Download:
pdf |
pdfSupporting Statement A
Longitudinal Investigation of Fertility and the Environment - NICHD
March 25, 2008
Germaine M. Buck Louis, MS, PhD
Chief & Senior Investigator
Epidemiology Branch
Division of Epidemiology, Statistics & Prevention Research
Eunice Kennedy Shriver National Institute of Child Health & Human Development
Department of Health and Human Services
6100 Executive Blvd., Room 7B03
Rockville, MD 20852
Tel: (301) 496-6155
Fax: (301) 402-2084
email: [email protected]
Table of Contents
A. JUSTIFICATION
A.1.
Circumstances Making the Collection of Information
Necessary .......................................................................................................
1
A.2.
Purpose and Use of the Information Collection.............................................
1
A.3.
Use of Information Technology and Burden Reduction................................
6
A.4.
Efforts to Identify Duplication and Use of Similar Information....................
7
A.5.
Impact on Small Businesses or Other Small Entities.....................................
8
A.6.
Consequences of Collecting the Information Less Frequently ......................
8
A.7.
Special Circumstances Relating to the Guidelines of 5 CFR 1320.5………… 9
A.8.
Comments in Response to the Federal Register Notice and Efforts to Consult
Outside Agency.............................................................................................. 11
A.9.
Explanation of Any Payment or Gift to Respondents.................................... 12
A.10.
Assurance of Confidentiality Provided to Respondents ............................... 12
A.11. Justification for Sensitive Questions............................................................. 15
A.12.
Estimates of Hour Burden Including Annualized Hourly Costs................... 16
A.13.
Estimate of Other Total Annual Cost Burden to Respondents or
Recordkeepers................................................................................................ 19
A.14.
Annualized Cost to the Federal Government................................................ 19
A.15. Explanation for Program Changes or Adjustments....................................... 20
A.16.
Plans for Tabulation and Publication and Project Time Schedule................ 20
A.17. Reason(s) Display of OMB Expiration Date is Inappropriate ...................... 25
A.18. Exceptions to Certification for Paperwork Reduction Act
Submission.................................................................................................... 26
References…………………………………………………………………………..
ii
27
LIST OF TABLES
Table A.12-1a. Estimates of Hour Burden by Gender of Respondents and Type of
Response .................................................................................................... 18
Table A.12-1b. Expected Distribution of Pregnancies and Drop Outs by Month of
Participation ............................................................................................... 19
Table A.12-2. Annualized Costs to Respondents .................................................... 19
Table A.13-1. Project Time Schedule...................................................................... 25
iii
LIST OF ATTACHMENTS
Attachment 1.a. Data Collection Schedule
Attachments 1.b.-o. Data Collection Instruments
1.b. Screening Instrument – Female Version
1.c. Screening Instrument – Male Version
1.d. Female Baseline Questionnaire
1.e. Male Baseline Questionnaire
1.f. Female Time to Pregnancy Journal
1.g. Male Time to Pregnancy Journal
1.h. Pregnancy Journal
1.i. Baseline Saliva Sample Instruction Sheet
1.j
High Fertility Saliva Sample Instruction Sheet
1.k. Semen Specimen Instruction Sheet
1.l.
Fertility Monitor Instruction Sheet
1.m. Pregnancy Test Instruction Sheet
1.n.
Blood Specimen Information Sheet
1.o. Urine Specimen Information Sheet
Attachments 2.a.-c. Brochure, Letter and Informed Consent Forms
2.a.
Example Letter of Introduction
2.b.
Study Brochure
2.c.
Informed Consent Documents (Michigan & Texas Research Sites)
Attachment 3. Institutional Review Board Approvals and National Institute of Health
Reliance Agreement (Michigan & Texas Research Sites and Data Coordinating Center)
iv
Attachments 4.a.-d. Other Supporting Documentation
4.a. Listing of Testing, Analytes & Metals to be Measured by Matrix and Required
Volume
4.b. Summary of Published Prospective Pregnancy Studies with Preconception
Enrollment
4.c. Summary of Public Comments and Response by the NICHD
4.d. Illustrations of Data Collection Devices
60-day Federal Register Notice (60-Day FRN)
30-Day Federal Register Notice (30-Day FRN)
v
JUSTIFICATION
A. 1. Circumstances Making the Collection of Information Necessary`
This is a request submitted by the Eunice Kennedy Shriver National Institute of Child Health and
Human Development (NICHD), the National Institutes of Health that the Office of Management
and Budget (OMB) approve under the Paperwork Reduction Act of 1995, clearance for the
NICHD to continue an epidemiologic study entitled “Longitudinal Investigation of Fertility and
the Environment”. The original information collection request was approved (OMB Clearance
0925-0543) following publication in the Federal Register on January 9, 2004, page 1589 and
December 2, 2004, page 70153. The 60-Day FRN for the current application was published in
the Federal Register on January 16, 2008, Vol.73, No. 11, p.2925. This research will be done by
the Division of Epidemiology, Statistics and Prevention Research consistent with the NICHD’s
mission to conduct basic, clinical and epidemiological research focusing on factors and processes
associated with human reproduction and development, thereby, ensuring the birth of healthy
infants capable of reaching full adult potential unimpaired by disabilities. The authority for the
NICHD to conduct research studies is contained in The Public Health Service Act, which
outlines the research and information dissemination mission of the NICHD in the area of child
health [42 USC 285g].
A. 2. Purpose and Use of Information Collection
The data collected as a part of this study will be analyzed to answer growing concerns about the
effect of persistent environmental chemicals, in the context of lifestyle and a couple’s previous
1
reproductive history, on human reproduction and development. Findings from wildlife
populations suggest that persistent environmental agents are adversely affecting reproduction and
development (Thomas & Colborn, 1992), particularly so-called endocrine disrupting compounds
(EDCs). These compounds are alleged to have estrogenic, anti-estrogenic, androgenic, or antiandrogenic activity and can interfere with the synthesis, secretion, transport, binding, action, or
elimination of the body’s natural hormones that are required for homeostasis, reproduction,
development, or behavior (Crisp et al., 1998). EDCs are reported to be capable of enhancing
(agonist) or inhibiting (antagonist) hormone activity. A recent summary has identified
exogenous substances that may affect sex hormone function in humans: 1) estrogenic effects of
high potency (e.g., diethylstilbestrol), medium potency (e.g., dietary phytoestrogens), and low
potency (e.g., bisphenol A, octylphenol and nonylphenol pesticides); 2) anti-androgenic effects
(e.g., pesticides) and 3) other effects (e.g., dioxin-like PCBs) (Joffe, 2003).
Recent findings suggest reductions in male fecundity, and disruption in mammalian oocyte
maturation and follicle physiology associated with exposure to organochlorine chemicals and
polychlorinated biphenyls (Faroon et al., 2001; Dallinga et al., 2002; Rozati et al., 2002; Hauser
et al., 2002; Pocar et al., 2003). Many human studies focusing on PCB exposure and
reproduction and development have relied upon proxy measures of exposure (e.g., fish
consumption) or retrospective collection of exposure and outcome data. Virtually no
prospectively collected longitudinal data exist regarding the effects of persistent environmental
chemicals on human reproduction and development in the context of lifestyle factors that are also
suspected of exerting deleterious effects on human reproduction and development. The absence
of convincing data is in sharp contrast to public belief that environmental chemicals adversely
2
impact human health (Sharpe and Irvine 2004; SIRC 2004). Further concern about EDCs has
resulted in a congressionally mandated response by the U.S. Environmental Protection Agency
(EPA) to form the Endocrine Disruptor Screening and Testing Advisory Committee (Brown
2003). The Agency for Toxic Substances and Disease Registry (ATSDR) is charged with the
preparation of toxicologic profiles of environmental chemicals at geographic sites of concern for
human health impact.
The NICHD investigators in conjunction with principal investigators (PIs) at two competitively
awarded research institutions are leading data collection and/or analyses. One of those two sites
has completed enrollment, leaving only one site at which data collection is still ongoing. A data
coordinating center is overseeing the collection of de-identified information from the research
sites and the two governmental laboratories quantifying exposures and male fecundity, i.e.,
National Center for Environmental Health and the National Institute for Occupational Safety and
Health, Centers for Disease Control and Prevention.
This research originally proposed to recruit 960 couples who are interested in becoming pregnant
and willing to participate in a longitudinal study. Given the uncertainty associated with the
percentage of U.S. couples planning a pregnancy at any point in time, our design assumptions
were conservative resulting in a higher estimated cohort size. To date, fewer than expected
couples were enrolled during the first three years of the project (n=350), predominantly due to
the fact that more couples were ineligible for participation than had been originally estimated.
Our statistical methodology and underlying assumptions have been revised in the context of new
information and empirical evidence generated from the LIFE Study as described throughout this
3
document. The current revised study plan is to enroll a total of 500 couples (i.e., 150 additional
couples for this OMB request), a sample size that will not compromise the main study objectives,
given that additional exposure data has relaxed some of our earlier assumptions. Prime reasons
for needing an OMB extension include: 1) our Texas site was required to temporarily stop
recruitment during Hurricane Rita and in the following year given extensive flooding in the
region and 2) one of our original three study sites was closed for budgetary reasons leaving two
remaining sites for available recruitment. One of the two remaining sites has completed
recruitment, given that they depleted their sampling framework. Data are being collected using a
combination of telephone interview for purposes of initial screening, in-person questionnaire for
baseline information, longitudinal collection of entries in a daily fertility journal or monthly
pregnancy journal, and electronic fertility monitors. The longitudinal capture of information is
necessary given the timed, highly interrelated aspects of human reproduction and development.
The data collection schedule is illustrated in Attachment 1.a. The study’s primary environmental
exposures include: organochlorine pesticides; polychlorinated biphenyls; polybrominated
diphenyl ethers; metals; perfluorinated compounds; cotinine; and phytoestrogens. A listing of
study compounds by biologic media is presented in Attachment 5.a. These agents are largely
persistent in the environment raising the likelihood of ubiquitous exposures for the American
public.
Nurses are instructing couples in the proper use of commercially available home fertility
monitors to aid couples in becoming pregnant and home pregnancy tests for the earliest
recognition of pregnancy. Both kits are being provided as a part of the research protocol free of
charge to participants and both are being used according to the manufacturer’s instructions. Of
4
added note is the acceptability of the kits by study participants and women can easily incorporate
use of the kits to aid them in becoming pregnant. Four types of biospecimens are being collected
as illustrated in Attachment 1.a. The nurses are collecting blood and urine specimens (see
Attachments 1.n. and 1.o.) from men and women and forward them to the laboratories at the
Centers for Disease Control and Prevention for toxicological analysis. Two semen samples from
male partners are being collected for use as a global measure of male fecundity as measured
primarily by sperm concentration, morphology and 24-hour motility and to quantify
contaminants in semen (see Attachment 1.k.). Two saliva samples are being collected from
women (see Attachments 1.i. and 1.j.) to measure cortisol levels as a marker of stress so that the
relation between environmental factors, stress and human reproduction can be assessed. Couples
are being asked to complete short daily diaries on lifestyle factors suspected of adversely
affecting the probability of conception or ability to carry a pregnancy to term (see Attachments
1.f. and 1.g.). These data will allow us to identify lifestyle factors (amenable to public health
intervention) that adversely affect reproductive and developmental outcomes and to adjust for
potential confounders when analyzing environmental chemicals. Additional questions on
menstruation and sexual intercourse are being asked on the female journal so that time-topregnancy is accurately measured and to help women know when to use fertility monitors (see
Attachment 1.l) and pregnancy test kits (see Attachment 1.m.). Women becoming pregnant are
being asked to complete a short journal each month about their health status and behaviors while
pregnant (see Attachment 1.h.) to aid in the analysis of our 4th and 5th hypotheses regarding
length of gestation and baby’s birth size. These data are critical for separating the effect, if any,
of environmental agents from lifestyle factors known to adversely affect fetal growth and
development.
5
We expect to write a number of scientific papers for a diverse audience, especially key papers
responsive to the five principal null hypotheses that there is no association between
environmental chemicals in the context of other lifestyle factors and: 1) time-to-pregnancy; 2)
infertility; 3) pregnancy loss; 4) length of pregnancy or gestation; and 5) infant’s birth size. Data
sharing is consistent with NIH policy http://grants.nih.gov/grants/guide/notice-files/NOT-OD03-032.html. These data will help to answer remaining questions regarding the association
between EDCs and human reproduction and development as previously discussed.
A. 3. Use of Information Technology and Burden Reduction
Research nurses are using programmed laptops to administer the baseline interview and to record
participants’ responses. All data are being uploaded via a web-based data management system
that allows monitoring of data collection and processing and extraction for analytic purposes.
The computer-assisted interviews have reduced burden and loss by tailoring the questionnaire to
the participants’ responses thereby avoiding unnecessary questions. The need for double data
entry has also been eliminated, which will reduce errors and generate a more valid dataset.
Furthermore, the web-based data management system has facilitated data sharing at the
conclusion of the study, as required by NIH policy.
All study participants have had the choice of providing their daily and monthly journal entries
either by mail-in card or by online web-based form provided they have access to an Internetconnected computer. Online forms were specially designed for respondents use to make data
6
entry simple and with a minimum of technical limitations. The study allows respondents to
switch between the two modes of journal entry at their convenience. Internet access in the target
population was estimated at 30% on the basis of surveys conducted by the Pew Internet &
American Life Project in September 2000 and August 2003 [www.pewinternet.org]. This
estimate took into account such factors as income level, educational level, age, race or ethnicity,
urban or suburban versus rural location, and interest in parenting. Respondents with Internet
access also have had the convenience of viewing via the study public web site information about
online journal entry and a summary of their journal data for the past week, similar to the mail-in
cards.
Digital fertility monitors are helping couples optimize their chance of conception and at the same
time record daily information on women’s fecundity as two reproductive hormones are tracked
by the monitor. The monitors minimize burden by electronically recording test results, which
can be uploaded by research nurses when they visit the home to provide additional test supplies.
Digital home pregnancy test kits are helping women recognize pregnancy as early as possible.
Retail sales of pregnancy test kits commenced in 1976 with approval of the Medical Device
Amendment of the Food, Drug, and Cosmetic Act and amount to 19 million tests (Pal, 2003;
Lipsitz, 2000). We are using a digital pregnancy test kit hat displays “pregnant”, “not pregnant”,
or “error” to avoid user misinterpretation; it has been documented to be one of the most sensitive
and reliable pregnancy tests currently available on the market (Cole et al., 2004).
A. 4. Efforts to Identify Duplication and Use of Similar Information
7
There is no duplication of information, since there are no prospective pregnancy studies relating
environmental exposures to reproductive and developmental end points in the context of
longitudinally collected lifestyle factors. Online searches for government-funded research
revealed no such studies, nor did our many personal discussions with other government agencies
or academic and private research institutions. This includes electronic searches of extramurally
funded grants and intramural research projects supported by the National Institutes of Health and
other research entities.
A. 5. Impact on Small Businesses or Other Small Entities
No small businesses will be recruited to participate in this study. Per the Small Business
Administration and 13 CFR, we have contracted with a small business, The EMMES
Corporation, to act as the data-coordinating center for the study.
A. 6. Consequences of Collecting the Information Less Frequently
For our study to be scientifically valid, we need data collection at critical windows reflecting the
highly interrelated and timed nature of human reproduction and development. Exposures
occurring outside critical windows may not exert an adverse effect underscoring the need for
collection of time-varying data (Wilson 1965; Selevan et al., 2000). We are focusing on
persistent environmental agents collected at baseline necessitating only one blood specimen.
Three urine samples will be collected mainly to look at changes in phytoestrogens and cotinine
that are expected to be more variable than persistent compounds. Two semen samples are
8
needed to globally assess male fecundity (WHO Manual, 1997); two saliva samples are needed
for the detection of changes in stress cortisol levels.
Longitudinal capture of exposure information on other study covariates while couples are
attempting pregnancy is needed to assess timing of exposures in relation to conception and
gestation. Support for this approach comes from recognition that periconceptional exposures
tremendously impact (un)successful human development such as the role of folic acid in
reducing neural tube defects or other recently identified periconceptional exposures (Wald et al.,
1991; Chapin et al., 2004). Retrospective recall of such exposures including chemicals is
reported to be inaccurate (Anwar 1993; McCauley 1998).
A. 7. Special Circumstances Relating to the Guidelines of 5 CFR 1320.5
Study participants are being asked to report more than on a quarterly basis. To determine
eligibility, selected individuals are being asked to take part in a five to ten minute screening
instrument (Attachments 1.b. and 1.c.). Once enrolled, each partner completes a baseline
interview (Attachments 1.d. and 1.e.) that takes approximately 20-25 minutes depending upon
response patterns. The daily time to pregnancy journals (Attachments 1.f. and 1.g) and the
monthly pregnancy journal (Attachment 1.h.) require approximately, two minutes a day and five
minutes a month, respectively. While this may seem ambitious, the literature strongly supports
that women/couples will complete daily diary information (and often collect daily urines for
hormonal assays) including sensitive information and withdrawal, if any, from the study tends to
occur in month one (Buck et al., 2004). We anticipate that 60% of couples will conceive within
9
three months, though it may be sooner given the use of fertility monitors to help identify the
fertile window. Assuming that pregnancy lasts nine months, most couples will be followed for
one year. The world’s literature relating to retention in prospective pregnancy studies with
preconception enrollment is summarized in Attachment 5.b.
Questionnaire data and biospecimens collected at baseline or before couples begin to attempt
pregnancy are critical for quantifying exposures before the study outcomes so that a temporal
relation can be established. The baseline interview serves another purpose in that the nurse also
uses that time to teach the couple how to use home fertility monitors and pregnancy test kits.
The collection of daily information on journals while the couple is attempting pregnancy is
essential for determining whether the woman’s menstrual cycle was at risk for pregnancy, which
is the basis for assessing time to pregnancy as well as infertility (defined as the inability of
couples to conceive with 12 months of trying). To estimate the incidence of infertility in this
cohort, we allow couples up to 12 months at which time many will seek medical care. In
addition, these data permit the longitudinal collection of information on lifestyle factors that may
exert an acute or chronic effect on human reproduction and development. Another reason for the
collection of daily information while couples are attempting pregnancy is so we can formally
evaluate the timing of exposures in relation to the menstrual cycle (i.e., proliferative or secretory
phase) and ovulation. Collection of monthly (ideally it should be more frequent) journal data
from pregnant women is our attempt to identify exposures or events that adversely affect fetal
growth and development so that we can control for important confounders when evaluating
environmental agents. Every effort is being expended to ensure the collection of internally valid
and, to the extent possible, externally valid data. Given that this is an etiologically oriented
10
study, interval validity is a critical aspect. Our ultimate success in recruiting couples from a
defined population will impact the generalizability of our findings, i.e., external validity.
A. 8. Comments in Response to the Federal Register Notice and Efforts to Consult Outside
Agency
Under the provisions of Section 3506(c)(2)(A) of the Paperwork Reduction Act of 1995, the
Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD),
the National Institutes of Health submitted to the Office of Management and Budget (OMB) a
request to review and approve the continuation of information collection for the current study.
This official notice requesting a continuation of the project was published in the Federal Register
on January 16, 2008 (Vol. 73, No. 11, page 2925) and allowed 60-days for public comment. In
addition, the 30-day notice was published in the Federal Register on March 21, 2008 (Vol. 73,
No. 56, page 15162) allowing 30 days for public comment.
Only one public comment, from a concerned citizen, was received during the 60-day comment
period. She indicated that she felt that the government was already “…bankrupt due to the
war…” and that the study should be stopped due to budget concerns.
We will consult with study participants on a monthly basis largely through the research nurses
who are checking in with couples to collect data or to provide additional test supplies. Any
issues can readily be brought to the nurse’s attention for communication to the Project Officer.
Furthermore, participants are able to contact us by telephone or e-mail, regardless of reason. We
11
are using these modes of communication to monitor participant burden and related issues during
the active phase of data collection.
A. 9. Explanation of Any Payment of Gift to Respondents
Each respondent receives a total of $75 cash remuneration for provision of biological specimens
during two home visits, including one blood specimen ($25), two urine specimens ($5 per
specimen), two saliva specimens ($20 per specimen for women only), and two semen specimens
($20 per specimen for men only). These amounts are well within incentives used in other
nationally representative studies (e.g., NHANES III) and individually less than the $50 amount
permitted by U.S. law for each blood withdrawal for transfusion, scientific or research purposes
[Title 24, Chapter 1, Section 30]. Several small non-cash incentives also are provided to
participants as shown in the data collection schedule in Attachment 1.a.
A review of the world’s literature on the use of incentives in prospective pregnancy studies
reflected a range of cash incentives from $10 every two months for daily urine collection and
storage to $500 for completion of an intensive protocol (see Attachment 5.b. for details).
Fertility monitors and pregnancy test kits are an essential component of the research plan and are
not being used as incentives. We are sensitive to the ethical principle of distributive justice
(Beauchamp and Childress 1994) given our interest in the inclusion of minority and socioeconomically disadvantaged couples in our study.
A. 10. Assurance of Confidentiality Provided to Respondents
12
Neither the NICHD nor the Data Coordinating Center (DCC) has any identifying information
about study participants. Each contracted research site is responsible for assuring study
participants’ privacy and the confidentiality of their data. Personally identifying information is
kept in either a secured project share drive or in a locked filing cabinet located in a locked office
or other restricted access room. Only research staff with signed confidentiality agreements have
access to the personally identifying information. None of these data are being shared with
outside parties.
Upon enrollment, participants are assigned a site-preferred participant ID (SPID) for use in the
DCC’s centralized data system. The SPID is comprised of a site identifier and a unique code
randomly generated by the data system. The only personally identifying information that is
included in the centralized data system includes ethnicity and date of birth. Participants’ data are
transmitted to the DCC from the research sites and laboratories using a password-protected
online distributed data entry system that uses Secure Socket Layer (SSL) encryption. All
electronic data are stored behind a firewall and are available only to individuals involved in the
project.
The Longitudinal Investigation of Fertility and the Environment protocol incorporates standard
operating procedures meant to ensure participants’ privacy and the confidentiality of their data.
Telephone calls to conduct eligibility screening interviews are being made from private offices at
each research site using landline telephones. Baseline interviews and biospecimens collection
are being done in the privacy of participants’ homes. Mail-in daily and monthly journals
13
completed by participants include only the SPID and numerical answers, which are meaningless
to someone who does not have access to the corresponding questions. As an alternative to the
mail-in method, participants are able to complete their journals online. Individuals who choose
this option can log in from the privacy of their home using a username and password provided by
the enrolling research site. Journal data, identified only by the SPID, are transmitted securely to
the Data Coordinating Center. Online journal forms provide additional security by limiting
access to previously entered data to only the current week and by automatically logging off the
computer when data are submitted. Participants are instructed how to ensure their privacy while
entering data online, including proper use of the secure data entry system. Only participants with
computer systems meeting minimum standards for data encryption and transmission are
permitted to use online data entry.
Each biological specimen is labeled with a bar code. No identifying information accompanies
shipped specimens other than the SPID. Blood and urine specimens are shipped by the sites to
the CDC’s National Center for Environmental Health laboratory for toxicology analyses.
Residual samples are stored frozen. Semen specimens are collected by male participants and
shipped directly to the National Institute for Occupational Safety and Health (NIOSH) laboratory
for andrology analyses then to CDC’s toxicology laboratory for analysis of seminal fluid. Saliva
samples collected by female participants are shipped for analysis of salivary cortisol. All
specimens are tracked via the DCC.
Institutional Review Board (IRB) approval for the conduct of this study has been obtained from
Texas A & M, RTI International, The EMMES Corporation, and the Eunice Kennedy Shriver
14
National Institute of Child Health & Human Development (see Attachment 3). A Certificate of
Confidentiality has been approved for this study under the auspices of the NICHD informs
participants that: 1) researchers will resist demands for information, other than from the
respondents themselves, which would identify the respondents but that the Certificate cannot be
used to resist demand for information from personnel of the United States Government that is
used for auditing or evaluation of Federally funded projects; 2) researchers might voluntarily
disclose respondent identity if it is necessary to protect the respondent or others from serious
harm, such as in situations of child abuse, reportable communicable diseases, possible threat to
self or others, or high levels of toxic chemicals in the environment (e.g., as suggested by lead
levels in blood); and 3) persons from the local Institutional Review Boards and the Federal
Office for Human Research Protections may review their information, though confidentially
[http://grants.nih.gov/grants/policy/coc/appl_intramural.htm].
In accordance with the Privacy Act of 1974, the study brochure indicates that the NICHD is
conducting this study as a part of its mission to sponsor research focusing on factors impacting
human reproduction and development (42 USC 285g). The informed consent documents
approved by the individual research sites outline how we plan to protect individuals’ privacy and
the confidentiality of their data. Further, the consent documents state that participation in the
study is entirely voluntary.
A. 11. Justification for Sensitive Questions
15
We are asking no third party questions. Two questions that are being asked may be sensitive for
some people: 1) frequency of sexual intercourse to quantify the length of time couples are
attempting to become pregnant and 2) frequency of ejaculations for the assessment of male
fecundity. Valid time-to-pregnancy requires counting only menstrual cycles at risk for
pregnancy or those where intercourse occurs five days before or on the day of ovulation (Dunson
et al., 1988; Wilcox et al., 1995). This information is critical for validity of time-to-pregnancy
and for couples seeking medical treatment if so desired.
All data inclusive of potentially sensitive information are stored free of personal identifiers
according to the IRB requirements at each research site and as fully described in Section A. 10.
All study participants are given an informed consent (see Attachment 2.c.), including full
disclosure of potential risks and benefits and other essential elements such as the voluntary
nature of research, privacy and confidentiality statements. Further, all consents are consistent
with the language requirements of the Certificate of Confidentiality.
A. 12. Estimates of Hour Burden Including Annualized Hourly Costs
We will be enrolling a total of 500 couples into the study (i.e., 150 additional couples). The
estimated number of response sets per respondent varies by gender, with women asked to
complete six sets and men three sets. This difference is attributable to menses and to pregnancy
occurring only in women. A single estimate of burden cannot be estimated, as length of
participation will vary based on gender and when and if a couple conceives (see Table A.12-1a
for the detailed calculations).
16
Specifically, the time interval for participation will range from nine months (for couples
becoming pregnant in the first month of trying) to 21 months for couples requiring a full 12
months to become pregnant plus nine months of pregnancy. The annual burden per male
respondent ranges from 2.10 to 5.48 hours per year, depending on when and if pregnancy is
achieved. This translates to a total annual burden of 1,050 to 2,740 hours among the 500 male
partners that will be recruited. For women, the annual burden per participant ranges from 3.28 to
9.90 hours, which corresponds to a total annual burden of 1,640 to 4,950 hours for the 500
female participants that will be recruited. Please note that 350 out of the total 500 couples have
already been recruited, so the additional burden hours that are being requested with this
collection extension are actually 70% lower than what is stated. These figures are conservative
estimates of participant burden (i.e., overestimates) in that the literature suggests that 60% of
couples are expected to become pregnant within one to three months, with 80% becoming
pregnant by six months. Further, the literature suggests that most study participants who drop
out of prospective pregnancy studies with preconceptional enrollment, drop out within the first
few months of participation (see Table A.12-1b).
17
Table A.12-1a. Estimates of Hour Burden by Gender of Respondents and Type of Response
Gender of Respondent
and type of Response
Female Partner of Couple
Screening instrument
Baseline questionnaire
Time to pregnancy and daily
early pregnancy journals
Fertility monitors
Pregnancy testing
Biospecimens collection
Number of
Respondents*
Frequency of
Response
Average Time per
Response
(hours)
Annual Hour
Burden**
500
500
500
Once
Once
30 entries/month
0.17
0.42
0.03
85
210
15
500
500
500
11 tests/month
Two tests/month
Blood: once
Urine: three times
Saliva: twice
One entry/month
-
0.12
0.07
Blood: 0.13
Urine: 0.05
Saliva: 0.10
0.08
-
60
35
65
25
50
40
15-210
Once
Once
30 entries/month
Blood: once
Urine: twice
Semen: twice
-
0.17
0.42
0.03
Blood: 0.13
Urine: 0.05
Semen: 0.33
-
85
210
15
65
25
165
15-210
Pregnancy journal
Total
Male Partner of Couple
Screening instrument
Baseline questionnaire
Time to pregnancy journals
Biospecimens collection
500
500
Total
500
500
500
500
500
Note: Duration of study participation is expected to range from one month to 21 months depending on when and if a
couple conceives (i.e., twelve months of trying to conceive plus nine months of pregnancy). Majority of couples are
expected to conceive in months 1-3 per past literature based on convenience sampling.
* We expect to enroll 500 respondents of which approximately 40 are expected to drop out following the baseline
visit and an additional 60 thereafter; however, burden estimates are not adjusted for these withdrawals.
** By design, burden is overestimated in that it assumes complete participation, which is unlikely. In addition, time
to pregnancy will vary across couples with the majority of couples becoming pregnant in first few months. Hence,
time to pregnancy coupled with use of fertility monitors and pregnancy kits are greatly overestimated.
18
Table A.12-1b. Expected Distribution of Pregnancies and Drop Outs by Month of
Participation
Baseline*
Dropout
this month
Cumulative
dropouts
Number
remaining
in the study
Pregnant**
Not
pregnant
40
One
10
Two
10
Three
5
Four
5
Five
12.5
Six
2.5
40
50
60
65
70
82.5
85
460
450
440
435
430
417.5
0
460
100
360
180
320
228
272
260
240
280
220
Months
Seven
2.5
Eight
2.5
Nine
2.5
Ten
2.5
Eleven
2.5
Twelve
2.5
87.5
90
92.5
95.0
97.5
100
415
412.5
410
407.5
405
402.5
400
292
208
300
200
304
196
308
192
312
186
316
184
320
180
* Month of baseline interview. Dropouts reflect attrition after completing baseline questionnaires but
before submitting any monthly diary data.
** Based on population-based studies reviewed in Buck et al., 2004
Table A.12-2. Annualized Costs to Respondents
Participant
Male partner
Female Partner
Number of
Respondents
500
500
Frequency of
Response
3
6
Hourly
Wage Rate
$10
$10
TOTAL
Respondent
Cost
$15,000
$30,000
$45,000
A. 13. Estimate of Other Total Annual Cost Burden to Respondents or
Recordkeepers
There are no additional costs to respondents or recordkeepers stemming from the
collection of information.
A. 14. Annualized Cost to the Federal Government
FY07 contract awards for participating sites are:
The EMMES Corporation
RTI International
Texas A&M University System Health Science Center
Subtotal
Interagency agreement costs:
CDC’s National Center for Environmental Health laboratory &
NIOSH’s Reproductive Health Assessment Section
Total
337,879
56,734
3,107,616
3,502,229
308,963
3,811,192
The NICHD Project Officer Dr. Germaine Buck Louis, has committed 30% effort to the
study. Additional contributions have included 50% effort from Dr. Courtney Lynch (who
has recently left the NICHD for academe), 15% effort from Dr. Aiyi Liu (biostatistics)
and 5% effort from Dr. Enrique Schisterman (biomarkers & methodology). The
estimated cost of Federal employees working on this project, including salary and fringe,
is $131,973.72. Additional cost for a bi-annual site visit at the remaining site is $1,000
annually.
A. 15. Explanation for Program Changes or Adjustments
This is a request for an information collection extension for an ongoing project (OMB
Clearance 0925-0543). This extension is needed given the initial delays in obtaining all
human subjects approvals in 2005. Also, recruitment was temporarily halted in our
participating Texas site following Hurricane Rita and extensive flooding in 2005. Based
upon past recruitment rates, the remaining couples needed for recruitment goals can be
recruited within the next 12-18 months.
A. 16. Plans for Tabulation and Publication and Project Time Schedule
Several papers will be published from this as developed from the analytic plan supporting
the five hypotheses. With univariate analysis, the association of each exposure with the
outcome will be analyzed to evaluate if that exposure is statistically associated with the
outcome, and to decide which exposures will be further considered in the multivariate
20
analyses. Potential confounders also will be considered in the model (e.g., change of the
$-coefficient estimated when confounders are entered into the model). The multivariate
model will evaluate simultaneously the exposures associated with the outcome, while
adjusting for potential confounder effects to search for the “best” group of exposures that
provide maximum prediction of the outcome.
Hypothesis One
The discrete time Cox’s proportional hazard regression model will be used to analyze
environmental exposures, relevant covariates such as past reproductive history and timeto-pregnancy (TTP) outcomes. We anticipate fitting a discrete time multiple events
Cox’s proportional hazard model that considers the joint distribution of multiple TTP
outcomes contributed by reproductive history. Another approach will be to put the
history TTP outcomes as independent variables into the model; results from this approach
need to be interpreted with caution, using conditional expectation arguments. Of
particular interest is the probability of being pregnant during a certain menstrual cycle.
These probabilities will be estimated from fitting the discrete time Cox’s proportional
hazard regression model, along with 95% confidence intervals (Cox and Oakes 1984).
The probability of delayed conception (>6 months) can also be estimated from the fitting
of the model. Meaningful interpretation of the results relies on correctly defined cycle 1
of a woman underscoring the importance of the baseline interview.
Other alternative methods can also be used to analyze the number of cycles required for
conception, and the probability of conception per-cycle. One useful method is to employ
21
the Beta-Geometric distribution to model fecundability (Weinberg and Gladen, 1986).
Effects of exposures and other covariates will be analyzed by modeling the mean
structure as a function of exposures and other covariates. Estimates of the effects can be
obtained using expectation-maximization (EM) algorithm.
Hypothesis Two
We propose two methods to analyze infertility and estimate the incidence rate and to
assess the exposures’ effects on infertility. First, let T be the number of cycles a woman
takes to getting pregnant (i.e., T is the discrete time to pregnancy). Then, the rate of
incident infertility is simply the probability of {T>12}, the survival function of the time
to pregnancy at 12 months. This probability and its association with exposures and
confounders can then be evaluated from fitting the discrete time Cox’s proportional
hazard regression model, or the Beta-Geometric model, for TTP, as discussed above.
Another method would be to fit a logistic regression model with infertility outcomes
(0=pregnant before 12 months and 1= being infertile) as the dependent variable and
exposures and confounders as independent variables. Fitting this model also will provide
estimates of the infertility rate and the odds ratio of each exposure as a measure of effect
on infertility. Technically, if the couples were followed indefinitely, the discrete time
Cox’s proportional hazard regression model approach, or the Beta-Geometric model,
would be more appropriate since it captures the survival (time to pregnancy) nature of the
outcome. We will evaluate the fit of both methods.
Hypothesis Three
22
Pregnancy loss, as a dichotomous outcome, may occur more than once during the 12month conception period. Furthermore, a woman may also report in her pregnancy
history previous pregnancy losses. These situations result in multiple dependent binary
outcomes. We propose using the generalized estimating equation (GEE) model with the
logistic regression model as the marginal model to analyze pregnancy loss outcomes. In
the model, exposures and confounders will be entered as independent variables and
pregnancy loss as dependent variables. We will fit separately the model assuming the
intercept to be a random effect or an unknown constant to be estimated. We will explore
several working correlation structure, the independent, the compound symmetry, and the
autoregressive structures in particular. Estimates of the $-coefficient or odds ratio will be
obtained along with their robust standard errors. The AIC criterion will be used to
evaluate model fitting with each working correlation structures. We also will treat
previous pregnancy loss (or other previous adverse pregnancy outcomes) as independent
variables in the model, to assess what these previous pregnancy outcomes may provide in
predicting the current pregnancy losses. We hypothesize that women with previous
pregnancy outcomes have increased risk of pregnancy loss as observed in the study
period.
Hypotheses Four and Five
Gestation will be analyzed both as a continuous variable (days from conception to birth
date) and as a dichotomous variable (short gestation versus normal). When treated as a
continuous variable, linear models will be used with gestation as a dependent variable
and exposures, confounders and other factors as independent variable. Data may be log-
23
transformed to stabilize variance and to achieve better normal approximation. When
input as a dichotomous variable, logistic regression models will be employed for analysis.
When previous gestation as reported in a woman’s pregnancy history is considered, the
GEE version of these models will be subsequently used to account for dependence among
gestation outcomes. Intercept parameter and working correlation structures will be dealt
with as described above for other outcomes. Alternatively, previous gestation will be
entered as dependent variables to assess their predictability of the current gestation. Birth
weight and other measures of birth size also will be analyzed using the above models
(linear and logistic regression models).
Lastly, our analytic plan includes model diagnostics, detection of outliers and influential
observations and missing data. Model fit will include examining the residuals of the
model. If a peculiar residual pattern occurs, we will reexamine the model by using
strategies such as adding a quadratic term into the model, adding interaction terms, trying
non-linear models, as may be suggested appropriate by the residuals. Outliers and
inferential observations will also be examined by comparing the fitting of two models,
one with and the other without the suspected observation or outlier. Significant change in
the model fitting by removing the observation signals abnormality and will be reported to
the Data Coordinating Center for quality assurance and to the study investigators and
Steering Committee for decision-making. Missing values are anticipated and introduce
uncertainty to the estimates, p-values and confidence intervals of the statistical analysis.
To evaluate such uncertainty, we will use the multiple imputation technique (Rubin,
1987, Schafer 1997). We will assume missing at random, unless otherwise informed. For
24
each item that has missing values, several imputations (three to five according to Schafer,
1997, Chapter 4) will be carried out, each generating a “complete” set of observations.
This then yields a set of p-values, estimates and confidence intervals for the parameter of
interest. These p-values, estimates or confidence intervals will then be combined to
create a single p-value/estimate/confidence interval for the parameter of interest (Schafer
1997).
Table A.16-1. Project Time Schedule
Activity
Time Schedule
(months)
1-2
1-18
1-18
1-18
19-24
19-24
Letters to eligible individuals
Recruit couples
Follow couples attempting pregnancy
Follow couples during pregnancy
Analyses (time-to-pregnancy)
Analyses (pregnancy loss, gestation, birth
weight)
Publications
24-36
Note: Assumes allowing couples up to 12 months to conceive and 9 months for average length of
pregnancy.
A. 17. Reason(s) Display of OMB Expiration Date is Inappropriate
The OMB expiration date is displayed in the upper-right hand corner of all study forms
near the control number. [New number and expiration data will replace the existing
number and date on all instruments and other study documents upon receipt from OMB.]
The statements referring to the Privacy Act and the Paperwork Reduction Act also appear
on all letters and brochures seen by study participants.
25
A. 18. Exceptions to Certification for Paperwork Reduction Act Submissions
There are no exceptions to the Certification requirements.
26
References
Anwar WA. Chemical interaction: enhancement and inhibition of clastogenicity. Environ
Health Perspect 1993;101(suppl 3):203-206.
Beauchamp TL, Childress JF. Principles of Biomedical Ethics. New York: Oxford
University Press, 1994: 326-394.
Brown V. Disrupting a delicate balance: environmental effects on the thyroid. EHP
2003;111:A642649.
Buck GM, Lynch CD, Stanford JB, Sweeney AM, Schieve LA, Rockett JC, Selevan SG,
Schrader SM. Prospective pregnancy study designs for assessing reproductive and
developmental toxicants. Environ Health Perspect 2004;112:79-86.
Chapin RE, Robbins WA, Schieve LA, Sweeney AM, Tabacova SA, Tomashek KM. Off
to a good start: the influence of pre- and peri-conceptional exposures, parental fertility,
and nutrition on children's health. Environ Health Perspect 2004;112;69-78.
Cole LA, Khantian SA, Sutton JM, Davies S, Rayburn WF. Accuracy of home
pregnancy tests at the time of missed menses. Am J Obstet Gynecol 2004;190:100-105.
Cox DR, Oakes D (1984). Analysis of Survival Data. Chapman & Hall: London.
Crisp TM, Clegg ED, Cooper RL, et al. Environmental endocrine disruption: an effects
assessment and analysis. Environ Health Perspect 1998;106:11-56.
Dallinga JW, Moonen EJC, Dumoulin JCM, Evers JLH, Geraedts JPM, Kleinjans JCS.
Decreased human semen quality and organochlorine compounds in blood. Human
Reprod 2002;17:1973-1979.
Dunson DB, Baird DD, Wilcox AJ, Weinberg CR. Day-specific probabilities of clinical
pregnancy based on two studies with imperfect measures of ovulation. Hum Reprod
1999;14:1835-1839.
Faroon OM, Keith S, Jones D, DeRosa C. Effects of polychlorinated biphenyls on
development and reproduction. Toxicol Indust Health 2001;17:63-93.
Hauser R, Altshul Larisa, Chen Z, Ryan L, Overstreet J, Schiff I, Christiani DC.
Environmental organochlorines and semen quality: results of a pilot study. EHP
2002;110:229-233.
Joffe M, Bisanti L, Apostoli P, Kiss P, Dale A, Roeleveld N, Lindbohm ML, Sallmen M,
Vanhoorne M, Bonde JP; Asclepios. Time To Pregnancy and occupational lead
exposure. Occup Environ Med. 2003 Oct;60(10):752-8.
Lipsitz R. Pregnancy tests. Sci Am 2000;283:110-111.
27
McCauley LA. Chemical mixtures in the workplace. Research and practice. AAOHN J.
1998 Jan;46(1):29-40; quiz 41-2.
Pal S. Home test market reaches $1.4 billion. US Pharmacist 1998;23(3).
www.uspharmacist.com
Pocar P, Brevini TAL, Fischer B, Gandolfi F. The impact of endocrine disruptors on
oocyte competence. Reproduction 2003;125:313-325.
Rozati R, Reddy PP, Reddanna P, Mujtaba R. Role of environmental estrogens in the
deterioration of male factor fertility. Fertil Steril 2002;78:1187-1194.
Rubin DB (1987). Multiple Imputation for Nonresponses in Surveys. Wiley: New York
Schafer JL (1997). Analysis of Incomplete Multivariate Data. Chapman & Hall: New
York.
Schafer JL (1997). Analysis of Incomplete Multivariate Data. Chapman & Hall: New
York.
Selevan SG, Kimmel CA, Mendola P. Identifying critical windows of exposure for
children’s health. Environ Health Perspectives 2000;108(suppl 3):451-455.
Sharpe RM and Irwin DS. How strong is the evidence of a link between environmental
chemicals and adverse effects on human reproductive health? BMJ 2004;328:447-451.
Thomas KB and Colborn T. Organochlorine endocrine disruptors in human tissue. In:
Chemically Induced Alterations in Sexual Development: The Wildlife/Human
Connection. T. Colborn and C. Clement, Eds. Princeton, NJ: Princeton Scientific
Publishing, 1992:365-394.
Wald N. Prevention of neural tube defects: results of the Medical Rearch Council
Vitamin Study. Lancet 1991: 338:131-137.
Weinberg CR, Gladen BC. The Beta-Geometric distribution applied to comparative
fecundability studies. Biometrics 1986; 42: 547-560.
Wilcox AJ, Weinberg CR, Baird DD. Timing of sexual intercourse in relation to
ovulation: effects on the probability of conception, survival of the pregnancy and sex of
the baby. New Engl J Med 1995:333:1517-1521.
Wilcox AJ, Weinberg CR, O'Connor JF, Baird DD, Schlatterer JP, Canfield RE et al.
Incidence of early loss of pregnancy. N Engl J Med 1988;319:189-194.
Wilson JG. Embryological considerations in teratology. In: Teratology: Principles and
Techniques. Wilson JG, Warkany J, eds. Chicago: The University of Chicago Press,
1965:256.
28
| File Type | application/pdf |
| File Title | untitled |
| File Modified | 2008-03-25 |
| File Created | 2008-03-25 |