Content uploaded by Joan Sabaté
Author content
All content in this area was uploaded by Joan Sabaté on Feb 25, 2020
Content may be subject to copyright.
COHORT PROFILE
Cohort Profile: The Adventist Health
Study-2 (AHS-2)
Terry L Butler,
1
* Gary E Fraser,
1
W Lawrence Beeson,
1
Synnøve F Knutsen,
1
R Patti Herring,
1
Jacqueline Chan,
1
Joan Sabate
´,
1
Susanne Montgomery,
1
Ella Haddad,
1
Susan Preston-Martin,
2
Hannelore Bennett
1
and Karen Jaceldo-Siegl
1
How did the study come about?
The Adventist Health Study-2 (AHS-2) began in 2002
with the goal of investigating the role of selected
foods to change the risk of cancer. AHS-2 is designed
to provide more precise and comprehensive results
than previous pioneering research among Seventh-day
Adventists,
1–6
a unique health oriented population
with diverse dietary habits.
The Adventist church, of 24 million adherents world-
wide, promotes a healthy lifestyle. Church members
are expected to be non-smokers and non-alcohol users,
and are encouraged to eat a vegetarian diet. Many
also avoid caffeine-containing beverages. However,
adherence to these recommendations is quite variable.
Adventists in North America are almost entirely a
non-smoking population. The vast majority are non-
drinkers and the small number who consume alcohol
do so infrequently. But they have a wide diversity in
dietary practices. Two previous longitudinal studies
in California showed a small percentage are total
vegetarians, many follow a lacto-ovo-vegetarian diet
or eat meat less than once per week (semi-vegetarian)
and about half have omnivorous diets similar to the
general population.
7
These studies in California, the Adventist Mortality
Study (AMS)
8,9
from 1960–66 and the first Adventist
Health Study (AHS-1)
10–13
from 1974–88 indicated
that Adventists had lower risks for most cancers,
cardiovascular disease and diabetes. Females lived 4.4
years and males 7.3 years longer when compared with
the general California population.
7,14
These studies
also showed the advantage of a vegetarian diet among
Adventists, found strong evidence that meat increased
risk of colon cancer
13
and coronary heart disease,
11,15
and that nut consumption reduced risk of coronary
heart disease.
11,12
Other significant associations
between cancers and other foods have also been
reported.
7,16–18
In the USA, it is estimated that there will be
1 444 920 new cases of cancer in 2007 (prostate
218 890, breast 180 510 and colon 112 340).
19
It has
been suggested that the lower incidence of prostate
and breast cancer in the Far East may be due to
the frequent consumption of soy.
20–23
Adventists, in
North America, with high levels and a wide range of
soy consumption offer better opportunities than
perhaps any other large group with a western diet
to investigate the role of soy and other nutrients, such
as calcium, in risk of cancers of the breast, prostate
and colon.
Thus, various characteristics of Adventists: the
diversity in soy consumption and other foods, the
virtual elimination of confounding from smoking
and alcohol, the findings of previous studies and
the potential to include a large group of African-
Americans resulted in the National Cancer Institute
funding Loma Linda University to undertake AHS-2,
with the goal to enroll 100 000 Adventists.
Table 1 provides an overview and compares some
characteristics of AHS-2 with the other Seventh-day
Adventist cohort studies in California.
What does it cover?
The broad scope of AHS-2 is to investigate the role
of various foods and nutrients, other lifestyle factors
and metabolic risk indicators that may be involved in
cancer causation.
The primary aims of the currently funded study are
to investigate the possible protective role for dietary
soy proteins and isoflavones on risk of prostate, breast
and colon cancers; to provide further evidence of
a possible protective role of vitamin D for these
cancers; to further evaluate a possible protective effect
of dietary calcium on risk of colon cancer, and a
possible hazardous effect on risk of prostate cancer;
and to help decide whether dietary linolenic acid
increases risk of prostate cancer.
* Corresponding author. School of Public Health, Loma Linda
University, 24785 Stewart Street, Room 203, Loma Linda,
California 9235, USA. E-mail: tbutler@llu.edu
1
School of Public Health, Loma Linda University, Loma Linda,
California, USA.
2
School of Medicine, University of Southern California, Los
Angeles, California, USA.
Published by Oxford University Press on behalf of the International Epidemiological Associatio n
ßThe Author 2007; all rights reserved. Advance Access publication 27 August 2007
International Journal of Epidemiology 2008;37:260–265
doi:10.1093/ije/dym165
260
The study will also investigate possible dietary
causes for the particularly high incidence of prostate
cancer in a population of African-American men, and
compare this incidence with that of a Caucasian male
population adhering to a broadly similar lifestyle.
Who is in the sample?
The sample includes Seventh-day Adventist church
members living in the USA and Canada who are
30 years and older, and who are sufficiently fluent in
English to complete the lengthy lifestyle question-
naire. Most participants are either Caucasian or Black/
African-American. There are smaller numbers from
other ethnic minorities.
There are about one million Adventists in the USA
and Canada. Although church membership records do
not contain data on age, race and gender, church
officials estimate that of those members who
regularly attend English language churches and are
30 years or older, approximately 90 000 are Black
members and 260 000 are non-Black members
(Figure 1).
Enrolment commenced in February 2002 and by
May 2007 more than 96 000 participants had com-
pleted the lengthy lifestyle questionnaire. Of this
number, about 25 500 are Black/African-American and
62 500 are females. There is a wide geographic spread
of participants (Figure 2).
The data presented in this article pertains to the
90 156 participants for whom information on the
reported variables has been checked and is currently
available.
Enrolment methods
The basic recruitment model targeted the two major
English-speaking congregation groups in the Seventh-
day Adventist church in the USA and Canada—1000
Black (African-American and Caribbean) congre-
gations and 3500 other congregations (mostly
Caucasian, but including some Latino and Asian).
For simplicity of designation these two groups of
churches are referred to as ‘Black’ and ‘non-Black
churches’. Recruitment was church-by-church and
staged by geographic region.
Local church pastors selected one or two church
members to be the study coordinators and chose a
Table 1 Characteristics of Seventh-day Adventist cohort
studies in the USA
Characteristic
Adventist
Mortality
Study
(AMS)
Adventist
Health
Study-1
(AHS-1)
Adventist
Health
Study-2
(AHS-2)
Geographic
region
California California 50 States and
Canada
Study
participants
22 940 34 192 96 194 (as of
May 31, 2007)
Congregations
(Adventist)
234 437 4500
Years of
follow-up
1960–76 1976–82
(incidence)
2002 (ongoing)
1976–88
(mortality)
Outcome
of interest
Mortality Incidence and
mortality
Incidence and
mortality
Demographics
Female (%) 64.6% 60.1% 65.1%
Mean age
(years)
50.9 54.3 60.2
Age range
(years)
35–90 25–90 30–112
White
(non-Hispanic)
(%)
100% 100% 65.3%
Black (%) 0% 0% 26.9%
Current cigarette
Smokers
Males (%) 1.7% 2.1% 1.2%
Females (%) 0.5% 1.1% 1.0%
Source population
1 Million Adventists in USA & Canada
From 4500 Congregations
350 000 Estimated Church Attendees
Aged 30+ Years and
Fluent in English
90 000 (Estimated)
Eligible Participants
from Black Churches
260 000 (Estimated)
Eligible Participants
from Non-Black
Churches
46 496
Participants Requested
and Received Lifestyle
Questionnaire (LQ)
114 505
Participants Requested
and Received Lifestyle
Questionnaire (LQ)
24 298
Participants from
Black Churches
Returned the LQ
71 865
Participants from
Non-Black Churches
Returned the LQ
Figure 1 Source of and number of participants (as of May
31, 2007) Black and non-Black Adventists in USA and
Canada
COHORT PROFILE: THE ADVENTIST HEALTH STUDY-2 261
promotion start date. Information resources (promo-
tional guidelines, brochures, video, posters and
announcements) were provided. Each church had a
suggested enrolment goal, based on membership, and
received weekly progress reports from the research
office during the 7–8 weeks of promotion. However,
the actual recruitment approach differed between the
Black and non-Black churches.
In non-Black churches, we used previously tested
promotional and motivational strategies similar to the
successful AHS-1 recruitment model.
24
Pastors and
coordinators received training by telephone and email.
Promotion consisted of brief presentations to the
congregation and the distribution of enrolment forms.
The completed forms were sent to the AHS-2 office
for processing and enrollees were mailed a question-
naire. On return of the completed questionnaire
participants in non-Black churches received a
personalized enrolment certificate and an AHS-2
ballpoint pen.
Based on pilot study results,
25
the process in Black
Churches was more personal. Most pastors and
coordinators attended regional training sessions con-
ducted by senior Black researchers, questionnaires
were personally distributed in church, group sessions
to complete the questionnaire were encouraged, and a
follow-up phone call was made to non-respondents 4
weeks after they had received a questionnaire.
26
Participants received $10 for returning a completed
questionnaire. In addition, coordinators and pastors
received a financial incentive of $200–$1000 pro-rated
on church size and percentage of local goal achieved.
Reminder postcards were mailed to all non-
responding subjects at 4, 7 and 10 weeks following
the distribution of the questionnaire. In year two, the
local recruitment model was supplemented with a
national advertising campaign and featured articles
in church magazines mailed to 300 000 Adventist
households, presentations at regional church convoca-
tions and on Adventist TV stations. All advertising
materials included a website and a telephone number
for enrolling. Finally, a single direct mail invitation
was sent to church members who did not respond to
the initial church promotion.
How often will study subjects be
followed-up?
The various phases and schedule of AHS-2 is found
in Web Table 1. The first 5 years of the study
concentrated on cohort assembly, collection of base-
line lifestyle and health history data, quality control
of the data and validation studies for the question-
naire.
27
Planned follow-up of the cohort during
the second and third 5-year phases, 2006–10 and
2011–15, will involve cancer incidence and mortality
outcome ascertainment and periodic re-measurement
of diet.
Morbidity and mortality follow-up
Follow-up of morbidity and mortality outcome is
accomplished in several ways: identified by partici-
pant self-report followed by obtaining hospital
records, linkage with State Tumor Registries and
linkage with the National Death Index in 2007 and
2010.
Tumour registry matching
To find incident cancers in the US and Canada-wide
AHS-2 population, we will computer match with every
state and provincial tumour registry during 2007–10.
This challenging process has never previously been
attempted. Because each state registry has its own set
of requirements and application process the approval
process is time consuming. However, with the support
of the North American Association of Central Cancer
Registries (NAACCR) and building on our pilot
experience with six tumour registries,
28
we expect
the process will be streamlined in the next few years,
thus also benefiting other national cancer studies.
A trained researcher/computer programmer travels to
each state and uses identical protocols and soundex-
based LINK-PLUS software (CDC) for the matching
process. The possible matches are then resolved with
software developed by AHS-2 and even with larger
registries this process was completed in 4–5 h. A
preliminary analysis suggests that we will miss only
2–3% of cases.
Hospital history form
As a back up to tumour registry matching, every
2 years we mail a four-page Hospital History Form
(HHF) to participants to collect self-report informa-
tion of hospitalizations and diagnoses of cancers,
heart attacks, stroke and diabetes during the previous
Figure 2 Distribution of participants by geographical region
as of May 31, 2007
262 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
2 years. Limited additional data on lifestyle and
demographic factors is also collected. In 2004, the first
HHF was mailed to the 37 000 participants who had
been enrolled for at least 2 years. The second HHF
mailed in 2006 included additional questions about
sun exposure.
Clinics in churches
Subject to the availability of funding, during years
2008–10, we plan to conduct clinics in 2000 churches
to collect fasting blood specimens and subcutaneous
fat from much of the cohort. Three mobile clinic
teams with two technicians each will travel from
church to church testing about 35 subjects each day.
The feasibility of this plan and the methods to
arrange clinics, schedule subjects, obtain, ship and
store biological specimens have been thoroughly and
successfully tested.
29
The bio-repository will enable us to use biological
variables as prospective exposures to complement
dietary analyses and provide additional information
about mechanisms and to correct biases resulting
from measurement errors using recently described
methods
30
.
What has been measured?
Main questionnaire
The comprehensive lifestyle and health questionnaire
was developed after extensive testing and validation
studies for diet and physical activity.
31–35
The
self-administered questionnaire is quite lengthy and
takes on average 1.25–3.5 h to complete. It consists
of nearly 2000 data fields divided into sections for
medical, diet, physical activity, female history, supple-
ment use and vegetarian food consumption (Web
Table 2).
The food frequency component is the largest and
most important part of the questionnaire. After
preliminary studies and analysis, 130 food frequency
items were selected to comprise the 26 nutrient/
phytochemical/mineral indices necessary to address
the major hypotheses, and to include all foods and
food groups commonly consumed. For each index we
identified 3–4 key variables.
Questionnaires are manually edited initially to
detect skipped pages, poor erasures or improperly
filled in dates. For moderate amounts of missing data
rescue telephone calls are made or the missing blank
pages are re-mailed. We then use the NCS 5000i
scanner to optically scan the questionnaire data to a
computer file and take an electronic image of each
page for archiving. A random 10% sample of all
subjects with missing data on any of 80 key variables
was telephoned to fill-in missing data. This can then
be used for accurate multiple imputation.
36
Calibration study
A calibration study of 950 participants (850 now
enrolled), to conclude in 2007, provides validation
data for self-reported diet, physical activity and sun
exposure. A two-stage random selection process, first
by church and then within church was used to
select 500 white and 500 black participants. Slight
over-sampling compensates for a 10% drop out after
selection. This study is used to provide estimates of
the validity of the food frequency data, and will assist
with the correction of biases due to measurement
error.
30
During a period of 1 year, each subject provides two
sets of three 24-h dietary recalls and two 1 week
physical activity recalls, 6 months apart. In the middle
of the dietary recalls subjects complete another full
food frequency questionnaire and attend a clinic at
their church. A mobile team of two technicians
process 5–10 subjects per day. An over-night urine
sample is collected and aliquoted, blood pressure and
body composition are measured. Blood is drawn and
a subcutaneous sample of fat is taken. For the last
500 subjects, we used a Minolta Chroma-meter to
measure sun exposed and unexposed skin tones.
This provides an index of tanning to help validate
sun exposure. Biological specimens are shipped back
to Loma Linda for aliquoting in straws and are frozen
in liquid nitrogen while awaiting analysis.
What is the rate of loss likely to be?
Adventists in the past have demonstrated their
willingness to be engaged in prospective health
studies and complete relatively lengthy question-
naires. In AHS-1 (1976–88), we were able to track
98.8% of cohort members over 12 years of follow-up.
24
In AHS-2, we devote considerable effort to minimize
attrition. A key element of cohort retention is to
ensure a participant’s sense of belonging to the
study through regular feedback and appreciation.
This includes: an annual newsletter; quarterly email
communication to participants with an email contact;
articles in church magazines; presentations at annual
church regional convocations and promotion of the
AHS-2 website.
To maintain current addresses for the participants,
national change of address services are used prior to
the mailing of the annual newsletter and the biennial
HHF.
For the first HHF mailed to 37 000 participants there
was a 90% return rate. Black members (80% return
rate) appear to be more mobile than white members
and require more effort to trace changes of address
and telephone numbers.
The Adventist cohort has some advantages that
assist us in follow up: we use the local church
congregation unit to communicate with and keep
track of members, membership lists are maintained
by each local church, and when a member moves
COHORT PROFILE: THE ADVENTIST HEALTH STUDY-2 263
from one Adventist church to another, a paper trail
of church transfer is usually created. Also, many
Adventists locate or retire in predominantly Adventist
communities.
What has AHS-2 found?
The study was started in 2002, and by early 2007
recruitment was almost complete. Web Tables 3–5
report some baseline lifestyle characteristics and
prevalence data. Incidence data will soon be collected
and will accrue to numbers adequate for analyses by
2010, when we expect to have 873 incident colon
cancers, 1187 breast cancers and 1098 prostate
cancers (246 in Black/African-American).
Web Table 3 reports demographic characteristics.
Females compose 65% of the cohort, the mean age is
60.2 years; 65.3% are non-Hispanic white and 26.9%
Black/African-American. Older persons are well repre-
sented with 2576 aged 85–99 years and 24 100 years
or older. Almost 100% are Seventh-day Adventist of
whom 63.7% were members of the Adventist church
by the age of 15 years. (See Web Table 3)
Web Table 4 describes selected lifestyle and
dietary characteristics. Notably, only 1.1% are current
smokers and 6.6% currently drink alcohol. Mean body
mass index was 27.4 for females and 26.8 for males.
Of particular interest is the wide diversity of dietary
status in this population. Based on the analysis of 27
relevant food questions, 4.2% are total vegetarian,
31.6% lacto-ovo-vegetarian, 11.4% include fish with
their otherwise vegetarian diet, 6.1% are semi-
vegetarian (eat meat <1 time/week) and 46.8% are
non-vegetarian. A wide distribution is also seen in the
consumption of other foods; for instance, 25% drink
soymilk several times per week and 66% eat nuts two
or more times per week (Web Table 4).
Web Table 5 summarizes various health conditions
prevalent at baseline; 83% report being in good or
excellent health, 3% had suffered a heart attack and
14.5% reported a previous diagnosis of a cancer, of
which 507 were colon or rectal, 1799 breast and 1209
prostate cancers (Web Table 5).
What are the main strengths and
weaknesses?
Main strengths include: participants’ responsiveness
and interest in health research; low levels of smoking
and alcohol consumption that reduce the effect of
confounding; support at all levels of church organiza-
tion and cooperation of local churches that enhance
the follow-up of participants; comprehensive food
frequency data; a wide diversity of diet; the wide
range of exposure to key variables—soy, vitamin D,
calcium, dairy foods and linolenic acid; broad
urban and rural geographic representation; large
numbers of persons of African descent; the large
and comprehensive calibration study to correct mea-
surement error and validate key exposures and the
planned bio-repository.
Two planned major components of the study are
pioneering and unique in scope: conducting 2000
church-based clinics across North America using
mobile teams of technicians to collect biological
measures and the plan to computer match with
60 state and provincial tumour registries. Both
components have been carefully planned and pilot
tested for AHS-2, and both will establish tested
methodologies that will benefit other large nation-
wide cohort studies.
Main weaknesses are: the length of the enrolment
questionnaire—this discouraged many from partici-
pating; the low proportion of males, especially Black
males; low representation of Hispanics largely because
the questionnaire was not available in Spanish;
a somewhat lower disease risk group, thus requiring
a little more time to accumulate incident cases of
cancers at some body sites.
Where can I find out more and what
is the potential for collaboration?
The study is still in progress and access to the data is
not yet freely available but the investigators would
welcome collaboration on specific projects. Further
details of the study are available from the study
website www.adventisthealthstudy.org or by contact-
ing the Adventist Health Study office at Loma Linda
University.
Supplementary data
Supplementary data are available at IJE online.
Acknowledgements
We recognize the commitment of the 96 000 study
participants in completing the lengthy lifestyle ques-
tionnaire and in follow-up. The cooperation and
support of church officials, and the thousands of
recruitment volunteers in churches throughout
US and Canada were invaluable for the success of
recruitment. We thank the primary investigators,
research staff, consultants and the pilot study
Tumor Registrars. The study is funded by a grant
from the National Cancer Institute (5R01 CA094594).
Conflict of interest: None declared.
References
1
Lemon F, Walden R, Woods R. Cancer of the lung and
mouth in Seventh-day Adventists: A preliminary report
on a population study. Cancer 1964;17:486–97.
264 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
2
Phillips R. Role of lifestyle and dietary habits in risk of
cancer among Seventh-day Adventists. Cancer Res 1975;
35(Suppl):3513–22.
3
Phillips R, Kuzma J. Rationale and method for an
epidemiologic study of cancer among Seventh-day
Adventists. Natl Cancer Inst Monogr 1977;47:107–12.
4
Phillips R, Lemon F, Beeson W, Kuzma J. Coronary heart
disease mortality among Seventh-day Adventists with
differing dietary habits: a preliminary report. Am J Clin
Nutr 1978;32(Suppl):S191–98.
5
Wynder E, Lemon F. Cancer, coronary artery disease, and
smoking: a preliminary report on differences in incidence
between Seventh-day Adventists and others. Calif Med
1958;89:267–72.
6
Wynder E, Lemon F, Bross I. Cancer and coronary artery
disease among Seventh-day Adventists. Cancer 1959;12:
1016–28.
7
Fraser G. Diet, Life Expectancy and Chronic Disease.Studies of
Seventh-day Adventists and Other Vegetarians. New York:
Oxford University Press, 2003.
8
Kahn H, Phillips R, Snowdon D, Choi W. Association
between reported diet and all-cause mortality. Twenty-one
year follow-up on 27,530 adult Seventh-day Adventists.
Am J Epidemiol 1984;119:775–87.
9
Phillips R, Garfinkel L, Kuzma J, Beeson W, Lotz T,
Brin B. Mortality among California Seventh-day
Adventists for selected cancer sites. JNCI 1980;65:
1097–107.
10
Fraser G, Beeson W, Phillips R. Diet and lung cancer in
California Seventh-day Adventists. Am J Epidemiol
1991;133:683–93.
11
Fraser G, Sabate J, Beeson W, Strahan T. A possible
protective effect of nut consumption on risk of coronary
heart disease. The Adventist Health Study. Arch Int Med
1992;152:1416–24.
12
Sabate J, Fraser G. The probable role of nuts in
preventing coronary heart disease. Primary Cardiology
1993;19:65–72.
13
Singh P, Fraser G. Dietary risk factors for colon cancer in
a low-risk population. Am J Epidemiol 1998;148:761–74.
14
Fraser G, Shavlik D. Ten years of life. Is it a matter of
choice? Arch Int Med 2001;161:1645–52.
15
Snowdon DA, Phillips RL, Fraser GE. Meat consumption
and fatal ischemic heart disease. Prev Med 1984;13:
490–500.
16
Mills P, Beeson W, Phillips R, Fraser G. Dietary habits
and breast cancer incidence among seventh-day adven-
tists. Cancer 1989;64:582–90.
17
Mills P, Beeson W, Phillips R, Fraser G. Cohort study of
diet, lifestyle, and prostate cancer in adventist men.
Cancer 1989;64:598–604.
18
Jacobsen BK, Knutsen SF, Fraser GE. Does high soy milk
intake reduce prostate cancer incidence? The Adventist
Health Study (United States). Cancer Causes Control 1998;
9:553–57.
19
Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ.
Cancer statistics, 2007. CA Cancer J Clin 2007;57:43–66.
20
Dai Q, Shu X, Jin F et al. Population-based case-control
study of soyfood intake and breast cancer risk in
Shanghai. Br J Cancer 2001;85:372–78.
21
Sonoda T, Nagata Y, Mori M et al. A case-control study of
diet and prostate cancer in Japan: possible protective
effect of traditional Japanese diet. Cancer Sci
2004;95:238–42.
22
Wu A, Wan P, Hankin J, Tseng C, Yu M, Pike M.
Adolescent and adult soy intake and risk of breast cancer
in Asian-Americans. Carcinogenesis 2002;23:1491–96.
23
Yamamoto S, Sobue T, Kobayashi M, Sasaki S, Tsugane S.
Soy, isoflavones, and breast cancer risk in Japan. J Natl
Cancer Inst 2003;95:906–13.
24
Beeson W, Mills P, Phillips R, Andress M, Fraser G.
Chronic disease among Seventh-day Adventists, A low-
risk group. Cancer 1989;64:570–81.
25
Herring R, Montgomery S, Yancey A, Williams D,
Fraser G. Understanding the challenges in recruiting
blacks to a longitudinal cohort study: the Adventist
Health Study. Ethnicity & Disease 2004;14:423–30.
26
Bell L, Butler T, Herring R, Yancey A, Fraser G. Recruiting
blacks to the Adventist Health Study: Do follow-up phone
calls increase response rates? Ann Epidemiol 2005;15:
667–72.
27
Montgomery S, Herring P, Yancey A et al. Comparing self-
reported disease outcomes, diet, and lifestyles in a
national cohort of black and white Seventh-day
Adventists. Prev Chronic Dis 2007;4:A62.
28
Yan R, Fraser G, Preston-Martin S. Sensitivity and PPV of
tumor registry matching for a nation-wide cohort study.
Am J Epidemiol 2006;163(Suppl):S93.
29
Chan J, Knutsen SF, Sabate J, Haddad E, Yan R,
Fraser GE. Feasibility of running clinics to collect
biological specimens in a nationwide cohort study-
Adventist Health Study-2: a brief report. Ann Epidemiol
2007;17:454–57.
30
Fraser GE, Yan R. A multivariate method for measure-
ment error correction using pairs of concentration
biomarkers. Ann Epidemiol 2007;17:64–73.
31
Knutsen S, Fraser G, Beeson W, Lindsted K, Shavlik D.
Comparison of adipose tissue fatty acids with dietary
fatty acids as measured by 24-hour recall and food
frequency questionnaire in Black and White Adventists:
the Adventist Health Study. Ann Epidemiol 2003;13:
119–27.
32
Knutsen S, Fraser G, Lindsted K, Beeson W, Shavlik D.
Validation of assessment of nutrient intake. Comparing
biological measurements of vitamin C, folate, alpha-
tocopherol and carotene with 24-hour dietary recall
information in non-Hispanic blacks and whites. Ann
Epidemiol 2001;11:406–16.
33
Myint T, Fraser GE, Lindsted KD, Knutsen SF,
Hubbard RW, Bennett HW. Urinary 1-methyl histidine
is a marker of meat consumption in black and white
California Seventh-day Adventists. Am J Epidemiol 2000;
152:752–55.
34
Singh PN, Fraser GE, Knutsen SF, Lindsted KD,
Bennett HW. Validity of a physical activity questionnaire
among African-American Seventh-day Adventists. Med Sci
Sports Exerc 2001;33:468–75.
35
Singh PN, Tonstad S, Abbey DE, Fraser GE. Validity of
selected physical activity questions in white Seventh-day
Adventists and non-Adventists. Med Sci Sports Exerc
1996;28:1026–37.
36
Fraser G, Yan R. Guided multiple imputation of missing
data: using a subsample to strengthen the missing-at-
random assumption. Epidemiology 2007;18:246–52.
COHORT PROFILE: THE ADVENTIST HEALTH STUDY-2 265
