EPA-HSRB-06-03

George Gray, Ph.D.

Science Advisor 

Office of the Science Advisor

1200 Pennsylvania Avenue, NW

Washington, DC 20460 

Subject:  June 27-30, 2006 EPA Human Studies Review Board Meeting Report

Dear Dr. Gray:

The United States Environmental Protection Agency (EPA or Agency)
requested the Human Studies Review Board (HSRB) to review scientific and
ethical issues addressing a human toxicity study involving one pesticide
active ingredient-chloropicrin; guidelines for conducting insect
repellant efficacy testing; protocols for conducting two insect
repellent efficacy studies; and protocols for conducting five
occupational handler exposure monitoring studies. At the Chair’s
request, the Board developed scientific and ethics criteria for new
protocols. The enclosed HSRB report addresses the Board’s response to
EPA charge questions for the Board’s consideration at its June 27-30,
2006 meeting.  A summary of the Board’s conclusions is provided below.

Chloropicrin

Scientific Consideration 

  

Ethical Considerations

There was not clear and convincing evidence that the conduct of the
research was fundamentally unethical (e.g., the research was intended to
seriously harm participants or failed to obtain informed consent).

There was not clear and convincing evidence that the conduct of the
study was significantly deficient relative to the ethical standards
prevailing when the study was conducted.

Insect Repellent Product Performance Efficacy Guidelines

Actions to Minimize Risks to Human Subjects

The consensus of the Board was that studies involving humans are
necessary to evaluate the efficacy of products to repel insects and
other arthropods. 

Risk identification and minimization is also essential.  In their
protocols, investigators should adequately identify risk to participants
and describe adequate steps they will take to minimize these risks.

Types of Toxicity Data That Should Be Generated

The consensus of the Board was that the minimum set of toxicity data
that should be routinely generated before an investigator conducts
repellent efficacy testing on human subjects with a new product is that
which will assure that subjects would not be at risk of permanent or
irreversible harm.  

Self–experimentation

It may not be a priori unethical or problematic from a scientific
perspective for a principal investigator to be a subject in his/her own
study IF:

The study was approved by an IRB in the same manner as was required for
most human subjects research;

Scientific issues: 

a. Principal investigator met all enrollment criteria;

b. The study was a well controlled trial with a justified sample size
adequate to answer the study question with statistical surety;

c. The principal investigator was one of many subjects, accounting for
normal human variability, and allowing results to be generalized to a
broad population; and

d. The outcome measure is objective and measured by another (blinded,
when possible) investigator;

A plan is in place to assure integrity and safety of the study while the
principal investigator was a subject

A plan is in place to ensure for study oversight if principal
investigator becomes incapacitated;

Participation of other research staff/employees should be allowed only
if the above criteria are satisfied and if issues of coercion/undue
influence can be addressed, which may or may not be possible;

The investigator justifies why he/she should be a research subject in
the study.

Negative Controls

The Agency should modify the guideline to say that negative controls
“may be” needed (instead of “are”) and that examples be given
both for when negative controls are needed and when they are not.  The
language on positive controls may also benefit from further expansion
and clarification.

Design of Studies to Support Assessment of Repellent Efficacy

The Board consensus was that the time to first confirmed bite, or the
time to first confirmed "intent to bite" (if ascertainable), has the
advantage of minimizing risk of vector-borne diseases.  However, for
some studies there is a statistical advantage for the use of relative
protection as an appropriate outcome measure. Since relative protection
procedures in field studies increases the risk of vector-borne diseases,
protocols must: (a) justify the level of risk by the probability and
social value of the benefits; (b) adequately identify all risks; (c)
present a description of adequate steps to minimize the risks; and (d)
provide consent materials that include information about the prevalence
and risks of any vector-borne diseases, consequences of contracting
disease, and alternative effective repellents outside of the research.

Minimum number of subject to evaluate the level of repellent efficacy 

It is critical that the proposed number of subjects be justified on the
basis of good research design. Because experiments to test effectiveness
of products to repel insect and tick bites are likely to vary in terms
of design, response variable, target population of interest, detectable
effect size and other important variables, requiring a specific minimum
sample size that guarantees sufficient accuracy in all cases might be
impractical. Instead, the guideline might require that registrants
present their own sample size calculations and that the methodology used
in the calculations be justified relative to the factors noted in the
bullet list above.

Compensating Research Subjects For Research-Related Injuries 

It is appropriate that sponsors of repellent efficacy research studies
should be required to assure that if a subject is injured as a result of
participating in a study, then the subject will not have to assume the
costs of medical care needed to treat such injuries.

Special Considerations in Informed Consent Materials

To comply with the human studies rule, consent information for
pesticides studies must include: (a) detailed information on the
procedure (e.g., number of insect bites or landings anticipated, nature
of apparatus or field context, length of time of exposure); (b) a clear
statement of the risks involved (e.g., discomfort from bites, risk of
vector-borne disease, medical consequences of the disease, treatments
available for the disease); (c) the voluntary nature of participation
(e.g., statements that eliminate the perception of coercion for students
or employees; specific instructions on how to signal desire to withdraw
from the study); (d) the fact that there was no immediate direct benefit
to the subject in participating as well as a description of alternative
available repellents; and (e) other steps outlined above. In addition,
informed consent information should be as detailed for experienced
subjects as for naïve subjects.

HSRB Protocol Criteria

Before the Board reviewed the presented human studies proposals, the
HSRB developed scientific and ethical criteria as a guide for its
evaluation of such studies.  Such criteria will be helpful for the
Agency, study investigators, and other members of the public to
understand the Board’s approach for the review of proposed human
studies.  

Study EMD-003 from Carroll-Loye Biological Research

	Scientific Considerations

The HSRB recognized three major limitations to the protocol as submitted
to the HSRB for review.  These limitations included: (1) the lack of a
clear rationale underlying the conduct of the study; (2) the lack of
identification and characterization of the formulations to be tested and
(3) the scientific design of the study.  Of these issues, the design of
the study was seen as the most significant shortcoming of the proposed
work.

Ethical Considerations

The Board concurred with the initial assessment of the Agency that the
study submitted for review by the Board failed to meet the ethical
requirements established in the Agency’s human studies rule (40CFR26).


The Board determined the proposed research described in these studies
did not comport with the applicable requirements of 40CFR26, subparts K
and Lsubpart K. The study documents submitted for review also failed to
comply with the requirements of 40CFR §26.1125, subpart M.  However,
the deficiencies noted, while significant, were not irreparable.

Study EMD-004 from Carroll-Loye Biological Research

Scientific Considerations

It was not clear whether new studies involving human subjects were
necessary; however, if the repellency had never been tested with North
American mosquitoes, the tests were probably necessary.

The potential benefits of the study were clear, i.e., that an effective
repellent would be available that would have either greater efficacy
and/or fewer drawbacks than what was currently approved. However,
empirical evidence or procedures to determine risks to subjects (e.g.,
of vector borne disease) were not adequate.

It was not clear if the stated numbers of subjects would be repeated in
both testing locations. The basis for the dose levels and formulations
was not provided. There were no controls with just the formulation
matrix without the repellent.

These issues would need to be addressed before the protocol could be
considered acceptable.

Ethical Considerations 

The Board concluded that the proposed research did not comport with the
applicable requirements of 40 CFR 26, subpart K.  

The proposed research does comport with 40CFR26 subpart L, as pregnant
women and children were excluded.  

Although the ethical concerns identified by the Board could be remedied,
there were sufficient questions raised about the adequacy of the
research design to cast doubt on whether the proposed research would
meet the criteria for IRB approval found under 40 CFR 26.1111(a) (1). 
In other words, absent a sound research design, any exposure of human
subjects to risk would be unnecessary and unjustifiable.

Occupational Handler Exposure Monitoring Studies

Scientific Considerations

The occupational handler exposure monitoring studies were components of
a large-scale exercise to create a contemporary database on occupational
exposure to agricultural pesticides. The undertaking is in itself likely
to be worthwhile in quantifying and improving our understanding of the
exposures and risks of pesticide handlers.  

The potential benefits are large and the risks appear to be relatively
modest. However, the materials supplied for HSRB review failed to deal
adequately with risks and benefits. None of these protocols can be
properly evaluated in regard to scientific validity because they lack:
(1) a developed rationale documenting the need for new data; (2) a clear
and appropriate plan for the handling of the data (including its
statistical analysis), and (3) an explanation of the uses to which the
data will be put. These points need to be addressed briefly at least in
each specific protocol and, more fully, in a separate and new
“governing document” that is not simply a generic description of the
planned activities.

Additional validation studies are recommended to determine the extent to
which dermal exposure measurements may underestimate true exposure.
Laboratory-based removal efficiency studies or field-based biomonitoring
studies could be conducted to achieve this goal. Such studies should be
published in the peer-reviewed literature. Broader participation of the
scientific community and of parties with a direct interest in the
database project, such as the labor community, would likely improve the
quality of the database and enhance the credibility of its use in risk
assessments.

The HSRB recommended that specific criteria for withdrawal from study
participation due to heat stress be included in these worker exposure
protocols, and that the protocols included a heat stress management
plan.  In addition, the length of each study should be truly
representative of a full workday, and each protocol should document the
basis for the proposed duration of the study.

The HSRB was gratified to receive the Agency’s response to its query
regarding the use of diazinon in the AHE37. It is the understanding of
the HSRB that the Agency would inform the AHETF that it needs to
identify a pesticide other than diazinon in this protocol to evaluate
exposures associated with open pour activities and applications using
open cabs, and that the Agency would ensure that future protocols comply
with the most current risk mitigation measures specified in IREDs and
REDs.

Ethical Considerations

The Board concurred with the initial assessment of the Agency that the
studies submitted for review failed to meet the ethical requirements
established in the 40CFR26. 

The Board determined the proposed research does not comport with the
applicable requirements of §40CFR26, subparts K and L subpart K.
However, the deficiencies noted, while significant, were not
irreparable. 

In conclusion, the EPA HSRB appreciated the opportunity to advise the
Agency on the scientific and ethical aspects of human studies research
and looks forward to future opportunities to continue advising the
Agency in this endeavor. 

Sincerely,

Celia Fisher, Ph.D. Chair

EPA Human Studies Review BoardNOTICE

This report has been written as part of the activities of the EPA Human
Studies Review Board, a Federal advisory committee providing advice,
information and recommendations on issues related to scientific and
ethical aspects of human subjects research.  This report has not been
reviewed for approval by the Agency and, hence, the contents of this
report do not necessarily represent the view and policies of the
Environmental Protection Agency, nor of other agencies in the Executive
Branch of the Federal government, nor does mention of trade names or
commercial product constitute a recommendation for use.  Further
information about the EPA Human Studies Review Board can be obtained
from its website at http://www.epa.gov/osa/hsrb/.  Interested persons
are invited to contact Paul Lewis, Designated Federal Officer, via
e-mail at lewis.paul@epa.gov.

	In preparing this document, the Board carefully considered all
information provided and presented by the Agency presenters, as well as
information presented by public commenters.  This document addresses the
information provided and presented within the structure of the charge by
the Agency.

United States Environmental Protection Agency Human Studies Review
Board

Chair

Celia B. Fisher, Ph.D., Marie Ward Doty Professor of Psychology,
Director, Center for Ethics Education, Fordham University, Department of
Psychology, Bronx, NY 

Vice Chair

William S. Brimijoin, Ph.D., Chair and Professor, Molecular Pharmacology
and Experimental Therapeutics, Mayo Foundation, Rochester, MN 

Members 

David C. Bellinger, Ph.D., Professor of Neurology, Harvard Medical
School, Professor in the Department of Environmental Health, Harvard
School of Public Health

Children's Hospital, Boston, MA  

Alicia Carriquiry, Ph.D., Professor, Department of Statistics, Iowa
State University

Snedecor Hall, Ames, IA 

Gary L. Chadwick, PharmD, MPH, CIP, Associate Provost, Director, Office
for Human Subjects Protection, University of Rochester, Rochester, NY 

Janice Chambers, Ph.D., D.A.B.T., William L. Giles Distinguished
Professor, Director, Center for Environmental Health Sciences, College
of Veterinary Medicine, Mississippi State University, Mississippi State,
MS 

Richard Fenske, Ph.D., MPH, Professor, Department of Environmental and
Occupational Health Sciences, University of Washington, Seattle WA 

Susan S. Fish, PharmD, MPH, Professor, Biostatistics & Epidemiology,
Boston University School of Public Health, Co-Director, MA in Clinical
Investigation, Boston University School of Medicine, Boston, MA 

Suzanne C. Fitzpatrick, Ph.D., DABT, Senior Science Policy Analyst,
Office of the Commissioner, Office of Science and Health Coordination,
U.S. Food and Drug Administration, Rockville, MD 

Kannan Krishnan, Ph.D., Professor, Département de santé
environnementale et santé au travail, Faculté de medicine, Université
de Montréal, Montréal, Canada *

KyungMann Kim, Ph.D., CCRP, Professor & Associate Chair, Department of
Biostatistics & Medical Informatics, School of Medicine and Public
Health, University of Wisconsin-Madison, Madison, WI  **

Michael D. Lebowitz, Ph.D., FCCP, Professor Emeritus of Medicine.
University of Arizona, Tucson, AZ 

Lois D. Lehman-McKeeman, Ph.D., Distinguished Research Fellow, Discovery
Toxicology, Bristol-Myers Squibb Company, Princeton, NJ  

Jerry A. Menikoff, M.D., Associate Professor of Law, Ethics & Medicine,
Director of the Institute for Bioethics, Law and Public Policy,
University of Kansas Medical Center, 

Kansas City, KS 

Robert Nelson, M.D., Ph.D., Associate Professor of Anesthesiology and
Critical Care, 

Department of Anesthesiology and Critical Care, University of
Pennsylvania School of Medicine, The Children's Hospital of
Philadelphia, Philadelphia, PA 19104

Sean Philpott, PhD, MS, Bioethics, Associate Director, Alden March
Bioethics Institute, Albany Medical Center, Albany, NY  12208-3479

Human Studies Review Board Staff

Paul I. Lewis, Ph.D., Designated Federal Officer, United States
Environmental Protection Agency, Washington, DC 

*  Recused from chloropicrin discussion and deliberation

**Not in attendance at the June 27-30, 2006 Public MeetingTABLE OF
CONTENTS

INTRODUCTION……………………………………………………
………………………….11

REVIEW
PROCESS………………………………………………………
…………………….12

CHARGE TO THE BOARD AND BOARD
RESPONSE……………………………………   14

1. 
Chloropicrin……………………………………………………
……………………………..14

2.  Insect Repellent Product Performance Testing
Guideline……………………………………21

3.  Insect Repellent Product Performance Efficacy
Studies……………………………………..43

4.  Study EMD-003 from Carroll-Loye Biological
Research……………………………………43

5.  Study EMD-004 from Carroll-Loye Biological
Research……………………………………47

6.  Occupational Handler Exposure Monitoring
Studies………………………………………   51

7. 
REFERENCES……………………………………………………
………………………….67

INTRODUCTION

On June 27-30, 2006, the United States Environmental Protection
Agency’s (EPA or Agency) Human Studies Review Board (HSRB) met to
address scientific and ethical issues surrounding a human toxicity study
involving one pesticide active ingredient-chloropicrin; guidelines for
conducting insect repellant efficacy testing; protocols for conducting
two insect repellent efficacy studies; and protocols for conducting five
occupational handler exposure monitoring studies.   

The Pesticide Registration Improvement Act (PRIA) requires that EPA
complete its decision-making process on certain types of applications to
register a pesticide product within specified amounts of time after
receiving the application for registration.  In addition, PRIA
established deadlines for EPA to complete “reregistration” of
pesticide active ingredients that are contained in pesticide products
initially registered before 1984.  Reregistration involves the
systematic reexamination of these older pesticides, applying
contemporary scientific and regulatory standards.  When a pesticide
active ingredient is approved for use on food, EPA combines
reregistration with the tolerance reassessment process mandated by the
Food Quality Protection Act of 1996 (FQPA).  

Chloropicrin is undergoing reevaluation in the reregistration process. 
As part of the review of the available toxicity data on chloropicrin,
EPA had identified a study involving intentional exposure of human
subjects which EPA intends to use in its risk assessment.  In accordance
with 40 CFR 26.1602, EPA sought HSRB review of this study.  

EPA regulates pesticides intended for use on skin to repel arthropod
pests.  As part of the application for registration of a new repellent,
EPA requires data to demonstrate that the product is effective.  The
Agency had developed a guideline for the conduct of such studies, and
presented it to the Board for comment.  The Agency had also received
protocols for two insect repellent efficacy studies, and as required by
the recently promulgated regulation, EPA is required to submit the
protocols to the HSRB for its review and comment.  See 40 CFR 26.1601.

In addition, EPA routinely considers the human health risks of
occupational handlers of pesticides in both its reregistration program
and as part of its review of an application for registration pending
under FIFRA and PRIA.  EPA has received five protocols for conducting
new research involving human subjects to collect data on the levels of
exposure received by people when mixing, loading, and applying
pesticides under various conditions.  In accordance with 40 CFR 26.1601,
EPA sought HSRB review of these proposed protocols.

For the human studies or guidelines under consideration, the Agency
provided the Board with the complete study report or associated
protocols and any supplements available to the Agency.  Similarly,
guideline documents were included with appropriate background
information.  Completed studies were assigned a unique identifier (e.g.,
the Master Record Identifier-MRID), which the Agency uses to manage
documents.  When a company submits multiple documents pertaining to a
single study, each document is typically assigned a unique tracking
number.

In addition, for each study, protocol or guideline to be evaluated, the
Agency provided a review of the ethical conduct.  Each ethics review
identified any deficiencies which were identified compared to
appropriate ethical standards.  EPA has intentionally deferred making a
final determination of whether the chloropicrin study satisfies the
ethical standards for acceptability in 40 CFR sections 26.1704 –
26.1706, pending the advice of the Board.

For most studies and protocols, the Agency develops documents, called
Data Evaluation Records (DERs), containing a scientific review.  The
Board was provided with one or more DERs for chloropicrin, the two
proposed insect repellent efficacy protocols, and each of the five
Agricultural Handlers Exposure Taskforce (AHETF) protocols.  DERs
contain summaries of the study design, methods and results, describe
potential deficiencies, and provide conclusions about the usefulness of
the study in risk assessment.  

In addition to the DERs, the Agency had prepared or included several
other background documents which address various elements of the issues
to be reviewed by the HSRB.  For example, for the AHETF protocols, a
number of types of documents had been provided including  transmittal
documents and the charge questions, general background information
pertaining to the manner in which the Agency completes exposure/risk
assessments, the AHETF protocols  and various documents that the AHETF
had developed related to the manner in which it intends to conduct
studies, the background documents related to the AHETF protocol review
by the Western Institutional Review Board of Olympia, Washington, and
the EPA science and ethics reviews of these protocols. 

  The Agency asked the HSRB to advise the Agency on a range of
scientific and ethics issues and on how proposed and completed studies
should be assessed against the provisions in 40 CFR 26 of EPA’s final
human studies rule.  This report transmits the HSRB’s comments and
recommendations from its June 27-30, 2006 meeting.        

REVIEW PROCESS

On June 27-30, 2006 the Board had a public face-to-face meeting in
Arlington, Virginia.  Advance notice of the meeting was published in the
Federal Register “Human Studies Review Board: Notice of Public Meeting
(71 Federal Register 32536 and 71 Federal Register 33747).  At the
public meeting, following welcoming remarks from Agency officials, Celia
B. Fisher, HRSB Chair, proposed a set of scientific and ethics criteria
consistent with the language of 71 Federal Register 6137 to guide Board
evaluation of completed studies.  The Chair’s scientific criteria
asked the Board to consider the following two questions: (1) did the
research design and implementation meet scientific standards and (2) did
the data generated by the study have implications for the Agency’s
Weight of the Evidence (WOE) review and, when applicable, aspects of the
risk assessment?  The Chair reviewed the Chair’s science criteria and
the Board’s criteria for scientific standards for human dosing studies
established at the Board’s May 2006 meeting.  The Chair’s ethics
criteria asked the Board to consider three questions: (1) did the study
fail to fully meet specific ethical standards prevalent at the time the
research was conducted; (2) was the conduct of the study fundamentally
unethical (i.e., specifically was there clear and convincing evidence
that the research was intended to seriously harm participants or failed
to obtain informed consent); and (3) was the conduct of the study
significantly deficient relative to the ethical standards prevailing at
the time (i.e., was there clear and convincing evidence that identified
deficiencies that could have resulted in serious harm based on knowledge
available at the time the study was conducted or the information
provided to participants could seriously impair informed consent).   

The Board then heard presentations from the Agency on the following
topics: scientific and ethical issues addressing a human toxicity study
involving one pesticide active ingredient-chloropicrin; guidelines for
conducting insect repellant efficacy testing; protocols for conducting
two insect repellent efficacy studies; and protocols for conducting five
occupational handler exposure monitoring studies. At the Chair’s
request the Board developed scientific and ethics criteria for new human
study protocols.

The Board heard oral public comments from the following individuals:

Chloropicrin

Robert Sielken, Ph.D., of Sielken and Associates and John Butala, Ph.D.
of Toxicology Consultants, Inc. on behalf of the Chloropicrin Task
Force.

Jennifer Sass, Ph.D. on behalf of the Natural Resources Defense Council.

Guidelines For Conducting Insect Repellant Efficacy Testing

Scott Carroll, Ph.D., on behalf of the University California at Davis
and Carroll-Loye Biological Research.

Protocols For Conducting Insect Repellent Efficacy Studies: Study
EMD-003 And Study EMD-004

Scott Carroll, Ph.D., on behalf of Carroll-Loye Biological Research.

Mr. Dan Giambattisto on behalf of EMD Chemicals, Inc.

Mr. Niketas Spero on behalf of Insect Control and Research, Inc.

Protocols For Conducting Occupational Handler Exposure Monitoring
Studies: Study AHE34, Study AHE36, Study AHE37, Study AHE38 And AHE42  

Victor Canez, Ph.D., Elliot Gordon, Ph.D., Mr. Curt Lunchick, and Mr.
Larry Smith on behalf of the Agricultural Exposure Handlers Task Force

Ms. Shelly Davis on behalf of Farmworker Justice Fund

In addition, the Board received written public comments from the
Agricultural Exposure Handlers Task Force, Carroll-Loye Biological
Research, the Farmworker Justice Fund,  the FMC Corporation, Toxicology
Consultants, Inc. and the Walter Reed Army Institute of Research, Vector
Control/Repellents Program.  

For their deliberations, the Board considered the materials presented at
the meeting, written public comments and Agency background documents
(e.g. pesticide human study, Agency data evaluation record (DER) of the
pesticide human study, weight of evidence review, ethics review,
pesticide human study protocol and Agency evaluation of the protocol).

CHARGE TO THE BOARD AND BOARD RESPONSE

Chloropicrin 

Charge to the Board

Chloropicrin is a non-selective soil fumigant whose primary toxic effect
is sensory irritation in which stimulated free nerve endings mediate
sensations and clinical signs in the nose, eyes, throat, and upper
respiratory tract.  Chloropicrin is a unique soil fumigant in that it is
also used as an indicator chemical or warning agent (2% or less by
weight in formulations).  The Agency is developing an assessment to
estimate inhalation risk to bystanders and workers from acute exposures
to chloropicrin.  

 	Scientific considerations 

The Agency’s “Weight of Evidence” (WOE) document and Data
Evaluation Records (DER) for chloropicrin described the study design of
the acute inhalation, human toxicity study.  The Agency had concluded
that the human toxicity study was appropriate for developing a point of
departure for extrapolation of inhalation risk to bystanders and workers
exposed to chloropicrin.  

Please comment on whether the study was sufficiently sound, from a
scientific perspective, to be used to estimate a safe level of
inhalation exposure to chloropicrin.  

Board Response to the Charge

Background of Study 

To determine a subject’s sensitivity for the detection and
characterization of feel to the human eye, nose, and/or throat produced
by chloropicrin vapors, as well as the odor threshold, healthy
volunteers (18 to 35 years of age, average 23 years) were exposed to a
range of vapor concentrations and exposure durations in a controlled
laboratory setting. The investigation consisted of three phases, very
brief exposures (Phase I) and more extensive exposures (Phases II and
III).  These phases are described in more detail below. 

	The study report cited Krieger (1996) as a review of the risks to
workers from exposure to chloropicrin in agricultural applications.  It
appeared that this reference was relied upon for basing concentration
and duration for the human sensory study. From this reference, a
time-weighted average of 0.1 ppm (100 ppb) was indicated to evoke no
response in humans. The report then indicated concentrations of 0.15 to
0.3 ppm would evoke concentration-dependent sensory detection via
chemesthesis, as well as reflex tearing and cough. Concentrations above
0.3 ppm would evoke an increasing degree of irritation. Odor was noted
as occurring at about 0.9 ppm.  The extended phases were focused on
concentrations of likely occupational relevance, both below and just
above 100 ppb, the ACGIH (American Conference of Governmental Industrial
Hygenists) threshold Limit Value (TLV) and OSHA Permissible Exposure
Limit (PEL). 

Statistical analyses including all subjects for each phase of the study
were provided in the study report. The EPA provided a logistical
regression when appropriate as well as an analysis for only those
subjects positively detecting chloropicrin for Phases I and II of the
study.   Additionally, the EPA provided a summary of the benchmark
concentration (BMC) analysis that was performed by TERA.

Phases of Study

Phase I: The objective of Phase I was the identification of chloropicrin
by odor (both nostrils, single sniff), eye feel (one eye, 25 seconds),
or nasal feel (one nostril, 7 seconds) at 356 ppb, 533 ppb, 800 ppb and
1200 ppb generated from a vapor delivery device. Phase I consisted of 62
subjects (32 male and 30 female) for odor and 63 subjects (32 male and
31 female) for eye feel. The same subjects participated for both odor
and eye feel. Confidence of feel was rated 1 to 5, with 1= very low, 3=
moderate, and 5= very high confidence. Severity of feel was not rated in
Phase I.   For Phase I, approximately 10% to 13% of subjects failed to
detect either odor or eye feel after momentary exposures to chloropicrin
over the range evaluated.  Approximately 13% (8 of 62) of subjects (5/30
female and 3/32 male) failed to detect the odor of chloropicrin over the
range of concentrations evaluated. Approximately 11% (7/ 63) (11%) of
subjects failed to detect eye feel (two male and five female) at any
concentration. The feel of chloropicrin in the nose was not a reliable
endpoint and was therefore dropped prematurely from the Phase I study by
the study director. The median concentration of all subjects for
detection of eye feel was 900 ppb, or 790 ppb for males and 1010 ppb for
females, although this gender difference was not significant. The median
for only those subjects detecting odor was 356 ppb while eye feel was
between 356 ppb and 533 ppb. 

Phase II: The purpose of Phase II was the detection of chloropicrin in
the eyes, nose, and/or throat during exposure to 50 ppb, 75 ppb, 100
ppb, or 150 ppb chloropicrin vapor in a walk-in chamber for 20-30
minutes. (Odor was not studied in Phase II).  A total of 62 subjects (32
male and 30 female) participated in Phase II. 12 of 30 female subjects
and 14 of 32 male subjects also participated in Phase I of the study.
Subjects responded “yes” for a positive feel or “no” for no
feel. A level of confidence to each event (eye, nose, throat) was also
recorded with 1= not certain, 2= moderately certain, and 3= very
certain. The first exposure in a day consisted of a known blank (air).
This exposure served to acclimate the subjects to the task in the
chamber. The subjects were asked to perform ratings as they would for
future blinded exposures. One female subject left the exposure chamber
after 16 minutes of chloropicrin at 75 ppb. An explanation for this
subject’s premature exit from the chamber was not provided.  At 150
ppb, this same subject along with another male in the chamber left the
chamber after 15 minutes.  On a separate day of testing, one female and
one male subject also left the exposure chamber after 15 minutes of
exposure to 150 ppb. Again, no explanation was given for these
subjects’ premature departure from the chamber. No subjects left the
chamber at 50 ppb or 100 ppb.  The results of Phase II indicated that
eye feel was more sensitive than either nose or throat feel. ANOVA
results provided in the study report indicated that concentration and
duration were significant (p<0.0001) for the eye response only.  As a
group, subjects differentiated 50 ppb chloropicrin in the eyes from the
blank after 20 minutes of exposure. Differentiation from blank occurred
after 5 minutes at 75 ppb, 3 minutes at 100 ppb, and 2 minutes at 150
ppb. There were no significant statistical interactions of response with
sex for the eyes, nose, or throat responses. On an individual basis,
binary detection indicators (yes/no) developed by the Agency were
combined by participant across dose levels. Using eye feel as a marker
of detection of the chemical, 20 of the 62 participants (32%) could not
detect chloropicrin at any concentration:12 of 30 (40%) females and 8 of
32 (25%) males failed to make progress toward eye feeling over a 30
minute period of exposure. In addition, 46/62 (74%) and 48/62 (77%)
subjects could not detect the chemical via the nose or throat,
respectively at any concentration, again indicating the greater
sensitivity of the eye. 

Phase III: The goal of Phase III was the detection of chloropicrin vapor
as evidenced by irritation to the eyes, nose and/or throat after 1 hour
(60 minute) exposures repeated over 4 consecutive days. Concentrations
tested included blank (air), 100 ppb, and 150 ppb. This phase included a
clinical exam of the eyes, nose and throat, as well as pulmonary
function testing with the outcome variable FEV1 (Forced Expiratory
Volume) and FVC (Forced Vital Capacity), rhinomanometry, and nasal
cytology. In addition, an assessment was performed based on ocular
cytology from samples of cells taken from the conjunctival membrane
inside the lower eyelid and from the concentration of exhaled nitric
oxide sampled from the lung (eNO) and nose (nNO). Subjects participated
in 3 cycles [(6 days per cycle) of 6 sessions, each beginning on Friday
and ending on the subsequent Friday] (no measurements taken on Saturday
or Sunday). Subjects remained in exposure chambers for 1 hour per
session on Monday through Thursday (4 consecutive days). The 3 cycles
included exposure to 100 ppb, 150 ppb, and just air (blank). The order
in which the subject was exposed to these concentrations was random to
prevent confounding. At least one week separated the end of one cycle of
exposures  and the beginning of another for each subject. Subjects rated
their symptoms in three setting: (1) severity of effect for eye, nose,
and throat while in the chamber (0= no symptom to 3= severe); (2) before
and after exposure in the chamber and at the beginning and end of each
week of exposures; and (3) at the beginning and end of a cycle of
exposure. Symptoms were rated using the Rhinconjunctivitis Quality of
Life Questionnaire (RQLQ), a series of 28 questions in seven domains,
where the subjects used a seven point scale from Not Troubled to Very
Troubled. The first two instruments referred to how the subject felt at
the time of rating, the RQLQ referred to how the subject felt over the
previous week. When in the exposure chamber, subjects rated symptoms (0
to 3) after 30 seconds, at 1 minute, and every minute until the end of
the exposure at 60 minutes. Every 10 minutes, study personnel read and
recorded the subjects’ blood oxygen saturation from a pulse oximeter
attached to the finger (data not included in report). A total of 15
males and 17 females participated in Phase III. Two females in Phase III
also participated in Phases I and II. One male in Phase III also
participated in Phases I and II and one male in Phase III participated
in Phase II only. 

Results of the Study 

For Phase II, one female subject left at 75 ppb and again at 150 ppb
with another male. On a separate testing day, one female and male left
the chamber prematurely at 150 ppb. 38%  (8 males and 8 females) of
subjects detected chloropicrin initially at 50 ppb and consistently up
to 150 ppb.  Subjects gave higher ratings to symptoms in the eye than to
those in the nose and throat.  Subjects gave nominally slightly higher
ratings in the nose than in the throat, but expressed no symptoms of
consequence at either site.  There was no indication of intensification
of symptoms based on subject scoring for any parameter on the
consecutive days of exposure. For the eye, the study report ANOVA
indicated Level of Exposure (p<0.001), and interaction of Level of 
Exposure by Duration of  Exposure was significant (p<0.001).  An ANOVA
also revealed an effect of Level of Exposure by Day (p<0.02). As a group
with all subjects included (even those not feeling), the analysis
provided in the study indicated the average rating of eye irritation at
100 ppb reached approximately 0.5 (1=mild) with 30 minutes to reach
steady state, which remained until the final minutes and then sometimes
regressed.  At 150 ppb, the average rating of eye irritation reached 1
(mild, symptom present, but minimal awareness, easily tolerated) with 20
minutes to steady state until fading slightly in the final minutes.  

On an individual level, the severity of ocular irritation reported by
subjects in Phase III varied from no symptoms to severe at both 100 ppb
and 150 ppb. Five of 17 females (29%) and 7 of 15 males (47%) rated no
eye irritation at 100 ppb while 3 of 17 females (18%) and 5 of 15 males
(33%) rated no eye irritation at 150 ppb. Nasal and throat irritation
was never reported above a “2" and mainly consisted of “0" or “1".
Scores of severe “3” ocular irritation were sporadic during the
first 30 minutes of exposure in 2 females and in 4 males at 100 ppb. 

The second half of the exposure to 100 ppb (31-60 minutes) revealed a
more consistent response in ocular severity (in 3 females and 5 males).
“Severe” (grade 1) was defined as a symptom that was hard to
tolerate and that could interfere with activities of daily living or
sleeping. At 150 ppb, 4 females and 3 males reported consistent severe
eye irritation beginning as early as 8 to 9 minutes of exposure until
the end of exposure at 60 minutes.  Moderate (grade 2) eye irritation
was also reported sporadically during the first 30 minutes by the same
individuals reporting severe eye irritation but with a more consistent
response in moderate eye irritation during the second half (31-60
minutes) of exposure. Two additional females and two additional males
reported moderate eye irritation during the second half of exposure that
did not report eye irritation during the first half of exposure. 
Results for the daily measurements (Cochran Q test) provided in the
study indicated the number of times a rating post-exposure exceeded a
rating pre-exposure for nasal congestion was not significant (Q=0.75)
while eye irritation (redness) was significant (Q=28.8, p<0.001).  Nasal
congestion and ocular erythema (redness) occurred more than the trivial
frequency. However, according to the report, the ocular irritation did
not translate into more prominent redness.  No biologically significant
changes were observed for the lower respiratory variables. 

For the lower respiratory variables (FVC, FEV1, eNO), ANOVA analysis
from the study report indicated a significant interaction of Level by
Order (p<0.05) for FVC (Forced Vital Capacity), with only 3% variation
in FVC.  FEV1 (Forced Expiratory Volume), averaged 93.6% before exposure
and 93.7% after exposure. The variation spanned 3% with no statistical
significance achieved.  Exhaled nitric oxide by the lungs (eNO) equaled
37.8 before exposure and 39.2 after exposure with no significance
achieved. Sex was not significant for any of interactions of the three
lower respiratory variables. Two upper respiratory alterations, nNO
(nasal nitric oxide) and flow, were observed for one-hour exposures that
occurred only day by day.  For the upper respiratory variables (nNO,
inspiratory flow, expiratory flow), nNO was significant for Level of
Exposure by Order of Exposure with 399 ppb before exposure and 425 ppb
after exposure (p=0.012).  Level of Exposure by Order of Exposure by Day
was not significant.  nNO increased 1% after exposure to blank, 10%
after exposure to 100 ppb, and 8% after exposure to 150 ppb.  The effect
of nNO did not continue from one day to the next.  Inspiratory flow and
expiratory flow equaled 450 and 415 mL/sec, respectively, before
exposure and 435 and 406 mL/sec, respectively, after exposure. 

Chloropicrin had a differential effect on flow.  Level of Exposure by
Order of Exposure  was nearly significant (p=0.087).  However, Level of
Exposure by Order of Exposure by Day was not significant. Flow decreased
2% after exposure to blank and increased 2% after exposure to 100 ppb
chloropicrin, however, flow decreased by 8% after exposure to 150 ppb. 
Sex was not significant in any of the relevant interactions for the
upper respiratory variables.  Physiological effects such as changes in
nNO and flow rate may indicate signs of nasal congestion and
engorgement. 

 	

Cell types and cell numbers from the Rhinoprobe samples were
approximately the same at the end of each cycle as at the beginning. 
For the RQLQ questionnaire results, nasal congestion was the only
parameter that reached a level where more than half of the subjects gave
a response above zero.  53% of subjects reported a non-zero response to
congestion after 4 days of exposure to the blank vs. 41% and 34% after
exposures to 150 ppb and 100 ppb, respectively. The average ratings
equaled 0.53, 0.34, and 0.41 for the blank, 100 ppb, and 150 ppb,
respectively, where a rating of 1 signified hardly troubled at all.
Watery eyes, sore eyes, and swollen eyes were scored higher by subjects
after exposure to either 100 or 150 ppb chloropicrin than to the blank. 
The Q test revealed significance for the sore eyes only (p<0.05). The
highest rating given after exposure to swollen eyes was 0.47. 

The LOAEL was determined to be 100 ppb, the lowest concentration tested,
based on eye irritation, increased nasal nitric oxide (nNO), and
differential effect on inspiratory and expiratory flow. A NOAEL was not
established in Phase III.

Critique of the Study

Strengths:

This was an excellent scientific study of eye, upper and lower
respiratory irritant effects at various concentrations over various
short term (i.e., acute effects) time periods.  The objective and
subjective measurements and the statistics were reasonable.  Most
importantly, Phase III of the study provided evidence of upper airway
(nasal) respiratory effects and established a LOAEL.  

Weaknesses:

Concentrations below 100 ppb were not investigated in Phase III so as to
compare with results from Phase II. 

HSRB Consensus and Rationale

Charge to the Board

Ethical considerations  

The Agency requests that the Board provide comment on the following:

a. Was there clear and convincing evidence that the conduct of the Cain
study was fundamentally unethical?

b. Was there clear and convincing evidence that the conduct of the study
was significantly deficient relative to the ethical standards prevailing
at the time the research was conducted?

Board Response to the Charge

The Cain study was conducted from 2002 through 2004. The study was
performed in La Jolla, California by researchers at the Chemosensory
Perception Laboratory of the University of California, San Diego. The
study sponsor was Chloropicrin Manufacturers Task Force, whose mailing
address is in care of Steptoe & Johnson, LLP, 1330 Connecticut Avenue,
NW, Washington, D.C. The documents provided by the sponsor specifically
state that the research was to be conducted with the approval of an
Institutional Review Board (IRB) at the University of California, San
Diego, and in compliance with the Human Subject’s Bill of Rights (a
provision of California law). The study was in fact reviewed and
approved by an IRB at that university, and the university had provided
documentation that it held a Federalwide Assurance with the Department
of Health and Human Services. The documentation provided by the
university’s Human Research Protections Program indicated that it
reviewed this study pursuant to the standards of the Common Rule (45 CFR
46, Subpart A) and determined it to be in compliance with that Rule.

Critique of Study

The Board concurred with the factual observations of the strengths and
weaknesses of the Cain study, as detailed by the EPA (USEPA 2006a).  The
Board wanted to comment on several specific aspects of the study:

1. The consent forms stated that chloropicrin “is used commonly to
fumigate fields for planting and as a warning agent in structural
fumigation.” It made no mention of prior uses of this compound for the
intentional purpose of harming and even killing people. As noted in one
of the documents supplied by the EPA (Prentiss 1937), chloropicrin
“appears to have been the most widely used combat gas” in World War
I. That reference further notes that “as a war gas [it] has a number
of desirable offensive properties,” and like phosgene gas, “it is a
lethal compound.” Some members of the Board concluded that in the
context of being asked to participate in a study to determine more
information about the harmful effects of this gas on human beings, the
subjects should have been informed about the historical use of
chloropicrin as a war gas. Others concluded that the dose levels and
other conditions were significantly distinct from the war use that
reference to those usages were not necessary for an informed
participation decision.

2. The consent forms, in describing the likely risks of participating in
the study, noted that “[e]xposure to chloropicrin in amounts greater
than anticipated in the studies have resulted in temporary tearing and
painful stinging eyes and nausea and vomiting that are completely
reversible after the exposure.” Some members of the Board believed
that this statement was inaccurate in describing the possible risks of
exposure to “greater” amounts of choloropicrin (which, as noted in
item 1 above, if in a sufficiently high dose, is well known to cause
substantial permanent effects, up to and including death). Those members
concluded that the consent form should have been more truthful in
describing the possible consequences of exposure to high doses of
chloropicrin (though it could also have explained why it would not be
the case that the subjects could ever end up being exposed to such high
doses as a result of participation in the study).

3. The consent form, in describing the purpose of the study, stated that
it was “intended to provide information regarding safe levels of
exposure.” This language might suggest to many prospective subjects
that the study was being conducted to see if it is important to create
increased restrictions on the use of this compound.  Some members of the
Board believed that the consent form should have explicitly stated that
this study was unlikely to lead to increased restrictions and, in fact,
its results, if they led to any regulatory changes, would more likely be
used to allow greater exposures of people to chloropicrin.  

HSRB Consensus and Rationale

The Board concluded that:

There was not clear and convincing evidence that the conduct of the
research was fundamentally unethical (e.g., the research was intended to
seriously harm participants or failed to obtain informed consent).

	There was not clear and convincing evidence that the conduct of the
study was significantly deficient relative to the ethical standards
prevailing when the study was conducted.

	The Board based these two determinations on its conclusion that this
study, based on the evidence presented, deviated from, but was not
significantly deficient relative to, the ethical standards prevailing
when the study was conducted.

Insect Repellent Product Performance Testing Guideline

Charge to the Board

	The U.S. EPA Office of Pesticide Programs requested that the HSRB
review and comment on the draft “Product Performance of Skin-Applied
Repellents of Insects and Other Arthropods” Testing Guideline in order
to determine what changes, if any, are necessary for the guideline to be
made consistent with the requirements for protection of human research
subjects set forth in 40 CFR part 26.  Below is a list of questions that
focus on these topics.

a. What actions should an investigator routinely take to minimize the
risks to human subjects exposed during laboratory and field research on
the efficacy of repellents?  

Board Response to the Charge

The Board began its review by emphasizing that human studies are
essential to assess the efficacy of repelling insects and other
arthropods. The only way to determine if these repellents are effective
is to test them on or near humans, since animals would not have the same
level of attractiveness to the arthropods as humans will.  

The following comprehensive list of conditions should be considered by
the investigator in order to minimize risks to human subjects. The
Board’s response to the Agency’s charge  focused primarily on the
ethical considerations and only secondarily on the toxicity data base
per se.  

(1) IRB approval is required before initiating any human exposures.

(2) Initial human tests should be conducted in a laboratory setting,
using insects and other arthropods which are known to be disease-free.

(3) Healthy volunteers should be selected who are not sensitive to
chemical reactions or drug/cosmetic allergies, and not allergic or
overly sensitive to arthropod bites and stings.  The subjects selected
should not be known to experience any adverse drug reactions or
allergies to other substances or toxins, and the selected subjects
should not be taking any drugs which might elicit an adverse drug
reaction (if the predictions from studies of metabolism of the repellent
suggest that drug-chemical interactions might occur at the level of
metabolism).

(4) If no known information about exposures in humans is available,
testing should begin with the low dose levels and if no adverse
reactions are displayed, rise gradually to the level of exposure
anticipated to be used in humans.  The formulations should be the same
as that expected to be applied on human skin.  If another formulation is
anticipated, such as a coil, then it should also be tested in laboratory
experiments.

(5) Although insect repellents would not be expected to yield adverse
effects, test subjects should be under close observation by an observer
trained to detect,  as well as listen to the subject, for any adverse
reactions.  If such adverse reactions are observed, this would be
grounds for terminating the exposure as soon and as completely as
possible.  

(6) The laboratory results would need to show a substantial likelihood
of repellency before field tests should be initiated because of the
possibility of subject exposure to disease-carrying arthropods.

(7) The field region for tests should have as low as possible incidence
of known disease-agent infested insects or arthropods (prior trapping
and microbial assays should document this minimal risk).

(8) The lowest possible number of untreated controls needed to ensure
scientific validity should be used in field tests.

(9) The overall toxicity, in terms of both the toxic effects and the
levels at which these toxic effects occur, should be determined from the
existing animal data base.  The search on existing animal data should
include acute, chronic, reproductive, eye and skin irritation and dermal
sensitization, so that the most reliable information on potential human
adverse effects is known.

(10) Any human data from controlled or inadvertent exposures, or from
routine uses in the past or in other countries should be accumulated. 
This information should be analyzed for evidence of direct toxic effects
or any adverse side effects, including allergic or sensitization
reactions.

(11) The test compound should be compared to data bases on similar
chemical classes of compounds so that educated predictions can be made
of types of toxicity that might be elicited in humans, the likelihood of
allergic or sensitization reactions, the likely disposition and
pharmacokinetics of the compound, including absorption, metabolism and
clearance.  In addition, the metabolism of the compound should be known
from in vitro tests using human liver samples, in order to predict the
toxicity or lack thereof of probable metabolites and to predict the
enzymes involved in the compound’s major routes of metabolism.  This
information on metabolism would be useful to predict any likely
interactions with drugs that an individual might be taking.

HSRB Consensus and Rationale

The consensus of the Board was that studies involving humans are
necessary to evaluate the efficacy of products to repel insects and
other arthropods. Risk identification and minimization are also
essential.  In their protocols, investigators should adequately identify
risk to participants and describe adequate steps they would take to
minimize these risks.

Charge to the Board

b. What types of toxicity data should be routinely generated before an
investigator conducts repellent efficacy testing on human subjects with
a new product?

Board Response to the Charge

In response to the question, the Board proposed a set of data that would
meet this requirement : 

The initial evaluation of a compound should include an analysis of
chemical structure that emphasizes the detection of possible adverse
effects.  The analysis can be informed by comparisons with repellents of
similar chemotypes for which toxicity data exist.  A variety of computer
applications and predictive models should be used to predict potential
alerts for metabolic activation, target organ toxicity or mutagenesis. 
This type of evaluation can include the comparison to similar chemotypes
of repellents for which toxicity data exist.

Acute (single dose) toxicity studies should be conducted, with emphasis
on the intended route(s) of exposure to the chemical.

Dermal and ocular irritation should be evaluated.  Characteristics of
the potential for dermal sensitization and nasal-pharyngeal
sensitization or triggering should be considered.

Absorption of the compound after administration by the route of intended
exposure, most likely dermal, should be determined.  This can be done in
laboratory animals, but may also be done using in vitro assessments of
percutaneous absorption in human skin or human skin surrogates.  If
conducted in animals, the study should include an assessment of the
routes of elimination of the compound.

The mutagenic and clastogenic potential of the compound should be
determined.  At this stage, this analysis could be an abbreviated
battery of in vitro genetic toxicology tests.

Some data on toxicity in a repeat dosing paradigm should be generated. 
This is particularly important if the compound is available
systemically.

If the compound is available systemically, its metabolic fate should be
investigated and it should be determined whether humans are likely to
metabolize the compound in a manner that is qualitatively or
quantitatively different from laboratory animals is recommended.

It should be noted that, although the Board has made these
recommendations for toxicity data, it did not specify the precise
methods by which the data set listed above should be generated.  The use
of animal models, validated in vitro methods or robust predictive tools
could be used in combination to generate the recommended data set. 
Furthermore, if an investigator desires to carry out multiple exposures
of a given compound in human subjects, then additional data, with
emphasis on subacute and/or subchronic toxicity and the assessment of
reproductive hazard, should be included in the toxicology evaluation.

In addition to the toxicity data summarized above, information on the
mode of action, potency and projected human dose is useful adjunct
information for assessing any potential risk associated with human
exposure.  Dose selection for efficacy studies in humans should be
justified on the basis of animal toxicity studies and/or other relevant
data (e.g., from in pharmaco-kinetic computer modeling, in vitro
studies, and human case series).  Present guidelines specify that the
amount applied should be up to the typical maximum dose applied by
consumers, with recognition that some clarification or comparison with
the toxicology benchmarks from animal studies will aid in dose selection
(e.g., lowest NOAEL from sub-chronic studies) and protect human health. 


HSRB Consensus and Rationale

The consensus of the Board was that the minimum set of toxicity data (as
delinieated above) that should be routinely generated before an
investigator conducts repellent efficacy testing on human subjects with
a new product is that which will assure that subjects would not be at
risk of permanent or irreversible harm.  

Charge to the Board

c. In private and university research laboratories, investigators
themselves have sometimes served as research subjects when assessing
chemicals for insect repellent activity. What scientific and ethical
issues would such a practice raise?  Under what conditions, if any,
would such a practice be acceptable?  

Board Response to the Charge

The topic of self-experimentation has been discussed and debated for
many years, and the debate is likely to continue. The scientific and
ethical issues presented in the assessment of insect repellents are not
different from the issues of self-experimentation in clinical research
in general. There is not a clear overarching answer; each study may
present a different situation. 

Arguments in favor of self-experimentation

There is a long and noble history of investigators experimenting on
themselves. Experiments on yellow fever, pernicious anemia, morphine and
cocaine as local anesthetics, H. pylori as the causative agent for
gastric ulcers, and many others have been instances of researchers using
themselves as research subjects. 

The Nuremberg Code, written in 1947 as part of the criminal trials of
the Nazi doctors, states “5. No experiment should be conducted where
there is an a priori reason to believe that death or disabling injury
will occur; except, perhaps, in those experiments where the experimental
physicians also serve as subjects” (USGPO, 1949).  If a researcher is
not willing to assume the risk of harm from research participation, how
can that same researcher ask anyone else to assume that same risk? 

Ethical research in compliance with 40 CFR 26 requires voluntary
informed consent.  Who better understands those risks than the
researcher? Who best understands the societal or scientific benefits of
the knowledge to be gained from the experiment?  There is no chance of
misunderstanding information in the consent process. 

If the ethical and scientific arguments against self-experimentation can
be addressed adequately, then self-experimentation seems quite
reasonable.

Arguments against self-experimentation

Ethical Considerations

Many have objected to self-experimentation on scientific and ethical
bases but their arguments can all be addressed with proper planning and
conduct of the research protocol. For this reason, self-experimentation
is not per se unethical or scientifically flawed per se. 

	One argument against self-experimentation is that researchers may take
unreasonable risks with their own health due to a blinding belief in the
importance of the research question, as well as and a personal
incentives such as of career advancement.  Addressing this concern
properly requires independent review of the study in order to ensure
that the risks are reasonable in relationship to the potential benefits
of the research. Therefore, the study must be IRB approved prior to its
conduct. As part of its review, the IRB must assure that ”risks to
subjects are reasonable in relation to anticipated benefits, if any, to
subjects, and the importance of the knowledge that may reasonably be
expected to result” as stated at 40CFR26.1111(a)(2).

 

Potential coercion of co-investigators and research staff by the
principal investigator is another area of concern. For this reason,
self-experimentation should be limited to the principal investigator in
most circumstances. Co-investigators and research staff (junior members
of the research team) should not be enrolled in a study if the principal
investigator has power or authority over them in the research setting or
in any other setting (e.g., classroom or other work environment). 
Situations such as these can lead to coercion or undue influence on
subordinates to participate in the research, and should be avoided,
except when there is an IRB approved protocol that would and allow for
truly voluntary participation. 

Scientific Considerations

There are many scientific issues that must be addressed in order for
self-experimentation to produce scientifically sound data that would be
useful and generalizable at the end of the experiment. 

One issue involves the type of outcome measure used in the research. If
the outcome measure is a subjective one, then the expectation of the
self-experimenter is likely to influence the results. This bias may lead
to an incorrect study conclusion.  To address this problem,
self-experimentation should occur only in research protocols with
objective outcome measures.  The investigator-subject should not assess
their own outcomes.  The outcome assessor should be blinded to the
subject’s identity, if possible.  In addition, the burden of proof is
on the principal investigator to demonstrate how their participation
does not introduce bias into the study results. 

Oversight of the study is another issue of concern in
self-experimentation. Since the principal investigator is responsible
for study oversight, this oversight can be compromised during the time
that the principal investigator is a research subject. Can the
experiment be safely completed, for example, if the investigator became
incapacitated while a research subject? This objection can be addressed
by identifying the person responsible for study oversight while the
principal investigator is a subject and if the principal investigator
becomes unable to resume study responsibilities. 

Many of the stories of self-experimentation in the history of medicine
have used a sample of one; the researcher was the only subject. These
studies thus lacked proper controls and did not account for
inter-individual variability.  Such studies were poorly designed to
answer a research study question with rigorous methodology. To address
this, a well-written protocol is required, which must have a sample size
that is adequate to answer the study question being asked. 

Concern also has been raised about whether investigators are thorough in
their evaluation of whether they meet all of the study’s inclusion and
exclusion criteria.  If researchers are convinced that they should be
and really wants to be subjects, they might not perform all screening
tests that are required by the protocol. This issue can be easily
addressed by having another investigator perform and assess the
screening results.  The principal investigator’s eligibility to
participate in the study should be assessed independently by someone
outside the research team, to avoid potential coercive influence of the
principal investigator on the sub-investigator. 

HSRB Consensus and Rationale

It may not be a priori unethical or scientifically problematic for a
principal investigator to be a subject in his/her own study IF:

1. The study was approved by an by an IRB in the same manner as was
required for most human subjects research;

2. The following scientific issues were addressed: 

a. Principal investigator met all enrollment criteria;

b. The study was a well controlled trial with a justified sample size
adequate to answer the study question with statistical surety
(Occasionally a study with a small sample size may be scientifically and
ethically appropriate if it is a pilot or feasibility study. However,
justification for the sample size chosen is still necessary, although
such justification may not be a statistical one.  In such a situation,
(c) does not apply);

c. The principal investigator is one of many subjects, accounting for
normal human variability, and allowing results to be generalized to a
broad population; and

d. The outcome measure is objective and measured by another (blinded,
when possible) investigator;

3. A plan is in place to ensure the integrity and safety of the study
while the principal investigator was a subject.

4. A plan is in place to ensure for study oversight if the principal
investigator becomes incapacitated;

5. Participation of other research staff and employees is prohibited
except in those cases where issues of coercion/undue influence can be
addressed, which may or may not be possible; and 

6. The investigator justifies why he/she should be a research subject in
the study.

Charge to the Board

d. Please comment on the scientific and ethical issues arising from the
use of (or decision not to use) negative controls groups in repellent
efficacy studies, in both laboratory and field studies.

Board Response to the Charge

Scientific Considerations

Negative controls (i.e., untreated/unprotected) are used in repellant
studies to show “biting pressure”.  This can be categorized as
sufficient, insufficient, or it can be quantitatively measured
(bites/minute over time).  Negative controls are also used in field
studies to confirm effectiveness that is shown in laboratory studies,

The use of a control group has been an essential characteristic of
repellent efficacy studies conducted in the laboratory, because a
comparison of the data from the treatment and control groups shows  a
measure of efficacy. The use of negative control groups in laboratory
studies appears to be a safe practice, since the insects involved are
known to be disease-free.  In contrast, the uncontrolled nature of field
studies means that the same assurances cannot be provided to
participants.  Because negative controls are not exposed to the
pesticidal active ingredient, there is no risk of toxicity from the
chemical.  The risk of harm and discomfort for subjects is primarily of
two types.  In laboratory and field studies, there is the discomfort of
the bite itself, which might include minor pain, itching and swelling. 
The discomfort experienced by humans is variable, some having negligible
reaction, others having a definite allergic response.  In addition to
this risk, field trials have the added risk of subjects acquiring a
vector-borne infection.  Fortunately, field procedures, such as capture
of insects just prior to biting can reduce such risk substantially.  In
addition, because negative controls are not exposed to the pesticidal
active ingredient, there is no risk of toxicity from the chemical.

Nonetheless, the Board failed to reach consensus regarding negative
control groups in  field studies of repellent effectiveness.  The basic
scientific justification for such controls is to confirm that "biting
pressure" exists.  If that is the only purpose, a single negative
control may suffice.  As some Board members suggested, biting pressure
might even be established through trapping or other methods that did not
involve an unprotected human subject.  Other Board members recognized,
however, that it may be important to establish a particular level of
biting pressure in order to compare the extent and duration of
repellency from trial to trial and compound to compound.  In fact, that
condition appears to be essential for accurate product labeling.  In any
case, since even one unprotected human subject could be at risk of
vector-borne disease, the use of negative control groups should not be a
default component in the design of repellency studies.  Instead it
should be justified in each protocol in which it is proposed.  

Ethical Considerations

By minimizing risk in the laboratory studies (e.g. screening for past
sensitive reactions, captive breeding of infection-free insects, and
mechanical aspiration at bite), the use of negative controls in the
laboratory should not be considered ethically problematic as long as it
is scientifically justified.  Steps can, and should be taken to minimize
risk in field studies. However, the risk of a significant life altering
infection can never be reduced to zero.  Thus, the potential benefit
from such studies must justify this risk.  The science must be sound and
alternative approaches - such as live trapping or laboratory studies -
must be shown to be inadequate.  The consent process must be truly
informed and subjects must be volunteers with the full right of
withdrawal.  These issues must be specifically and completely addressed
in the study protocol.

HSRB Consensus and Rationale

The HSRB suggested that the Agency modify the guideline to say that
negative controls “may be” needed (instead of “are”) and that
examples be given both for when negative controls are needed and when
they are not.  The language on positive controls may also benefit from
further expansion and clarification.

Charge to the Board

e. Please comment on the scientific and ethical issues raised by the
design of studies to collect data sufficient to support assessment of
repellent efficacy using the two different efficacy metrics: time to
first confirmed bite (TFCB), and time providing x% protection of treated
subjects from bites relative to untreated controls (RP). 

Board Response to the Charge

The distinction between efficacy and effectiveness is useful in
answering the questions about the Insect Repellent Product Performance
Testing Guideline.  Although the efficacy of a repellent can be
established using laboratory techniques, the effectiveness of a
repellent can only be established in the field under actual use
conditions.

Scientific Considerations

A particular study design can either minimize risk to all subjects
enrolled in a study (for example by using only laboratory mosquitoes or
ticks to eliminate the possibility of vector-borne diseases, excluding
those who might adversely react to the insect bites), and/or minimize
overall risks by reducing the number of “at risk” subjects to the
lowest number possible while maintaining scientific integrity,
pretesting insects to confirm probable absence of vector borne diseases,
and/or utilizing techniques to remove insects prior to bite when
feasible. Risk minimization strategies will depend upon knowledge of
variability in subject attractiveness, the effectiveness of the
repellent, the interaction of biting pressure to insect hunger and
subject attractiveness, characteristics of the test environment, and the
scientific reliability of generalizing insect performance from the lab
to the field.

A question was raised about the provision of prophylactic antibiotics or
use of a preventative vaccine in order to minimize risks.  The
difficulty with this approach is that the effectiveness of these
interventions would need to be established, and the dangers associated
with such treatments would become part of the overall assessment of
whether the risks of the research are offset by the importance of the
knowledge to be achieved.  The measurement of pre-exposure and post
exposure antibodies, unless done in a context to only include subjects
who are immune to the vector borne disease of concern, does not minimize
risk other than documenting the presence of the subject's immune
response.  Overall, the best approach is to choose a study design that
either eliminates or minimizes the risk of vector-borne diseases.

Ethical Considerations

	The time to first confirmed bite, or the time to first confirmed
"intent to bite" (if ascertainable), has the ethical advantage of
minimizing the risk of exposure to vector-borne diseases.  However,
based on the background materials for the meeting, the use of relative
protection can be an appropriate outcome measure based on statistical
advantage.  Relative protection could thus would be an appropriate
outcome measure for a laboratory based efficacy study.  As long as there
was a sufficient "biting pressure", one could then do a measure of the
duration of relative protection for protected subjects in a field study
and compare it to laboratory based measurements.  This could also be
done using time to first confirmed bite.

Participation in insect repellent research offers no direct benefit to
subjects when their exposure to insects or arthropods is for the purpose
of the study given the presence of existing insect repellents on the
market.  As such, the sponsors of repellent efficacy research are
obligated to provide insurance to cover possible future medical costs
that result from injury or illness experienced by the subjects as a
consequence of their participation in the research.  It is less clear
whether sponsors would have an obligation to provide for lost income in
such instances. As noted previously protocols must justify the level of
risk by the probability and social value of the benefits, adequately
identify all risks, and present a description of adequate steps to
minimize the risks.

	The informed consent materials also must provide information about the
prevalence and risks of any vector borne diseases (if applicable), the
consequences of acquiring such a disease as a result of the research,
and the availability of effective insect repellents outside of the
research.

HSRB Consensus and Rationale

The Board consensus was that the time to first confirmed bite, or the
time to first confirmed "intent to bite" (if ascertainable), has the
advantage of minimizing risk of vector-borne diseases.  However, for
some studies there is a statistical advantage for the use of relative
protection as an appropriate outcome measure. Since relative protection
procedures in field studies increases the risk of vector-borne diseases,
protocols must: (a) justify the level of risk by the probability and
social value of the benefits; (b) adequately identify all risks; (c)
present a description of adequate steps to minimize the risks; and (d)
provide consent materials that include information about the prevalence
and risks of any vector-borne diseases, consequences of contracting
disease, and alternative effective repellents outside of the research.

Charge to the Board

f. Please comment on appropriate approaches for estimating the minimum
number of subjects needed to evaluate the level of efficacy of a
repellent in laboratory and field studies. 

Board Response to the Charge

Introduction 

As written, the current draft of the Guidelines suggests that six should
be the minimum number of research subjects in laboratory or field
experiments where the efficacy of an insect repellent is investigated.
It is not clear from the Guidelines whether the number refers to the
entire experiment or to the number of subjects allocated to each
treatment under consideration. The Board argues below that establishing
a single sample size for all types of experimental designs and
objectives is not the most appropriate approach.

	Critique 

Correctly estimating the sample size that is needed in an experiment
(conducted either in a laboratory or in the field) is important to
ensure reliable inferences about the treatment under study. Sample size
calculations can be carried out using several approaches, but the two
most common ones (at least in terms of usage) are based on:

Power calculations: sample size is chosen to guarantee that tests of
hypotheses reach a pre-determined power.  Power is defined as one minus
the probability of incorrectly failing to reject the null hypothesis of
no treatment effect.  In other words, power is the probability of
finding a difference if such a difference is “true”.  That is, in
under-powered experiments, investigators have a high chance of not
detecting a “true” treatment effect.

Confidence interval calculations: sample size is chosen so that the
100(1-α)% (for α typically chosen to be 0.05) confidence interval
around a treatment effect estimate is sufficiently small. The narrower
the confidence interval, the more reliable the point estimate of the
treatment effect size. 

While smaller than needed sample sizes result in under-powered studies
and wide confidence intervals for true effect sizes, excessively large
samples are not desirable either. First, resources are wasted when
samples are larger than they need to be.  While it is always possible to
increase the power of a study by increasing the sample size, at some
point the cost of obtaining an additional observation outweighs the
potential gains in power. Second, very large sample sizes may result in
statistically significant results that have no practical implication. 
Finally, including more subjects in an experiment than is required for
statistical reasons may unnecessarily place subjects at risk.

Both the power of a test and the width of the confidence interval around
a point estimate depend on various design and data attributes,
including:

Sample size: power increases as sample size increases; the width of
confidence intervals decreases as sample size increases.

Variance across experimental subjects: the smaller the variability in
the response across experimental subjects, the larger the power and
narrower the confidence interval for a given sample size. 

The size of the effect that needs to be detected: in experiments in
which detecting a very small difference between two treatments or
between a treatment and a control, the minimum required sample size for
achieving a certain power or for keeping confidence intervals to a
desired width will be larger than when the difference to be detected is
larger.  In other words, the smaller the difference sought between two
groups, the larger the required sample size, all other factors being
held constant.

Whether the design calls for replicate measurements obtained from the
same individuals in the study (which induces correlation across
measurements) or for “true” replication involving different
individuals observed under the same conditions:  correlation across
measurements (repeat measures in the same individual) in general
decreases the power of an experiment.

The appropriate approach for estimating the minimum required sample size
in insect repellency studies in the laboratory or in the field will
depend greatly on the design of the study.  Factors to be considered
when estimating sample size include the following:

Whether the experiment was conducted in the laboratory or in the field:
a larger sample size will typically be required for experiments
conducted in the field because uncontrollable factors that may affect
the response increase the variance across test subjects.

The number of treatments (e.g., potency formulations or modes of
application of an insect repellent) included in the study.

The presence of control subjects, and whether the same volunteers will
serve as both controls and experimental test subjects (as in experiments
in which one arm of each subject is treated with a repellent while the
other one is not). Experiments in which the same subjects act
simultaneously as controls and as tests require smaller sample sizes
(all other factors being equal) than those studies in which different
individuals act as controls and tests.

Whether the design calls for repeated measurements on experimental
subjects.

The outcome variable of interest: this relates to the between-subject
variability mentioned earlier. The variance across subjects might be
larger for some outcome variables than others. For example, the
between-subject variance might be expected to be larger when the outcome
variable is repellency of a product over a long period than when the
product’s repellency over a shorter period is of interest. Thus, the
minimum sample size for adequate study reliability (either in the power
or the width of confidence interval senses) would be larger in long-term
studies than in short-term ones.

The presence and potential effect of confounders that cannot be easily
controlled via the experimental design. For repellency studies, for
example the intensity of a person’s odor from carbon dioxide emissions
contribute to the attractiveness of the person to blood-seeking
mosquitoes. The sample must be large enough to ensure that the
variability in the general population of consumers of the product is
represented in the study.

The heterogeneity of the target population from which the sample is
drawn: if the product is meant to protect all individuals (e.g., all
ages and both genders) then the minimum sample size might need to be
computed within population strata, to ensure that each population
sub-group is adequately represented in the sample and that inferences
about the effectiveness of the product can be reliably drawn for the
entire population.

The heterogeneity of environments in which the product is expected to be
used: if the product is to be used in a variety of environments (e.g.,
open fields, forests, marshes, and the typical backyard) where a
different concentration of insects and ticks can be expected, the
environment must be included as a factor in the experimental design.  In
laboratory conditions, field insect and tick concentrations can be
mimicked by varying the density of insects and ticks in experimental
cages. The larger the number of environments in the study design, the
larger the minimum sample size needed to achieve the desired level of
inferential accuracy.

Actual calculation of minimum required sample size typically requires
estimating the sample variance of the point estimate of interest. Point
estimates, in turn, follow different sampling distributions depending on
the quantity that is being estimated. In repellency studies, various
outcomes are of interest and these differ in the distributional
assumptions that can be justified:

When the outcome or response variable is the time to first confirmed
bite (TFCB), an appropriate distribution for the response might be the
exponential distribution (or the more general gamma family). A point
estimate of the mean response is given by the sample mean of the
response variable, but construction of a confidence interval for the
true mean response must be based on the correct sampling variance
calculation. 

When the outcome variable is relative protection, the product passes the
efficacy test if treated subjects receive 95% fewer bites than control
subjects. Because the number of bites can be best modeled as a Poisson
random variable, a point estimate of the number of bites under different
treatments and a standard deviation around that point estimate must be
estimated under that Poisson model. A normal approximation to the
Poisson would be reasonable only when the number of bites anticipated
for each subject is large, a situation not likely to be encountered in
practice. 

HSRB Consensus and Rationale

It is critical that the proposed number of subjects be justified on the
basis of good research design. Because experiments to test effectiveness
of products to repel insect and tick bites are likely to vary in terms
of design, response variable, target population of interest, detectable
effect size and other important variables, requiring a specific minimum
sample size that guarantees sufficient accuracy in all cases might be
impractical. Instead, the guideline might require that registrants
present their own sample size calculations and that the methodology used
in the calculations be justified relative to the factors noted in the
bullet list above.

Charge to the Board

g. Please comment on whether or not investigators should have an ethical
obligation to provide subjects of repellent efficacy research with
insurance to cover possible future medical costs or other losses that
result from injury or illness experienced by the subjects as a
consequence of their participation in the research.  

Board Response to the Charge

The broad issue of compensating research subjects for research-related
injuries, together with the somewhat narrower one of paying for the
costs of medical care for such injuries, has received substantial
analysis. The report by the National Academy of Sciences on intentional
dosing studies (NAS 2004), the principles of which Congress specifically
required to be reflected in the EPA regulations on such studies,
directly addressed this issue.

As the NAS Report notes:

Debate continues in the United States about whether compensation should
be provided for research-related injuries. The Common Rule requires only
that when research involves more than minimal risk, information should
be disclosed regarding whether medical treatment and other compensation
will be provided for research-related injuries. Many critics of the U.S.
policy believe there should be more than disclosure of information about
compensation and call for the provision of medical care for
research-related injuries without cost to the participants and, in
addition, for compensation for lost wages, disabilities, and death.
These claims are based on the belief that research participants,
whatever their motivations, accept risk on behalf of society. When
participants are injured, justice, fairness, and gratitude mandate, at a
minimum, the provision of needed medical treatment without cost to the
participant. Further study is needed regarding the provision of other
types of compensation. (NAS 2004.)

Based on this analysis, the NAS Panel examined the ethical issues
associated with intentional human exposure studies, adopting the
following as one of its Recommendations:

Recommendation 5-5: Compensation for Research-Related Injuries 

At a minimum, sponsors of or institutions conducting intentional human
dosing studies should ensure that participants receive needed medical
care for injuries incurred in the study, without cost to the
participants.  In addition, EPA should study whether broader
compensation for research-related injuries should be required.  (NAS
2004)

The Board agreed with the reasoning and recommendations listed in the
NAS Report with regard to a research subject not being required to bear
the costs of medical care needed to treat injuries incurred as a result
of participating in a research study.

Indeed, the conclusions of the NAS Report reflect a growing consensus
that the provision of such free medical care should be adopted as a
requirement for many categories of research studies. For example, in
Volume 1 of its 1982 Report, Compensating for Research Injuries, the
President’s Commission for the Study of Ethical Problems in Medicine
and Biomedical and Behavioral Research, concluded that “compensation
of injured subjects is appropriate to the research enterprise. A program
to assure compensation is thus a desirable policy goal for a just and
compassionate government, both as the sponsor of most biomedical and
behavioral research and as the means through which society acts on
matters of common interest, such as the search for new biomedical
discoveries” (at page 64). That Report did not specifically call for
the adoption for such a program, since it concluded that it did not have
enough information about whether subjects were indeed already receiving
such compensation, and about the costs and other practicalities relating
to adopting a program.

More recently, in its 2001 report on Ethical and Policy Issues in
Research Involving Human Participants, the National Bioethics Advisory
Commission (NBAC 2001) reviewed the literature on this issue, concluding
that a “comprehensive system of oversight of human research should
include a mechanism to compensate participants for medical and
rehabilitative costs resulting from research-related industries. The
inclusion of this mechanism has long been justified on ethical
grounds” (at page 123). It echoed the President’s Commission’s
call for a study of the need for a compensation program. 

Similarly, in 2003, the Institute of Medicine, in Responsible Research:
A Systems Approach to Protecting Research Participants, commenting that
“[b]ecause the contributions of science benefit society as a whole, it
seems indisputable that society is obligated to assure that the few who
are harmed in government-sponsored scientific research are appropriately
compensated for study-related injuries. . . . . The same argument
applies to privately funded research, perhaps even to a greater extent,
as the economic survival of a company depends largely on the
availability of participants to test new therapies, drugs, and other
products. Because the participants are ultimately contributing to the
profits of the company, any costs that result from the research should
be the responsibility of the sponsor” (at pages 188, 190). The
Institute of Medicine report also reviewed international standards
relating to this issue, pointing out that Guideline 13 of the Council
for International Organizations of Medical Sciences (CIOMS) requires
that subjects be equitably compensated for “any temporary or permanent
impairment or disability.” The report concluded that although laws
vary, “most [nations] make some provision for compensation” (at page
189).

These arguments have special import in the context of the intentional
dosing studies that this Board will be reviewing, including repellent
efficacy research. These studies will almost never produce any direct
benefits for study participants. On the other hand, there is frequently
the possibility that subjects will suffer significant injuries as a
result of their participation. In the repellent efficacy studies, for
example, subjects may be at risk or contracting a serious vector-borne
illness as a result from insect bites received during the course of the
study.  Given the lack of direct benefits to subjects, and the
possibilities of very significant harm, the justification for requiring
sponsors to cover the costs of medical care for research-related
injuries is heightened.

Three important points also need to be mentioned regarding issues raised
by the wording of the charge to the Board. First, the Agency asked for
comments regarding whether “investigators” should be required to pay
for the costs of such medical care. In most cases, it would be most
appropriate for that obligation to be imposed upon the sponsors of
research, who are usually the most immediate beneficiaries of the
research, rather than the investigator. The investigators should only
have this obligation when there is no external study sponsor (i.e., when
they are effectively acting as the sponsor of their own study).

Second, the Agency’s charge spoke of requiring that subjects be
provided with “insurance” to cover the relevant medical costs. The
Board believed that sponsors should be provided with some degree of
flexibility in demonstrating how they will cover the medical costs of
subjects. A sponsor that has sufficient assets, for example, might be
able to contractually commit itself to pay for these costs. Given the
possible substantial administrative costs of having a sponsor purchase a
special type of insurance for subjects, it does not appear appropriate
to rule out other ways for assuring that a subject’s medical costs are
covered.

Third, the Agency’s charge raised the possibility of requiring payment
for “other losses” beyond the costs of medical care. Payment for
such “other losses” (for example, the cost of lost wages when a
subject is not able to work for a period of time) is a more complicated
and controversial issue than covering medical expenses. With regard to
this issue, the Board agreed with the conclusions of the NAS Report that
further study should be required in order to better evaluate whether
requirements to cover such “other losses” should be imposed.

HSRB Consensus and Rationale

	For the reasons discussed above (including justice, fairness and
gratitude), the Board concluded that it is appropriate that sponsors of
repellent efficacy research studies should be required to ensure that if
a subject is injured as a result of their participation, then the
subject will not have to assume the costs of medical care needed to
treat such injuries.

Charge to the Board

h. Please comment on any special considerations that should be addressed
in the informed consent materials provided people who are candidates to
become subjects in insect repellent efficacy research. 

Board Response to the Charge

The general requirements for informed consent are outlined in 40 CFR
26.1116 of the Agency’s final human studies rule.  A basic element in
seeking informed consent is that the subject should be told that the
study involved research and given an explanation of the purposes of the
research, the expected duration of the research, a description of the
procedures to be followed, and an identification of any procedures that
are experimental.

The informed consent should begin with a clear statement that this study
is research, and a statement as to whether the product being tested is
approved and marketed in the test formulation or still in the
experimental stage.  Information about the potential efficacy of the
product against the test insect should also be provided.

For the insect repellent studies, it is especially important to be very
clear about the experimental set-up, either the laboratory or the field,
and what the expectations are for the subject.  Because there appears to
be a tendency in these research studies to use “seasoned” subjects
(i.e., those who are in the scientific field or have participated in
these studies before), a detailed explanation of the procedures might
seem to be unnecessary to the investigator.  

Nonetheless, the written details of the experimental procedure must be
sufficient to inform a potential subject who has never anticipated in
this type of study and to remind one who has done so.  A video,
PowerPoint presentation, or photographs might help the subject to
visualize what will occur to him/her during the study.  If it is a
laboratory study, it may help to have the subject place their arm into
the cage.  If the subject is expected to use an aspirator, training on
its use should occur prior to the beginning of the study.  A
demonstration of what a landing and probing feels like might be
appropriate because the dermal sensitivity of individuals will vary.

The length of time that the study would take should be clear in the
informed consent document, including whether the test would be repeated.
 The process for randomizing subjects to the test or experimental group
should be included.  In field studies that take the entire day, it might
be explained whether food would be provided to the subjects.  

A clear discussion of the stopping rules should be included, especially
for the field studies.  For example, to discontinue participation, does
the subject raise their hand, return to the van, or find a study monitor
to express a desire to stop.  

Another basic element of informed consent is a description of any
reasonably foreseeable risks or discomforts to the subjects.  Inclusion
of a Material Data Safety Sheet is not sufficient to adequately inform
the subject as far as all the inherent risks and benefits of study
participation.  For insect repellent studies, three types of risks are
reasonably foreseen. 

The first is the risk of being bitten.  The informed consent document
should give an estimate of the potential number of bites that a subject
could receive in the control and experimental groups.  A statement that
the researcher strives for each subject to receive few to no bites is
not sufficient.  Additionally, should a subject have an allergic
reaction to the insect bites, medical procedures and remedies that would
be present should be clearly described in addition to any available
follow-up treatment (e.g. will subjects be given an antibiotic or
steroid cream for their bites?

The second risk is that of a sensitivity or allergic reaction to the
experimental product.  A synopsis of the animal studies and any human
data should be given to the subject along with a discussion of the
theoretical risk of a reaction occurring.  Emergency care procedures
should be presented for a subject who has a reaction to either the bite
itself or the repellent.

The third and most serious risk is the potential for acquisition of
vector-borne illness as a result of insect bites.  The severity of these
illnesses should be clearly explained, even if the researcher believes
the risk is minimal to non-existent due to either through the use of
disease-free insects in a laboratory study or the selection a
disease-free zone for the field study.  The risk of insect-borne
diseases might not occur to a subject who normally does not work in the
field.  The investigator should make sure that the subject clearly
understands both the risks of disease transmission and what symptoms to
look for with any potential insect-borne diseases.  One suggestion might
be to test the subject, either verbally or in writing, about their
understanding of the procedures and the risks. 

Another basic element of the informed consent process is a clear and
complete description of any benefits to the subjects or others that are
reasonably expected to result from the research.  It should be very
clear in the informed consent document that there are likely to be no
direct benefits to the subject as a result of study participation.  The
only potential benefit is to society at large to have an arsenal of
insect repellents available.  Additionally payment for participation in
a research study can not be considered a benefit of the study.

Each research participant should be told the extent, if any, to which
confidentiality of the records identifying the subject should be
maintained.  The researcher should be careful to only include those
organizations that have jurisdiction over the study and might therefore
have the right to inspect the records.  It is equally important that
access to the records is limited to as few individuals as possible and
that strict confidentiality procedures be developed and are strictly
adhered to. 

Since insect repellent studies should be classified as research
involving more than minimal risk, the subject should be clearly told if
the researcher will cover medical treatment if an injury occurs (this
issue was reviewed by the Board in more detail in response to question
g.) , including not only treatment during the research study but
long-term care, if needed (e.g. in those circumstances in which a study
subject contracts an insect-borne disease).  A phone number should be
provided to volunteers so that they can obtain additional information
about the risks and benefits of study participation, and their rights as
study subjects.  

The subject should clearly be told that the study is voluntary and
refusal will not result in any loss of benefits or privileges.  For
studies that use only one control group, it is especially important for
a subject assigned to the control group to understand that they can
withdraw even if this withdrawal might invalidate the study.  It should
also be clear that the subject does not need to give a reason for
withdrawal from the study.  The consequences of a subject’s decision
to withdraw from the study should be addressed, including how it will
affect any payment for participation in the study.  

Students or employees used as research subjects in this study are
considered “vulnerable subjects” because they might feel coerced
into participating either by their supervisor, thesis advisor, or even
fellow students/employees.  It should be clear that participation in
these types of studies is neither a condition of employment nor an
academic requirement for students.  An explanation of whom to contact if
the subject feels coerced should be provided.  This contact should not
be associated with the investigator and the subject should be guaranteed
anonymity.  Employees who report directly to the investigators or study
sponsors, and students of the investigator should be excluded from the
study.  

The subjects should also be told that they will be informed if any new
information is found during the course of the study that might affect
the subject’s willingness to participate.  Additionally the subject
should be told that they will be informed if it is found that either a
test site, or a laboratory strain of insects used is discovered to have
a higher level of disease than previously thought.

The informed consent document should be written in a language
understandable to the subjects and the subjects should be informed about
any potential conflicts of interests that the researchers have. 

HSRB Consensus and Rationale

The consensus of the HSRB was that informed consent should comply with
all of the requirements of 40 CFR 26.1116 of the Agency’s final human
studies rule.  To comply with the human studies rule, consent
information for pesticides studies must include: (a) detailed
information on the procedure (e.g., number of insect bites or landings
anticipated, nature apparatus or field context, length of time of
exposure); (b) a clear statement of the risks involved (e.g., discomfort
from bites, risk of vector-borne disease, medical consequences of the
disease, treatments available for the disease); (c) the voluntary nature
of participation (e.g., statements that eliminate the perception of
coercion for students or employees; specific instructions on how to
signal desire to withdraw from the study); (d) the fact that there was
no immediate direct benefit to the subject in participating as well as a
description of alternative available repellents; and (e) other steps
outlined above. In addition, informed consent information should be as
detailed for experienced subjects as for naïve subjects.

Charge to the Board

i. Does the HSRB recommend that the draft guideline be revised? If so,
please explain what aspects or sections might improve with revision.

Board Response to the Charge

The Board recommended that the Agency refer to Board responses to
previous questions to address revisions to the draft guideline.  

Review of HSRB Protocol Criteria 

Before the Board reviewed the presented proposed human studies research,
the Board developed science and ethics criteria as a guide for its
evaluation of such studies.  Reference to such criteria would be helpful
for the Agency, study investigators, and other members of the public to
understand the Board’s approach for the review of proposed human
studies.  The relative emphasis placed by the Board on each criterion
may be applied case-by-case and may vary with the nature of the chemical
product, study design, and participants.  Specific studies may also call
for additional criteria. A list of the science and ethics criteria
identified by the Board are provided below:

 Science Criteria

The criteria for the evaluation of the scientific quality of studies
involving human subjects was based on a series of questions which the
Board agreed needed to be addressed by the details provided in the study
protocol.

1) Is a valid scientific question addressed by the study?

2) Are existing data adequate to answer the scientific question?

3) Are new studies involving human subjects necessary to answer the
question?

4) What are the potential benefits of the study?

5) What is the likelihood that the benefits would be realized?

6) What are the risks? Are they serious or irreversible?

7) Is the purpose of the study clearly defined?

8) Are there specific objectives/ hypotheses? 

9) Can the study as described achieve these objectives or test these
hypotheses?

10) What is the sample size and how is it derived?

11) What is the basis for the proposed dose levels and formulations in
the study?

12) Is there a plan allocating individuals to treatment?

13) Can the findings from this study be generalized beyond the study
sample? 

14) Is there a justification for the selection of the target population?

15) Are participants representative of the population of concern? If
not, why not?

16) Are the inclusion/exclusion criteria appropriate?

17) Is the sample a vulnerable group?

18) Will the measurements be accurate and reliable?

19) Are measurements appropriate to the question being asked?

20) Are adequate quality assurance procedures described?

21) Can the data be statistically analyzed?

22) Is the statistical method appropriate to answer the question?

23) Are point estimates accompanied by measures of uncertainty?

24) Do laboratory conditions simulate real-world conditions?

25) Are field conditions representative of intended use?

26) Does the protocol include a stop rule plan, medical management plan,
and a safety monitor?

Ethics Criteria

The criteria for the ethical acceptability of environmental research
protocols involving human dosing and intentional exposure proposed by
the Board are grounded in the general criteria for IRB approval found in
Subpart K of the Agency’s human studies rule (40 CFR 26.1111 and
1116).  This approach is similar to that taken by the National Academy
of Sciences (2004) in formulating the criteria for scientific and
ethical acceptability (recommendation 5-1) and participant selection
(recommendation 5-2).

Scientific Validity and Social Value

One of the most important criteria for the ethical review of protocols
in fact is scientific.  The research design must be sound (i.e.,
scientifically valid) and the risks of the research must be reasonable
(or balanced) in relation to the importance of the knowledge that may
reasonably be expected to result.  Absent a sound research design, the
prospect of the research generating usable knowledge is severely
diminished.  Although the risks to research participants may be balanced
against anticipated benefits to these same subjects, much environmental
research (such as intentional exposure studies) will not offer any
direct benefit to the research subjects themselves  (See NAS
recommendation 3-1.).

The justifiable risks to which research subjects may justifiably are be
exposed should be directly proportional (i.e. reasonable or balanced) to
the importance of the knowledge expected to be gained.  In other words,
the information to be gained from the research study must be "worth
knowing".  The NAS took this approach in recognizing that scientific
accuracy alone is insufficient justification for exposing research
subjects to anything more than "no identifiable risk."   For example,
there must be a "reasonable certainty of no harm" to research subjects
if the only benefit of the research is to improve the scientific
accuracy of extrapolating animal to human data (NAS recommendation 4-1).
 As such, a research protocol needs to describe the benefits of the
knowledge that may be obtained so that the reasonableness of the risks
can be judged against the importance of that knowledge.  In addition to
improved scientific accuracy of risk assessment, such benefits may
include a more stringent regulatory standard, new public health measures
that could be adopted, or new products that may protect public health.

Federal regulations state that an IRB should not consider the possible
effects of the research on public policy when evaluating those research
risks that fall within the responsibility of the IRB (40 CFR 26.1111 (a)
(2)).  Nevertheless, the public policy implications of the knowledge
that may result from the research does affect the importance of that
information.  Further, the Human Studies Review Board is not limited by
Subpart K in evaluating the ethical acceptability of a proposed research
study.  As such, scientific accuracy alone may be an insufficient
justification for the importance of a research project.  The protocol
should address the potential benefit of improved scientific accuracy,
and to whom this benefit would accrue.  As recognized by the NAS in
recommendation 4-2, studies that may have a potential public health or
environmental benefit could involve a somewhat higher level of risk
while not causing any lasting harm to research subjects.

Minimizing Research Risk

The research should not expose any human subjects to unnecessary risk
(40 CFR 26.1111 (a) (1)).  This ethical principle has a number of
important corollaries.  First, the use of human subjects must be
absolutely necessary in order to answer an important scientific question
that could not otherwise be answered by using animal models.  In
addition, any intentional dosing studies can only be justified if
observational studies would neither answer the question nor be feasible.
 Admittedly, the judgment of feasibility may be ethically difficult
especially if the only consideration is time and expense.  Second, the
elimination of unnecessary risk means that there is no way to answer the
scientific question that involves less risk if human subjects are to be
used.  Third, the scientific protocol should involve no additional
exposure of study participants to risk unless absolutely necessary.  The
ethical responsibility for "using procedures already being performed on
the subjects" translates, in the environmental context, to studying
those situations in which human subjects are exposed to environmental
toxins as part of their usual activities without increasing their
exposure to those same toxins.  In addition to the ethical priority of
animal over human studies, there is an ethical priority for
observational research over intentional dosing research involving
environmental toxins if scientifically appropriate.  Whether a study
meets the scientific and ethical criteria necessary to justify the
exposure of human subjects to potential risk can only be evaluated in
the context of a given research protocol if the investigator and/or
sponsor specifically addresses alternative means of obtaining the
desired data.

Equitable Selection of Subjects

The selection of subjects should be equitable (40 CFR 26.1111 (a) (3)). 
In practical terms, this means that the selection of subjects should
reflect the scientific purposes of the research and not the availability
of a particular population.  This ethical criterion may be especially
problematic in the context of environmental hazards research.  Often the
exposure to environmental hazards in the workplace or at home is greater
for those who are either socioeconomically or educationally
disadvantaged.  As such, subject selection based purely on scientific
design may be insufficient protection for the research subjects, with
additional safeguards.  The need for such safeguards must be assessed
within the specific context of a particular protocols based on an
in-depth knowledge of the community within which the research will take
place.  The ability to "minimize the possibility of coercion or undue
influence" (40 CFR 26.1116) may require the involvement of
representatives from the community from whom the research subjects will
be drawn and within which the research will take place.  A research
protocol also must include specific measures for assuring the equitable
selection of subjects, including recruitment practices, incentives
(financial or otherwise), impact on employment, and the possibility of
retaliation.  In addition, any incentive for participation (whether
financial or otherwise, such as time off from work) should not be
included in the analysis of risks and potential benefits of the
research.

Informed Consent

The information that is included in the informed consent process and
documentation should include all of the information found under the
general requirements for informed consent in 40 CFR 26.1116.  There are
a few specific features of the informed consent information that are
worth highlighting in the context of environmental research.  First, EPA
regulations do not allow for a waiver of either informed consent or the
written documentation of informed consent.  Second, the informed consent
information must include the identity of the pesticide and its mode of
action if the research involves intentional exposure of subjects to a
pesticide (40 CFR 26.1116 (e)).  Given the vulnerability of the research
subjects that are likely to be enrolled in environmental research (as
discussed above), the default position for any research on environmental
toxins (whether observational or intentional) should be that the risks
of any potential pesticide exposure be included in the informed consent
information.  However, if the risks of the toxins are not part of the
research, but instead are part of daily work life, this should be made
clear.  Third, the alternatives to research participation (40 CFR
26.1116(a)(4)) should include all steps that might minimize the risk of
exposure to environmental hazards, up to and including removing oneself
from that environment.  Fourth, as noted previously, the HSRB supports
the view that research subjects should receive needed medical care for
research related injuries at no cost to themselves (consistent with NAS
recommendation 5-5).  As such, the oft-used informed consent template
statement that “no program of compensation is available” would be
unacceptable in human dosing or pesticide exposure research.  The HSRB
acknowledges that the determination that any given injury may be
research-related might be difficult when the protocol combines
observational or interventional procedures with non-research related
exposure to environmental toxins.  Nevertheless, the principle of
providing medical care for research related injuries at no cost to
research subjects must be affirmed. Fifth, the voluntary nature of
participation must be carefully and explicitly described during the
consent process. Investigators, study sponsors and pesticide registrants
are obligated to ensure that neither employment status nor economic need
creates a coercive context for study participation. Finally, the process
and documentation of informed consent needs take into account special
circumstances that may arise in the context of any given research
setting, including language barriers, literacy, comprehension,
employment status, and the confidentiality of screening tests such as
for pregnancy.

Subject Safety

The research protocol must also discuss provisions for assuring the
safety of subjects enrolled in the research, both during and after the
research has been completed.  This obligation goes beyond simply
"monitoring the data collected" to include procedures for collecting
real-time exposure data to the environmental toxins during the research,
and procedures for intervening should the health of research subjects be
at risk from the environmental toxins (regardless of whether the
exposure is intentional or not).  The protocol should describe in detail
any procedures for reversing experimentally-induced harms.

IRB Approval

The HSRB believed that the ethical analysis of a research protocol
requires information concerning the potential risks to human subjects,
measures proposed to minimize risks, the nature and magnitude of all the
expected benefits, and to whom they may accrue, alternative means of
obtaining information comparable to what would be collected for the
proposed research, and the balance of risks and benefits of the proposed
research (40 CFR 26.1125(a)).  Further, the HSRB believes that an IRB is
unable to make the determinations required under 40 CFR 26.1111 absent
this information.  As such, the HSRB expects this information to be
found in the protocol submitted to the responsible IRB.  Although an IRB
may be able to gather this information from other sources, the lack of
this information in the protocol and the lack of a substantive
discussion of these issues in the IRB minutes would raise doubt about
the adequacy of the IRB review.

Insect Repellent Product Performance Efficacy Studies

Study EMD-003 from Carroll-Loye Biological Research

Charge to the Board

a. Does the proposed research described in study EMD-003 appear likely
to generate scientifically reliable data, useful for assessing the
efficacy of the repellent? 

Board Response to the Charge

The protocol submitted for review by the HSRB outlined studies to
evaluate the efficacy of IR3535 as a tick repellent in human subjects. 
The protocol described a laboratory study in which the movement of the
Western black-legged tick (Ixodes pacificus) up the forearm was to be
determined.  Studies in humans are required to assess the efficacy of
such repellents because laboratory animals differ in their
attractiveness to the pest, and therefore do not provide an accurate
assessment of efficacy in humans.  A more general protocol (CL-001),
which provided additional information relevant to study conduct, was
also submitted for review in combination with protocol EMD-003.

applied (as a lotion, aerosol or spray) and the evaluation of repellency
against mosquitoes was indicated in the rationale provided for the
study.  Furthermore, the protocol indicated that the dose to be applied
was 1 mg formulation/600 cm2, when in fact, the authors of the protocol
intended the applied dose to be 1 gram/600 cm2.   These mistakes were
not considered to be fatal errors in the protocol, but suggested a lack
of attention to the details of protocol preparation and review by the
investigators.  Staffers from the USEPA provided comments on the
numerous shortcomings of the proposed study, and the HSRB fully
concurred with these weaknesses.

IR3535 is commercially available, and there is a large amount of
toxicology data suggesting that it is a compound of low toxic potential.
 Therefore, human subjects are unlikely to be at risk of experiencing
adverse effects relative to exposure to the proposed formulations.  

The HSRB recognized three major limitations to the protocol as submitted
to the HSRB for review.  These limitations include: (1) the lack of a
clear rationale underlying the conduct of the study; (2) the lack of
identification and characterization of the formulations to be tested and
(3) the scientific design of the study.  Of these issues, which are
discussed in more detail below, the design of the study was seen as the
most significant shortcoming of the proposed work.

With respect to the clear rationale for the conduct of the study, the
HSRB understood that all new formulations must be evaluated for
efficacy, and that such studies must be conducted in human subjects to
be valid.  However, the investigators failed to identify what was new
about the formulations being studied and failed to identify the
potential benefit of the formulations.  This shortcoming was considered
to be minor and could readily be addressed by providing such additional
information in the study protocol.

With respect to the formulations to be evaluated, the investigators
provided tables listing the percent of active ingredient along with
incipients used to formulate the spray, aerosol and lotion to be used in
the study.  However, there was no additional information regarding when
the formulations would be prepared relative to study execution, whether
the formulations would be characterized analytically to confirm active
ingredient composition, and whether the stability of the formulations
was to be determined.  This information is critical to the overall valid
execution of the study and could be remedied by providing such detail in
the protocol.

The major limitation with the scientific conduct of the study concerned
the study design and data collection.  In particular, the protocol
outlined a study using six test subjects for each product formulation,
with two additional subjects serving as a negative and a positive
control.  No information was provided to justify the group sizes used in
the research. Given the nature of these studies to evaluate tick
repellency, the HSRB considered that a test in which each subject served
as his own control (using one arm for the untreated, negative control
and one arm for the test) was a more appropriate design that would also
be more powerful statistically and more likely to generate reliable
results.  

Additional questions were raised by the HSRB concerning how subjects
would be trained to accurately and consistently collect information
regarding the number of ticks crossing or repelled from the arm skin. 
The protocol defined that a crossing is scored by the movement of a tick
by at least two centimeters toward the elbow starting from a line at the
wrist, and that subjects select a new tick from a pool of unused,
prescreened ticks every 15 minutes.  There was no information made
available to the Board as to how subjects were trained and qualified to
establish that they could collect accurate data on tick movement.  The
Board agreed that such information was important for establishing good
quality control of the data collection concerning repellency.

HSRB Consensus and Rationale

Overall, the HSRB concluded that there were numerous technical
deficiencies in protocol EMD-003, and the information provided in the
general protocol (CL-001) did not make up for the deficiencies in the
specific protocol.  Therefore, the Board concluded that the available
protocol did not warrant moving forward with the study.

Charge to the Board

b. Does the proposed research described in Study EMD-003 from
Carroll-Loye Biological Research appear to comport with the applicable
requirements of 40 CFR part 26, subparts K and L?

Board Response to the Charge

Background

The study proposed was to evaluate the efficacy of a compound known as
IR3535 as a tick repellent in human subjects. The study is to be
conducted by Carroll-Loye Biological Research, a private research
laboratory in Davis, California by using healthy volunteers and a
controlled laboratory environment. Two protocols were submitted for
review, a general protocol (CL-001) that provided considerable
background information about tests of insect repellency in general, and
the protocol for analysis of the efficacy of IR353 as a tick repellent
evaluated here.

For this protocol, the efficacy of IR3535 as a tick repellent would be
determined by placing Western black-legged ticks (Ixodes pacificus) on
IR3535-treated and –untreated forearms and measuring the speed and
distance that moving insects would penetrate into the treated area. 

Strengths and Limitations

The Board concurred with the factual observations of the strengths and
weaknesses of the study, as detailed in the EPA’s Initial Ethics
Review (USEPA 2006b). This study, it was argued, would provide critical
data on the efficacy of IR3535 as a tick repellent. IR3535 is
commercially available and has been used as a repellent in Europe for
years with no evidence of toxicity, so the subjects enrolled in this
study were unlikely to be at increased risk of experiencing adverse side
effects upon exposure. The ticks used for the study also were bred and
raised in a laboratory environment and are considered to be
pathogen-free, minimizing the risk of vector-borne diseases.

The Board concluded, however, that given the deficiencies noted by the
Agency, the proposed research described in Protocol EMD-003 did not
comport with the applicable requirements of §40CFR26, particularly
subpart K. Carroll-Loye Biological Research and the IRB of record also
failed to obtain or to provide all of the documents necessary to be in
compliance with the requirements of §40CFR §26.1125, subpart M. The
IRB, for example, refused to release copies of the minutes documenting
the discussion of EMD-003 protocol, preventing the Board from evaluating
whether or not they considered fully the issues listed under the EPA’s
Final Human Studies Rule as part of their review.  Furthermore, the
protocol and supplementary documents submitted to the Board were poorly
written which, while not a fatal flaw in and of itself, should have
precluded IRB and HSRB review and approval.

The HSRB recognized several significant limitations to the protocol, as
submitted to the Board for review. There was, for example, lack of a
clear rationale justifying the conduct of the study as designed. The
Board’s concerns about this are discussed in detail with respect to
the companion protocol submitted by Carroll-Loye, EMD-004, but it was
felt that absent any clear scientific rationale for conducting such a
study, exposure of human subjects to the risks inherent in this protocol
would be unnecessary and unjustifiable. 

Additional limitations of the study protocol provided to the Board can
be grouped into two broad categories: (1) concerns about equitable study
subject selection and recruitment; and (2) questions about whether or
not the documentation and process of study subject enrollment was
sufficient to meet prevailing standards of voluntary informed consent.

Subject Recruitment

The Board expressed concern about the potentially coercive nature of
study subject recruitment. Although the study is to be conducted by
Carroll-Loye Biological Research, a private research laboratory in
Davis, California, the Principal Investigator of the study and Co-Owner
of the research laboratory, Dr. Scott P. Carroll, also is an adjunct
faculty member of the Department of Entomology at the University of
California, Davis. As the majority of research participants will be
recruited from the University’s student population, including from Dr.
Carroll’s own department, the protocol and consent documents need to
be altered to define clearly the mechanisms in place to prevent any
coercive enrollment, as well as the additional concerns listed below.

Voluntary Informed Consent

The Board believed that the protocol and consent documents, as provided,
lacked sufficient information to ensure that all study participants were
adequately informed about the risks, benefits and alternatives to
participation in the study. It was unclear, for instance, that
participation in the study would have no direct benefit for volunteers
or that the study was being conducted solely for development of data
required for regulatory review by EPAmarketing research.  The major
risks of participation in the study also needed to be more clearly
identified in the informed consent form and in supplementary documents
provided to study subjects. For example, one additional risk that the
study investigators may have failed to consider arises from the plan to
pre-screen female volunteers in order to exclude any subjects who may be
pregnant. In accordance with the newly promulgated provisions in the
EPA’s final human studies rule (§40CFR26.1701 - 26.1704), minors and
pregnant women are explicitly excluded from participation, the latter
being confirmed by requiring all female volunteers to undergo a
self-administered over-the-counter pregnancy test on the day of the
study. Because many of the volunteers are undergraduate or graduate
students at a nearby college, the unexpected revelation that a subject
may be pregnant could have a profound psychological or social impact;
Dr. Carroll also may have a professional relationship with these
students through his affiliation with the University. These risks should
be specifically addressed, and the Board recommended that a separate
consent document for female volunteers be prepared that addresses these
risks as well as the safeguards established by study investigators to
ensure that the results of over-the-counter pregnancy tests would be
kept private.

Study investigators also may wish to provide either a more detailed
explanation of the study protocol – including a detailed method for
manipulating the ticks used in the experiment and a clear description of
the study’s duration. 

Finally, it was felt that the informed consent documents should be
re-written to: (a) comport with the reading and comprehension level of
the likely subject population; and (b) clarify the section on
compensation for research related injury.

HSRB Consensus and Rationale

The Board concurred with the initial assessment of the Agency that the
study submitted for review by the Board failed to meet the ethical
requirements established in the Agency’s final human studies rule
(§40CFR26). 

The Board determined the proposed research described in this study did
not comport with the applicable requirements of §40CFR26, subparts K
and L. The study documents submitted for review also failed to comply
with the requirements of §40CFR §26.1125, subpart M.  However, the
deficiencies noted, while significant, were not irreparable.

Study EMD-004 from Carroll-Loye Biological Research

Charge to the Board

a. Did the proposed research described in Study EMD-004 from
Carroll-Loye Biological Research appear likely to generate
scientifically reliable data, useful for assessing the efficacy of a
test substance for repellent ticks? 

Board Response to the Charge

Introduction

The Board began its review noting that this protocol addresses repelling
insects, not ticks.

The Board concluded that the proposed research should generate
scientifically useful data for assessing efficacy.  Protocol EMD-004
describes a test of the efficacy of 3-[N-butyl-N-acetyl]-aminopropionic
acid, ethyl ester (IR3535) to repel mosquitoes in field experiments.  It
describes the formulation and dose of the repellent and the number of
replications (6-10 for each formulation). The components of the three
formulations were provided by the Agency.  There would be one to two
untreated controls and one to two positive (DEET-treated) controls.  Two
locations would be used, in or adjacent to the Central Valley in
California and the Florida Keys. The experiment would be double-blinded.
The compound has a very low toxicity profile in animal tests and has
been used in Europe for over 20 years as a repellent for many years
without reports of adverse effects in humans.

General Scientific Criteria

The scientific question was stated (i.e., to test the efficacy of IR3535
in repelling mosquitoes).

It was not clear whether existing data were adequate to answer the
question.

It was not clear whether new studies involving human subjects were
necessary; however, if the repellency had never been tested with North
American mosquitoes, the tests may be necessary.

The potential benefits of the study were clear, i.e., that an effective
repellent would be available that would have either greater efficacy
and/or fewer drawbacks than what was currently approved.

It was likely that the benefits would be realized (i.e., efficacy as a
repellent) because there was a long positive history on this compound
from its European use.

The risks were not specifically noted.

The most likely relevant risk would be disease transmitted by the
mosquitoes, if the mosquitoes carried pathogens, and some mosquito-borne
diseases (e.g., West Nile virus-mediated disease) were serious. The
protocol did not indicate the likelihood of the mosquitoes in the two
test areas to be carriers of disease organisms that could be transmitted
to humans.  However, using the fewest number of untreated controls would
provide the least risk of disease to the participants. The protocol did
not indicate whether all the inert ingredients in the formulations are
GRAS compounds or have documented lack of toxicity at the exposure
levels anticipated.

Study Design Criteria

The purpose of the study was clearly defined (i.e., efficacy testing).

There were specific objectives/hypotheses (i.e., that IR3535 is an
effective repellent).

The study as described can test this hypothesis.

The sample size and how it was derived was not clear, but seems to have
been taken from the guidelines. The number of subjects listed in section
9.1.3 of the protocol listed potentially more subjects than in the table
in section 8.3.2. It was not clear if the stated number of subjects
would be repeated in both locations. The basis for the dose levels and
formulations were not provided. There were no controls with just the
formulation matrix without the repellent.

There was a plan allocating individuals to treatments.

The findings from this study can probably be generalized beyond the
study sample.

Participation Criteria

There was partial justification for the selection of the target
population.

The participants were representative of the population of concern.

The inclusion/exclusion criteria were appropriate.

The sample was not a vulnerable group.

Measurement Criteria

The measurements were expected to be accurate and reliable.

The measurements were appropriate to the question being asked.

Quality assurances issues did not appear to be addressed.

Statistical Analysis Criteria

The data should be able to be analyzed statistically if the efficacy
with time was the subject of the analysis and the comparisons are made
across time.  However, if there is only one untreated control (which
would be more protective against possible disease transmission), then
there would be difficulties with statistical analysis with comparisons
to the untreated control.

The statistical method seems to be appropriate.

Measures of uncertainty were not addressed.

Laboratory and Field Conditions

No laboratory experiments were proposed in this protocol, probably
because of the data already available due to the compound’s long
previous use.

The field conditions were representative of the intended use.

The protocol did not include a stop rule plan, medical management plan,
and a safety monitor.

HSRB Consensus and Rationale 

It was not clear whether new studies involving human subjects were
necessary. If the repellency had never been tested with North American
mosquitoes, however, the tests were probably necessary. The potential
benefits of the study were clear, i.e., that an effective repellent
would be available that would have either greater efficacy and/or fewer
drawbacks than what was currently approved. However, empirical evidence
or procedures to determine risks to subjects (e.g., risks of contracting
a vector-borne disease) were not adequate. It was not clear if the
stated number of subjects would be repeated in both testing locations.
The basis for the dose levels and formulations were not provided. There
were no controls with just the formulation matrix without the repellent.
Therefore, the Board concluded that some of the more critical
deficiencies in information identified above would have to be adequately
addressed before this protocol could receive a positive recommendation.	

Charge to the Board

b. Did the proposed research described in Study EMD-004 from
Carroll-Loye Biological Research appear to comport with the applicable
requirements of 40 CFR part 26, subparts K and L?  

Brief Overview of the Study

The proposed study would evaluate the efficacy of three different skin
applied formulations of an already registered and marketed (in Europe)
insect repellent IR3535.  There would be two study sites, one located in
central California and the other located in the Florida Keys.  The test
compounds would be administered to a standardized skin surface area,
with a comparison to one positive control and one negative control.  The
subjects allocated to the intervention groups would be blinded to the
treatment.  The chosen outcome measures are "percent reduction in the
rate of alightments" and "complete protection time."  The protocol
stated that there would be 6 to 10 subjects per treatment group, with
one subject per control group.  However there was no discussion of
sample size justification.  As discussed below, the protocol lacked any
discussion of risks.

Ethics and Regulatory Compliance

Subpart K of the Agency’s final human studies rule requires that the
investigator submit to the EPA all information that pertains to the IRB
review of proposed research (40 CFR 26.1115a) as well as additional
information specified in 40 CFR 26.1125, if not already included in the
IRB documentation.  The information requested under 40 CFR 26.1125
includes a discussion of the potential risks to human subjects, the
measures proposed to minimize these risks, expected benefits if any and
to whom, alternative means to obtain comparable information, and the
balance of risk and benefits of the research.  In addition, subject
information sheets and approved written informed consent agreements
should be provided, along with any information about recruitment and the
presentation of this subject information.  Finally, the investigator
should provide copies of all correspondence with the IRB, including
official notification of IRB review and approval.  

In the case of this protocol, the principal investigator made a request
to the reviewing IRB (Independent Investigational Review Board Inc.
located in Plantation, Florida) for the documents required under 40 CFR
26.1125.  The response from the IRB, dated May 12, 2006, did not include
the minutes of IRB meetings at which the protocol was discussed.  As a
result, the Board was unable to assess whether the IRB discussed or was
even aware of the controversial issues raised by this protocol.  The IRB
did provide templates of two different forms, the EPA Protocol Checklist
and Research Evaluation Form.  Although these forms were fairly
comprehensive, the Board was not provided with copies of the forms used
for the specific protocol review and thus cannot assess whether or not
the forms were used or the content of the IRB analysis and discussion. 
The membership roster of the IRB was included.  Although the membership
was diverse and meets the regulatory requirements, there was no
scientific member that appears to have sufficient expertise in the
scientific issues involved in field testing of insect repellents to
assure that the IRB was qualified to make an adequate assessment of this
protocol.  The scientific and ethical assessment may have been adequate,
but the lack of IRB minutes made this determination impossible.  In
effect, the IRB response was to provide procedural documentation of the
IRB's compliance with 40 CFR 26.1115 but to withhold any substantive
documentation that this procedural compliance resulted in an adequate
ethical and scientific review of the submitted protocol.  As such, the
proposed research failed to meet the requirements of 40 CFR part 26,
subpart K.

The investigator, to his credit, remedied some of these deficiencies in
a supplementary document submitted to the EPA as part of the Board's
review.  This undated document was presumably written after the IRB
review.  In this document, the investigator addressed the potential
risks to human subjects, the measures proposed to minimize these risks,
the nature and magnitude of all expected benefits of the proposed
research and to whom they would accrue, the balance of risks and
benefits of the proposed research, and alternative means of obtaining
information comparable to what would be collected through the proposed
research.  

Several observations are in order.  First, none of this material,
including the discussion of risks and benefits can be found in the
protocol submitted to the IRB.  The absence of this information in the
protocol further compounds the uncertainty created by the absence of
minutes  showing how the IRB made the determinations required under 40
CFR 26.1111.  The information about the potential risks to human
subjects, the measures proposed to minimize these risks, the nature and
magnitude of all expected benefits of the proposed research and to whom
they would accrue, and the balance of risks and benefits of the proposed
research should be part of the research protocol submitted for initial
IRB review.  Otherwise, the IRB lacked sufficient information to make an
appropriate assessment of the proposed research.  This was not to say
that the protocol would fail to meet the criteria for IRB approval, only
that there was no evidence that the IRB had sufficient information or
expertise to make these determinations.  Second, the protocol did
minimize the risk of vector-borne diseases by limiting the untreated
control group to  a single subject who was experienced in field biology
or entomology.  The risk was minimized further by using an outcome
measure that does not require biting, but rather preparatory activities
on the part of the mosquito after lighting on the subject followed by
aspiration and removal of the mosquito.  However, this approach raised
concerns about the scientific adequacy of the protocol design.  Third,
the investigator addressed the question of alternative means of
obtaining information by arguing that the protocol was designed in
compliance with previous EPA guidelines for registration of these
products.  The Board did not take a position on whether the protocol was
indeed in compliance with these previous guidelines.  However, there was
sufficient discussion by the Board of the new draft EPA guidance on
"Insect Repellent Product Performance Testing" to cast doubt on the
scientific adequacy and necessity of the approach taken in this
protocol.

Finally, the Board agreed with the ethical deficiencies noted by the EPA
(USEPA 2006c)   With the exception of amending the protocol to include
the applicability of additional standards of ethical conduct and the
process of informing appropriate regulatory authorities of any
amendments or deviations from the approved protocol, all of these
deficiencies related to the informed consent discussion and document. 
These included a more accurate discussion of subject assignment, a more
extensive discussion of the risks (with specific information about the
risk of vector borne diseases), the correction of an important
typographical error in the pregnancy section, a clarification of the
section on compensation for research related injury, a clarification of
the lack of direct benefit to research subjects and additional
information under the heading of confidentiality.  The Board also
discussed the topic of pregnancy testing and whether there should be a
separate consent for such testing.  As many of the research subjects may
be in a professional relationship to the principal investigator (such as
graduate students or colleagues), a protocol and consent document needs
to discuss how the confidentiality of such pregnancy testing would be
protected.

HSRB Consensus and Rationale

The Board thus concluded that the proposed research described in Study
EMD-004 from Carroll-Loye Biological Research did not comport with the
applicable requirements of 40 CFR part 26, subpart K.  The proposed
research does comport with subpart L, as pregnant women and children
were excluded.  Although the ethical concerns identified by the Board
could be remedied, there were sufficient questions raised about the
adequacy of the research design to cast doubt on whether the proposed
research would meet the criteria for IRB approval found under 40 CFR
26.1111(a) (1).  In other words, absent a sound research design, any
exposure of human subjects to risk would be unnecessary and
unjustifiable.

Board Response to the Charge

Occupational Handler Exposure Monitoring Studies

Charge to the Board

The Agricultural Handlers Exposure Task Force (AHETF) had submitted
protocols for five pesticide exposure studies that are part of a larger
research program the AHETF is conducting.  The premise of the AHETF
research program is that data can be used generically by various
stakeholders (e.g., applicants, registrants, EPA, and others) for
calculating exposures for the occupational handlers of pesticides.  The
scope of the AHETF research program was very broad in that it intends to
address exposures related to many job functions in agriculture and also
to assess generally the impacts of various parameters on exposure (e.g.,
How do changes in the pounds of pesticide handled or acres treated
affect exposure levels?).  The protocols submitted for HSRB review
described studies to measure exposures for five specific scenarios. 

The Agency believed these studies had the potential to improve EPA’s
ability to assess the risks of using pesticides because the data would
reflect current agricultural practices, equipment and techniques and
would allow for more refined exposure estimates.  Further, the
monitoring techniques to be used for these studies have been
standardized for use across the AHETF research program. These more
refined and reliable data would allow the Agency to estimate better how
worker exposure levels are affected by changes in various factors such
as the amount of active ingredient handled, type of application
equipment used, application rate used, volumes handled, and personal
protective equipment used.

	It should be noted, however, that the use of the data generated in this
study by the EPA and other stakeholders would depend upon the nature of
the results.  For example, the adequacy of the field or laboratory
quality control data may dictate that correction factors are applied to
adjust monitored exposure levels to account for losses from field
samplers or low performing analytical methods.  

1. AHETF Closed System Mixing/ Loading of Liquids Protocol (AHE34) 

a. Does the proposed research described in Study No. AHE34 from the
Agricultural Handlers Exposure Task Force appear likely to generate
scientifically reliable data, which would be useful, together with other
data, for assessing the potential levels of pesticide exposure received
by people when mixing, loading or applying a liquid pesticide with
closed systems? [Note:  In a few cases, corresponding application events
are also to be monitored; the same question applies to those elements of
the study.]

b. Does the proposed research described in Study No. AHE34 from the
Agricultural Handlers Exposure Task Force appear to comport with the
applicable requirements of 40 CFR part 26, subparts K and L?  

2. AHETF Airblast Application to Trellis Crops in the West Protocol
(AHE36)

a. Does the proposed research described in Study No. AHE36 from the
Agricultural Handlers Exposure Task Force appear likely to generate
scientifically reliable data, which would be useful, together with other
data, for assessing the potential levels of pesticide exposure received
by people when making an airblast application of a pesticide to a
trellis crop under conditions found in the western United States? [Note:
 In a few cases, corresponding mixing/loading events are also to be
monitored; the same question applies to those elements of the study.] 

b. Does the proposed research described in Study No. AHE36 from the
Agricultural Handlers Exposure Task Force appear to comport with the
applicable requirements of 40 CFR part 26, subparts K and L?  

3. AHETF Airblast Application to Trellis Crops in the East Protocol
(AHE37)

a. Does the proposed research described in Study No. AHE37 from the
Agricultural Handlers Exposure Task Force appear likely to generate
scientifically reliable data, which would be useful, together with other
data, for assessing the potential levels of pesticide exposure received
by people when making an airblast application of a pesticide to a
trellis crop under conditions found in the eastern United States? [Note:
 In a few cases, corresponding mixing/loading events are also to be
monitored; the same question applies to those elements of the study.] 

b. Does the proposed research described in Study No. AHE37 from the
Agricultural Handlers Exposure Task Force appear to comport with the
applicable requirements of 40 CFR part 26, subparts K and L?

4. AHETF Closed Cab Airblast Application to Orchards Protocol (AHE38)

a. Does the proposed research described in Study No. AHE38 from the
Agricultural Handlers Exposure Task Force appear likely to generate
scientifically reliable data, which would be useful, together with other
data, for assessing the potential levels of pesticide exposure received
by people when making an airblast application of a pesticide to orchard
crops? [Note:  In a few cases, corresponding mixing/loading events are
also to be monitored; the same question applies to those elements of the
study.] 

b. Does the proposed research described in Study No. AHE38 from the
Agricultural Handlers Exposure Task Force appear to comport with the
applicable requirements of 40 CFR part 26, subparts K and L?  

5. AHETF Fixed-Wing Aerial Application Protocol (AHE42)

a. Does the proposed research described in Study No. AHE42 from the
Agricultural Handlers Exposure Task Force appear likely to generate
scientifically reliable data, which would be useful, together with other
data, for assessing the potential levels of pesticide exposure received
by people making an aerial application of a pesticide from fixed-wing
aircraft? [Note:  In a few cases, corresponding mixing/loading events
are also to be monitored; the same question applies to those elements of
the study.] 

b. Does the proposed research described in Study No. AHE42 from the
Agricultural Handlers Exposure Task Force appear to comport with the
applicable requirements of 40 CFR part 26, subparts K and L?  

Board Response to the Charge

For the Board’s review of the agricultural handler protocols, the
Board decided to focus  its analysis addressing the common strengths,
limitations and overall conclusion of the five protocols.  

Scientific Considerations

Study Overview

The pesticide handler exposure study protocols submitted to the HSRB
were part of a larger project that was initiated in December 2001 by the
Agricultural Handler Exposure Task Force (AHETF). The project will
produce a generic agricultural handler exposure database (AHED™).  EPA
and other regulatory agencies would use this database to calculate
pesticide handler exposures across a wide range of work conditions.  All
of the protocols follow a similar pattern. They involve personal
measurements of inhalation and dermal exposure among a group of workers
who conduct what is referred to as a “scenario”; that is, the study
participant would conduct a specified job task with specified equipment,
handling a particular product formulation that contains one of six
pesticides.

These studies can be referred to as scripted, so as to distinguish them
from purely observational studies. Workers are asked to conduct their
work activities under a set of scripted conditions similar, but maybe
not identical, to those they experience in their normal work activities.
The overall plan for the exposure database had been discussed with
regulators from EPA, California EPA, and Health Canada on a regular
basis. The presentation of these protocols to the HSRB is the first
independent scientific review of the task force project.

The task force had proposed 33 handler exposure scenarios, distinguished
by equipment type, work task, and pesticide formulation. The task force
has already conducted or initiated 14 exposure studies, and has
sponsored four studies. In addition to the five protocols presented to
the HSRB, the task force planned to conduct approximately 40 additional
studies over the next several years.

Each of the protocols focuses on one primary exposure scenario, but all
of the protocols include more than one scenario. The five protocols
reviewed by the HSRB are summarized in Table 1.  

) *

	AHE34	AHE 36	AHE37	AHE38	AHE42

Scenario

1	Closed system mix-load of a liquid	Open cab trellis crop airblast
application	Open cab trellis crop airblast application	Closed cab
orchard crop airblast application	Closed cockpit fixed-wing aircraft
application

N	10	5	5	8	7

Scenario

2	Open or closed cab airblast application	Closed cab trellis crop
airblast application	Closed cab trellis crop airblast application	Open
pour mix-load of a liquid	Open pour mix-load of a liquid

N	0-3	4	4	not specified

	0-3

Scenario

3	Closed cockpit fixed-wing aircraft application	Open pour mix-load of a
liquid 	Open pour mix-load of a wettable powder

Closed system mix-load of a liquid

N	0-3	0-4	4

0-3

Scenario

4	Closed cockpit rotary-wing aircraft application	Closed system mix-load
of a liquid



	N	0-3	0-4



	Total N	13	13	13	not specified	10

Chemical	Malathion 8

(80% a.i.)	Malathion 8

(80% a.i.)	Diazinon 50% WP	Carbaryl 

4lb a.i./gal	Chlorothalonil

6lb a.i./gal

Location	CA	West (CA)	East (NY)	FL and GA	Pacific NW



The task force studies are using six different pesticides: three
organophosphorus insecticides (acephate, diazinon, malathion), one
carbamate insecticide (carbaryl), one organochlorine fungicide
(chlorothalonil), and one triazine herbicide (simazine). A description
of the selection criteria for these compounds was provided as a part of
the task force documentation package. These selection criteria did not
include the toxicity of the compounds, and toxicity was not discussed in
the Agency reviews of the protocols.  However, oral comments from task
force representatives at the June HSRB meeting indicated that the
formulations of these pesticides have been selected such that all fall
into EPA toxicity categories III or IV; i.e., relatively low toxicity
formulations. It was also stated that all workers in these studies wear
long-sleeve shirts, long-legged pants, as well as socks and shoes. 
Protective equipment, such as chemical-resistant gloves and eye
protection, are provided to the workers if required by the pesticide
label.

The database to be developed from the task force studies is intended to
supersede an existing database – the Pesticide Handler Exposure
Database (PHED). This database was developed in the late 1980’s and
early 1990’s through the compilation of existing data. These data were
drawn from both registrant-sponsored studies and studies published in
the scientific literature. The studies included in PHED used a different
method for measuring dermal exposure. This method, known as the “patch
technique” (deposition coupons distributed over body regions attached
to the outer layer of clothing), has served as the standard method for
such studies since the 1960’s. When coupled with a hand rinse
technique, it provides an estimate of exposure to all body surfaces.

The documents submitted by the AHETF in support of the proposed exposure
studies consisted of the following:

Cover letter dated May 24, 2006

Analytical method validation reports for 5 of the 6 pesticides (missing
simazine)

List of 33 exposure scenarios

Description of selection requirements for surrogate compounds

32 standard operating procedure (SOP) documents

A generic field exposure monitoring protocol

5 exposure study protocols: AHE34, 36, 37, 38, and 42

IRB documents related to each protocol

In addition, the HSRB received an EPA review for each protocol, EPA,
Office of Pesticide Program guideline documents, and several general
documents on pesticide handler exposure. Finally, the AHETF provided
public comments (AHETF 2006) containing comments on the EPA review of
the five protocols.

Critique of Study

General Scientific Criteria

The primary aim of these studies is to generate personal measurement
data on pesticide handlers suitable for use in an agricultural handler
exposure database. The notion that such a generic database for pesticide
handlers can be developed is supported by substantial scientific
evidence. This evidence indicated that occupational pesticide exposure
in agriculture is largely process rather than chemical-dependent
(excluding chemicals with high volatility). Thus, if sufficient data can
be collected on the key variables that influence exposure, then a
database can be developed to estimate exposure for a wide range of
exposure scenarios. A major concern of the HSRB was that these protocols
included too many variables, and that, even when combined with the full
complement of studies proposed, the database would be inadequate for
meaningful exposure estimates.

The process that has guided the exposure database project had some
significant limitations. The database project has been developed over
the past 5 years by a pesticide industry task force with the input of
regulatory staff from EPA, California, and Canada. Such an ambitious
undertaking would have benefited from an initial independent scientific
peer review, particularly of the study design and statistical plan, as
the HSRB finds itself raising some fundamental questions mid-stream in
the project. Input from the labor community would also have enhanced the
project regarding procedures such as subject recruitment, selection of
pesticides to be used in individual studies, and informed consent. The
purpose of the project, after all, is to develop data to estimate worker
risks. It seems reasonable to give those who would be taking the risks
an opportunity to contribute to the design of the project.

In regard to justification for new human studies, the Agency currently
uses an existing generic pesticide handler exposure database, known as
PHED. It is recognized that new data have not been added to PHED in a
number of years, and that the existing data have a number of scientific
limitations.  However, the inadequacy of PHED was not documented in the
protocols. The Agency had not provided a compelling justification for
these new human studies in the materials provided.

Benefits of the study were not described in the protocols. However, the
AHETF comments (AHETF 2006) and the EPA review documents provided some
general information regarding the role of a handler exposure database in
EPA’s regulatory process.  It was not possible to determine the
likelihood that the benefits would be realized, since the protocols did
not include a description of the full database and how it would be used.

		Study Design Criteria 

The purpose of these studies was clearly defined. The objective was to
collect high quality personal measurement data for use in a generic
exposure database. The protocols reviewed by the Board should be able to
produce such data.

Approximate sample sizes were presented in the protocols and discussed
in more detail in the Agency reviews of the protocols. Within each
protocol, the sample sizes for particular scenarios were quite small.
All of the protocols contained multiple scenarios, with sample size per
scenario ranging from 1-10 (see Table 1). The inability to define
exactly how many samples would be collected in each proposed study was
understandable, since the task force was attempting to take advantage of
‘real-world’ conditions. Weather, logistical challenges, and grower
decisions regarding pest management can all affect the number of workers
available for a given study. The effort to study exposures under
realistic conditions required expenditure of significant resources, and
was viewed by the HSRB as highly commendable. The HSRB understood that
these protocols should not be viewed as “stand-alone” studies, since
data from these studies would be combined with other similar studies.
Presumably all of the data collected in these five studies would be
allocated to one of the 33 exposure scenarios outlined by the task
force. Under these circumstances, there was insufficient information for
the HSRB to evaluate the adequacy of the sample size.

In regard to dose levels, participants would be handling varying amounts
of pesticides under variable exposure conditions. The HSRB presumed that
the conditions outlined in the protocols and reviewed by the Agency all
fall within parameters on the label. Actual dose during these studies
would likely be lower than normal, due to the wearing of a whole-body
cotton garment, and strict observance of label instructions.

Participation Criteria 

Participants are referred to as ‘replicates’ both in the AHETF
protocols and in the Agency reviews. This term is problematic from a
scientific perspective, since it is used to refer to both a series of
independent observations (e.g., three persons doing the same thing one
time) and a series of repeated measures (e.g., one person doing the same
thing three times). This language needs to be altered such that
different terms are used for repeated measures on one person as compared
to observations on unique individuals. The AHE34 protocol, for example,
indicated that “ten different mixer/loader workers (or replicates)
will be monitored . . . each mixer/loader replicate should be performed
by a separate worker.”  This type of awkward description could be
eliminated through use of unambiguous terminology.

The protocols indicated that the participants would be “experienced
workers” recruited through their employers, but there was no detailed
description of the recruitment procedures, nor were there clear
inclusion/exclusion criteria other than age and pregnancy status. It was
hard to tell whether the workers who volunteer for a protocol exposure
study would be representative of the typical worker.  Random sampling
from a group of eligible workers would improve this aspect of the
protocols.

Measurement Criteria 

AHETF investigators are using three different methods to measure skin
exposure: cotton garments (whole body dosimeters), hand rinse, and
face/neck wipes. The cotton garments should be able to capture pesticide
that would normally be deposited on skin.  However, no method for
preventing or monitoring garment breakthrough was presented. If
breakthrough occurs, the dermal exposure measurements would
underestimate true exposure. The hand rinse method and face/neck wipe
both measure the amount of material that can be removed from the skin at
the particular time of the sampling. This amount is some fraction of the
total material deposited on the skin, since some of the material would
have been absorbed into the skin. This method is likely to underestimate
the true exposure. Published laboratory and field studies have indicated
that the fraction of the amount deposited on skin that can be removed by
rinsing or wiping can be quite variable, depending on the nature of the
chemical, its formulation, skin characteristics, and the length of time
the chemical has been in contact with the skin.  In particular, the
face/neck wipe method may seriously underestimate exposure to these
surfaces. This method was not among the methods presented by the Agency
in its 875 guidelines (Occupational and Residential Exposure Test
Guidelines: OPPTS 875.1100 Dermal Exposure – Outdoor), and has not
been validated. The accuracy of these measurements could be improved
through the conduct of laboratory removal efficiency studies. The 875
guidelines do not require removal efficiency studies, but they do
indicate that investigators should address this concern. If such method
validation studies are contemplated, the HSRB recommended that they be
conducted as independent scientific studies published in the
peer-reviewed literature.

In summary, all of the methods for dermal exposure measurements have the
potential to underestimate exposure. The study investigators should
acknowledge this problem in the protocols, and explain what steps, if
any, they have taken to improve or verify the accuracy of the
measurements.

The protocols stated that hand rinse and wipe samples may be collected
multiple times during the work period (e.g., prior to eating, whenever a
worker would normally wash hands), and that this would vary from worker
to worker. The protocols did not explain how multiple measurements from
a single worker would be combined. They also did not discuss whether or
not samples across workers with different rinse/wipe regimens can be
considered comparable. For example, can the amount of pesticide
recovered in a single hand rinse from a worker at the end of the study
period be put in the same database as that from a worker who had four
hand rinses across the study period? The answer would seem to be
“no” from a sampling perspective, given the nature of dermal
absorption processes.

The quality assurance components of the protocols are of high quality.
There was substantial documentation regarding the reliability of
analytical methods available for each of the sampling media to be used.
There were detailed standard operating procedure documents for field and
laboratory quality assurance activities.

Statistical Analysis Critera 

An inadequate statistical analysis plan was provided in the protocols.
The HSRB identified this deficiency as the most critical scientific
limitation for these protocols.  There was a need for a more
professional and comprehensive treatment of statistical issues in the
analysis of data, and in the design of individual protocols. Chief among
these issues was the question of statistical power. It is critical to
address the HSRB’s concern that the present design calls in most cases
for a single observation per experimental condition. In other words, it
appears that the present studies are intended to be parsed in terms of
formulation, container size, frequency of worker activity, equipment,
air temperature, wind speed, relative humidity, amount of cloud cover,
rainfall, crop, amount of material handled, rate of application, acreage
treated, and geographic location (along with other possible qualifiers).
As a result, the number of variables to be evaluated appears to approach
or even exceed the total number of subjects for a given scenario. One
may hope that some useful information might yet emerge from a properly
performed analysis of the full data set coming from studies involving
different chemicals, sites, and conditions. What is needed now, however,
is a cogent and thoughtful discussion, in the protocol, of just what can
be accomplished along these lines, and an explanation of how it can be
accomplished. Further thought may lead to the conclusion that the
current data-gathering plan is in fact overly optimistic in regard to
the issue of statistical power. It would then be essential to
restructure the plan and change the study design to ensure that the
enormous effort in this large and important project would not be wasted.

If the goal of these studies was to estimate the distribution of
exposures across a variety of application scenarios, it would be
important to include true repeated measures for at least some of these
scenarios to assess the extent of within-worker variability.

Laboratory and Field Conditions

The protocol states that the field conditions were selected to be
representative of real-world use of pesticides. In this regard,
participants are experienced workers, are allowed to wear their own
clothing, and conduct normal work activities in an actual agricultural
setting. Studies are distributed across the U.S. and across the year in
an attempt to develop a range of exposure conditions. While laudable in
scope, it is important to restate the concern that the large number of
variables included in these studies may prove extremely problematic for
analysis. 

The issue of potential heat stress was discussed at length at the HSRB
public meeting. The HSRB concluded that the protocols should include
explicit criteria for halting a study due to heat stress risk.

The protocols stated that workers would be monitored “during a period
of time representative of a full day’s work”. The protocols also
indicated that monitoring times would  conform to a “typical”
workday.  However, none of the protocols defined the typical work period
for the specific tasks to be studied.  Instead, the protocols stated
that monitoring time “will involve work periods with a target of 4
hours.” This language suggested that even four hours of monitoring
might not be achieved in some cases, and there was no indication that
workdays as long as 8 or 9 hours would ever be monitored.  It was not
clear to the HSRB that these studies would necessarily reflect a full
workshift or a “typical” workday. Many factors can influence the
length of the workday, including weather conditions, and the need to
“get the job done” due to pest pressures or the stage of crop
development. Fatigue is an important factor that can affect exposure,
and usually inattention occurs at the end of the day. The HSRB
recommended that the protocols document the time of a typical workday
(or the range of these times) for each scenario, and that the monitoring
time be based on this information.

	Special Concerns Regarding Use of Diazinon in Protocol AHE37

Protocol AHE37 involves handling a wettable powder (50WP) formulation of
diazinon. Workers would be monitored during open cab trellis crop
airblast applications and open pour mixing-loading operations. The HSRB
was concerned that these practices are not consistent with current
Agency policy.  The Agency’s May 2004 interim registration eligibility
document (IRED) for diazinon stated that engineering controls are
required during handling.  The “IRED Facts for Diazinon” states,
“All application equipment must use lock and load engineering
controls. All wettable powder formulations must be packaged in
water-soluble bags. Closed cabs are required for all ground equipment,
except for applications to apples.” The IRED Executive Summary further
stated, “Occupational exposure to diazinon is of concern to the
Agency.  For agricultural uses of diazinon, most mixer/loader/applicator
risk scenarios currently exceed the Agency’s level of concern (i.e.,
MOEs are less than 100 for dermal exposure and MOEs are less than 300
for inhalation exposure).  Taking into consideration both the risks and
benefits of these uses, EPA has determined that most agricultural uses
may continue with the adoption of the following mitigation measures: . .
. engineering controls for mixers and loaders and closed cabs for
applicators for all application scenarios . . .” The above statements
indicate that open pouring of diazinon is not permitted, and that open
cab airblast applications are not permitted in trellis crops.

HSRB Consensus and Rationale

The five studies presented for HSRB review were components of a
large-scale exercise to create a contemporary database on occupational
exposure to agricultural pesticides. The undertaking is in itself likely
to be worthwhile in quantifying and improving our understanding of the
exposures of and risks to pesticide handlers. The potential benefits are
large and the risks appear to be relatively modest. However, the
materials supplied for HSRB review failed to deal adequately with risks
and benefits. None of these protocols can be properly evaluated in
regard to scientific validity because they lacked: (1) a developed
rationale documenting the need for new data; (2) a clear and appropriate
plan for the handling of the data (including its statistical analysis),
and (3) an explanation of the uses to which the data would be put and
adequate sample sizes and protocols for repeated measures to
appropriately estimate exposures within individuals and between
scenarios. These points need to be addressed, at least briefly, in each
specific protocol and, more fully, in a separate and new “governing
document” that is not simply a generic description of the planned
activities.

Additional validation studies are recommended to determine the extent to
which dermal exposure measurements may underestimate true exposure.
Laboratory-based removal efficiency studies or field-based biomonitoring
studies could be conducted to achieve this goal. Such studies should be
published in the peer-reviewed literature. Broader participation of the
scientific community and of parties with a direct interest in the
database project, such as the labor community, would likely improve the
quality of the database and enhance the credibility of its use in risk
assessments.

	The HSRB recommended that specific criteria for cessation due to heat
stress be included in these worker exposure protocols, and that the
protocols included a heat stress management plan.  In addition, the HSRB
recommended that the length of each study should be truly representative
of a full workday, and that each protocol should document the basis for
the proposed duration of the study.

	The HSRB was gratified to receive the Agency’s response to its query
regarding the use of diazinon in the AHE37. It is the understanding of
the HSRB that the Agency would inform the AHETF that it needs to
identify a pesticide other than diazinon in this protocol to evaluate
exposures associated with open pour activities and applications using
open cabs, and that the Agency would ensure that future protocols comply
with the most current risk mitigation measures specified in IREDs and
REDs.

Board Response to the Charge

Ethical Considerations

Background

These five studies are part of a series of studies that are to be
conducted by the AHETF, a coalition of 19 pesticide registrants that was
formed in December 2001 to share resources in the design, evaluation,
and development of a proprietary agricultural mixer/loader and
applicator exposure database for use in regulatory risk assessments.

The study protocols were designed by AHETF investigators after joint
discussions with the US EPA, Health Canada, and the California
Department of Pesticide Regulation, in accordance with the
recommendations of such guidance documents as: 1) US EPA, Occupational
and Residential Exposure Test Guidelines, Series 875.1000 through
875.1600 (1996); and 2) US EPA, Working Draft - Occupational and
Residential Exposure Test Guidelines, Series 875 Group-B,
Postapplication Exposure Monitoring Test Guidelines Version 5.4 (1998).
The supporting and supplementary study documents also assert compliance
with the Good Laboratory Practice (GLP) Standards established by the
1972 amendment to the Federal Insecticide, Fungicide, and Rodenticide
Act (FIFRA) (§40CFR160). Finally, these protocols were reviewed and
approved by the Western Institutional Review Board (WIRB) of Olympia,
Washington, prior to submission to the Agency.

The aims of these studies are to provide critical exposure information
for individuals who mix, load, and apply agricultural pesticides. 
Agricultural producers (“growers”) would be recruited by the study
coordinators several months prior to initiation of the study; in
exchange for their participation in the study, each grower would
receive, free of charge, an amount of liquid pesticide equivalent to the
normal quantity mixed and loaded into closed-mixing systems and spray
rigs for the duration of the study (expected to be a single day).
Participating growers also may be asked to recruit other growers and
pesticide applicators into the research study.

 

Study investigators would recruit agricultural handlers on-site;
volunteers would receive $100/day for their participation in addition to
their regular pay. Voluntary informed consent would be solicited by
study investigators, and will be documented using a standardized
informed consent form. Because the study participants would be recruited
from a pool of experienced agricultural workers who routinely mix and
load liquid pesticides as part of their normal duties, the AHETF had
argued that participation in this study presents a negligible increase
in pesticide exposure risk to volunteers. In accordance with the newly
promulgated provisions in the EPA’s Final Human Studies Rule
(§40CFR26.1701 - 26.1704), minors and pregnant women are explicitly
excluded from participation, the latter being confirmed by requiring all
female volunteers to undergo a self-administered over-the-counter
pregnancy test on the day of the study.

Dermal exposure to pesticides would be ascertained through hand rinses
and face/neck wipes, as well as the use of long cotton underwear – as
a surrogate for skin – to be worn under the study participant’s
clothing. In addition to the long underwear, all participants would be
required to wear long sleeved shirts and long pants, shoes plus socks,
in accordance with accepted worker protection standards. Volunteers may
wear their own clothing provided they are freshly laundered;
alternatively, the AHETF would provide freshly laundered clothing. Any
personal protective equipment (PPE) that may also be required, such as
chemical resistant gloves and protective eyewear, will be provided. At
the conclusion of the four-hour study observation, the long underwear
would be removed and subjected to laboratory analyses to estimate
whole-body dermal pesticide exposure. Study participants would also be
asked to wear OSHA Versatile Samplers (OVS) outfitted with glass
filters, XAD-2 sorbent, and tygon tubes to measure inhalation exposure. 
The tubes would be attached to the volunteer’s collars with the
openings positioned in their breathing zones. By using such
state-of-the-art monitoring techniques, the AHETF argues, this study
would provide critical exposure information for individuals who mix/load
liquid agricultural pesticides. 

Strengths and Limitations

These studies would provide critical exposure information for
individuals who mix/load and apply agricultural pesticides. It is also
believed that the monitoring techniques proposed for these studies
represent the current state-of-the-art. However, the Agency also
recognized that use of the data resulting from this studies would take
careful scrutiny and may require a number of adjustments depending upon
the results. Finally, the overall design of these studies should be
considered in the context of the goals of the AHETF which are to develop
a broad-based database that can be generically used as a predictive tool
for estimating exposures to pesticide handlers and that the
interpretation of the results of these studies may or may not
necessitate the need for additional monitoring data.

The Board concurred with the factual observations of the strengths and
weaknesses of the studies, as detailed in the EPA’s Initial Ethics
Review of the AHETF Template Protocol and each individual study
protocol. The Board concluded that, given the deficiencies noted by the
EPA, the proposed research described in the AHETF Template Protocol and
each individual study protocol do not comport with the applicable
requirements of §40CFR26, subparts K and L. Furthermore, the AHETF and
WIRB failed to provide all of the documents necessary to be in
compliance with the requirements of §40CFR §26.1125, subpart M.

Although public comments from several members of the AHETF helped
assuage some of the Board’s concerns, the members of the HSRB believed
that further comments about this protocol were warranted. The comments
below are grouped into four broad categories: (1) whether the study was
designed to adequately minimize risk to study participants; (2) whether
the documentation and process of study subject enrollment was sufficient
to meet prevailing standards of voluntary informed consent; (3) whether
study participants would be adequately compensated in the event of a
study-related injury; and (4) whether appropriate alternatives to
participation are provided.

Minimization of Risks to Study Participants

This study proposes to measure dermal and inhalation exposure to liquid
pesticides by agricultural handlers who usually perform pesticide
mixing, loading, and application as part of their daily routine.
However, it was unclear to Board members, given the semi-scripted nature
of the protocol provided, as to whether or not study participants would
be exposed to greater quantities of these compounds than would normally
occur. Are the studies proposed purely observational in nature, or are
study investigators intervening by requesting that study participants
use different types and quantities of pesticide, or different mixing,
loading, and application methods, than they normally would? If the
latter is true, the assumption that this study represents a negligible
increase in pesticide exposure risk to volunteers may be unfounded.
Several Board members also expressed concern that the additional
requirements for donning and removing the equipment used to measure
pesticide exposure may inadvertently lengthen the participant’s normal
work day. If so, this should be clearly described during the consent
process, as should the question of whether the $100 paid for study
participation is expected, in whole or in part, to compensate for the
extension of the work day.

. Although pesticide mixing instructions and Material Safety Data Sheets
are made available to study participants, given that many agricultural
workers may not be fluent in English (or may even be illiterate), a
clear plan for ensuring that volunteers are properly educated in
minimizing their exposure to these compounds should be included.
Furthermore, study investigators may want to make arrangements to
provide volunteers with the results of the study following completion.

One of the greatest risks to study participants is heat-related illness,
given that dermal exposure to pesticides will be determined by asking
volunteers to wear long underwear in addition to their normal protective
equipment (e.g., long sleeved shirts and long pants, and other
applicable protective gear). Although study coordinators are expected to
be vigilant for signs of heat-related illness among volunteers, in order
to minimize the risks posed to the study participants the protocol also
should include: a) explicit starting and stopping criteria based on a
quantifiable measure like ambient temperature or heat index; and b) a
clear description of the symptoms of heat-related illness in the
informed consent documents. There should also be a clear plan for
reporting any heat-related illness (or, for that matter, any other
adverse event) to the study investigators, Western IRB, and the EPA.

Because some of the study participants may be undocumented immigrants,
measures to ensure strict confidentiality should be developed. Many
undocumented workers, for example, may be loathe to report any adverse
study-related event requiring medical attention or hospitalization if
they believe that their illegal status will be reported to immigration
authorities. Alternatively, study investigators may wish to require
documentation of citizenship or immigration status as part of the
inclusion criteria for recruiting study participants. In addition,
because many pregnant day-laborers may fear job loss in the event that
their employer learns of their condition, extra care should be taken to
keep the results of over-the-counter pregnancy tests private. 

Voluntary Informed Consent

Several Board members felt that the AHETF protocol, as provided, lacked
sufficient safeguards to ensure that all study participants were
adequately informed about the risks, benefits and alternatives to
participation in the study. For example, it was felt that the informed
consent documents provided were written at too high of a reading and
comprehension level. Given the sociodemographic characteristics of the
farm worker population, many of the study participants may have limited
education, may speak English as a second or even a third language, or
may even be illiterate. Study investigators should develop a clear
consent document which – in addition to including a more detailed
description of risks (including the risks of the pesticides being
handled) as described previously, as well as a clear distinction between
what comprises research versus normal activities – is written at a
lower grade-level and translated into the various languages likely to be
spoken by study participants. A brief oral test of comprehension should
also be developed, with volunteers required to demonstrate a clear
understanding of the purposes and the risks of the study prior to
enrollment. 

The Board also expressed concern about the potentially coercive nature
of the study, given the potential for study participants to believe that
there is a direct relationship between study investigators and growers.
Absent additional safeguards, the “gift” of study pesticide to the
growers may contribute to undue influence on employees to participate in
the research. Western IRB, in its initial review of several of the AHETF
protocols, recommended that “extra care” be taken during the
recruitment and consent process to minimize coercion or undue influence
on study participants. However, no documentation was provided to the
HSRB as to how the AHETF addressed WIRB’s concern. For example, there
was no evidence to suggest that AHETF researchers solicited the help of
the farm-workers themselves or other community leaders to ensure that
study participants would not be covertly or overtly coerced into
participating in the study. The rights of participants to withdraw from
the study at any time also should be emphasized. It is unclear from the
informed consent or other study documents, for instance, as to whether
volunteers are entitled to receive monetary payment even if they chose
to withdraw during the course of the study. Although the Board was
reassured during the discussion that sufficient alternate work was
available, the protocol also failed to specify that workers would still
be paid for a day’s labor even if they refused to participate in the
research.

Compensation for Injury to Study Participants

The study protocol and informed consent documents state that: “If [a
study participant is] injured as a result of being in this study,
treatment will be available from a health professional at a nearby
medical facility.  The costs of such treatment will be covered by the
AHETF.  This does not cover any injuries resulting from [the
volunteer’s] normal activities.” Given the nature of the study
design, however, it is unclear whether a distinction between injuries
resulting from normal work activities versus participation in this study
can be made. Two of the symptoms of heat exhaustion, for example, are
dizziness and loss of coordination – will study coordinators be able
to distinguish between an accidental injury caused by clumsiness versus
an injury resulting from potentially-unrecognized symptoms of
heat-related illness? In light of these concerns, the Board recommends
that the AEHTF cover medical treatment for all participant illness and
injury occurring during the study period (i.e., the day of the test).

Alternatives to Participation

As noted above, the design of this study involves collaboration between
the researchers and growers in which the growers receive, free of
charge, a particular pesticide that they are required to apply to their
fields on the day of the study. That arrangement will lead in many
circumstances (except of the few coincidental instances when the grower
had already planned to use that chemical on that day) to a change in the
pesticide being applied by the grower.

The following question thus arises: What alternatives are offered to
agricultural handlers working for that grower who choose not to
participate in the study? One option is that they could be offered the
choice of applying that pesticide that day, but not needing to
participate in any other study procedures (such as wearing the long
underwear). Some members of the Board believed that if that is the only
alternative to participation, then this aspect of the study would not
comply with 40 C.F.R. Part 26, Subpart K. A primary purpose of the EPA
rule is to prevent a person from being intentionally exposed to a
pesticide without their voluntary informed consent. The EPA emphasized
this point when it promulgated the final version of its rule, commenting
that the term “research involving intentional exposure” covers
“any research on a substance, unless the subjects of the research
retain complete control over whether, when, and how they are exposed to
the substance.” 71 Fed. Reg. 6138, 6146 (2006). 

Some members of the Board accordingly concluded that for agricultural
workers who had pre-existing expectations of earning money working for
the grower on the day of the study (either as employees or as
independent contractors with contractual expectations of working that
day), the protocol must provide them alternatives for earning that same
amount of money that do not require them to apply the pesticide used in
the study. Acceptable alternatives could include applying some other
pesticide they have in the past applied, performing some other task they
regularly perform, or being paid their expected earnings without needing
to work. Absent such alternatives, the protocol would appear to be
inappropriately coercing such persons into applying the study compound
or else losing the money they expected to earn that day.

HSRB Consensus and Rationale

The Board concurred with the initial assessment of the Agency that the
studies submitted for review failed to meet the ethical requirements
established in the 40CFR26. 

The Board determined the proposed occupational handlers exposure studies
do not comport with the applicable requirements of 40CFR26, subparts K
and L. However, the deficiencies noted, while significant, were not
irreparable. 

REFERENCES

AHETF 2006.  Comments to the Human Studies Review Board (HSRB) Regarding
Five Protocols Submitted for Review.  June 20, 2006.

NAS 2004.  Intentional Dosing Studies for EPA Regulatory Purposes:
Scientific and Ethical Issues. National Academy Press. 

NBAC 2001.  Ethical and Policy Issues in Research Involving Human
Participants.  National Bioethics Advisory Commission. 
www.bioethics.gov

Prentiss AM.  1937.  Chemical in War: A Treatise on Chemical Warfare. 
McGraw-Hill Book Company, Inc.  New York.  

USEPA 2006a.  Human Studies Review Board.  Weight of Evidence
Determination for Chloropicrin.  June 7, 2006

USEPA 2006b. Ethics Review of Protocol for Human Study of Tick Repellent
Performance.   June 9, 2006.

USEPA 2006c. Ethics Review of Protocol for Human Study of Mosquito
Repellent Performance  June 9, 2006.

USGPO.  1949.  Trials of War Criminals before the Nuremberg Military
Tribunals under Control Council Law No. 10, Vol. 2, pp. 181-182.
Washington, D.C.

Proposed Final Draft v. 1 Dated August 28, 2006; Do Not Cite or Quote 

  PAGE  1  of   NUMPAGES  67 

See OPP memorandum.  One way to address OPP’s comment would be to
delete this sentence.

See OPP memorandum.  One way to address OPP’s comment would be to
delete this sentence.

The protocol actually described testing with three different
formulations of the test material:  an aerosol, a lotion, and a spray.

>

A

`

c

Ì

Ù

Ú

ô

õ

Ë

Ì

Ù

Ú

ô

õ

	



Ô

Õ

Ú

Ü

Ó

Ô

„@

„Ð^„@

 hqj

 hqj

  hqj

 hqj

 hqj

@

 hqj

h¢

h¢

h

 hqj

[

f

Ž

’

©

«

¶

¾

Ã

Ù

ð

ô

5

K

N

 hqj

A

B

¸

¹

5

6

J

K

h

 hqj

 hqj

 hqj

 hqj

 hqj

 hqj

…

Œ

OJ

QJ

OJ

QJ

 hqj

OJ

QJ

OJ

QJ

 hqj

hA-

 considered again by the HSRB, the Agency will discuss the overall
research plan which will include a discussion of the number of
monitoring events to be collected in order to address each scenario.

The Agency wants the research to be performed in a manner that minimizes
risk and produces data on exposure that reflects actual use practices,
including a reasonable range of care in handling the test material. 
While training subjects should ensure that they follow applicable
labeling requirements, providing special training may lead subjects not
to behave in a typical manner.  The Agency would appreciate HSRB comment
on this tension.

