit".

.

and

be,

company headquarters simultaneously within one

kilometer and within a two day period of tirae following
use of Agent Orange.

the

�31

As an alternative, although complex and more costly,
such

a

procedure

could

be

performed by utilizing actual

distances and time following exposure in an algorhithm which
would have a decreasing probability of exposure according to
both increasing distance and time following the Agent Orange
appl icatio&gt;n.
:

This

technique,
|

however,

implies

greater

accuracy than is probably warranted.
Another possible refinement to such a mechanism

for

developing the exposure likelihood index would be to utilize
an arbitrary weighting scheme in which &amp; second
of

the

between

herbicide

within a defined period of time, such as

four

twelve

application,

and
would

potential than the
defoliation,

be
first

to

weeks

assumed

following

the

first

to have a greater exposure

application

because

of

partial

particularly of the highest canopy as a result

of the first application.
applied

application

produce

a

The arbitrary weighting could
higher

be

index for troop exposure in

proximity to a second spraying than to a first spraying.

�H) Identification of Individuals in Companies
The fourth step in the development of
likelihood

index

would

step

reports
relating
the

is

the

company

show,

fqrr

any

to

temporary

assignments.

All

The
level

given

individuals

company,

exposure

be to place individuals within the

company on a day-by-day basis.
this

the

primary
Morning

date,

document
Report.

significant

for

These
events

including transfers in and out of
duty

assignments

and

I?

and

R

of these assist in placing an individual

at the company level on a given

day.

In

order

to

fully

define the presence or absence of all individuals in a given
company for a specified short time period such as one
the

Morning

Reports

for

that

company

reviewed for up to six months preceding

would
and

week,

have to be

following

the

specified time period to identify all material relevant to a
given individual.
company

This is

commanders

the

result

frequently

did

not

of

the
learn

fact

that

of medical

evacuations or other reasons for absence until months later,
and

that personnel orders were often changed after issuance

to the company.
supplemented

The review of the Morning Reports should be

by the use of a variety of other records, such

as Personnel Type Order Files, Special Orders, Line of
Files,

Welfare

which must be

Fund

Files,

investigated

Duty

and possibly some other files
for

content

such

as

Reporting Files and General Finance and Fiscal Files.

Command

�5) . I n d i v i d u a l E x p o s u r e L i k e l i h o o d I n d e x
After completion of the roster of individuals in
given

company

and the dates of presence in the company for

each individual, the
exposure

likelihood

fifth
index

step

in

development

company

exposure

day-by-day

likelihood

of

the

would be to relate this file to
'

the

a

exposure

'

'

index,

A

cumulative

index can be built for qach individual

by summing the number of times he was actually present in

a

company at a time the company was potentially exposed taking
into account transfers to other companies, leaves etc.

6) Construction of an Exposure Likelihood Gradient
Once the cumulative exposure
the

individual

individuals can
exposure.

soldiers
be

This

ranked
ranked

individuals to be assigned

have

likelihood

according
list

can

to
be

for

calculated,

been

index

the

their
used

level

to

of

identify

to the low and high likelihood of

exposure cohorts who are at the extremes of this gradient or
spectrum.

While it

detailed

gradient

moderate

and

estimation,

high,
it

might
of

be

possible

to

have

a

more

likelihood of exposure, such as low,
in

order

to

provide

dose-response

is perhaps desirable in this study, because

of the imprecise nature of the exposure likelihood index, to
identify
index.

only

two groups, those at the two extremes of the

These groups would provide
likelihood

of

the

exposure,

maximum
and

achievable

separation

on

minimize the

possibility

of misclassification between the two levels.

�R e c o n r.i e n d: e d.. P r o c e d u r g s
Mr.

Richard Christian (Chief of

Rulenaking

Branch,

Department

of

the

the

Research

a.nd

Army) has the most

experience and expertise regarding the records of both Agent
Orange

use

and

troop

movements.

Therefore, we recommend

that a set of criteria be sent to the Agent

Orange

Working

Group for each of the steps outlined above to be transmitted
to Mr.

Christian and his

record

searches,.

group

who

will

do

the

A summary of the steps, some of the types

of records available and the outcome of each step
in

Table 1.

exposure

is

given

In order to insure that the limitations of the

likelihood

developing

actual

the

index

analysis

which
and

will

be

important

in

interpreting the results are

thoroughly understood, we recommend that a member of the investigative staff of the one coordinating center work closely,
with

Mr..

Christian

information

and

and

lists

his

from

staff
which

membership in the two cohorts will
responsibility

for

in

obtaining

the

the final selection of
be

made.

Placing

the

obtaining this information on the Agent

Orange Working Group will, facilitate as rapid a response
possible

and

considerable
developing

will

also

M'F.

coordinating

the

expertise and experience with
a

benefit

of

their

the

problem

of

protocol for the study of the possible health

effects, of Agent Orange.
from

provide

as

Upon

obtaining

this

information

Christian and the Agent Orange Working Group the
center can establish the

cut

points

for

the

�35

gradient

of

exposures

based

on

the range of values, the

number of troops entered into the gradient and the number of
individuals needed for each cohort.
In

order

to

reduce

'

costs

and

reduce

the

time

i

required

to

complete

the

exposure

likelihood

recommend that after the completion of the
records

the

locations

of

all

index, we

review

identified

uses

of

KACV

of Agent

!

Orange, including those in the HERBS tapes, be plotted on
computer

map

of

South

Vietnam

intervals so that

areas

of

identified

and

further

directed to the most
separation

between

for raonth-by-month time

high

efforts

likely

and

likelihood

usage

areas

for

can

achieving

high and low exposure groups.

involving

investigated.

low

be

and record reviews can be

previously mentioned, the identified
accidents

a

aborted

maximal
Also, as

missions

and

Agent Orange should also be thoroughly

The entire process of developing an

exposure

index could be greatly simplified if one or more

of these situations is

clearly

linked

to

exposure

of

a

sizeable number of groundtroops.
If, after the completion of work oh the
of

the

exposure

likelihood

index,

identify comparable groups with clearly

development

it is not possible to
different

exposure

likelihoods, then the study should not be implemented.

�Table 1

Summary of procedures for selection of participants

Selection procedure
nodal points

Records available

Activity

HERBS Tapes, fixed
wing abort mission
records, likelihood of perimeter
spraying, eg MACV
records, chemical
unit records

Review for areas
frequently sprayed
and never sprayed
by dates.

Identify areas
frequently exposed
-and never exposed
(or infrequently)

B. Selection of companies for development of exposure likelihood
index

Military movement
records, eg,
Battalion S3
records, situation
reports

Review for time
and place of company headquarters
in identified
areas

Identification of
companies for exposure likelihood
index development

C. Determination of
company specific
exposure likelihood index

A § B above

Integrate agent
Orange use and
company location
data

Company specifics
exposure likelihood index

D. Selection of
possible participants

Morning reports,
personnel records

Review for company
members present
during company time
of interest

Identification of
personnel for person specific exposure likelihood index

E. Determination of
person specific
exposure likelihood index

C fT D above,
records of other
companies in which
individual served

Integrate day by
day company specific exposure index
and personal day by
day presence or absence in company

Person specific
exposure likelihood
index

F. Selection of
participants for
study

Exposure likelihood index determined in E above

Order potential
participants by exposure likelihood
index, select from
appropriate percentile ranges

Selection of
veterans to be
followed up and
invited to participate

A.

Selection of areas
of highest and lowest air spraying
and exposure

Outcome

�,

E s t a b l i s h i n g c o h o r ts

Group C o;:. p a r a t i 111
I u r i n ~ t h e c e v c 1 c ;:;. t r. t c f
index.

the

the

exposure

r e c o r d s of battalions selected in areas of high

and lou[ probability of exposure (for which
will

bd

abstracted)

rcneral c o m p a r a b i l i t y

The

initial

achieve

concurrent

cuch

exposure GronTp~sj.

use

antifungals.

in

searched

important
of

as
high

should
possible

and

oif

low

confounderc.
attempt

on

to

other

Agent Orange

tine

period, types of coisbat activities

such

things

an tTn aTi

as

riot

control

insect

in the Delta region would

agents,

repellents

For example a battalion which provided

patrols

records

for evidence of

possible

battalions

the

conpany

Other concurrent exposures should include

are^"3,

insecticides,

jungle

on

be

conparability

exposures

Geographic

should

selection

as

and known

likelihood

and
mostly

be more likely to

be exposed to insecticides, insect re 'p'eTl.-arrrrsl .and antifungal

agents

than one which operated in the less heaviTjf forested

areas of the Central Highlands.

Final Criteria for Selection of Individuals

and

Individual

Conparability
Once the individuals with high and low
been

have

identified according to the exposure likelihood iTiaex%

�the i n u i v i d u a l personnel records
abstracted

for

in

Saint

Louis

LUEL

infcrcation on other V i e t n a m service, prior

cedical

history,

history,

1JOS

enlistment

characteristics,

Classification,

service

and discharge status.

Until

the completion of work on the exposure likelihood index,
will

not

be

oe

possible

to

specify

any

type

sampling plan because the nature and size of

it

of possible

the

potential

groups are unknown.
To estinate the effect
Agent

Orange,

the

two

on

comparison

comparable as possible on other
affect

health

health

outcome.

A

groups

factors
variety

of

exposure

to

oust

as

which
of

potentially

denographic

personal factors, including age, race, socioecononic
and

educational

level,

and

status

have been clearly established in a

wide variety of epidemiologic studies as factors
influence

be

health status.

which

nay

In addition, volunteers have beer,

shown in several epideniologic studies to have

a

different

health status than nonvolunteers.

Proa our discussions with

DOD personnel,

likely

differences
rifle

it

between

companies

cocpanies.

appears

There

the

and

very
types

those

also

were

that

there

were

of individuals who made up

who

made

probably

up
icajor

headquarters
differences

between individuals in nonconbat units and in coabat

units.

We would expect there to have been major differences between

�those individuals Who served one tour of duty .in Vietnam and
those who reinlisted for multiple tours of duty.

Therefore,

comparison of volunteers tc draftees, of ooabat to noncoibat
units

or

of

reinlistees

to

one-term

servicemen

probably be confounded by

differences

personal

as well as other possible factors

characteristics

in

would

demographic

and

associated with these different types of units.
The development of an exposure likelihood index will
be

complex

of duty.

even for an individual who served only one tour

It would be nuch more difficult

for

duty.

The individual who served more than one tour of

also

Orange

much
sometime

individual

who

individual

construct

index

is

an

to

core

who served more than one tour of

likely

during
served

to

his
only

likelihood

of

Vietnam
one

exposure,

service

tour.

may

well

be

impossible.

while the
have

the

identification of a
likelihood

Officers

represent the extreme of this problem since they
likely

the

could

duty
the

than

Thus,

comparable group with multiple tours but low
exposure

duty

have been exposed to Agent

individual serving multiple tours of
highest

an

of

probably
were

most

to have multiple tours of duty and to have served in

a variety of both combat and noncorabat situations.

�40

For the reasons outlined above,

our

reconaendatior.

is to restrict the study to men serving 13 months or less in
Vietnam and 3 years or less in the
men

woulc

offer

the

greatest

arced

services.

These

probability of achieving a

clear separation between high and low likelihood of exposure
groups
and

which

are

decographic

potentially

also comparable on a variety of personal
characteristics,

confound

the

comparison

However, all menbers of a company
could

be

which

or

otherwise
of

other

these

might
groups.

relevant

unit

included in the initial exposure likelihood index

and followed through records of their service in Vietnam
develop a complete exposure likelihood index for them.

to

This

would allow investigation of the possiblity of identifying a
large

enough

group

of high and low likelihood of exposure

individuals comparable in other

factors

among

those

with

multiple tours of duty, including officers, for inclusion in
the ultimate study.

This would,

of

course,

require

much

more extensive and costly record review.
If a sufficient number of individuals is
in

the

identified

high and low likelihood of exposure groups it eight

be possible to individually match each
exposure

high

likelihood

of

soldier to a low likelihood of exposure soldier on

several important characteristics particularly age, race and
socioecononic status at entry into the armed services.
matching could be done by
would

add

considerably

existing
to

the

computer
cost

of

Such

programs

but

the study.

Any

�41

possible confounding

factors

should

be

examined

and

if

necessary handled by stratification or adjustment techniques
in the analysis of study results.
We have
cohorts

of

considered

woaen

the

possibility

in the study.

of

including

Final judgement should be

made after the review of records for the exposure likelihood
index.

DOD

has

records on the
However,
snail.

we

been

number
have

unable
of

wonen

to

supply Us with accurate

who

served

in

Vietnau.

been told that the number was relatively

Therefore, a meaningful study of then is not

likely

to be feasible.
If there are sufficient numbers of non-combat troops
who have high and low likelihoods of exposure, consideration
should be Given to stratifying the high and

low

likelihood

of exposure cohorts into combat and non-conbat groups.

This

would allow examination of the joint effects of Agent Orange
exposure and other factors associated with combat.

�42

5.

Exacinatior. Procedures

a.

Introduction

The

detailed

Appendices

C-G

literature

review

recommended

for

the

pilot testing.

data

there

are

a

wide

information

While the most well

established health effect in humans appears to
chloracne,

variety

of

be
other

that

species.

renal

These

abnormalities,

effects,

and

a

include

iinmunologic

more

liver disorders,

effects,

reproductive

variety of cancers (particular soft tissue

sarcomas in

humans).

studies

human

on

effects

of

effects

suggested from human studies and established in one or
animal

in

(summarized in Section II of this protocol)

have been considered in constructing
forms

included

Because

of

the

few

epiderniologic

effects, and particularly because of the

unique nature of the Agent Orange exposure in South Vietnam,
there

exists

the possibility of a variety of unanticipated

health effects as well.
the

study

For these reasons, the

of

cohorts must include methods for detecting those

conditions which have been suggested by
well

testing

the

literature

as

as methods for detecting unanticipated health effects.

�43

The procedures to collect data free all

Berbers

of

the high and low exposure likelihood exposure cohorts should
include 1) a questionnaire to elicit demographic factors and
occupational and residential history, an exposure history to
chemicals including herbicides, a Vietnafc
and

history,

medical history.
2) a complete physical
i.
examination including a detailed neurologic examination, 3)
a

a

service

complete

series

of

laboratory

tests

designed to complement the

physical examination, H ) a psychological test battery and 5)
a battery of neuropsychological tests.
The
instruments

following

subsections

describe

and test batteries which we recommend.

more extensive than would be practical for
study

but

the

should

a

large

various
All are
cohort

be completely tested in the pilot study.

After the pilot study, these data collection procedures
be modified as necessary.

None of the data collection forms

have been put in a self-coding
instruments

should be.

format

although

the

final

The range of responses in the pilot

test should be examined before final
schemes are completed.

can

revisions

and

coding

�44

The
information

procedures
collected

should

be

standardized

on standard forms.

and

the

The adherence to

standard procedures should be monitored by the study office.
Information
different

collected

can

procedures.

be

Both

validated using a number of
the

participant

and

those

members

of the study staff responsible for data collection,
&lt;(
the physical examinations and laboratory tests must be

blinded

as

to

the

exposure

status

of

Details of the standardizing and validating
given in section III B 9.

the participant.
procedures

are

�45
b.

Suggested Questionnaire

A su^tje ste d questionnaire s u i t a b l e for pilot testing
follows

belov;.

In

the

development

questionnaire for the pilot tect, wt

of

have

this

proposed

drawn liberally

from the questionnaire developed for the Australian Veteran:
Herbicide Studies anc the United
("Ranch

Kano")

Study,

States

Air

Force

as well as from other studies which

sought inf orr.iatiori on the general health history
Me

have

also

added

questions

of

tales.

and sections which we feel

provide important additional information and
some

Health

have

modified

questions to make then uore appropriate .for our tarjet

population.
formatted

The
by

questionnaire

has

been

reviewed

the UCLA Survey Research Center.

has extensive experience and expertise

in

and

This center

the

development

and administration of interview schedules.
Questions were drawn froa
Herbicide

Study

first was that
expertise

in
we

Veterans

felt

it

appropriate

to

draw

on

The
the

of the Australian group which has taken more than

Orange

this

draft

questionnaire

for

their

study, and the second is to make our study as

comparable as
comparisons

Australian

particular for two major reasons.

two years to develop
Agent

the

is

feasible

to

that

study

so

that

might be possible between the two studies.

some
One

of the ipajor probleras likely to face studies of the possible
health

effects

of exposure to dioxin is the need for large

sample sizes, especially

for outcomes which

are

rare.

By

�using

'couparable

data

collection

techniques

for the two

studies, it cay be possible to overcome sorae of
size

problems

by

cautiously

obtained in a similar manner.
two

coisbinin^

the

data

sample

which

was

Similar findings between

the

studies would, of course, increase the probability that

a finding is real.
The questionnaire has been administered to
tc

identify

poor

questions,

etc.

subjects

On the basis of these,

appropriate changes have been nade where indicated.
The draft questionnaire has been designed to

elicit

1) standard denocra'phic inforiaation, residence'., military and
exposure histories, 2) information on

possible

confounding

factors, 3) pre-existing and familial nedical conditions, 4)
conditions and diseases reported in

the

literature

to

be

associated'with Agent Orange, and 5) conditions and diseases
which nay be realted to exposure but which have not yet been
reported.
The section on demographic and personal factors
been

has

designed to provide 1) information on which to compare

the two cohorts (eg, years of education, father' s, occupation
etc),

2) information about exposures to hazardous chemicals

including defoliants
other

activities,

participants

frota

occupational,

recreational

and

3) a military service history and i{) the

perception of his exposure status

to Agent Orange in Vietnau,

with

regard

�47

The section on health status has
provide

information

on

conditions anci diseases
since

Vietnan

service.

existing

designed

to

health

perceived

been

status

anu

prior

to

and

occurring

V/e have included specific sections,

on diseases reported in the literature to be associated with
Agent

Orange

but

have

also

asked

screen for conditions not reported but
with

exposure

problems.
uay

usace.

conparison

health
These

status such as smoking, drinkinr and
questions

nay

are

included

to

permit

between cohorts of factors affecting health, and

enhance

the

possible

adverse

predisposing

effects

factors

of exposure.

provide the opportunity for validating reported
and

associated

as autoimmune diseases and respiratory

also to allow assessment of
which

possibly

Questions have also been included on habits which

influence

dru~

such

questions designed tc

To

conditions,

to uncover unreported conditions, ciedical release forns

should be obtained for each source of uedical care including
the current physician.
The section on military service has been included to
provide

independent

information (frou that of the military

record) of companies and

areas

served

a

and

to

provide

in

which

framework

on

the
which

participant
to report

activities involving exposure to Agent Orange in Vietnam.

�The questions on perceives exposure statue- to

Ajer.i

Oran-e have beeu i n c l u d e d primarily tc, p e r m i t uc to e v a l u a t e
the extent to which responses to
been

biased

(consciously

the

or

health

history

unconsciously)

have

by

the

respondent's perception of his exposure status (see Analysis
Section
also

for details.) The responses to these questions will
provide

the

opportunity

uisclassificaticn.

For

to

instance,

identify

if

lar^e

rcajor

numbers

veterans froa a particular time and geographic area
to

be

unexposed,

report

of

thought

the sake information about their

exposure to A^ent Orange, this v;ould constitute

presunptive

evidence of undocumented Ajent Orange use.
Since a najor question
effects

of

Agent

Orange

about

exposure

the

possible

concerns

reproductive

effects, considerable data on this area should be
In
a

order to accurately
questionnaire

(also

health

^thered.

assess the nale effects we feel that
validated

froa

adninistered to the spouse is mandatory.
the

nain

nedical

records)

A suc^ested spouse

questionnaire

follows

questionnaire.

questionnaire

attempts to ascertain a complete reproductive

history alor.3 with information on

all

confounding

procedure

factors.

The

best

important

This

potential

(although wore

difficult, costly and perhaps less acceptable) would

be

interview all former spouses as well as current spouses.

to

�49

c.

Physical Examination

The physical examination was designed to screen
possible

abnormalities

examination

is

Australian

study

questionnaire.
UCLA

School

all

from

for

the

that
same

organ

of

terms

Medicine,

we

have

systems.

developed
reasons

In consultation with Dr.

Australian form.
medical

adapted

in

for

be

the

slightly modified the

Before pilot testing, the spelling of
should

the

Dennis Cope of the

"Americanized".

The

the

form

for

The

recording the physical examination follows below.
was

This

for
as

for

form

designed to require the physician to specifically check

normal findings as well as abnormal findings to maximize the
proper

completion of the physical examination.

physical
physician.

examination
The

can

be

completed

by

The general
any

general

neurologic examination, however, should be

conducted by a trained

neurologist.

As

detailed

in

the

quality control section, all physicians from the examination
centers should be given a five-day training program

in

the

administration of this particular examination to standardize
the exam procedures and the conduct of
the maximum extent possible.

the

examination

to

�50

The last pcrtior. of the p h y s i c a l exsr fcrc
the

examining

rec.uirer

physician to succarize his findings for each

organ system and to express his level of certainty of

those

findings.

group

Use

of

this

t e c h n i q u e would

allow

comparisons on both certain and suspected abnormalities.

As

detailed in Section III B Hi, the exacining physician shoula
be responsible for explaining any abnormal findings
veteran

and

to

the

for urging or providing appropriate diagnostic

or therapeutic follow-up.

�d.

Laboratory tests

The recommended laboratory tests are listed below in
Tables

2

and 3.

Those procedures indicated by an asterisk

are those which must be completed in the examination center.
All

other

laboratory

specimens

central laboratory facility.
sampling

Procedures

be
for

analyzed at a
blind

to

complement

The laboratory procedures were

the physical examination, especially

for organ systems suggested to be affected by Agent
and

to

assist

in

detection

of

procedures

were

Orange,

subclinical or impending

conditions not easily found on physical exam.
the

split

and validation of the laboratories are detailed in

the quality control section.
chosen

should

In

chosen to provide a general

addition,
screening

battery for all organ systems for which laboratory tests are
useful.
The procedures
and

to be followed for drawing

collection of specimens must be standardized.

of laboratory personnel at each examination
proper
center.
pernit

techniques

must

be

provided

by

center

blood

Training
in

the

the coordinating

The schedules of examinations should be designed to
mailing

receipt by
specimen

that

of

specimens to the central laboratory for

laboratory

collection.

A

within

responsible

twenty-four

laboratory

specimens.

to

expedite

hours

of

individual should be

designated at both the data collection centers
central

of

handling

and

at

the

and shipping of

�52

Rationale for Laboratory 'Procedures
1) Studies on the toxicity of TCDD in anictls have
show:: that the following or^cr. systecsc are cacaceci:
a)

Liver:
Hepatic necrosis, liver
enzyme
chances, hypoproteinecia, hypercholesterolecia,
hypertriglyceridesia.

b)

Reticuloendothelial systec: Thycic atrophy,
altered cellular immunity, decreased lymphocyte
counts.

c)

Henopoietic System: Anemia, throcbocytopenia,
leukopenia, pancytopenia.

d) Endocrine system: Heciorrhage and atrophy of
adrenal cortex, hyppthyrcidisn
e)

Renal:

Increase in blood urea nitrogen

f)

In addition,
statistically
significant
increases
in
hepatocellular
carcinomas
(liver) and squamous cell carcinomas of
the lung were found.

2) Studies on the toxic effects of TCDD in man have
shown that the following organ systems may be damaged:
a)

Liver: Porphyria cutanea tarda.
Increased
levels of transaniinase and of GGTP. Enlarged,
tender liver, hyperlipideoia.

�53

b)

Renal:
Hemorrhagic
pyelonephritis

c)

Endocrine system:

c)

Reproductive system:

cystitis,

focal

Hypothyroidisr.
Infertility

3) Based on the reports of toxic effects in animal
and
human
exposures, the following organ
panels are recommended:
a)

Hemopoietic

b) Reticuloendothelial
«

c)

Renal

d)

Endocrine

e)

Meuromuscular

4) Hecopoietic screening should include:
a)

heraatocrit

b)

Hemoglobin

c)

RBC indices

d)

Erythrocyte sedimentation rate

e)

Platelet count

f)

Prothrombin time

�5) Reticuloendothelial system:
a)

White blood cell count

b)

Selective use of
determination

quantative immunoglobulin

6} Hepatic screen:
a)

SCOT

b)

GGTP

c)

Bilirubin, Total and Indirect

d)

Cholesterol

e)

HDL

f)

Triglycerides

g)

Urine prophyrins

7) Renal screen:
a)

Urinalaysis

b)

BUN

8) Endocrine screen;
a)

8.AM cortisol

b)

FTI

c)

Fasting plasma glucose

d)

2-hour post prandial glucose

$4

�:

9) Reproductive system:
a)

Testosterone

b)

55

Semen analysis
I

Additional
Performed

Diagnostic

on All

Tests

•

;

which

should

be

Subjects

Electrocardiogram
on step stool

- resting and

following

exercise

Blood Pressure - Right arm sitting
Chest X-ray - AP and lateral
Visual acuity
Nerve conduction testing - ulnar motor latency,
ulnar fast velocity, ulnar slow velocity, peroneal
motor latency and fast velocity, sural sensory latency
Routine spirometry - FEVj , FVC
Height

'

•

i
Weight
These

are

general

cardiovascular,

respiratory

be

at

performed

the

screening

procedures

for

and nervous systems.
examination

the

All must

centers.

The

electrocardiogram, chest x-ray, nerve conduction testing and
spiriometry should all be interpreted at a central

facility.

�e.

Psycholocic Tests

The recommended

psychological

follows was developed by Dr.

test

which

Carol Aneshensel of UCLA.

reasons fcr consideration of each of the
and

battery

3

test

their potential shortcomings are discussed.

batteries
At cininuc

we reconaend that the KllPI and the SCL-90 be conducted.
MHPI

is

a

self-administered

administered

by

questionnaire.

the

test

interviewer

The

SADS-RDC

and

adninisters
provide

much

definitive diagnostic data and could be extrecely
However,
least

this

the

The

the SCL-9D can be

who

would

Tne

the
more

valuable.

test must be administered by someone with at

equivalent

training

of

a

masters

degree

in

clinical psychology or medical social work and would thus be
more costly.

We recommend that the SADS-RDC be evaluated in

the pilot test and the decision on its use in the full study
be based on that experience.
f.

NeuropsycholO£ic Tests

The neuropsychologic test battery
Dr.

Francis

J.

Pirozzolo,

Baylor University Medical
recommended

Chief

Center.

was

designed

by

of Keuropsychology at
This

test

battery

is

as a means to screen for subtle central nervous

system toxic effects such as might occur

from

exposure

to

Agent Orange.

The series of tests was designed to represent

the

generally

most

neuropsychologic

test

accepted
methods.

and

standardized

It was developed with the

goal of producing a test battery which could be administered

�s:
by a trained technician and does not require the services of
a neuropsychologist.
menber

of

responsibility
technicians,

the

However, a neuropsychologist must be a
coordinating

for

training

supervising

the

Center's
the

staff

examining

continuing

and

have
center

quality of their

work, and interpreting of the results of this test battery.

�-

•

6.

61

Possible confounding factors

Confounding
those

which

may

factors

epidemiology

are

distort the apparent relationship between

two variables under study.
relationship

in

between

For example, in a study

an

occupational

exposure

of
and

the
lung

caner, cigarette smoking (a known cause of lung cancer) is .&lt;a
potential

confounding

factor.

These

confounding factors

must either be similarly distributed in the different
groups

or

data.

be

appropriately handled in the analysis of the

Similarity of

alternative

study

since

distribution
analytic

is

the

more desirable

techniques may be difficult to

apply and only partially effective.
In the proposed study potential
factors

include:

exposure

picloram or cacodylic acid;

to

other

confounding

herbicides

such as

use of licit and illicit drugs;

exposure to insecticides, and riot control agents such as CS
tear gas;
fungi);

exposure
water

to

diseases

supply,

length

(eg,
of

malaria,

Vietnam

cutaneous

service,

and

multiple tours of service in different geographic regions of
Vietnam;

pre

or

post

Vietnam

(including

TCDD

exposure);

experience;

educational

occupational

classification

service,

multiple

occupational

combat

level;

versus

race;

(MOS).

history

non-combat

rank, and service

Length

of

Vietnam

tours of service and rank can, be bandied

by the eligibility criteria specified in section III B 4.
list,

more

complete,

A

of the potential confounders we have

identified is included in Appendix I.

�62

These potential confounders are
the

moment

and

seasonal

assumed

at

to

be fairly equally distributed among similar
/
units operating at about the same time in similar geographic
circumstances.

Questions

on

confounders are included in the questionnaire.

the potential
All of these

i

should

be

more

thoroughly

investigated

development of the exposure likelihood index,
the

Army

records

of

to

both

the

through

their use or occurrence, and through

search of the DTIC files.
found

during

Any of the potential

confounders

be unequally distributed between the two exposure

likelihood cohorts should be examined in the analysis

phase

for the need for stratification or adjustment procedures.

�63

7.

Tracing, of men. selected for stucj.y

All possible effort cust be
locate

every

man

selected

for

made

to

identify

this

study.

and

This

is

particularly important because both the exposure status
the

possible

adverse

difficulty in tracing.
and

outcome

may

be

and

connected

with

Therefore, differences between

high

low likelihood of exposure cohorts in the proportion of

cohort

members

association

located

or

create

could

either

ail artifactual

obscure.
one.

a

real

For example, a

potentially highly exposed combat veteran may be both harder
to

find

and more likely to have suffered an adverse effect
i-

than an accounting clerk stationed

solely

in

Saigon

potentially

low

in the U.S.

If it is easier to locate the unexposed

the

exposure who returned to a business 'career

unaffected,

then

arty

relationship

possible adverse effect could be
hand,

if

the

with

of

obscured.

and/or

exposure
On

the

and
other

combat veteran is concerned about his health

and has filed a claim with the VA he may be easier to locate
than

the

former

file

clerk

physician for medical care.

who

In this

has

gone

type

spurious association might be observed.

of

to a private
situation

a

(The examples given

are deliberately extreme to illustrate the

problem

and

do

not indicate any specific opinion or information on expo-sure
likeli&amp;ood, or adverse effects from exposure.)

�64

Tracing operations are expensive, time consurjing and
nay

be

considered

by

sone

breaches of confidentiality.
here

is

as

invasions

privacy or

The tracing protocol

detailed

designed to make maximum use of existing military,

veterans, and public
specialized

record

data * 'bases

sources

found

first,

going

on

to

which may provokfe more privacy and

confidentiality' concerns only after
been

of

unsuccessful.

The

the. first

sequential

steps

steps

in

have
the

tracing protocol are detailed below:
a) Abstraction of existing military records of
selected

serviceman

.the record.
including
number

and/or

birthplace;
religion;
date

of

aliases

known

social

security

sex;

race

name in full,

through the record;
number;

service

birthdate

and

and/or any ethnic identification;

marital status, including

date

of

information,

marriage if available, full name of spouse and any

address information;
information,

full

subject and address;
address

in as much detail as is available from

This information should include:
any

each

of

next of kin data,
name

of

including

next-pf-kin,

parental data - same

serviceman at induction;

date

of

relationship

to

as

next-of-kin;

address at discharge;

service details, including dates, assignments,

job

in

disciplinary

service

actions;
weight,

and
eye

training,
physical
and

information,

description,

hair

characteristics and any
states.

medical

color,

permanent

titles,

including

height,

distinguishing

physical

or

progressive

medical

�These data, and any other information
included

in

the

military

found

to

be

record

and useful for tracing,
i
»
should be abstracted onto master coding forns.
The forms
should

be designed so that identifying_information, such as

names,

addresses

and

service

numbers

will

separately

fron service or medical details.

maintained

in

the

data

base

management

be

stored

Linkage can be
system

through

arbitrary number coding.
Data
selected

to

specified

above

for

extraction

have

been

provide 1) tracing routes (addresses, names of

spouse, kin, etc.), 2)

information

through

and in data bases (birthdate, social

data

sources

necessary

for

security or serial number, etc.), 3) indicators for
tracing

activity

(medical

information,

disciplinary action, etc.), and
verify

H)

tracing

further

service training,

information

useful

to

that the veteran located is the correct one (height,
:•

{

'•

eye color, permanent distinguishing characteristics, etc.).
Some members of the selected exposure
be

"located"

at

this

service, those still in
military

prison

step,

including

service

sentences.

and they will be transferred

and

cohorts

will

those who died in

those

still

serving

Their identification is direct
to the study file described

the last paragraph of (b) below.

in

�66

j j s t a b l i s h p e n t of a c o n p u t e r b a s e d d a t a
system

incorporating

data

collected

in

step

grovisior.^ for a d d i t i o n of information derived
tracing

procedures.

or

from

the

further

provide

computer

hard copy lists of selected subgroups, ordered

as required, for use by agencies or organizations
in

with

This systec 1) would provide the basis

for tracing procedures, 2) can be used to
readable

one

tracing,

3)

can

assisting

keep track of tracing procedures
!

already used, the procedure currently in progress and,
appropriate,

when

a success indicator along with status and fact

of location of veteran.
To assure
identifying

privacy

information

and
and

confidentiality,
the

actual

necessary

location

of the

traced veteran should be incorporated in a separate conputer
file for study use;

the actual address should not reside in

the tracing file.
c) Utilization of Veterans

Administration

records.

All VA data sources, data bases and benefit providers should
be queried for the most recent contact with members
two

cohorts.

and

the

Data forms must be provided for recording of

the specific contacts,
veteran

of

the

searched include:

date

the

most

location

of that information.

benefits - such

disability, especially pensioners;
Orange register;

recent

complaints

as

of

Records to be

education,

housing;

medical services;

registers;

any other veterans services provided.

the

death

Agent

benefits;

�The ordering of these searches should be
both

during

and

after the pilot test in consultation with

the VA to caxinise the
existence

of

determined

VA

data

yield
and

of

any

given

search.

the file location of that data

should be identified in most cases

from

Beneficiaries

Identification

Records

(BIRLS) file.

Other VA computer files may only need

searched

occasionally.

The

and

Those

the

Veterans

Location System

veterans

deceased in BIRLS can be eliminated

from

and

to

be

identified

as

further

tracing.

The parameters which determine the usefulness of any type of
VA file such as the date of the file or . the
(eg,

type

of

file

pension versus educational benefit) should be examined

in the pilot test.
d) J? oo per a ting V e te ra ns orga ni zations ; Veterans not
located

through

steps

a

or c above, or whose most recent

identified location is no longer current, should
t

through

cooperating

contact rosters.
date

of

*

location

Veterans

be

sought

'

organizations

To avoid privacy problems,

membership or
the

fact

and

could be obtained from the organization.

Actual initial contact could be

made

by

the

organization

with study contact only after consent by the veteran.

�68

Unit-related military service organizations, such as
the

"Big

Red One," can be queried for location of veterans

eligible for membership in t-he sane way as the more

general

veterans organizations.
e) Letters to the veteran at the last known address.
This

address

could be the induction address, the discharge

address, or an address obtained through a,

c

or

d

above.

While this letter should be used to locate or to confirm the
location of the veteran, it can also be structured to advise
the

veteran

of

the study, his selection for participation

and to initiate the recruitment

process

by

enclosing

the

initial contact letters (see Section III B 8).
The first and second mailings of this letter
be

by

ordinary

mail

address correction.
mail,

return

with a prepaid request for notice of

The third letter should be by certified

receipt

requested.

referred for field check by study
agency

such

as

should

Equifax.

Ho'n
staff

Returned

responses

can

or

tracing

by

a

be

mail should result in

further tracing steps.
f) A field check should be done in

those

which mail is neither answered nor returned.

cases

in

The first step

in such a check should be by telephone, utilizing both usual
and

reverse

directories

to

obtain telephone numbers.

this procedure is not effective, a field visit to the
should

If

house

be made by study staff or a contract organization to

ascertain if the veteran does indeed live

there.

If

not,

�the field check could then be extended to neighboring houses
to gain any information
present

whereabouts,

possible

about

the

veteran,

his

the date of his most recent residence

there, etc.
It should be
re-utilized

as,

noted

that

steps

e

and

f

can

be

a followup for all later step's resulting in

an address for the veteran.

They can be repeated each

time

an address is obtained.
g) Letter i to spouse or next of ^kin.
should

proceed

in

This

tracing

the same way as step e if the spouse or

next-of-kin address is different from that used

in

step

e

for the veteran, and if the veteran has not been located.
h) Motor vehicle operators
yehicle

registration.

Motor

registration

vehicle

and

bureaus in the last

known state of residence, nearby states, sunbelt states
other

states experiencing
operators

licence

information

including

address.

(or

vehicle

registration)

The computer tracing data

system should include information on states queried so
is

not

and

large population growth should be

queried for

inquiry

potor

repeated.

verified through step e above.

Addresses

obtained

that

should be

�70

i ) Retail or e c i t c h e c ks t.hro u ^h EC;uif ax , _ _ o r
^rga r;i z&amp; ti ons.

sini.l_ar

Again, locations obtained should be checked

through step e procedures.
j)

.Social
.date,

employment

Security
disability

and

system

the

Unfortunately,
sufficiently
current.

the
out

It

is

of

date

current,

..System

for
that

the

however,

checks

for

retirerjent
still

_status.

employed

information
for

last

those

is

is
not

receiving

benefits, and even 3 year old information may be more useful
than 10-15 year old information.
be

The information will

also

useful to ascertain employment record and therefore some

measure of stability, hence locatability.
k)

Internal Revenue Service for filing and

information only.

This file is current within one year plus

time to computerize.
for

this

study,

address

If the NIOSH waiver

the

can

be obtained

IRS files should be one of the first

tracing resources utilized, ie, for those not located

after

step 3 above.
1)
sear c h e s.

FBI (and other law enforcement agencies)
These

record

types of agency records could be searched

for contacts with the selected veterans, present location if
serving

sentences

or

on

probation

address and date of information if not.

status,

most

recent

�71

n)
be

should

Na t i o n.a 1 Peat h

I nci e x .

While

the

BIRLS

file

virtually complete through 1 9 6 1 , it is likely to

be less complete in subsequent years because of the recently
enacted law reducing the burial benefit.

Thus, for veterans

who cannot be located by the above steps, a
National

search

of

the

Death Index 'should be performed to identify deaths

not in the BIRLS file.

Note that this source should be used only for

those

not located in steps a-1 above.
n)

Advert is ing

for

w here g bou t s,

particularly

in

publications likely to attract the group not located through
steps a-E

above.

This group can be

characterized

through

information in the computer derived from the military record
and the results of the
advertising

campaign

tracing

steps

developed

and

at

an

that

appropriate
time.

Such

advertising could, for instance, take the foro of television
spots,

advertisements in major newspapers or advertisements

in so called underground newspapers.
such
still

The size and scope

of

a campaign would also depend on the number of veterans
remaining

considerations.

unlocated,

and

on

confidentiality

�8;

Subject contact procedures

After the

completion

of

tracing

procedures

(see

Section III D 7 for details) the subjects will be identified
as either a definitely identified and located individual
a

located

individual

of

uncertainty as to whether
Each

subject

a

follow-up

correct

the

the

purpose
of

contact

the

of

coordinating
the

individual

will

study,
and

individual.

center

which

encourages

be made by telephone.

the

that

a

A suggested

Along with

this

from the coordinating center, it would be helpful to

enclose

a

letter
the

encouraging
United

participation

States,

from

President

of

obtained.

For any subjects for whon a telephone

not

with some

specifies

draft of this letter is shown in Figure 2.
letter

but

identified in either category should be first

the

participation

nace

is

it

contacted by a letter from
outlines

correct

or

the

if such a letter can be
number

is

available a postcard should be enclosed with the letter

which asks for the current

telephone

nunber

so

that

the

individual can be contacted.
At the tine of nailing of the letter, the individual
data

collection

center

notified of the mailing

closest
and

to

given

the subject should be
the

address and telephone nunber (if known).
period following the mailing

of

the

subject's

current

Within a 10-14 day-

introductory letter,

each subject should be contacted by telephone by a specially
trained staff member of the data

collection

center.

This

�73
!

staff ce:-ber should be familiar with the study so that he or
£h-3 ear.- answer- questions that the subject' nay h a v e about the
study.

The

telephoner

icfcrmation • at

the

should

time

of

also

ascertain

telephone

verifying

contact

for

an&gt;

contact,

the

subjects whose identity is in doubt.
At

the

interviewer

tine

should

of

this

try

telephone

to schedule an appointment for the

subject to be examined at the data collection
interviewer .

must

be

transportation in that
assist

as

familiar
geographic

necessary

in

the

region

arranging

lodging for the subject.
participation,

with

center.
mechanisms

to

If

be

telephone

and/or

interviewer should try to ascertain the

the

subjects
number,

to

In the case of veterans who refuse

veteran

is

willing,

overcome
a

the

who

return

procedures

a

the

shortened

questionnaire should be administered by phone at that
For

of

able

transportation

reasons for this refusal and possible ways to
objections.

The

postcard with
^
above would be

time.
their
used.

Subjects who do not return the postcard must be contacted by
field visit.

The field visit should be conducted in similar

fashion and for the same purposes as the telephone contact.

�74

FIGURE 2
Suggested Initial Contact Letter
Dear Subject:
As you are aware, cany veterans who served in
from exposure to Agent Orange.
State*

is

also

South

The governnent of the United

concerned about this problem and has asked

that a study be done to ascertain what

health

effects

nay

have occurred froni Agent Orange exposure so that appropriate
treatment

and

compensation

can

be

provided.

We

are

conducting that study for the governnent.
Your name has been selected by a scientific
to

process

represent veterans who served in South Vietnacu

Many of

the veterans chosen we believe not to have been
Agent

Orange,

Agent Orange.
of

exposed

while many others we believe were exposed to
You may have been chosen to represent

these groups.

either

As the attached letter from the President

of the United States indicates, this study is considered
be

of

vital

concern

particularly the
Therefore,
study.

to

to

veterans

to

the people of the United States,
who

served

in

South

Vietnan.

we strongly encourage your participation in this

�75

In the next few weeks, a nenber of
center

closest

the

examination

tc your hont will call to explain the study

t$&gt; ypu jjin pore detail and answer any questions you may
about

your

participation.

Briefly,

consist of a questionnaire which will
illnesses

and

hospitalizations,

have

the examination will
ask

about

you

about

your

your reproductive

experience, about your service in South Vietnam, and about a
variety

of other questions which may relate to your current

health.

Whether you have been

£roup

believed

selected

to

represent

the

to have been exposed to Agent Orange or the

sroup believed not to have been exposed to Agent Orange, you
will

be given the opportunity to express your own knowledge

about possible exposure of which the study may not be aware.
The

rest

of

the

examination

will

physical examination with blood
variety

of

special

test

and

consist of a complete
urine

procedures

testing

and

a

designed to evaluate

different body systems.

.

.

All information which you give in this study will be
kept

strictly

confidential,

to

group analysis of study results.
be

part

be

utilized only for the ,

The information

will

not

of your VA records unless so requested by you, and

will not be disclosed to any government agency or any

other

agency or individual without your expressed written consent.
You will, of course, be notified promptly of any abnormality
which

is

detected

during

abnormalities are found, you will

the
be

examination.
given

If

assistance

any
in

�76
securing further evaluation of the problec as necessary
either

th-::

physician.

Veterans
You

aay

sent to the Veterans

Administration
have

IMA Supervisor

your

private

the results of this exaninatior,

Administration

physician upon your written request.
Sincerely yours,

or

fron

or

your

own

private

�Q u£lit i" control

9.

The ultimate value of the results of

data

analysis

from any study is d e p e n d e n t upon the quality of the raw data
going into that analysis.

Potentially

daLia^in^

errors

can

occur in the data set at any sta^e of data collection.

Such

errors are frequently the result of nisunderstandinss,

lack

of

care

in

the

recording

of

data

application of individual judgement.
errors

The

inappropriate

possibility

for

is particularly strong when multiple data collection

centers are utilized as in this study.
errors

or

can

be

achieved

through

The control of

the

such

use of a variety of

techniques, such as rigid criteria for the

recruitnent

and

training of study staff raenbers, the careful standardization
of all
exposure

study

procedures,

status

blinding

of

staff

as

to

of individual subjects, and utilization

the
of

procedures for double checking at least a sacple of the work
of staff njeabers.
each of the

Individual quality control procedures for

critical

steps

discussed in the following

in

the

paragraphs.

data

collection

are

�78
a.

Selection of cohorts

Part
likelihood

of

the

index

tasi;
will

situations in which the
Orange

can

of
be

constructing
to

select

likelihood

of

the

for

exposure

study

exposure

those

to

Agent

be tiost clearly documented and those units with

the most complete and careful records of troop locations and
individuals present.

In addition, when the actual selection

of individuals for study
personnel
checked

records
with

accomplished,

(predoninantly

the

unit

Such

consistency.

is

a

records

the

from St.
fron

individual

Louis) nust be

South

Vietnam

for

comparison cat) be nade by a computer
i

prosrau which can be developed during the pilot study.
prosran

would

match

against

those

fron

Individuals

with

dates
the

and units fron personnel files

South

Vietnan

discrepancy.

locate less obvious
explained

by

unit

records.

discrepant information in these two files

should be rechecked for obvious errors which
the

This

these

Further
errors.
checks,

eliminated fron the study.

checking
If

the

these

would

explain

would be necessary to
discrepancy
individuals

is
would

not
be

�79

b.

Record abstraction

The basic data set for the early phases of the stucy
will be based on abstracts of the arr.y ur4t records from SVi.
and the personnel files.
trained-

study

Abstracting

should

be

done

by

staff menbers who review the records by hand

and record the information on standard forms.

A

sample

of

these abstracts chosen by a predetermined sanpling algorithm
should

be

conplete

reassigned

to

reabstraction.

another
The

record

abstract

abstractor

records

fron each

abstracter can be compared in a siiaple list projraa
computer

for

any inconsistency.

for

on

the

Inconsistencies should be

referred to the abstracting supervisor who

can

investigate

the source of the error and bring the error to the attention
of the appropriate abstracter.
substantive

errors

forws after

two

on

more

attempts

Any abstracter found to have
than five percent of abstract

at

retraining

that

individual

should not be retained in this aspect of the study.
Additional quality control can be
the

application

of

careful

abstracters.

These

intellectual

abilities

the abstracting

standards

criteria

should

achieved

for recruitment of

include

deconstrated

at a level sufficient to understand

procedures

and

consistent clerical abilities.

demonstrated

careful

for

careful

abstracting,

and

The traininc program for the

record abstractors should include a full discussion
need

through

of

the

the effect on the study of

errors in abstracting, instruction in

writing;

numerals,

a

�80

careful

step-through

of each forr.i to be e n c o u n t e r e d in the

records, supervisee; abstracting
records

which

problems

to

illustrate
be

independent

htive'

in

of

been

a

selected

series

of

the types of forras and potential

encountered

aostractinc

records which

of

a

abstracting,

and

the

series of selected training

carefully

preabstracted

by

the

abstract supervisor.
c.

Tracing

The primary difficulty which night be encountered in
tracing is the identification of the wrong individual.
problem can be minimized by the

use

of

matching

This

criteria

utilizing a short series of question responses obtained from
the potential subject at the time of the
contact.

Information

criteria include birth
Social

Security

which
date,

nunber

would
full

or

name,

service

assignment and date in SVN and any
characteristics.

A

positive

be

initial telephone
useful

parents'
number,

distinguishing
match

for

for

such

names,

and

unit

physics!

purposes

of

determination

of

proceeding with recruitment could be

the

any

The rate of Eisnatches

three of these six data points.

detected at the time of examination should be monitored.

If

a mismatch rate of greater than one percent is detected, the
system should be reviewed to determine the cause of the high
mismatch rate.

�81

d.

Interviews

F; i c i d criteria for the recruit cent
and

of

interviewers

a standardized training program must be, developed.

manual used by the UCLA Survey Research Center
in

A p p e n d i x J ,•

is

The

included

The manuals frora the University of Michigan

Survey Research Center (4) and the Public Opinion Center (5)
uay

also

be

helpful.

Further

information

on

training

procedures and supervision of interviewers can be

found

in

the discussions and in the references included in references
6 through

14

examination

bibliography.

center

should

The

interviewers

undergo

a

two

should include:

2) explanation of

3) an

questionnaire

interview

five

The

training

the

need

for

exact procedures in conduct of the interviews and

recording of data;

manner;

training

1) discussion of the study design,

rationale and procedures;
following

each

week

session conducted at the coordinating center.
program

from

and

opportunity

to

procedures

comment

"subjects"

(each

at

least

of the questionnaires so

completed should be gone over with the
interviewer training staff).

the

in a constructive

and 4) supervised practice interviews of

volunteer

on

interviewer

by

the

�82
Each questionnaire completed by the interviewers must
be

c h e c k e d by a coder/editor for accuracy and completeness;

errors should be brought to the attention of the interviewer
and

the

interviewer

supervisor.

(If

the

number

of

interviewers in an examination center is more than three, an
interviewer
The

supervisor for that center should be selected.)

interviewer

retraining

as

supervisor

should

be

' responsible

for

necessary and for releasing interviewers who

repeatedly are unable to complete interviews accurately.
the

In

early stages particularly, and froia time to time during

later stages, the interviewing supervisor should sit in on a
randomly

selected

interview to observe the conduct of that

interview and the adherence to the expected protocol.
The interviewers nust be kept blind to the
status

of

the

exposure

individual subject as far as possible.

The

Questionnaire has been designed to help maintain blinding by
placing

the

exposure

questions

towards

questionnaire following the collection of
data.

the

end

outcome

variable

A final standardizing technique for the interviewers

should be to train one or more "dunny" subjects who
scheduled,

unknown

study subjects.
a

of the

predetermined

to

the

can

be

interviewer, for interviews as

These individuals should be trained to give
set

of

answers to be checked against the

recorded answers by the interviewer and in addition they can
observe the interviewer for aspects of the actual conduct of
the interview.

�The actual data in the questionnaire
have

several

questions

quality

are

itself

control procedures applied.

repeated

during

the

course

should
Severe.!

of

the

questionnaire and can be used to obtain a measure of subject
reliability.
data

Several questions in the questionnaire ask for

which is recorded in the personnel records (eg, dates,

units,

locations

questionnaire
for

served

in

in

SVK).

The

data

from

the

the personnel record files can be checked

consistency.

If

major

discrepancies

are

found

a

specially trained telephone interviewer should recontact the
subject to gather additional information which can

help

to

determine the source of the discrepancy.
Validation of medical data in the questionnaire will
rest

in

part

on the examination program.

report of shrapnel wounds of the leg
checking

results

of

validations
conditions

in

be

should

the
done

the

can

physical
of

be

validated

exam..

the

questionnaire.

For instance, a

Additional

reported
Each

medical

subject

will,

according to the proposed questionnaire, be asked to sign
release

of

tedical

medical

a

information form and will be asked for

the nane and address of the
reported

by

diagnosing

conditions.

physician

for

all

The diagnosing physician in

the case of all diagnoses of interest to the study should be
sent

a

diagnosis
diagnosis.

questionnaire
and,

of

asking

course,

for

for

the

details

verification

of
of

that
that

Each subject will also be asked for the name and

�address of his regular physician.
sent

a

general

questionnaire

conditions and

which

interest

the

to

questionnaire along
reported

which

includes
study.

with

This physician should

a

list

The

the

asks

about

of

use

specific

medical

conditions

of

be

this

of

general

questionnaires

to

diagnosing physicans would allow the estimation of

both under and over reporting of conditions.
e.

Coding and editing

The

individuals

doing

coding

and

editing

of

questionnaires, including the validating questionnaires fro:.:
physicians, should have the
for

recruitment

sane

standardizing

procedures

and training as for the record abstracters

in Section b above.

In addition a

random

sample

of

five

percent of the questionnaires should be receded and reedited
by a separate coder-editor

and any discrepancies

brought

to

the attention of the supervisor.
f.

Physical examination

The

physician

or

physicians

in

each

examining

center, to be assigned to this study, should be selected for
their interest and abilities in physical
physicans

oust

All selected physicians

should

be

a 5 day training course at the coordinating center or

the examination center for the pilot
course

These

be kept blind to the exposure status of the

subject being examined.
given

diagnosis.

must

emphasize

the

test.

necessity

This

training

for

carefully

standardized conduct of the physical exam and must review in

�85

detail

each

examination

(15) d e v e l o p e d

by

the

procedure.
Australian

The examiners manual
study

tear,

v a l u a b l e for developing this training program.
program should also include the independent
several

trained

"dummy"

subjects

so t-hat

should

be

The training

examination

of

inconsistencies

among physicians can be detected and discussed.
The actual conduct of the physical
be

standardized

by

exanination form.
all

normal

.parameters.

and

physical
editor

use

of

the

standard

This

procedures

requirement
are,

in

for

completeness.

Any

detected

dummy interview subject used to check

procedures.

followed.

incomplete

on the physical
attention

supervising physican from the coordinating center.

be

of

examination form must be reviewed by a coder

examination forms should be brought to the

also

physical

should help to
fact,

sections, apparent errors or inconsistencies

can

can

The form includes a specific checkoff

ensure that all exan
Each

the

examination

used

on

the

of

a

The same

interviewers

as a check on the physician examination

This individual can be moved fom site

to assure intersite consistency.

to

site

�g.

Laboratory tests

Laboratory
objective

the

most

outcome data to be gathered in this study.

To bo

of caxiraal v a l u e ,
standardized.

tests

represent

hov;ever,

All

between

of

Lust 1 ; bte

they

very

collection

those
and

not

affected

analysis

at

least

ensure

specimens collected foa
addition,

a

random

comparability

different

five

split

reliability

samples.
or

laboratories

for

examining

submittal,

In

the

form

of

This would allow the calculation of

repeatability

of

the

procedures.

for

validation

of

the

The

This validation should consist of
as

brought

study

into

laboratory
the

regular

subject specimens, to the study laboratories

of standard specimens analyzed at the reference
The

centers.

of

of the Centers for Disease Control, or another

study

procedures.

analysis

percent sample of all laboratory

standard laboratory as appropriate, should be
this

tine

This procedure

specimens should be sent to the laboratory in
blind

by

of the speciuer:,

should be performed at a central laboratory.
would

carefully

laboratory procedures which can be done

at a central laboratory, ie,
delays

spiie

laboratory.

results from the study laboratory would be conpared for

accuracy to the value from the reference

laboratory.

�87
For those procedures

which

must

be

done

in

the

laboratory of the examining center, the laDoratory personnel
should be carefully trained by a reference laboratory
standard

procedure

to be followed.

specimens of the randon 5
submitted

to

the

percent

examining

in

a

Wherever possible, the
saaple

center

should

laboratory

also
as

be

blind

A

sanples
standard

for

reliability

specimens

for

testing,
those

laboratory for validity testing.

along

tests

from

with
the

known

or

reference

�h.

Key entry

Key entry of data
ra:.,

based

be

done

using

on

site

output

(either

readable

study staff at a
data

entry

through

would

to

a

be

the

computer

terminals linked to a
computer

level.

entry
than

facility.

Uon

supervisor.

1$)

with

central computer) or by
Centralized

much easier to supervise.

100?

key

rates (ie, greater
require

by

central

should be done at a
referred

edit

This can be done by trained key entry

central computer, or onto a local compatible
portable

an

on a data management programming system such

as SAS, EAi:iS or SIR.
operators

should

following

Verifying

matches

should

be

Unacceptable error
retraining

replacement of the key entry operator.

should

Part of the

training of interviewers, coders and editors should

include

instruction and practice in numeral and character writing to
remove one uajor source of key entry error.
i.

Computer editing

Following key entry and verification, data
computer

edited

should

necessary

(key

be

for acceptable values of variables and for

cross-variable inconsistencies
problems

must

be

printed

operator,

subject) for resolution.

and incompatibilities.
out

coder,

and

referred

interviewer,

back

Such
as

examiner,

�89
,

i

10. Data Ilanageuent
The data niana^enient protocol for
size

and

project.
and

cost

is

a

a

study

of

this

crucial factor in the success of the

It includes study control,

data

capture

designs

documents, establishment of data bases, quality control

and preparation for analysis.
necessarily

parallel

The phases of data management

those

of the protocol as a whole and

are divided, here, for purposes of discussion as follows:
a.

exposure

.livelihood

index'

development

and

identification of exposure groups
b.

•identification

of

individuals

in

selected

exposure groups
c.

location of individuals

d.

recruitment, scheduling of examinations

e.

interview examination, follow-up advisories

f.

collection of validation uaterial

g.

analysis

h.

notification

of

assumed

exposure

status

and

study results
These are several general principles and assumptions
underlying

this

proposed

data management protocol.

assume that data management operations will be
central

computing

facility

equipped

with

main-frame computer such as the IBM 3033.

based
a

1) We
in

a

large scale

With hundreds

of

variables

on at least 12,000 subjects, data base management

and

analysis

data

would

be

impractical

on

a

smaller

�90

computer.

The

computer

should

be

appropriate peripheral equipuent and
capacity
entry.
system

to

interact

the

including;

sys.tem

with

the

with remote, non-hard-wired terminal

2) There will be
euployed

equipped

a

such

primary

as

RAKIS

data

base

II.

is

a

Princeton, NJ).

proprietary product of Mathenatica, Inc.

(RAMIS

management
II

A

description of features of RAKIS II from the 1980 version of
the Users Manual is included in Appendix K.
incorporating

these

features

could

be

Another

systea

substituted;

the

/following discussion of data management is predicated on the
use

of

a system equipped to handle hierarchical

maintenance of logical relationships among
elements,logging

of data transactions

access of non RAMIS files.

data

manipulated

little as possible prior to computerization.

5)

The

data

must

be

capable

different data sources.
be

capable

and

as

4) Self-coding

fully

as

possible.

management system must be capable of updating

and editing the data base files.
system

files

and the potential for

3) Data will be

data capture documents will be used as

files with

of

6)

The

data

management

of linking the various files fron
7) The data management system

must

limiting access to different files to those

individuals with a need to know.

�91

If any of these assumptions are not appropriate
the

facility

designated for the study, certain adjustments.

will then, be needed in the protocol in terzas
IBli

for

compatibility,

management programs

designing
or

adapting

of

and

writing

this

developing
data

protocol

base

to

other

existing systems, etc.
In the following discussion, sections a
the

use

through

f,

of the computer in assisting in data collection is

discussed.

Because

complications

of

involved

the

large

saiaple

size

and

the

in determining a, sample frame, this

computer use is advisable.
a.

Exposure

identification

of

Xikelihoodinde&gt;;

exposure

groups.

data for this activity already exist
form

in

the

HERBS

tape.

development

Much of the necessary
in

Additional

computer

readable

data on helicopter

spraying activities, ground spraying, herbicide storage
air

or

ground

accidents

involving

and

Agent

and

Orange must be

identified and computerized by location and date of exposure
to match the KERBS tape format.
After identification
areas

in

SVK,

companies

of

high

operating

tine-space areas should be identified.
these

companies

coordinates.

and
in

low
the

time-space
specified

Locational data from

should be. computerized with time and space

�92
The time-space files of exposure

and

company

data

fci-.e-id be- compared to assign day-to-day exposure likelihoods
for the companies,

(see Section III B 3 above).

Identification and inspection of the
records

is

documents

required

car.

management

be

before

information
information

so

We

data

recommend

Arny

capture

that

data

with the study observe the military

specialists

the needed

appropriate

developed.

specialists

existing

as

they

locate

the

needed

as to develop instruments that will capture

inforraation -

minimizing

both

hampering

the

abstractor or slowing the computerization process.
b.

Identification

^exposure groups.

of

individuals

in

selected

Company records, including morning reports

and operation report lessons learned,

etc.,

for

companies

selected will have to be reviewed for personnel acquisitions
and losses, both temporary and permanent.

Names and service

or

social security numbers should be computerized, along
'
'
\
with dates and character of listed
events
(arrival,
departure,

leave,

illness,

MIA

or KIA, etc.) relevant to

presence or absence in the area at a specific

tine.

These

data should be linked to the data file from step A above for
each individual and his exposure
presence

and

his

company's

calculated

daily

based

on

exposure index.

his
After

calculation of the index for each person, individuals should
be

ranked

by

exposure

likelihood

and

follow-up or exclusion as per study protocol.

identified

for

�ft basic computer file
followed

should

be

created

of

persops

selected

to

be

at this step, and used as the

base fil 51 throughout the study.

This

file

should

include

name, servici or social security number, tine in company anc
exposure

index.
A second fiTe

dates

should

also

for follow-up.

of

day-by-day

exposure

levels

and

be maintained on all persons identified

This file should

questionnaire

be

linked

obtained

on

later

exposure

to
so

the
that

comparison of assumed and perceived exposure can be made.
c.
in

B

Location of individual

above

should

be

The basic file created

used to identTf; individuals to be

located.

Tracing procedures are detailed above

III B 7.

The conputer data management

in

section

should be used

to direct the tracing procedures to be used, ;o

keep

track

of procedures which have been used, to record the outcome of
tracing

and

to

transfer

located

individu

-n?

into

the

appropriate files for follow-up:
1) Located f r e e - l i v i n g - r e c r u i t for
examination

inte~FvTev^ and

�94
2) Located in institution depending

on

nature

of

appropriate

institution,

(the

follow

computer

up
ccr,

identify these for hunan decision)
3) Located,

deceased

-

follow

up

for

av&amp;ilat&gt;lfc

medical records.
All locational
loss

to

follow

findings,

including

acceptance

of

up should be transferred back to the basic

file (E) as provisional endpoints (located,

living)

or

as

codified

as

final endpoints (deceased, unable to fo,llow).
These subfiles
information

is

can

received

be

utilized

from

and

further

procedures

and

activities.
Information for the location
gathered

as

The St.

described

in

procedures

should

be

the training section (III B 7).

Louis records will provide the

major

quantity

of

this information which should be abstracted onto self-coding
forms.
of

The design of these must be based on both the format

the

files

files in the St.

and

the customary procedures in using those
Louis

center.

(neither

part

of

this

i

necessary

information

is available to us at this writing.)

The information to be abstracted is specified in section III
B 7.

�•These dgjta should be
file-

created in E.

and direct
strategy

the

underway should

appropriate
above.

be

to

the

computer,

The computer can then scan the material

next

-outlined

transferred

step

Regular

available,

as

in

the

updates
should

searching

of

searches

information

on

levels of success at each step, number of $teps necessary to
Iocs to, etc.
The outcome of each step for each individual should
.1
be entered into the computer record as it becomes available.
Individuals successfully traced, should

be

entered

into the appropriate follow-up sub-file.
d.

Recruitment &gt;

interviews.

The

follow-up

scheduling ... o.f

examinations,

sub-files developed in C above

i

should be used to direct recruitment

and

closest

be identified for each

examining

facility

individual,

and

that

site

recruitment

and

scheduling.

should
notified
The

scheduling.

to

The

proceed

with

file on each individual

should be kept open and the site queried on a regular

basis

until recruitment is completed and the exanination scheduled
for compliant individuals.
noted

in

the

Refusal

or

deaths

should

be

recruitment file and transferred back to the

basic file as an endpoint achieved.

�96

Following scheduling of examination, sites should be
queried

regularly,

using

a

system

preparation system of RAKIS II,
completed

until

like
the

the

reports

examination

and data received at the central facility.

information

is

transferred

to

completion

received,
the

next

transferred

the

individual

file,

back

with

to

the

basic

After

should

information
file

is

be

on the
as

a

provisional endpoint.
e

The

*

Interview,

interview

and

examination,

insofar

advisories.

examination schedules are described and

included in section III B 5.
self-coding

follow-up

as

The forms are designed

possible.

to

be

Information from these

forms can be key entered centrally or through remote on-site
terminals,

utilizing

System of RAHIS II.

a

system

In the pilot

like the Record Management
study

other

methods

of

data entry such as key punching or optical scanning could be
investigated.
terminals

In the case of key entry, the

with

self-prompting

should also be investigated.
from

cards,

separately
systeu
verified

direct

The

entry

of

smart

and editing routines

choice

of

entry

system

disk entry or tape should be evaluated

for each major source of data.

should

use

monitor

intake;

and

This

management

entry should be 100%
, \
(or reentered) to assure a minimum error level.

�Programs nust be developed and linked
canageiient

system

to

identify

t o " , the

individual

responses

findings or patterns of responses or findings which
follow-up.

These

prograns

collaboration with clinician

should
members

be
of

cs.t,o

require

developed

the

or

in

coordinating

center at the start of the study so that they are responsive
to the group being examined, VA resources and
medical

current

practice.
f.

that

good

Collect ion o f v a1 id a t i o n ra a t e r i a 1.

follow-up

is

In the event

required for reported health events not

verified in the examination procedures or for

ascertainment

of time of initial diagnosis of a disease entity, individual
records can be identified for such follow-up.
follow-up

should

query

the

of

follow-up

is

be

computer

examination

assisted.

center

The need

for

Again, periodic

should

be

made

until

either completed or declared to be impossible

to complete.
The initial data capture fora for
should

be

these

a letter to the physician or hospital requesting

specifics on the particular event and/or date
As

the

of

interest.

study progresses, connon areas of inquiry or comnon

diseases nay
^

follow-ups

\,

developed.

emerge;

specific

data

forms

can

then

be

�98

Analysis.

A cajor function of

successful

data

nanajeiuent is to preserve anfi. present the data for analysis.
The fully linked hierarchical files
each

individual,

fron

his

will

be

likelihood

validation of his childhood asthma.

available on

of

exposure

Clearly,

this

is

to
too

long, too complex, and too detailed a file for most analytic
evaluation.

The data management system

selection

of

variables,

and/or

characteristics

subsets
of

for

of

should

variables,

of

permit

values within

individuals according
use

the

to

in specific analyses.

record

These data

subsets can be created as needed according to specifications
for

individual

analyses.

directly accessible by a

These

variety

data
of

subsets

statistical

Dust

be

packages

such as BMDP, SAS or SPSS.
h. Notification
study results.

of

assumed

exposure

and

As noted in the section on notification (III

B 1*0, there is an obligation to advise the
the

status

participant of

results of the study and what they might mean to him as

an individual.

Part of this notification would include

likelihood

exposure, and part might depend on responses

of

and findings in the interview and examination.
based

data

The computer

management system should allow determination of

an individual's report content.
ordered

his

Personalized reports can be

for all participants, or for those determined to be

at some designation of "high risk".

�99
11.

Data analysis

Analysis should be carried out in a series of
designed

1)

steps

to identify the presence of missing or suspect

values, 2) to evaluate the reporting bias present due to the
participant's perception of his Agent Orange exposure, 3) to
investi£ate which statistical methods are
for

analysis,

likelihood
cohorts,

4)

of

appropriate

to confirm the comparability of the high

exposure

and

most

finally

and
5)

low

likelihood

of

exposure

to determine whether there exist

diseases or indicators of bad health

associated

with

high

exposure to'Agent Orange.
s.

Quality of Data and Management... of

Missing

and

.Unreasonable Values
The first step of analysis should be to assure
the

data

gathered

are

of

the

that

highest quality possible.

Errors can occur at any of the steps between record location
and

putting

the

data into machine readable form.

techniques that are

easy

detect these errors.
consistency.
interview

For

are

available

example,

additional

internal

consistency

reported

diseases

information

to

The data must be screened for internal

examinations

provide

physician.

implement

several

questions

schedule require consistent answers.

of laboratory
should

to

Several

and

obtained

of

and

the

data

the

The results

physical

opportunities

on

examination

for checking the

collection.

In

addition,

conditions can be compared with the
from

the

attending

or

current

�100

The data must be scree.ned
values

for

missing

values

and

w h i c h are either lower or higher than is reasonable.

Whenever possible, these values should be replaced with data
fron

one

of

the

other

data

sources (entry, service and

discharge records;

vital statistics;

medical

physical examination;

history;

interview

schedule;

and/or laboratory

test results) .
Information

on

the

sane

factor

obtained

from

multiple sources should be compared.

Inconsistencies should

be

the

reviewed

without

knowledge

of

status

of

the

individual and a technique established whenever possible for
resolving inconsistent information.
b.

Relationship of Participant Reported Exposure to

Agent Orange and Health Outcomes
A major concern in this study
will

is

that

individuals

know their exposure category and, therefore, will give

biased responses (either consciously or unconsciously) based
on

that

suspected

exposure.

There

are

individuals for whon health outcomes can be

four

groups of

compared

which

raay provide an estimate of the amount of bias resulting from
suspected exposure to Agent Orange.

The high likelihood

of

exposure cohort will contain individuals who report exposure
to Agent Orange and individuals who do not
to

Agent

Orange.

report

exposure

Likewise, the low likelihood of exposure

group will also contain individuals who

report

or

do

not

�101

report

exposure

compared

(as

Comparison

to

Agent

described

Orange.

below)

Health outcoues can be

for

these

exposure

estimate

groups.

of findings between those reporting exposure but

not verified to have exposure with those
had

four

of

but
the

not

verified

to

have

reporting exposure would provide an

amount

of

bias

present

due

to

the

participants perception of his exposure status.
c.

Statistical Methods A p p l i c a b l e to This Stuoy

Univariate
packaged
These

descriptive

statistical

plots,

SAS

and

SPSS

can be applied to all data.
simple

histogram

univariate

Commonly available packages such

(measured

as

can be used for these descriptions as

statistics.
and

plots,

and common summary measurements for the

well as for later analyses suggested by the results
descriptive

major

include

variables of interest.
BMDP,

using

routines

descriptions

box-whisker

statistics

multivariate

These

should

techniques.

of

the

include

both

Discrete

data

on nominal and ordinal scales) can be analyzed by

simple chi-square analysis and by log-linear

analysis.

For

interval (continuous) data, logistic regression analyses can
be used with a dichotomous outcome.
the

The appropriateness

of

logistic model can be examined by using goodness of fit

criteria.

�102

d.

Conpa ra bi11ty of Cohorts

The validity of
dependent

in

the

historical

cohort

cohorts

is

large part on the assumption of cocparability

of the two cohorts with different exposures.
two

design

will

hopefully

Although

the

be chosen in a manner to make

then as similar as possible on all inportant measures except
exposure,

it

is

extremely

inportant

to

differences that night affect the apparent

look

for

relationship

any
of

exposure to health outcomes.
A first step should be to compare the response rates
between the two cohorts.

Sone information will be available

from

and

entry,

service

non-respondents.

Using

discharge

information

,

comparisons between respondents and

records

froc
t

.these

non-respondents

for
sources
within

each cohort can be made.
Demographic

characteristics

such

as

years

of

education, father's income, and characteristics of childhood
residence

should

carefully
studied to see if
any
/
between the cohorts emerge. The two cohorts

differences

be

should also be compared for the prevalence of diseases or
I
conditions present before entry into the military service.
Other characteristics which should be
prevalence
hazardous
Vietnam,

of

include

the

familial diseases or conditions, exposure to

materials
and

studied

type

(including
of

herbicides)'

outside

of

military service including years of

service and distribution of ranks.

�103

Echcvior&amp;l differences car. also

be

examined.

For

c;:&amp;:.:;-ie,

J

th* cohorts ccr. be eccptrei with respect to health
••&amp; i
h£tits such as seeking, drinl:ir.£, use of sirijuer.a, end the
practice

of

the

seven

health h&amp;bits associated with gooo

health status &amp;nd longevity (as outlined by
if

Ereslow)

(16),

the data on these htbits prior to Victr.an service cen be

re-li&amp;bly obtained.
Statistical techniques
incluco

log-line&amp;r

or

can

Or£r.£;c.
the

be

these

conparisor.s

can

analysis and logistic regression^ which

c&amp;r. point out pre-existing
with

for

used

variables

that

are

associated

as predictors for exposure to Agent

For exanple, usir.^ exposure versus non-exposure

outcone

variable,

a

logistic

as

regression can hslp to

determine if any differences exist between exposure to Ajent
£.r^;c and dessogrephie characteristics at enlistment.

E xpo u r e

The
should
Further

clao

behavioral
be

discussed

above

exaained in the post-exposure lino period.

analyses

exposure

and

adjusted

according

of

health

characteristics.

characteristics

association

between

Ai;ent

Orange

outcomes nay need to be stratified or
to

cohort

differences

in

thcc«

�104

A number of health outcomes can be cx£.n;i ncd.

Tr.csc

include the presence of specific diseases or concitions, the
results of psychologic testing, the
tests,

the

perception

occupational status.
either

adjusted

of

results

general

of

laboratory

health zr;C the current

The search for health outcomes can

be

for or stratified on any differences found

between the two cohorts.
The

screening

relationship

procedure

for

investigating

of specific diseases or health conditions with

exposure to Agent Orange should include a screening
health

outcomes

solicited.

Special

emphasis

of

conditions

reported

to

confounding

be

diseases

in the literature to be associated

with Acent Orange exposure.
given

all

should

placed on looking for differences in prevalence of
or

the

Careful consideration should be

factors

which

night

decrease

likelihood of obtaining statistical significance for s.

the
true

difference.
One set of the statistical
discrete

analyses

use

a

outcome variable (presence or absence of disease).

Here either a logistic regression or a
can

should

log-linear

analysis

be used to determine relationships between exposure and

disease.

When some or all of the independent variables

continuous,

are

the logistic regression should be the method of

choice.

When all variables are categorical either technique

can

used,

be

although

the

log-linear

appropriate for studying general

relations

anlaysis
among

is nore
all

the

�105
variables,

Any

factors found to be important in the study

of the conparability of the two cohorts can be
the

equations.

stepwise

The

fashion.

variables

are

method

strongly

keeps the set of variables
While

life

should

point

associated

used

as

out

which

with disease and

concise

as

possible.

table methods are commonly used in longitudinal

studies, they were designed for,
situations

intc

others can be entered or deleted in a
This

most

forced

in

which

appropriate

to,

outcome variable is common.

the

and

a,re

The

majority of outcome variables in this study are too rare for
meaningful

use

of

life

interest in some fairly
divorce

table

techniques^
i
outcome such

common

post-discharge,

for

example,
&lt;

If

there is

as

rate

of

life table analyses
i

could be useful.
Much of the laboratory data will provide
variables
cohorts.

which

can be studied for differences between the

For these

procedures

variables

many

classical

statistical

are available including simple t-tests and their

non-paranetric analogues as well as multivariate
studying

syndromes

rather

techniques that can
among

continuous

variables

be

used

include

components analysis.

than

for

single

be

outcomes.

study

to

factor

These can

tests

interrelationships

analysis
used

to

and

Other

principal

describe

relationships and to avoid multicollinearity problems.

the

�106

Comparison of laboratory test results
changes

early

nay

expressed

which

have

clinically d e f i n a b l e disease.
to

not

yet

been

indicate

Parameters which can be

as

used

estimate differences in the mental well being of the two

cohorts

include

psychological

comparisons

scales,

the

of

the

results

of

scores
the

on

the

perception of

general health questionnaire and comparison of

the

current

occupational status of members of the two cohorts.
i
Significant results found in any of the preceedin£
'i
analyses
should
be checked by closely exarnininc the
univari&amp;te results and by
exists
the

in

data

determining

subgroups of the samples.
should

include

whether

the

effect

Final presentation of

explanations

understandable

at

different levels of statistical sophistication,
f.

Estimated Sensitivity of the Study

Finally,

the

level

differences should be noted.

of

probability

the

probability

of

being

reach

statistical

able

to

differences given the prevalence of the disease or
and,

the

sample

size

of

the

population

consideration age, etc.) should be made.
levels

of

significance

observed

For those conditions for which

a difference was noted but which did not
significance

of

All

find

a

condition

(taking
reporting

into
of

should take into consideration the

effect of multiple comparisons on the same populations.

�12.

Sample size

Although the actual available number of subjects for
the two exposure cohorts cannot be determined until the work
on the exposure
sample

size

likelihood index

4

be

completed,

the

requirements for various possible outcomes can

be estimated and an optimal
and

can

have

been

sample size chosen.

Figures

3

drawn to indicate needed sample sizes to

distinguish between two proportions Pj(for low exposure) and
&gt;
P2(for high exposure)with a fixed risk ratio of ?2/?is2 or 3
respectively.

(That is, the outcome is twice or three times

as common, respectively in one exposure likelihood group than
in the other.) These ratios were chosen to be representative
of

the

lower

levels

of

risk

which

might

therefore, to represent the most

conservative

estimates.

given

We

feel

that,

the

be seen, and
sample

size

potential

for

cisclassification in this study, risk ratios of less than
should

not

be

considered.

Figures 1 and 2 also utilize a

fixed alpha of 0.01 (probability of a type I
one-sided test of Ho:P 1 =P 2 with njsn2=n.

error)

and

exposure

group

a

type II error.
example,

gives

the

While the vertical axis represents the

sample size needed for each group.
represents

a

The horizontal axis

of each graph represents the PjOr incidence expected in
low

2

Each line in the

graphs

different level of beta, the probability of a
In other words, each line of Figure

3»

for

the needed sample size for a given beta for

an alpha = 0.01 and a risk ratio of 2.
sample sizes for a risk ratio of 3.

Figure 4 gives

the

�108

We

chose

the

alpha

level

of

0.01

potential expense of the study and because
of

the

questions

certainty before
sample

size

was

to

because
the

of the

seriousness

be answered dictate a high degree of

results

are

declared

significant.

The

computed using the arc-sin transformation

for variance stabilization.
„ * 2 &lt; z l-g * 2l-a'2
(2 arcsinyFT - 2 arcsin/FI)
Lemeshow, Hosmer and Stewart (17) have shown that the above
sample

size

formula

provides a close approximation to the

exact results for very
known

to

underestimate

small

proportions

although

a

is

n somewhat for larger proportions.

The sample sizes given in Figures 3 and 4 are
for

it

those

needed

test of equality of proportions when a simple random

sample is taken from each group.

For Pj = 0.1 the necessary

sample size to detect a doubling in the risk (?2/?. = 2) with
95% power is less than 400 in each group (for tripling, less
than 120).

Hence, the plots do not extend beyond 0.1 on the

horizontal axis.

�109

Using

the

incidence of

an

outcome

variable

of

i

interest

one

can

read froc the g^aph the necessary sacple

size for &amp;. desired level of beta.
interested

For instance, if one were

in a comparison of total mortality, which in the

general population represents a level of somewhat less
1J,
a

than

to detect a doubling (P /P « 2) of total mortality with
beta

of

0.05

(or

a

power

of

95$.)

one

approximately 1,500 subjects in each group.

would

need

If on the other

hand one were interested in a cancer which might occur at

a

level of about one in a hundred thousand (P?* 10 ;, a sample
size of about 1.5 million per group
detect a tripling (^/P,* 3) of
The examples above are
yearly

incidence.

In

tne

would

be

required

rate with 95? power.

based

on

a

comparison

a

period.

group of initially 21 year old men followed for

about 12 years, one would expect (from U.S.
cumulative

mortality

of

about 2%.

life tables)

a

From Figure 3, then, a

sample size of about 2,300 would be needed in each group
detect

of

a follow-up study such as this, one

can use the cumulative incidence over the follow-up
Assuming

to

to

a

doubling in risk at the specified levels of alpha

and beta.

Likewise in this follow-up period the cancer with

a yearly incidence of one in a hundred thousand would have a
cumulative incidence of a little over one in ten thousand
-4
(Pj * 10
). To detect a tripling of this incidence at the
specified alpha and beta, one would
about

150,000

veterans per cohort.

need

(from

Figure

4)

As a final example, if

�110

one were interested in a disease with a yearly incidence

of

about one in one thousand, the expected cumulative incidence
would be approximately

one

in

a

hundred

sample size of about 4,500 per group would

(F ^ =10

-2

A

required

to

plotted

be

).

on

detect a doubling in the rate.
As can be seen,
log-log

paper,

the

when

lines

vertical distance between
P.&lt;0.01.

If

other

the

are
them)

graphs

are

"parallel"
and

(have constant

almost

straight

for

lines are desired, they can be plotted

very simply after calculation of 3 or

4

points

using

the

formula given above.
The graphs can also be used to estimate
that

the

effect

practical constraints on sample size would have on the

power of the study.

Using again the example of total yearly

mortality, if the available sample size were 2,900 per group
instead of 4,500, the power would decrease from 95$ to about
80%

(the beta would increase from 0.05 to 0.20).

in the case of the cancer the
500,000

per

group

sample

size

Again, if

available

were

instead of 1.5 million, the power would

drop from 95% to 50$.
The effect on power of
smaller

sample

size

the following table.

being

forced

to

accept

a

than recommended can be summarized in

�Ill
TABLE 4

Available
Sample Size as a
% of recommended

Pow?r

100$

9.5Sb
t

1

95%

93.9

90

92.5

85

90.9

80

69.6

75

86.7

70

84.0

65

80.9
f,

60

77.3

50

68.5

no

57.3
'

35

50.9

This table holds for either risk ratios of 2

or

3

and

is

that

is

derived directly from the arcsin formula:

Z

l-a

where f is the fraction of the recommended
actually available and y

is the

sample

resulting power.

�n:
These sanple size estimates do not take into account
non-response
and
diminution
of
respondents
due
• '
'
,
insufficient .data, errors in collection, etc. They also
not

take

into

account

the

to
do

use of multivariate analyses.

Therefore, they should probably be increased by at least 20$
to

ensure

that

the final number of usable participants is

sufficient.
Given these sample size considerations, we recommend
a

sample

size

of

cohort size would
cohort,

to

not

be

less

than 6,000 per cohort.

sufficient,

detect at

losses

from

recommended

the

after

levels

of

This
the
alpha

and beta a doubling of the risk of a disease with an expected
yearly incidence of one in one thousand in the
cohort.
detect

With 6,000 men per
a

tripling

(approximately
age

20-29)

in

30/100,000
even

with

group
the

exposure

it should be possible to

total

yearly

low

cancer

incidence,

incidence
all

some non-ascertainment.

males,

�1121

Figure 3. Sample size needed per group, by frequency of outcome, to detect
doubling of risk.

Based on a • .01, ?2/Pi -2,
H t f P 2 - ^ , H A :P 2 &gt;P,.
Each line represents a level of /3
(1-/3 is the power of the study).

�H2ii

Figure 4. Sample size needed per group, by frequency of outcome, to detect
tripling of risk.
10

Based on a • .01, P2/Pi-3,
H0:P2 -P,,H A :P 2 &gt;P,.
Each line represents a level of £
(1-/3 is the power of the study).

�113

13«

Organization
i

'

Exact specification of the organizational
of

the

study

requires

a decision on whether the Veterans
t
i

Administration or soiae independent contractor
the

study.

structure

will

conduct

The major factors in this decision appear to be
i

greater efficiency and lower cost for a VA run study on
one

the

hand, and serious questions about veteran participation

and belief in the ultimate results, if the VA

conducts

the

*

study, on the other hand.
a

vested

interest

important

data,

in

The VA has been accused of having
a

negative

study

putcome.

Some

particularly on veteran participation, can

be gathered in the pilot test of the protocol.

We

believe

that the VA hospitals can be utilized as .examination centers
with an independent coordinating center responsible for
conduct

of

the study and the data analysis.

the

We outline in

this section an organizational structure for this

approach.

The proposed structure would incorporate the cost savings of
a VA study with
independently

the
run

generally
study.

If,

higher
after

credibility
the

of

an

pilot test, it

appears that the veterans will not accept examinations at VA
hospitals,

the

examination

centers

will

have

to

be

contracted.

Probably the best approach would be to identify

well established and recognized non-university clinics which
could perform the examinations.
the

recognized

expertise

and

These
may

centers
be

more

would

have

likely

to

participate in a project in which they would have little

or

�114

no

opportunity

for

publications than

would a university

center.
For purposes of cost reduction, advantage should

be

taken of the large number of regionally located VA hospitals
for selection of examination centers.
chosen

The centers should be

• •

*!*

to

ensure

a

geographic

distribution such that no

v e t e r a n is required to travel more than H hours to
center.

In

should

addition,

also

be

hospitals.

as

feel

medical

because

of

center

affiliated

that the Air Force approach of having

only one examination center would be
study

a

many of the centers as possible

university

We

reach

impractical

for

this

the expected much larger sample size and

the resultant enormous transportation and housing costs.
In each hospital, a minimum staff must be designated
to perform all study examination procedures.
'•s

•

.

'

•

a dedicated staff

will

ensure

examinations

and

data

collection.

include

following:

the

consultant,

a

nurse,

a
one

capable of performing those
facility

and

-

better

necessary.
etc.

a

neurologist

more laboratory technicians

tests

done

a trained interviewer.

additional

of

The minimum staff must

at

the

examining

Depending on the size

of the expected veteran population to be
center,

.

standardization

physician,
or

Designation of

examined

at

each

staff can be designated to the study as

Back-up staff in

case

of

illness,

should also be designated and trained.

vacations,

�115

Subject participation in any sucfo study is
by

ensuring

pleasant

surroundings

and

Therefore, in each center a special
hired

or

dedicated

totally

examination

coordinator.

responsible

for

are promptly
procedures

are

pronpt attention.

staff
the

This

greeting

examined

to

individual

and

ensuring

that

a

manner.

No veteran should

be

required

portion

of

exanination

additional

staff

be

be

and

to

also

be

laboratory

wait

for

any

responsible

forwarding

member

all

coordinated and efficient

headquarters all data collection forns.
an

should

for more than 15 minutes.

snaller centers, this person can
checking

must

the veterans, ensuring that they

in

collecting,

member

study to serve as an

performed

the

enhanced

would

to
In

be

the

larger

required

In
for

study
centers
for this

function.
A
contract

coordinating
to

a

well-respected

proven epideraiolgic
staff.

center

expertise

should

be

established

research' organization with
included

in

the

permanent

This group must have total authority over all study

personnel including those in the examination
coordinating

group

should

be

headed

by

epideuiologist and an experienced manager.
this group should also contain:

centers.

and

an

individual

standardization procedures.

The

an experienced
At

a

minimum,

a statistician in charge of

data management, an individual in charge of veteran
efforts,

by

in

charge

of

tracing

examination

Each of these individuals

will

�require a support staff.

The size of the support staff will

d e p e n d or: the ultimate sample size
such

and

other

intangibles,

as the- difficulty encountered in tracing.

factors should be

identified

in

the

pilot

These other
test

of

the

responsible

for

protocol.
The coordinating center should
all

subject

be

identification and tracing.

They can delegate

to the examining centers the actual contact
of

the individual subjects.

also be responsible for:
the

examining

and

The coordinating center should

1) ensuring proper coordination of

centers, 2) ensuring standardization of data

collection procedures, 3) collection, coding,
management

scheduling

of

study

data,

*l)

cleaning

and

analysis of study data, 5)

notification of centers if abnormal examination findings are
discovered,

and

final

notification

of

veterans

of

appropriate study results.
An independent scientific overview
also

add

to

the

coranittee

eventual credibility of the study.

committee could also oversee the conduct of the pilot
The

membership

include:
eminent

of

this

a

physician

questionnaire
should

neet

a

data
design
on

management
and

test.

one

specialist in physical

examination procedures, a specialist in clinical
procedures,

This

committee, if established, should

at least one eminent epidemiologist, at least
statistician,

might

expert,

administration.

and

laboratory
an expert in

The

committee

a regular basis with the senior members of

�117

the coordinating center to review
Such

meetings

should

be

progress

of

the

held quarterly at a mininiuD, ar.d

probably more often at the

beginning

of

operating

oversight

coianittee

budget

of

the

the

study.

sufficient to allow nenbers to travel as needed
examining

centers

or

the

properly

oversight
serve

its

The

should
to

be

various

to the coordinating center to rev lev;

study procedures in progress.
that

study.

cornnittee
function

compliance with the protocol.

Care must be taken to
has
of

ensure

sufficient

funding to

independently

assessing

�118

14.

Notification of subjects

We

believe

initially

that

individual

subjects

should

bv.

unaware of their exposure status according to the

study definition and of the
because

possible

anticipated

outcomes

of the possibility of conscious or unconscious bias
i

on the part of the participants and/or the interviewers.
addition,

the implications of possible exposure will not be

understood
analysis

In

until
of

the

study

.completion

results.

We

of

the study and -the
\
believe, therefore, that

notification of the participants of their

assumed

exposure

status should be reserved until the completion of the study.
At that time the individuals should
exposure

status

be

notified

of

their

and also be given information on the study

results and the possible implications of their exposure.
The more immediate question is that
illness

or

abnormality

of

significant

which may be discovered during the

examination period and which is unknown

to

the

individual

',

(i.e., is unreported in the medical history).

The examining

ph.ysican should be responsible for notifying the participant
of

any

abnormality

found

at the tine of the examination.

This physician should also be responsible for
laboratory
notifying

results
the

perforaed

participants

appropriate follow-up.

at
as

the
needed

checking

all

examining hospital,
and

encouraging

�119

As outlined in the section on data
/ data

fron

each

examination

(including

;

performed
directly

laboratory
to

analyses

computer

within

findings.

which

one

and

should

month

checked

for

then

missed

reported

accuracy

be

checked

then

be

to the individual examining center.

or

abnormalities

laboratory analyses.

and

for

Any identified abnormalities which were

allow the center to follow-up any abnormalities
have

be

of the tine of examination.

unreported in the medical history should
inmediately

centrally

*

The data fron each individual should
abnoraal

the

the

the coordinating center) should be keypunched,

entered into the
consistency

management,

reported

This would
they

might

found in the centrally done

The coordinating center responsibility

for notification of the examination centers should rest with
a designated physician who will review the
individual

with

examining centers
direct

contact

an
will

identified
have

the

records

abnormality.
responsibility

of

any

While

the

for

the

and follow-up with individual subjects, the

coordinating center should be responsible for overseeing the
notification to ensure that it is properly accomplished.

�120
15.

Pilot testing of protocol

A study of the size, complexity and probable cost of
this

proposed

study

pilot testing.
test

the

should

not

be done without adequate

Pilot testing has several purposes:

overall

feasibility

of

individual components of the study;
the

study

data to

procedures

be

and

c)

study

and

to

of the

b) to revise and refine

so as to maximize the quality of the

collected

operations;

the

a)

and
to

the

efficiency

of

the

study

estimate the costs of the overall

study operation and of individual study components including
identification

of compromises between cost and data quality

which

necessary.

may

be

selection
should

of

only

Pilot

testing

should

include

high and low exposure cohorts and, therefore,
be

conducted

following

the

satifactory

development of an exposure likelihood index.
The pilot test should include testing of
and
done

procedures
on

likelihood

a

necessary for the final study.

sample
index.

of

men

We

recommend

likelihood of exposure

identified

group

likelihood

of

exposure

not

on

and

considerations

based

measures

since

outcome

of

adequate

of

participation

rates

and

This can be

the

exposure

200

men

from

the high

200

men

from

the

low

This number is

comparison

of

outcome

measures should not be compared in
The

comparison

forms

in

group be examined.

groups in this pilot test.
for

all

is

sample

size,

differences
large

enough

in
to

however*

is

tracing

or

ensure

a

�reasonable; likelihood

n i g h t , be erfcb~un t er '

of

identifying major problems which

eg, any problem

affecting

more

than

randomly

frora

a b o u t 55? of t h e s u b j e c i population.
The 400 individuaTa should be chosen
the\ potential

subjects

in

(200i per group!) and should
to

phen

be

traced

and

contacted

protocol proceedurela given in sections III B 7

and III I 8.

variety

each exposure likelihood group

We recommend the simuTTal

application of a

of the most promising tracing

ra, including

a search for fanily members, particularly parents,
of

the

VeterSTT3 Administration, records and the IRS recoras7.

such siculta'nec'U-a application
pilot

test

L"

subjecftrsi

and least cost

can

of

tracing

methods

to

use
By
all

the procedures with the highest yield
&gt;e

identified

and

the

best

tracing

protocol can be establishe'cll.
he subjects shoulc be
center

invited

to

an

exaninatior.

o undergo all exanrinaliui: procedures as detailed in

&gt;ection :il B 5.

A single exaninaTTol center .with excellent
should

be

elected.

The entire program of stafi training can then

be

eveloped

and

and

ranspor

nearby

pilot

tested

housing

as

facilities

well)

While the use of a

ingle examination^center would have hignr transportation and
i
ousing

costs,

11

ould provide the besfi

concentrated effort on a single center
test

of

study

&gt;rocedures

having to deal with inte~rcente^ variabilityv

without

�122
The data should be recorded on the

developed

draft

i

study

forms

procedures
schemes

ancJ

coded

recoanended

and

computer

in

and

managed

that

according

section.

Final

of

the

decided upon.
screening,

responses
The data

search

appropriate
procedure,

for

univariate
the

are

or

coding

range

and

known and the final content

analysis

should

outliers

and

analysis.

consist

missing
The

of

data

values

and

of

each

cost

frequency of non-traceable individuals, the

frequency of nonparticipation and the frequency
data

the

entry formatting should be completed

after exacination of the collected data when the
nature

to

nonreporting

for

each

study

of

missing

variable should be

estimated.
Upon completion of this pilot phase,
possible

to

to

refine

study

procedures

to

of

should

full study.

be

further

value

account for problems

identified in the conduct of the pilot study.
it

should

refine the protocol to eliminate collection of

data which occurs too infrequently to be
and

it

In

addition,

be possible to estimate the ultimate cost of the

�123
We recomuend that the overall results of
study
the

the

pilot

be reviewed to decide on the merit of c o n t i n u i n g w i t h
full

study.

untraceable

In

particular if

individuals

is

greater

the

proportion

thtn

20';

in

either

cohort, or if there is a difference of raore than 10$ in
proportion

of

the

traced in the two cohorts, we recoaiaend that the

full study not be done.

Likewise if the rate of refusal

to

participate among veterans is more than 20£ of those traced,
or there is a differential in refusal rate of
10*

greater

between the two cohorts, the study would be unlikely to

be successful.
Overall,

a

These criteria can be applied independently.

combined

non-traceability

greater than 30? or a differential
greater

and refusal rate of

between

the

groups

for

instance,

criteria

for

not

proceeding

'esources.

be

considered

with the full study unless a

solution to this problen can be identified.
cost

spread

less than the difference between 70? in

one group and &amp;5% in the other group) should

the

of

than 15? (ie, it should be possible to collect data

on at least 70% of the members of the cohorts with a
of,

than

Decisions about

of the study will depend upon the availability of

�124

16. Timetable
After, acceptance of the protocol, selection

of

the

coordinating cente?, hiring of staff and receipt of security
clearances, we estimate the timetable as follows:
a) development of the exposure likelihood index - 12
months
b) selection of cohorts - 6 months
c) pilot test

\t

planning (concurrent with b)
conduct - 3 months
analysis and revision of protocol - 6 months
d) full protocol
selection of examining centers (concurrent with c)
hiring and training staff - 3 months
tracing and recruiting subjects
and conduct of exams - 24 months
analysis - 12 months
This
acceptance
analysis.

represents
of

the

a

protocol

to

of

5

completion

is expected to take 3 years.

will be

1/2

years
of

It should be noted that the conduct of

protocol

needed

including

to

complete

developing

the

selecting the cohorts for the
staff,

total

from

the final
the

full

The rest of the time

necessary

preliminary

steps

exposure

likelihood

index,

study,

hiring

and

training

pilot testing the study and revising the protocol as

necessary.

�125

17.

Future follow-up of subjects

V.'c. have not] d e s i g n e c a

•
fcTTtfvi-u up

future

pt»O£o.sed s t u d y ,

warrant

who

of

shoul

regular

participate

for
in

the
this
is

coctpletion of tnTO p r o p o s e d s t u d y and
underway.
for

follow-

could be done fairly

At a miniauri we r e c o m m e n d

mortality be conducted

inexpenslVeT5,
of

accurately

1

although
from

exclusions

yearly query of th« National Death Index.
information

protocol

further follow-up

d e c i s i o n on v

other stu
that

specific

could

be

gained

by

totally

not

systeq, by a

Much

additTona

also obtaining the death

certificates so thalp the distribution?
could be examined.

the

This

of

causes

of

deat$

�126

REFERENCES

1. National Acadercy of Sciences: Report of the Committee on the
Effects of Herbicides in South Vietnam, Part A - Summary and
Conclusions, Nat'l Acad. of Sciences, Washington, D.C., 1974.
2. Young, A.L., et al. The Toxicology, Environmental Fate and
Human Race of Herbicide Orange and Its Associated Dioxin.
USAF, OEHL Technical Report TR-78-92, 1978.
3. Veterans Administration: Review of Literature on Herbicides,
Including Phenoxy Herbicides and Associated Dioxins, Volumes
I and II, JRB Associates, 1980
4. Interviewer's Manual, revised edition, Survey Research Center
Institute for Social Research, University of Michigan, 1976.
5. Field Interviewing Manual, Public Opinion Center, Dayton,
Ohio, 1972.
6. U.S. Department of Commerce, Indexes to Survey Methodology
Literature, April 1974.
7. U.S. Department of Health, Education and Welfare, Health
Interview Survey Procedure, 1957-1974, Vital and Health Statistics
1(11), April, 1975.
8. U.S. Department of Health, Education and Welfare, A Summary
of Studies of Interviewing Methodology, Vital and Health
Statistics 2 6 ) March 1977.
(9,
9. American Marketing Association, Special Issue: Recent Developments
in Survey Research, Journal of Marketing Research, August, 1977.
10. U.S. Department of Health, Education and Welfare, Experiments in
Interviewing Techniques: Field Experiments in Health Reporting,
1971-1977, NCHSR Research Report Series, November 1977.
11. Belloc, N.B. and Breslow, L., Relationship of Physical Health
Status and Health Practices, Preventive Medicine 1:409-421, 1972.
12. Breslow, L. and Enstrom, J.E,, Persistence of Health Habits and
Their Relationship to Mortality, Preventive Medicine 9:469-483, 1980.
13. Lemeshow, S., Hosraer, D.W., and Stewart, J.P., "A Comparison of
Sample Size Determination Methods in the Two Groups Trial Where
the Underlying Disease is Rare", Communications in Statistics,
BIO, No.5, 1981.

�</text>
                  </elementText>
                </elementTextContainer>
              </element>
            </elementContainer>
          </elementSet>
        </elementSetContainer>
      </file>
    </fileContainer>
    <collection collectionId="30">
      <elementSetContainer>
        <elementSet elementSetId="1">
          <name>Dublin Core</name>
          <description>The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.</description>
          <elementContainer>
            <element elementId="50">
              <name>Title</name>
              <description>A name given to the resource</description>
              <elementTextContainer>
                <elementText elementTextId="4687">
                  <text>Alvin L. Young Collection on Agent Orange</text>
                </elementText>
              </elementTextContainer>
            </element>
            <element elementId="41">
              <name>Description</name>
              <description>An account of the resource</description>
              <elementTextContainer>
                <elementText elementTextId="49809">
                  <text>&lt;p style="margin-top: -1em; line-height: 1.2em;"&gt;The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.&lt;/p&gt;&#13;
&lt;p&gt;For more about this collection, &lt;a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a"&gt;view the Agent Orange Exhibit.&lt;/a&gt;&lt;/p&gt;</text>
                </elementText>
              </elementTextContainer>
            </element>
          </elementContainer>
        </elementSet>
      </elementSetContainer>
    </collection>
    <itemType itemTypeId="1">
      <name>Text</name>
      <description>A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text.</description>
      <elementContainer>
        <element elementId="52">
          <name>Box</name>
          <description>The box containing the original item.</description>
          <elementTextContainer>
            <elementText elementTextId="19210">
              <text>062</text>
            </elementText>
          </elementTextContainer>
        </element>
        <element elementId="53">
          <name>Folder</name>
          <description>The folder containing the original item.</description>
          <elementTextContainer>
            <elementText elementTextId="19211">
              <text>1686</text>
            </elementText>
          </elementTextContainer>
        </element>
        <element elementId="54">
          <name>Series</name>
          <description>The series number of the original item.</description>
          <elementTextContainer>
            <elementText elementTextId="19214">
              <text>Series III Subseries III</text>
            </elementText>
          </elementTextContainer>
        </element>
      </elementContainer>
    </itemType>
    <elementSetContainer>
      <elementSet elementSetId="1">
        <name>Dublin Core</name>
        <description>The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.</description>
        <elementContainer>
          <element elementId="39">
            <name>Creator</name>
            <description>An entity primarily responsible for making the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19208">
                <text>Spivey, Gary H.</text>
              </elementText>
              <elementText elementTextId="19209">
                <text>Roger Detels</text>
              </elementText>
            </elementTextContainer>
          </element>
          <element elementId="40">
            <name>Date</name>
            <description>A point or period of time associated with an event in the lifecycle of the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19212">
                <text>January 22 1982</text>
              </elementText>
            </elementTextContainer>
          </element>
          <element elementId="50">
            <name>Title</name>
            <description>A name given to the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19213">
                <text>Protocol for Epidemiologic Studies of Agent Orange</text>
              </elementText>
            </elementTextContainer>
          </element>
          <element elementId="49">
            <name>Subject</name>
            <description>The topic of the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19215">
                <text>Agent Orange Exposure Study</text>
              </elementText>
              <elementText elementTextId="19216">
                <text>ground troops exposure</text>
              </elementText>
              <elementText elementTextId="19217">
                <text>cohort study</text>
              </elementText>
              <elementText elementTextId="19218">
                <text>study protocol</text>
              </elementText>
            </elementTextContainer>
          </element>
        </elementContainer>
      </elementSet>
    </elementSetContainer>
    <tagContainer>
      <tag tagId="1">
        <name>ao_seriesIII</name>
      </tag>
    </tagContainer>
  </item>
  <item itemId="2711" public="1" featured="0">
    <fileContainer>
      <file fileId="1495">
        <src>https://www.nal.usda.gov/exhibits/speccoll/files/original/cf559abc9397fcc381c86059207b4e20.pdf</src>
        <authentication>b55a20462fc1f2aeba9e56b7c99f5c16</authentication>
        <elementSetContainer>
          <elementSet elementSetId="4">
            <name>PDF Text</name>
            <description/>
            <elementContainer>
              <element elementId="60">
                <name>Text</name>
                <description/>
                <elementTextContainer>
                  <elementText elementTextId="63444">
                    <text>Item ID Number

°1684

Author

Spivey, Gary H.

Corporate Author
Roport/Artldo TltlO

Review

of the Draft Protocol for Epidemiological Studies
of Agent Orange

Journal/Book Title
Year

000

°

Month/Day
Color
Number of Images

D

22

Descriptor Notes

Wednesday, June 06, 2001

Page 1685 of 1688

�SUBJECT:

Review of the Draft Protocol for Epidetniological Studies of Agent
Orange
Submitted by Gary H. Spivey, M.D., MPH, Principal Investigator
Rogert Detels, M.D., MS, Co-Principal Investigator
Division of Epidemiology
School of Public Health
University of California
Los Angeles, California

Attached please find the individual comments of members of the Science Panel
of the Agent Orange Working Group. Basically, every member expresed concern
about the lack of details in the protocol to the point that it is not possible
to constructively review the proposal.
.
The following paragraphs taken from comments submitted by individual members
highlight these concerns:
General Comments
1.

"While we certainly appreciate Dr. Spivey1s concern that release of
certain specifics of his anticipated protocol might induce bias in the
eventual study, we cannot provide an effective analysis of a protocol
without such information. We suggest that at least a small subcommittee
of the Science Panel be supplied with all of the details of the protocol
and that the report of this subcommittee be held in confidence and not be
released to the general public. We believe that an informed evaluation
is absolutely essential before any further action is undertaken to
initiate any subsequent studies."

2.

"The section on proposed outcome measures is particularly weak. The
statement that an examination will be done because '...the veterans will
expect a physical exam1 is inappropriate. The inclusion of special
examinations for individuals with recognized disease unrelated to Agent
Orange, for example, an examination of the eye backgrounds and peripheral
pulses in subjects with a history of diabetes mellitus is of questionable
value in such a protocol. At the same time the protocol ignores entirely
the neurological examination, which both animal and human data suggest
may be of importance.
"Statements such as the one included on page 9 which opines that chloracna
is a 'self-limiting skin condition' raise further questions about the
authors' full understanding of the potential health effects of dioxins.
Chloracne can be a severe skin condition that in some individuals is
persistent for years even following discontinuation of exposure. The
statement on page 18 that 'Chloracne is the only established health
outcome associated with dioxiri exposure1 is not justified."

.3...

"It is clear that the current UCLA protocol is inadequate. Therefore, a
study is yet to be designed and conducted. Overall, it is our opinion
that two important factors must be present for the design and conduct of

�a study. First, it is critical that adequate epidemiologic expertise be
• available within the Group or Agency which assumes responsibility, and
second, there must be continuous interface with and cooperation from the
DOD and VA so that details of records and activities during the Vietnam
War are accessible to the researchers.
"Finally, any delay dependent upon further review of this UCLA protocol
should be avoided due to its incomplete nature. Any further review
should be postponed until an appropriate scientific protocol based upon a
complete iteration of exposure data and veterans' data is available."
4.

.

"In summary, prior to any further attempts to design a study on Vietnam
veterans, it is recommended that the Veterans Administration review the
morbidity data they have collected thus far, that the Department of
Defense establish information on exposure data and determine what the
sizes of prospective cohorts might be, and that the Veterans Administration
embark on a mortality study. Since any outside group is unfamiliar with
the record keeping system of the military, it would be redundant, wasteful,
and time-consuming to have outside groups do this preliminary work for
the military."
Specific Comments

Exposure
1.

"I am deeply troubled by this aspect of the report. On page 43, the
authors correctly surmise, 'We have not identified a mechanism which
would document actual exposure.1 Over the past year in our Committee, as
well as the Agent Orange Working Group in the White House, we have wrestled,
frankly unsuccessfully, with trying to establish some mechanism for
documenting exposure. I recall clearly our meeting with the members of
the National Academy of Sciences and their comments regarding any proposed
epidemiological study on Agent Orange exposure in Vietnam. The take-home
message was, 'If we cannot scientifically validate and document exposure,
we cannot do a scientific epidemiological study.1 Although Spivey's
approach suggests a mechanism by which we might overcome this problem, I
suspect we are justifiably due some criticism for the grouping approach.
I am now persuaded that we will never be able to do an epidemiology study
on individual veterans per se, but must examine military units serving in
specific spray areas. There is now some hope from recent DOD activities
that we might be able to document some segments of the military population
in Vietnam exposed to Agent Orange. Every effort then must be made to
work closely with Mr. Christian and his associates in DOD in meticulously
reviewing records and films to establish some case for exposure. I
recommend we do not fund any additional feasibility studies until a
thorough and comprehensive search and cataloging of available DOD records,
films, and reports are completed."

2.

"In conclusion, I atn not convinced that significant ground troop exposure
to 2,4,5-T containing herbicide occurred as a result of aerial application.
Other uses of the herbicide most likely represented a greater exposure.
Additionally, the study must address the question of did the Vietnam

�conflict participant incur a health decrement risk over and beyond that
which was expected and secondly, if a risk was incurred, is it service
connected? This protocol requires greater examination of the exposure
criteria and further discussion and refinement."
Use of Terminology
"Definition of Antipersonnel gas: Riot agents such as GS and CN used in
Vietnam were not antipersonnel gases since they do not kill or incapacitate
for an extended period of time. Both CS and CN have been used throughout
the world by civilian police to control riots of civilians and in prisons
without causing fatalities. This improper definition should be corrected.
"The substitution of 'riot control agents' in place of 'antipersonnel
gases' is suggested."
Conclusion
The members of the Panel had many other specific comments and only some of
their major concerns were quoted here, -ft—a-ppears that" the present proposal
is inadequate and it is recommended that a course of action be developed that
will not cause any further unnecessary delays in attempting to answer questions
about health issues of Vietnam veterans. A specific protocol should be -"Allayed
in which the size of the cohorts and their perceived exposures are characterized
and which will serve as the basis for the studies.

�&gt;;&amp;W&gt;*&amp;^i(&amp;X^
**iic«S»S*v.^

A
The major concern we have with this draft document is the lack of detail provided
by the authors in describing the protocol itself. This lack of detail makes
an informed critical review impossible. The investigators state that a detailed
protocol is not provided due to their inability to examine all relevant military
records on which the studies would be based and also because of potential bias
they fear might result from premature public release of a detailed study design.
The quality, scope, and availability of military records on Vietnam veterans in
addition to the ability to locate individuals in a time - space frame in Vietnam
have previously been major issues with regard to developing an epidemiologic
protocol. In our view, an adequate protocol requires a clear and detailed
evaluation of those military records that are available. While a preliminary
feasibility study might be indicated to evaluate the procedure of establishing
cohorts with differing exposure levels, a detailed and clear understanding of
what records are available may make this unnecessary. A thorough evaluation of
those records that exist, what the records contain, and how they might be used in
establishing appropriate cohorts must be performed before any protocol can be
properly reviewed.
We also strongly disagree that the full protocol should be withheld due to
potential bias. It is important that the full protocol receive adequate peer
review due to the importance of the investigation. The publicity which currently
surrounds the study has already influenced those individuals who will eventually
be included. A full presentation of the protocol in our view would have little
further adverse impact. To withhold pertinent details may cause far more harm
by seriously damaging public confidence in the credibility and independence of
the study.
...
,.
..--..
We do not agree that the "historical cohort study" should be limited to draftees
and one term enlisted men. Excluding individuals with longer service will
undoubtedly exclude some individuals with the greatest potential exposures.
Inadequate detail is provided about cohort selection by the authors. Full and
extensive discussion of the HERBS data, of the nature of troop movements through
Vietnam, and the pattern of likely exposures are critical in evaluating the
protocol. There is a fair amount of information offered on environmental persistence of agent orange, but no assessment of how the data affects presumed exposure to
ground troops. The concept of establishing low to high exposure groups based on
a "time - place - company exposure grid" is too vague as presented and it is
not clear that the authors have a fundmental understanding of the core issues
which need to be addressed in establishing such cohorts.
The sections in the protocol&gt;on potential confounding variables and control
groups are both non-specific and short on detail.
The section on proposed outcome measures is particularly weak. The physical
examination in our view can be a significant factor in this study in determining
ill health, since the potential end organ toxicity for a number of organ systems
can be identified and specifically evaluated through a physical examination. The
statement that an examination will be done because "...the veterans will expect
a physical exam" is inappropriate. The inclusion of special examinations
for individuals with recognized disease unrelated to agent orange, for example
an examination of the eye grounds and peripheral pulses in subjects with a
history of diabetes mellitus is of questionable value in such a protocol. At
the same time the protocol ignores entirely the neurological examination, which
both animal and human data suggest may be of importance.

�Statements such as the one included on page 9 which opines that chloracne
i is a "self-limiting skin condition" raise further questions about the authors'
! full understanding of the potential health effects of dioxins. Chloracne can
I be a severe skin condition that in some individuals is persistent for years
] even following discontinuation of exposure. The statement on page 18 that
"chloracne is the only established health outcome associated with dioxin exposure"
!
. is not justified.
The authors have devoted considerable effort to giving a review of basic epidemiologic principles and a superficial review of toxicity information. While this
information may be of interest as introductory material, the critical
task of-the contract was to establish a thorough, detailed, and scientifically
"~"fc.
defensible protocol.
"•••;'s;»;;i'-h:'"ii'...; :::j;PK.::^::':^
:^.Ex2rr::;v::r^

^^

It is our opinion that the current draft protocol is inadequate as presented and
that a major effort will be required to develop the protocol to a point where
further peer review can be meaningful.
;: We look forward to working with the Science Panel in responding to the VA on this
important proposed study.

••" " "

. .
1. While we certainly appreciate Dr. Spivey's concern -that
release of certain specifics of his anticipated protocol might
induce bias in the eventual study, we can not provide an effective
analysis of a protocol without such information. We do not support
blind trust in the correctness of Dr. Spivey's deductions and
selections. We suggest that at least a small subcommittee of the
Science Panel be supplied with all of the details of the protocol
and that the report of this subcommittee be held in confidence and
not be released to the general public. We believe that an informed
evaluation is absolutely essential before any further action is
undertaken to initiate any subsequent studies.
2. The following comments are provided with references to paragraphs and stamped page*numbers at the bottom of each page:
a. Pg. 007, Para. B, Method:
HERBS data. This has already been
defoliation missions with separate
Veterans Administration has such a

Calls for Computer mapping of
accomplished for both cropland
map overlays by year. The
set of overlays.

b. Pg. Oil, Last paragraph:' Our records show that all fixed
wing missions using herbicides blue and white were not stopped
until 31 October 197.1. The last helicopter mission with Herbicide
Orange was recorded on 9 June 1970. All fixed wing defoliation
missions with Orange ceased on 16 April 1970. The same paragraph
neglects to mention that.herbicides were also sprayed along the
sides of rivers, roads, and communication lines to prevent cover
to the enemy.

�c. Pg. 020, 1st Para.: For possibly highly exposed personnel
he might also wish to include chemical unit personnel at battalion
level charged with spraying base camp perimeters, riverine personnel
who sprayed the edges of rivers, and other personnel involved in
cleaning up herbicide spills after accidental releases.

d. Pg. 040, Ques. 1 posed: The discussion here is extremely
limited and does not mention the fact that many personnel both
Orange and non-Orange exposed may have had prior exposures to TCDD
through agricultural or home use during childhood and adolescence.
There is no reason to believe that commercially available herbicides
did not also contain TCDD. In fact, during the period when these
young men were children the TCDD concentrations in such compounds
may have been higher due ..to lack of perfection in the manufacturing
process. We believe this should be looked into and also considered.
If such is the case then an "unexposed" guaranteed population of
young adult males of that period may be impossible to find. We are
now researching the use of 2, 4, 5-T at posts and stations here in
the United States and overseas. Because of the long latent period
for carcinogenic effects, perhaps those persons claiming effects
such as testicular cancer may be reaping exposures which took
place while in childhood here in the United States and having
nothing to do with Vietnam service. Similarly, prior exposures
to other toxic environmental hazards may now be showing up as
diseases in the Vietnam veteran population. No mention is made
of possible other exposures to dioxin containing substances such
as pentacholophenol wood preservatives and hexachlorophene used
in surgical germicidal soaps. Thus unexposed Orange troops may
have gained dioxin exposure from other sources. Likewise, a
veteran who served in Vietnam say from 1967 to 1968 has had 12
years in which (depending upon his civilian work environment) he
may have been exposed to several other toxic and hazardous
compounds that are now producing disease symptoms and which have
no relation to Herbicide Orange in any way.
e. Pg. 041, Ques. 2, Other Exposure: Attachment 1 to these
comments lists possible other exposures encountered by military
members serving in Vietnam. It is by no means an exhaustive
list. We agree that these additional sources of exposure to
various substances and disease entities may be very important
confounding factors in relation to sorting out primary effects
caused by exposure to Herbicide Orange.
f. Pg. 043, Ques. 5: Comparison to veterans who served in
Korea and World War II may be exceptionally difficult since many
if not most of the personnel records were destroyed in the St.
Louis Records Center fire.
g. Pg. 047, 1st line, 1st paragraph: Suggest substitution
of "riot control agents" in place of "antipersonnel gases."

�^ f ' &amp; s ^ S f
Q e S ^ &amp; S ^ ^

;

h. Pg. 049, Para. E.I.: Why will the study be limited to
draftees and one term enlisted men? Why not include young troop
commander officers and senior enlisted personnel who were in the
field? Follow-up physical examination records for career
enlisted and officer personnel would probably be much more
extensive and easily obtainable than from civilians who left the
service at the end of one enlistment.

i. Pg. 052, Para. 3, Estimation of Exposure: We see no
reason to construct a second exposure index involving a long term
environmental persistence as many.studies have been already
directed at herbicide persistence, especially considering the
high rrtobility of the troop populations while in Vietnam and
decay studies of herbicides and TCDD such as Seveso. Also supporting this contention would be the long terra studies conducted at
Englin^Air Force Base at the loading hardstands where there was
exceptionally high concentrations of Herbicide Orange spilled and
trapped in the ground.
j
j. Pg. 052, last para.: What is proposed in this last paraj graph and onthe following page can amount to millions of dollars
,;• of labor time to locate a defined population with respect to the
:
HERB tapes. The already accomplished Battalion studies conducted
.1 by the Army and the Marine Corps have pointed out how laborious
and time consuming such a massive effort would be. The major
fallacy of such a study would be that the-major source of
contamination and/or exposure to ground troops may not be from
fixed wing Ranch Hand spraying. This is because of entrapment
of the herbicide in triple layer tree canopys, rapid absorbtion
of the herbicide into plant leaf tissue, vaporization of the
spray above the jungle canopy, rapid decay of dioxin due to
photo degradation (less than 6hr.half-life), and low or nonexistent volatility and very low solubility of dioxin in water.
A probable source of much higher contamination would be from
spraying of herbicides around the perimeters of camps and fire
bases through the use of helicopters, truck mounted decontamination spray units, backpack hand sprayers, and Buffalo turbines.
These types of perimeter spraying are very poorly documented at
best and only a tiny fraction of all helicopter spray missions
are documented in the HERBS tapes. Couple this lack of data as
to amounts, type and dates of perimeter spraying with the already
proven difficult task of placing troops with respect to physical
location by each day of the year at a specific grid location and
we have a very "iffy" situation at best. What the author
advocates in no way seems to be a solution. Finally after 18
months neither the Science Panel nor anyone else has been able to
establish just what constitutes an "Exposure" to Herbicide Orange.
k. Pg. 053, Para. 4, Establishing Cohorts: In this discussion we assume the -author is referring to single or multiple
exposures from Ranch Hand flights over or near the selected
company. There are many problems to this method of counting on a
company and its personnel being at a certain place at a certain
time with respect to an overflight by Ranch Hand aircraft. To
name a few, we have the following:

�(1) In the case of Marine records the troops were carried
on the Morning Reports (MR) as either Killed in Action (KIA),
Wounded in Action (WIA), Missing in Action (MIA), present for
duty, or transferred. Hence minor disabilities were not entered
' in the MR and consequently a Marine may have remained in base
camp while his platoon was out on patrol.
(2) Companies did not necessarily operate all together
and stay together. Marine units very frequently operated in
small 'patrols and platoon size units. Documentation as to their
day-to-day locations are very poor. In one Marine battalion study
for a .two month period only 4 exact unit locations could be
determined for an operational area of 10km long along route 9 and
width of 6.5km back from each side of this. road. We know that
patrols roamed all over this area to protect the security of the
.route from enemy ambushes but it is impossible to fix company,
platoon, or squad locations during this time.
(3) In the case of the 1st. Cav. units, platoons would
, be detached from a company for a period of time and assigned to
'• another infantry battalion far from the operating location of
thie 1st. Cav. parent company.
(4) In air mobile operations, it was not uncommon to
have helicopters drop a platoon of men into a landing zone and
then the platoon would go into a search mission through the
jungle with no documentation as to their day-to-day location.
Finally one or more days later the platoon would reassemble at
the landing zone and be airlifted out. In the mean time the
rest of the Company would be at the base camp or operating in
other areas.
(5) We have also found that there are errors in the
morning reports regarding the duty status of individuals. It
has been observed that as long as six months elapsed before a
correction was made to pick-up or drop an individual. Similarly
some people may have been detached from the units for short
periods of time (several days) without this showing up in the
Morning Report. We know of one case of an individual whose body
was in the mortuary in Tan San Nhut and his unit was unknown for
quite a while. When his unit was contacted they thought he was
still with them until they made a detailed search. He was, of
course, being carried on the Morning Report as present for duty
although he had been dead for several days. We cannot trust the
validity of these combat records in all cases.
(6) There was also a very high thruput of.personnel in
the Army battalions studied. For a battalion of about 971
persons over 2,300 personnel served for varing periods of time
in the one year period which was studied.
1. Pg. 063, last para.: It seeins wise to mention that a
preliminary review cost for pulling a record at the St. Louis
Records Center and determining minimum locator information will
be $5.17 per record. More detailed research could run the bill
up and overall costs of a major search could be very very expensive.

�m. Pg. 069, HERBS Tape Mapping: The National Academy of
Sciences report "The Effects of Herbicides in South Vietnam"
Part A, Summary''on page IV-104 shows just such a mapping as is
desired to be accomplished. We understand that it took the
computer at least 5 twelve hour runs to produce these maps.
Why should they be done over as they were made directly_from the
HERBS hard copy reports and should be available. This is a very
costly process to duplicate. As may be seen from the referenced
page, each spray track is shown on the map with the date of the
mission, number of gallons sprayed and the type of herbicide.
(Attachment 2 enclosed)

^

~ -.;; ^sssssz?..*'

n. Pg. 070, Major portion of page: This is an idealized
approach and is not at all typical to many units operating over
there, especially those which were air mobile, or to Marine .
units. We determined this when we looked at the same Marine
battalion that GAO had said was right under a spray track. In
one of the battalion studies no detailed unit records were found
for an entire six month period.
~
o. Pg. 089, 1st new para.: As mentioned earlier detailed
personnel records for Korean veterans may not be available
because of the St. Louis records center fire. We understand
reconstruction (costly) of about 35% might be possible. One
major factor has been overlooked in a possible comparison
between Vietnam and Korean veterans and that is Korea was not
in a jungle environment and units operated as units. Winter
fighting was common in portions of the Korean war while Vietnam
was in a jungle environment with many different types of disease
exposures and small unit operations involving living off the
land. Korea was not a s&gt;. unpopular as Vietnam.
p. Pg. 249, Definition of Antipersonnel gas: Riot agents
such as CS and CN such as used in Vietnam were not antipersonnel
gases since they do not kill or incapacitate for an extended
j
period of time. Both CS and CN have been used throughout the
j
world by civilian police to control riots of civilians and in
i
prisions without causing fatalities. This improper definition /
should be corrected.
""'
q. Pg. 249, Definition of Battalion: An idealized definition,
not. necessarily true in Vietnam type combat operations.

�As you requested, we have reviewed the "Draft Protocol for Epidemiologic
Studies of Agent Orange", G.H. Spivey and R. Detels, U.C.L.A., submitted to
the Veterans Administration (contract V101 ( 3 P-842).
9 )
It is our understanding'that the purpose of the contract was to procure an
epidemiologic protocol outlining specific plans for the study of Vietnam
Veterans exposed to Agent Orange. The protocol submitted, if it is a final
product, does not fulfill this goal. It is diffuse and nonspecific and does
not outline an approach particular to the problem being addressed. In fact,
numerous epidemiologic approaches to the study of veterans, are outlined, as
are numerous hypotheses which could be tested. Not included, however, is a
clear discussion of benefits and drawbacks of various approaches, and a
clear recommendation for a study addressing this issue.
The authors of the protocol note a number of problems which they encountered
including access to military data and the issue of bias if specific plans
for a study of Veterans a_re publicly discussed. To these problems should be
added the very limited time which the contractor had to fulfill the
assignment. The question of scientific bias, however, should not continue
to be a deterrant to the development and outline of specific plans for a
study of the Vietnam Veterans. In fact, any protocol for such a stiidy
should be specific and receive the widest review possible before a study is
initiated.
In response to Senator Cranston's letter, it is clear that the current UCLA
protocol is inadequate. Therefore, a study is yet to be designed, and
conducted. Overall, it is our opinion that two important factors must be
present for the design and conduct of a study, regardless of what Group or
Agency specifically carries out these tasks. First, it is critical that
adequate epidemiologic expertise be available within the Group or Agency
which assumes responsibility, and second, there must be continuous interface
with and cooperation from the DOD and VA so that details of records and
activities during the Vietnam War are accessible to the researchers.

k

ael?end&lt;int

" ° further review of this UCLA protocol
P°

should be avoided due to its incomplete nature. In uarHo.lar-P
the National *ca&lt;toay of Seizes should be P^tponed^t 1 «

�This review represents my own personal assessment of the protocol
and does not represent necessarily a.
(Department
jposition. Several fundamental comments are provided in
the way of general criticism. More specific, perhaps less
important, thoughts are then attached, Atch 1.
a. Confidentiality of Protocol Design
Since premature public disclosure could potentially jeopardize
the validity of the study, it is agreed that certain specific
details of the protocol design requires restricted public
access. However, this scientific panel must obtain a more
detailed discussion of the study design. The Veterans
Administration must develop a mechanism which preserves the
validity of the study while providing adequate scientific peer
review. Functionally it would be preferable to have the detailed
protocol made available .to this panel with a follow on discussion
with Dr Sprivey and the other principal designers some weeks
later.
b. Definition of the Fundamental Question to Be Answered.
The Vietnam veteran through some form of "collective wisdom"
perceives himself to have suffered an abnormal, unexpected health
decrement either in himself or his progeny following his
participation in the Vietnam conflict. One cannot determine why
this "collective wisdom" has singled out Agent Orange as the
causative .factor. However, the "collective wisdom" of a group of
persons occupationally exposed has often times proved correct.
Subsequently,'the scientific community has then through rigorous
examination determined the cause of the observed decrement. I
have not come to a conclusion regarding Herbicide Orange, the
"collective wisdom" of the veteran may prove to be correct on
both counts i.e. health decrement and cause. However, close
examination of the "occupational environment" of the Vietnam
conflict does not easily reduce to a single factor of exposure
but rather reveals a multiplicity of exposures which singularly
or in combination could produce the decrements described. The
fundamental question to be answered by the VA study is: . Did the
Vietnam conflict participant incurr a health decrement risk over
and beyond that which was expected and' secondly if a risk was
incurred is it service connected? The Congress, perhaps naively,
tried to address this question by directing the VA to conduct an
epidemiology study. Approximately a year later the Congress had
gathered additional information on the Vietnam experience. While
not yet an act of Congress, the newly proposed legislation has
enjoyed wide acceptance and support. . The so called broadening of
the VA study to include other possible causitive factors was
supported by this'scientific panel.

�Dr Spivey has concluded that "the original Agent Orange question,
# 1 page 30, addressed with suitable safeguards for confounding
by other exposures and factors, would seem to be the most
appropriate for this study". His discussion of this narrowly
definded question leave's the impression that a true decrement may
go unrecognized and that sequential study of other narrowly
defined exposures may be necessary to determine the existence of
a health risk. The Government and the veteran wants an answer to
the broader question in terms of what diseases and health
decrements are in excess of the norm and is the excess, if it
exist, service connected. For the purpose of prevention of
future disease the identification of the causative factor becomes
useful.
Dr Spivey's discussion of the broader question, fno. 4, page 33,
implies that the discrimination of an effect would not be
decernable. A properly .designed study could examine the
existance of disease excess in both broad categories of impacted
systems and specific disease end points. Identification of
exposure factors could be examined by use of a regression
matrix. The use of multivariate statistics may be ideal in this
study design instance. Further, a group of reasonably informed
individuals could develop a list of major independent variables
(occupational and environmental exposures) which given present
state of the art knowledge represent potential risks to human
health singularly or in combination.
c. Exposure Criteria
The Spivey design if adopted would be largely dependent on an
exposure which at this reading admittedly lacked definitive
objective criteria. Sole reliance on the Herbs tape and troop
headquarters location assumes that the aerial spraying of
herbicide was the principal route of exposure to 2,4,5-T
containing herbicides. This method functionally ignores the
application of herbicide by means other than aircraft and
helicopter. Additionally, it fails to observe the possible
presence of other 2,4,5-T containing herbicides 'which may have
been used by local commanders. The various agencies were free to
procure these herbicides through the Federal supply system; the
General Services Administration (GSA) regularly negotiated
contracts with private companies. Only the tactical herbicides
were controlled and even then local commanders obtained these
herbicides probably by barter. A list of commerically available
2,4-D and 2,4,5-T herbicides is attached, Atch 2. It is not
known how many of these were procured by the GSA. Attachment 3
lists some Federal Stock Numbers (FSN's) which were 2,4-D and
2,4,5-T containing herbicides other than Herbicide Orange. If
Herbicide Orange is the causative factor for the health
decrements claimed, the aerial application was not necessarily
the significant route of exposure. While the quantities applied
aerially were massive, the ground troop proximity was not
sufficiently close to represent a significant exposure. A more
likely exposure would have been 'the local application of
herbicide to base perimeters and associated undesirable
vegetation. Here some ground troops were immediately adjacent to
the spray. This is not to say that in certain instances that the
Ranch Hand aircraft application of herbicide to base perimeters
didn't occur. There is reason to believe that base perimeter
spraying by Ranch Hand did occur.

�Lastly/ if the Herbs tapes are to be used to establish exposure
then a consistent objective exposure model needs to be
developed. Such a model must consider wind direction, altitude
of dispersal, air temperature, particle size, volitility of the
herbicide and the type of vegetation targeted. If ground troops
were not directly below -the spray path, then the above parameters
become important in determining what residual material may have
been inhaled by personnel located some distance away from the
spray path. Obviously the greater the distance the less the
concentration of exposure. However, a simple plume dispersion
model may not be appropriate since local vegetation would
intercept the mist/vapors and incident radiation may degrade the
level of TCDD. It is suggested that local weather information be
used if available. Failing that, simplifying assumptions could
be made using seasonal information. Development of this
methodology may not prove definitive but would most likely
describe reasonable bounds for exposure.
d'. Proportionate mortality study
This short term study proposed by Dr Spivey's group has merit.
The proportionate mortality study should be undertaken even
though the results may be less than definitive.
ln conclusion, I am not convinced that significant ground troop
exposure to 2,4,5-T containing herbicide occured as a result of
!, aerial application.
Other uses of the herbicide most likely
represented a greater exposure. Additionally the exposure to
other occupational and environmental factors present in Vietnam
represent equivalent in greater risks of chronic disease. The
study must address the question of did'the Vietnam conflict
participant incurr a health decrement risk over and beyond that
which was expected and secondly, if a risk was incurred, is it
service connected? This protocol requires greater examination of
the exposure criteria and the futher discussion and refinment of
\question number four, page 33.

3 Attachments
1. Specific Comments
2. Formulations Containing
2,4,5-T

�^

Specific Comments
1.

Design of a Study Using Vietnam as the Exposure Criteria

The concept of a factorial experimental design could be used
to establish as independent variable matrix. This matrix could
have incorporated within it the Agent Orange exposure as well as
exposure to such factors as Dapsone, pentachlorophenol, other
2,4-D and 2,4,5-T herbicides, Malithion, illicit drugs, alcohol
consumption, riot control agents, DEBT and smoking.
Variables
which are not quantitative or quantifiable could have digital
descriptors applied as indicators of exposure.
The dependent variable matrix could access categories of
disease or specific disease end points. The independent matrix
could then be used to predict disease outcome via miltivariate
regression. If significant relationships were found,
examination would be required to determine plausibility before
causation could be assigned.
The control group of non Vietnam veterans would have
similiar independent and dependent matricies developed. A
comparison of the two dependent matricies would examine the
question of health decrement over and beyond what was expected.
All persons in the study matrix would receive a study
questionnaire and have their military and private medical
records examined. A random number of personnel in the large
matrix cells would be selected for physical examination. The
objective physical examination findings would flesh out and
validate the independent" and dependent matricies. Should
significant health decrements be found in certain matrix cells,
additional cell members would be physically examined to validate
the initial findings. Mortality analysis of the full cohort
would be undertaken.
2.

Personnel to be studied

While the greatest involvement of personnel did occur in the
time period 1965-1972, the greatest concentration of TCDD in
2,4,5-T containing herbicides occurred prior to 1965. Any
exposure criteria uniformly applied across the years of
involvement would indicate that .early year personnel would have
been exposed to greater levels of TCDD. Also, it would seen
advantageous to include people who had multiple year exposure.
Personnel who voluptered to go to Vietnam should be included
in the study. If desired the differentiation of volunteer/
draftee and enlis.ted/officer personnel could be entered into the
independent variable matrix. The illicit drug and alcohol
habits of the four possible groups may have been different.

�:£i:ji!iKVfcV^^

Note: The Marines had a policy of. no hard liquor while in
Vietnam. It is unclear if this was for all Marines.
3.

Cost of Study

At ,the present time cost is not a consideration of study
design. Only after the best possible scientific design has been
established must a judgement be made to accept the anticipated
cost. In this case it could seem cost acceptance becomes the
ultimate burden of the Congress after the Veteran's
Administration describes what can or cannot be done
scientifically.
4.

Assessing Herbicide Orange Exposure by Questionnaire

This technique frequently used in an occupational setting is
extremely risky in this case. Since the same aircraft sprayed
White and Blue, the recollection of a study participant that a
plane or helicopter "flew near or over him spraying something"
will not establish exposure to Herbicide Orange. The additional
observation that a camouflaged aircraft was used doesn't narrow
the possibilities.
The first Ranch Hand planes sent to Vietnam for defoliation
had bright silvery skins.
5.

Active Duty Deaths

One should not assume that an active duty death occurring
within one year was due to traumatic battle injury.
Additionally, the fact o_f contributory cause may be important if
the immune system were compromised. Lastly some deaths may have
occurred as a result of disease only.

�i^/w^.^&amp;rtSi^i&amp;fcC'rt.U.i.^yS^^

67346

.

Available Formulations Containing
' as of February 1, 1973
Arnchem Products, Ambler•, Pennsylvania 19002

Emulsamine 2,h,5~T
Alkjclamine salt

$8.9% act. ing.

33.7# a.e.

3 Ib/gal

Envert-T
Butoxyethanol ester

33.9* act.- ing.

Trinoxol
'
Butoxyethanol ester

59.7# act. ing. 1*2.9$ a.e. k Ib/gal

Trinoxol Super 6
Butoxyethanol ester

2h.3% a.e. 2 Ib/gal

•
8 . $ act. ing.
15

58.5£ a.e. 6 Ib/gal

57.2£ act. ing.

lil'.Qfc a.e.

h Ib/gal

. £8.3^ act. ing.

ijl.9^5 a,e«

h Ib/agl

Weedone 2,li,?-T ,3?pcial Air Snray Formula
Butoxyethanol ester
£8.7^ act. ing.

1^2,2^ a.e.

it Ib/gal

Weedar 2,h,?-T
Triethylamine salt
Weedone 2,!t,^-T
Butoxyethanol eater

gjlt-D and 2,&gt;-t,£-T mixtures
Dinoxol
Butoxyethanol-ester
Emulsamine HK •'
Alkylamine salts
./' '
'• Dinoxol Super 6
Butoxyethanol esters
Emulsaverb 100
acid 't anune salt
F^nulsavert 2h8
acid &amp; amine salt

2,i;~D 3 . a.i. 21.75? a.e. 2 Ib/gal
1^
2,U,5-T 30.2£ a.i. 21.735 a.e. 2 Ib/gal
2,U-D
2,^,5-T

3 . a.i, 16.93? a.e. 1.5 Ib/gal
1W
'29.5^ a.i. 16.5^ a.e. 1.5 Ib/gal •

2,h-D
2,fc,5-T

Ii3.?q' a.i. 29283 a.e.
la.755 a.i. 29.8* a.e.

3 Ib/gal
3 Ib/gal'

2,U-D
2,U,5-T

21.255 a.i. ..11.9^ a.e.
21.255 a.i. 11.92&amp;ra.e.

1 Ib/gal
1 Ib/gal

2,li-D
2,h,5-T

13.25C a.i.
6.1^ a.e.
19.3^ a.i. 12.255 a.e.

^Ib/gal
1 Ib/gal

�, / '' •^V;V\ •',:l'c7;'.; ;^
' •-

:.^..^^i^^^'^'';^-''^^^^

:.'•'•....-•' ' ' . . • '••", . !•' •.': • ^ .V.-«/.' '•

3.'

~ *- "
1

Envertf-.DT
Butoxyethanol ester

G7347

_ 2,h-D 1-7.1% a.i.
2,U,5-T ' 17.03.a.i.

12.2$ a.e.
12.2g ae.

lib/gal
1 Ib/gal

j'

Weedar Amine BK •
Dimethylamine salt
2,ij-D
Trimethylamine salt 2,U,5-T
V/eedone BK 6k
Butoxyethanol ester
Weedone IBK
Butoxyethanolester

2,/i-D
2,U,$~T
2,h-D
2,U,^rT

2k.5$ a.i. 2 . $ a.e. 2 Ib/gal
0[
2 . a.i. 2 . $ a.e. 2 Ib/gal
8^
01
19.7^ a.i. lh.2^ a.e. 1.33 Ib/gal'
lil.^- a.i. 28.6^ a.e. 2,6? Ib/gal
31.1^ a.i. 21.1$ a.e. 2 Ib/gal
29.7^ a.i. 21.k% a.e. 2 Ib/gal

Dow Chemical Co . , Midland^ ^ M i chiga n o 86ItO

Esteron 2h? Concentrate
Propylene glycol butyl ether ester

92.5'^ a.i. 60.2^ a.e'.

Esteron 2l
Propylene glycol butyl ether ester

69.2% a.i, U5&gt;»0$ a.e. k Ib/gal

Heddon
Propylene glycol butyl ether ester

20.1$ a.i. 13.3$ a.e. 1 Ib/gal

Veon
Trimethylamine salt

$6.1% a.i. 1 0 2 i.e. k Ib/gal
:,$

Verton 2T
Propylene glycol butyl ether ester

37.2$ a.i. 2lu2$ a.e. 2 Ib/gal

6 Ib/gal

* V

''

' 2 &gt; Ij-D and 2, Itf 5&gt;-T mixtures

Brush Killer LV 2-2
Isooctyl esters

3k.7% a.i. 23.03 a.e.
33-K a.i. 23.03 a.e.

2 Ib/gal
2 Ib/gal

6£.0$ a.i. 'k$,3% a.e,

k Ib/gal

Esteron Brush Killer
Propylene glycol butyl either esters
(PGBE)
2,h-D
•
36.0'&lt;a.i.
22.2^ a.e.
2,U,?-T
3U.1J6 a.i. 22.2% a.o.

2 Ib/gal
2 Ib/gal

Brush Killer LV liT
Isooctyl ester

Tippon 2-2
PGBE esters

2,h~D
2,U,5~T
2,U,$-T

2,h-D
2,ii &gt; £~T

36.3^ a.i.
3luW a.i.

22.1$ a.e.
22.k% a»e.

2 Ib/gal
2 Ib/gal

�.. ,
..;'•'• .;• i'Vii%'t'v'»'4'-$'*i'^'1:'1'1'' •;' V
.'...,• . • ; '••j.j ,:**^&gt;wfrn. !."..'--•'•••
•&gt; •

— 3 —

67543

Veori Brugh Killer
Dimethylamine
. Trimethylamine

2,4-D
2-,4,5-T

24.3% a.i.
2.8.2% a.i.

20.2% a . e .
20.2% a.e.

2 Ib/gal
2 Ib/gal

Verton CE
(PGBE)

2,4-D ' 36.0% a.i.
2,4,5-T
34.17o a.i.

22.2% a . e .
22.2% a.e.

2 Ib/gal
2 Ib/gal

10.3% a.e.
41.3% a.e.

1 Ib/gal
4 Ib/gal

TORDQN and 2,4,5-T Mixtures
Tordon 155
Isooctylester Piclorara
PGBE
.
2,4,5-T

15.1% a.i.
63.4% a.i.

Thompson-Ha yward
2.4.5-T
BED-WEED- LV-6

2,4,5-T

64.0% a.i.

44.4% a.e;.

4 Ib/gal

2,4,5-T

IsoocyCl. ester

83.5% a.i.

58.8% a.e.

6 Ib/gal

22".2% a . e .
22.2% a . e .

2 Ib/gal
2 Ib/gal

DED-WEED LV-9

Isoocytl ester

2.,4-D__ and 2,4, 5-X mixture
DED WEED LV-33
Isooctylester
Isooctylester

2,4-D
2,4,5-T

33.5% a.i.
31.9% a . i .

Transvaal Inc. , Jacks£nville, Arkansas

7207/?

2.4,5-T
Brush-Rhap A-4T
Triethylamine salt

2,4,5-T

. 57.0% a.i.

40.8% a.e.

4 Ib/gal

Brush-Rhap LV-4T
Ethyl hexyl ester

2,4,5-T

65.3% a.i.

45.4% a.e.

4 Ib/gal

Brush-Rhap LV-6T
Ethyl hexyl ester

2,4j5-T

87.0% a.i.

60.4% a.e.

6 Ib/gal

Brush-Rhap A-2D-2T
Dimethyl amine salt
Triethylamine salt

2,4-D
2,4,5-T

24.7% a.i.
28.67, a.i.

20.5% a.e.
20.5% a.e.

2 Ib/gal
2 Ib/gal

Brush-Rhap LV-2D-2T
Ethylhexyl ester
Ethylhexyl .ester

2,4-D
2,4,5-T

34.7% a.i. - 23.0% a.e.
33.1% a.i.
23.0% a.e.

2 Ib/gal
2 Ib/gal

�This document entitled "Draft Protocol for Epldemiological Studies of
Agent Orange" was prepared under the direction of Dr. Gary II. Spivey,
University of California, Los Angeles, California, Veterans Administration Contract V-101(.98)3P-842. It consists of three parts. The
first 19 pages are primarily an introduction. The second 65 pages
which are entitled "Research Methods and Proposed Protocol" primarily
represent a discussion of the difficulties normally faced in epidemiology
studies, and the rest of the document is a literature review covering
about 141 pages. In addition, an outline of a proposed cursory physical
examination is also attached.
From the information provided in this draft, it is not possible to
constructively criticize any proposed study since insufficient information
is provided to determine how this study is to be conducted. Apparently
the authors of this proposed protocol had a number of difficulties;
among them, the inability to obtain exposure data partly because they
did not have any security "clearance. Basically, the authors of this draft
document recommend that cohorts be established for follow up, that they
receive a basic physical examination, and that a questionnaire be administered.
The proposal in the document on how to determine exposure appears rather
cumbersome and will probably not be very rewarding. I, therefore, suggest
that battalions be identified and that the basic exposure of these
battalions be determined. The individual soldiers should then be assigned
to their respective battalions. In addition, the amount of time each
veteran spent in Viet Nam should be determined. It could then be
established on a subsample how great the variation is for the amount of
time that the different veterans actually spent within the battalion versus
the' amount of time spent in Viet Nam. If it turns out that the amount of
time per soldier spent away from his battalion in Viet Nam does not differ
a great deal, it is unnecessary to check for the presence or absence of all
soldiers in the morning report since the exposure for all of them will be
roughly the same. It is recommended that this be determined by the Department
of Defense under the supervision of the IIHS working..group and be reviewed
by an outside advisory group.
Since the sprayed Agent Orange is persistent in the environment, it is
assumed that, in addition to the exposure to spraying missions, the
soldiers had additional exposure while in sprayed areas. It is recommended
that it be determined whether battalions who were sprayed on were also
usually battalions that were in sprayed areas. In addition, all battalions
that were exposed to aborted missions should be identified and their special
additional exposure determined.

�While the Department of Defense is determining exposure levels of Viet Nam
veterans and identifying veterans who could be grouped into cohorts, it is
recommended that the Veterans Administration review the morbidity data
which have been collected in different Veterans Administration Hospitals
from Viet Nam veterans to determine whether any obvious clustering of
certain symptoms and signs can be identified, or any disease patterns
which are out of the ordinary. Concurrently with these efforts, a prospective and retrospective mortality study should be .conducted. Approximately twenty people should be trained to locate death certificates of
Viet Nam era veterans. Information about deaths of Viet Nam. era veterans
will have to be obtained from different sources since particularly in the
early part of the Viet Nam War, social security numbers were not used as
serial numbers. However, by using a multifaceted approach, it should be
possible to locate better than 95% of the death certificates within an
eight-month's period. Once death certificates have been located, the Viet
Nam veterans have to be separated from veterans who served in other areas
during the same time period. In addition, all casualties can be separated
before a retrospective mortality study is done. In addition to this
retrospective mortality study, a prospective mortality study can also be
initiated at the same time. A more detailed outline of this study will
be provided later.
. .
Since the Veterans Administration and the Department of Defense are most .
familiar with their own records and since this is merely a search for
records, no obvious bias could be introduced into this part of the study
or into the interpretation of the analysis of death certificates, particularly if the raw data are made available to a review group. There is,
therefore, no reason why this should be done by a group outside of the
military, particularly if outside help is solicited in areas where no
internal expertise is available.
It is stated on pciges 64 and 65 of the present draft document that from
the Viet Nam era veterans during the period from 1965 to 1972, there are
now 130,000 deaths and that approximately one-third (i.e., A3,000) of the
soldiers are expected to be Viet Nam veterans. These numbers should be
rechecked since there have been statements in the press that a total.of
about 50,000 veterans were killed in action in Viet Nam. One would have
to assume that there should be many more death certificates than 43,000.
In summary, prior to any further attempts-to design a study on Viet Nam
veterans, it is recommended that the Veterans Administration review the
morbidity data they have collected thus far, that the Department of Defense
establish detailed exposure data and determine what the sizes of prospective
cohorts might be, and that the Veterans Administration embark on a
mortality study. Since any outside group is unfamiliar with the record
keeping system of the military, it would be redundant, wasteful, and time
consuming to have outside groups do this preliminary work for the military.

�\

As requested at our last Advisory Committee meeting on Health Related Effects
of Herbicides, I have reviewed the Draft Protocol submitted by Dr. Gary H.
Spivey and Dr. Roger Detels of the Division of Epidemiology, School of Public Health, University of California, Los Angeles, California. I have read the
report, but I am not professionally qualified to comment on many of the
medical and epideraiological aspects of the Draft Protocol. It is my understanding that the Department of Health and Human Services (HHS) will provide
sufficient epidemiological expertise to comment on those segments of the report.
Dr. Spivey does make a good case, however, for the historical cohort design in
his proposed epidemiological protocol.
I am qualified to comment on two issues in the report, i.e., (1) Review "of
Environmental Behavior of Agent Orange, and (2) Exposure.
ExperimentalStudies
The section dealing with environmental studies is understandably brief and
should be so for a report: of th'is nature. The authors have done an acceptable
job of reviewing the major environmental issues. However, since the magnitude
of the literature is substantial, they have only treated it in a superficial
manner. Its major problem is that it presents a non-critical evaluation of
the literature. Consequently, it presents selected references without really
commenting on the relevance of such data to the proposed epidemiological study.
There are numerous typos in this section that should be corrected in the next
typing.
Exposure
.
^
I am deeply troubled by this aspect of the report. On page 43, the authors
correctly surmi.se, "We have not identified a mechanism which would document
actual exposure." Over the past year in our Committee, as well as the Agent
Orange Working Group in the White House, wo. have wrestled, frankly unsuccessfully, with trying to establish some mechanism for documenting exposure.
I recall clearly our meeting with the members of the National Academy of
Sciences and their comments regarding any proposed epidemiological study
on Agent Orange exposure, in Vietnam. Th2 take-home message was, "If we
can not scientifically validate and docur.cnt exposure, we csn not do a scientific epidemiological study." Although Spivey'a approach .suggests a mechanism
by which we might overcome this problem, I suspect we are justifiably due some
criticism for the grouping approach. I am now pursuaded that we will never
be able to do an epidemiology study on individual veterans per se, but must
examine military units serving in specific spray areas. There is now some
hope from recent DOD activities that we might be able to document some segments :
of the military population in Vietnam exposed to Agent Orange. Every effort then
must be m'ade to work closely with Mr. Christian and his associates in DOD in
meticulously reviewing records and films to establish some case for exposure.
;

�I believe Dr. Spivey and his associates hav.e made a start on identifying the
kind of epidemiology study needed and the information necessary to begin the •
project. I am not impressed, however,, that"we liave a working program that
could be used for the basis of funding a large epidemiological study mandated
by Congress. We have a long way to go, and a major problem is the quality of
the record of documented exposures that can be elicited from the DOD Record
I Center. I recommend we do not fund any additional feasibility studies until a
I thorough and comprehensive search and cataloging of available DOD records, films,
| and reports is completed.

&lt;v

I

.

H
As a concise detailed protocol this document is clearly only a preliminary
outline, primarily, as the authors stress repeatedly, because they have yet to
be able to assess the data sources on which the studies will depend. As a
result, the document consists mostly of extensive reviews of various aspects
of the Agent Orange/dioxin'issue, its history, and public reaction. It also
presents much material describing epidemiologic prinKJ^les and techniques.
The latter is well written, clear and concises a good statement of principles
and issues underlying the choice of cohort study approach.
Considering the lack of access as yet to data sources, the proposal as
presented seems quite reasonable: a retrospective cohort study, preceded by
some feasibility work and several preliminary studies using existing overall
veteran mortality/morbidity data. The cohort study itself would be a massive
effort, the details of which cannot be fleshed out until the investigators
examine first-hand the materials they will need to use. Until they get to
that point, however, one cannot expect them to produce a protocol document of
the sort we were able to construct for our birth defects/veterans study (given
the fact that data sources and methodology were all at hand and familiar to
us). Unfortunately, from our experiences with the "Smoky" cohort followup, I
would be pessimistic that the investigators will have easy and prompt access
to data or that tracing cohort members will go smoothly.

�</text>
                  </elementText>
                </elementTextContainer>
              </element>
            </elementContainer>
          </elementSet>
        </elementSetContainer>
      </file>
    </fileContainer>
    <collection collectionId="30">
      <elementSetContainer>
        <elementSet elementSetId="1">
          <name>Dublin Core</name>
          <description>The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.</description>
          <elementContainer>
            <element elementId="50">
              <name>Title</name>
              <description>A name given to the resource</description>
              <elementTextContainer>
                <elementText elementTextId="4687">
                  <text>Alvin L. Young Collection on Agent Orange</text>
                </elementText>
              </elementTextContainer>
            </element>
            <element elementId="41">
              <name>Description</name>
              <description>An account of the resource</description>
              <elementTextContainer>
                <elementText elementTextId="49809">
                  <text>&lt;p style="margin-top: -1em; line-height: 1.2em;"&gt;The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.&lt;/p&gt;&#13;
&lt;p&gt;For more about this collection, &lt;a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a"&gt;view the Agent Orange Exhibit.&lt;/a&gt;&lt;/p&gt;</text>
                </elementText>
              </elementTextContainer>
            </element>
          </elementContainer>
        </elementSet>
      </elementSetContainer>
    </collection>
    <itemType itemTypeId="1">
      <name>Text</name>
      <description>A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text.</description>
      <elementContainer>
        <element elementId="52">
          <name>Box</name>
          <description>The box containing the original item.</description>
          <elementTextContainer>
            <elementText elementTextId="19192">
              <text>062</text>
            </elementText>
          </elementTextContainer>
        </element>
        <element elementId="53">
          <name>Folder</name>
          <description>The folder containing the original item.</description>
          <elementTextContainer>
            <elementText elementTextId="19193">
              <text>1684</text>
            </elementText>
          </elementTextContainer>
        </element>
        <element elementId="54">
          <name>Series</name>
          <description>The series number of the original item.</description>
          <elementTextContainer>
            <elementText elementTextId="19195">
              <text>Series III Subseries III</text>
            </elementText>
          </elementTextContainer>
        </element>
      </elementContainer>
    </itemType>
    <elementSetContainer>
      <elementSet elementSetId="1">
        <name>Dublin Core</name>
        <description>The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.</description>
        <elementContainer>
          <element elementId="39">
            <name>Creator</name>
            <description>An entity primarily responsible for making the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19190">
                <text>Spivey, Gary H.</text>
              </elementText>
              <elementText elementTextId="19191">
                <text>Rogert Detels</text>
              </elementText>
            </elementTextContainer>
          </element>
          <element elementId="50">
            <name>Title</name>
            <description>A name given to the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19194">
                <text>Review of the Draft Protocol for Epidemiological Studies of Agent Orange</text>
              </elementText>
            </elementTextContainer>
          </element>
          <element elementId="49">
            <name>Subject</name>
            <description>The topic of the resource</description>
            <elementTextContainer>
              <elementText elementTextId="19196">
                <text>Agent Orange Exposure Study</text>
              </elementText>
              <elementText elementTextId="19197">
                <text>ground troops exposure</text>
              </elementText>
              <elementText elementTextId="19198">
                <text>cohort study</text>
              </elementText>
              <elementText elementTextId="19199">
                <text>AOWG</text>
              </elementText>
            </elementTextContainer>
          </element>
        </elementContainer>
      </elementSet>
    </elementSetContainer>
    <tagContainer>
      <tag tagId="1">
        <name>ao_seriesIII</name>
      </tag>
    </tagContainer>
  </item>
</itemContainer>
