MilCohort Anthrax Validation Mar07 AJPM

Anthrax Vaccination in the Millennium Cohort
Validation and Measures of Health
Besa Smith, MPH, Cynthia A. Leard, MPH, Tyler C. Smith, MS, Robert J. Reed, MS, Margaret A. K. Ryan, MD, MPH,
for the Millennium Cohort Study Team
Background: In 1998, the United States Department of Defense initiated the Anthrax Vaccine Immunization Program. Concerns about vaccine-related adverse health effects followed, prompting several studies. Although some studies used self-reported vaccination data, the
reliability of such data has not been established. The purpose of this study was to compare
self-reported anthrax vaccination to electronic vaccine records among a large military
cohort and to evaluate the relationship between vaccine history and health outcome data.
Methods:

Between September 2005 and February 2006 self-reported anthrax vaccination was
compared to electronic records for 67,018 participants enrolled in the Millennium Cohort
Study between 2001 and 2003 using kappa statistics. Multivariable modeling investigated
vaccination concordance as it pertains to subjective health (functional status) and objective
health (hospitalization) metrics.

Results:

Greater than substantial agreement (kappa⫽0.80) was found between self-report and
electronic recording of anthrax vaccination. Of all participants with electronic documentation of anthrax vaccination, 98% self-reported being vaccinated; and of all participants
with no electronic record of vaccination, 90% self-reported not receiving a vaccination.
There were no differences between vaccinated and unvaccinated participants in overall
measures of health. Only the subset of participants who self-reported anthrax vaccination,
but had no electronic confirmation, differed from others in the cohort, with consistently
lower measures of health as indicated by Medical Outcomes Study 36-Item Short Form
Health Survey for Veterans (SF-36V) scores.

Conclusions: These results indicate that military members accurately recall their anthrax vaccinations.
Results also suggest that anthrax vaccination among Millennium Cohort participants is not
associated with self-reported health problems or broad measures of health problems severe
enough to require hospitalization. Service members who self-report vaccination with no
electronic documentation of vaccination, however, report lower measures of physical and
mental health and deserve further research.
(Am J Prev Med 2007;32(4):347–353) © 2007 American Journal of Preventive Medicine

nthrax vaccine was first broadly administered to
military personnel during the 1991 Gulf War,
when as many as 150,000 United States service
members received one to two doses.1,2 Beginning in
1998, the U.S. Department of Defense (DoD) initiated
the Anthrax Vaccine Immunization Program, which
began to vaccinate all 2.4 million U.S. service members,

starting with those assigned to high-threat areas.1,3
Because of a low supply of the anthrax vaccine and
other complications, the DoD scaled down vaccination
efforts in 2000, but still attempted to vaccinate service
members deployed to specific high-risk areas.1,4 Numerous public anecdotal reports questioning the safety
of the anthrax vaccine have prompted much research
on the subject. Service members’ immunizations, including the anthrax vaccination, were captured elec-

From the Department of Defense, Center for Deployment
Health Research, Naval Health Research Center, San Diego,
California
The Millennium Cohort Study Team is composed of Paul Amoroso, MD, MPH (Army Research Institute of Environmental Medicine, Natick MA); Edward J. Boyko, MD, MPH (Seattle Epidemiologic
Research and Information Center, Veterans Affairs Puget Sound
Health Care System, Seattle WA); Gary D. Gackstetter, PhD, DVM,
MPH (Department of Preventive Medicine and Biometrics, Uniformed Services University of the Health Sciences, Bethesda MD, and
Analytic Services, Inc., ANSER, Arlington VA); Tomoko I. Hooper,

MD, MPH (Department of Preventive Medicine and Biometrics,
Uniformed Services University of the Health Sciences, Bethesda MD);
Gregory C. Gray, MD, MPH (College of Public Health, University of
Iowa, Iowa City IA); James R. Riddle, DVM, MPH (Air Force Research
Laboratory, Wright-Patterson Air Force Base OH); and Timothy S.
Wells, DVM, PhD (Air Force Research Laboratory, Wright-Patterson
Air Force Base OH).
Address correspondence and reprint requests to: Besa Smith,
MPH, DoD Center for Deployment Health Research, Naval Health
Research Center, P.O. Box 85122, San Diego CA 92186-5122. E-mail:
[email protected].

Introduction

A

Am J Prev Med 2007;32(4)
© 2007 American Journal of Preventive Medicine • Published by Elsevier Inc.

0749-3797/07/$–see front matter
doi:10.1016/j.amepre.2006.12.015

347

tronically beginning only in 1998, forcing early studies
to rely on paper records or self-report of anthrax
vaccination history,5–9 underscoring the importance of
investigating the veracity of anthrax vaccination data.
Although previous studies have used self-reported
anthrax vaccination data, the accuracy of self-report
for anthrax vaccination has not been well established.6,10 –14 Most studies on adult vaccinations to date
have shown that self-report is a moderately reliable
method to determine an individual’s vaccination status.10,15–18 Research examining the validity of adult
influenza and pneumococcal vaccination data indicate
self-report to be highly sensitive (70%–100%) and
moderately specific (22%– 89%).15,17,19,20 These studies
also show that self-reported influenza vaccination is
more sensitive (94%–100%) than other self-reported
vaccinations such as the pneumococcal vaccination
(70%–97%). This may be due in part to shorter recall
time, because the influenza vaccine is administered
annually while the pneumococcal vaccine is usually
administered only once.15,19,20 Using limited anthrax
vaccine data from the 1991 Gulf War, the sensitivity of
self-reported anthrax vaccination among Gulf War veterans was 73.9%.10 However, among Gulf War veterans
with a documented anthrax vaccination, self-reported
morbidity, including functional impairment, hospitalization, and limitation of employment, was higher among
those who self-reported receiving the anthrax vaccine
compared with those who reported not receiving or not
knowing if they received the vaccine. Additional research
suggests that 1991 Gulf War service members who used a
vaccination record to self-report data were more likely to
report receiving an anthrax vaccine compared with those
who did not use a record, indicating some may have
forgotten that they received the vaccine.12 These studies
suggest that self-reported anthrax vaccination data may be
subject to recall and reporting bias. The purpose of this
study is to compare the concordance of self-reported
anthrax vaccination from a large, population-based
cohort study with electronic vaccination records. Further, baseline measures of health will be described
within the cohort to evaluate the relationship between
subjective and objective vaccine history and health
outcome data.

Methods
Population and Data Sources
The Millennium Cohort Study was launched in 2001 to
evaluate risk factors related to military service that may be
associated with long-term adverse health outcomes.21,22 The
77,047 participants used for these analyses enrolled in the
study between 2001 and 2003, and were demographically
representative of the invited sample. Participants deployed to
the first Gulf War between 1990 and 1991 were excluded from
the study population (n⫽9251) because the electronic vaccination database was not launched until 1998, and during the

348

Gulf War it was common to receive anthrax vaccination. Of
the remaining 67,796 participants, 67,018 (98.9%) indicated
whether they had ever received an anthrax vaccination, and
they form the basis of this study. These analyses were conducted between September 2005 and February 2006.
Demographic data were obtained from the Defense Manpower Data Center, and included gender, birth date, highest
education level, marital status, race/ethnicity, pay grade,
service component (active duty or Reserve/Guard), service
branch (Army, Navy, Coast Guard, Air Force, and Marines),
occupation, and previous deployment experience to Southwest Asia, Bosnia, or Kosovo. Electronic anthrax vaccination
status was determined from records maintained by the Defense Eligibility and Enrollment Reporting System (DEERS)
at the Defense Manpower Data Center, Monterey Bay CA.
Hospitalization data were obtained from the computerized
databases of standardized discharge diagnoses for hospitalizations within the Military Health System and for hospitalizations billed to the DoD by nonmilitary facilities. These
databases contain hospitalization summaries including dates
of admission and discharge, and up to eight individual
discharge diagnoses for each encounter, which are uniformly
coded across U.S. military services. These data were linked by
unique identifier to Millennium Cohort Study participants.
The baseline Millennium Cohort questionnaire consists of
67 items, including questions regarding vaccination, functional status, mental health, and other morbidity. Two questions regarding anthrax vaccination were asked. The first,
“Have you ever received the anthrax vaccine?” was followed by
“If YES, how many shots of the anthrax vaccine have you
received?” To compare differences in functional and mental
status, the Medical Outcomes Study 36-Item Short Form
Health Survey for Veterans (SF-36V) was used.23–26
The SF-36V uses standardized scoring algorithms to assess
eight health constructs: physical functioning, role limitations
caused by physical problems, bodily pain, general health,
vitality, social functioning, role limitations caused by emotional problems, and mental health.

Statistical Analyses
Vaccination concordance status was separated into four categories: respondents who self-reported not being vaccinated
with concurrence from the electronic records, respondents
who self-reported being vaccinated but electronic records
showed no vaccination, respondents who self-reported not
being vaccinated but electronic records showed at least one
anthrax vaccination, and respondents who self-reported receiving vaccination with concurrence from the electronic
records.
Descriptive measures were calculated to evaluate selfreported vaccination status compared with electronic vaccination records. Percent self-reporting vaccination among all
participants with electronic documentation of anthrax vaccination, and percent self-reporting no vaccination among all
participants with no electronic record of vaccination are
reported.
The kappa statistic was used to measure the degree of
nonrandom agreement between self-reported anthrax vaccination and electronic anthrax vaccine records.27 Definedagreement levels included “greater than substantial agreement” as a kappa (␬) between 0.8 and 1.0, “substantial

American Journal of Preventive Medicine, Volume 32, Number 4

www.ajpm-online.net

agreement” (␬⫽0.6 – 0.8), “moderate agreement” (␬⫽0.4 –
0.6), “fair agreement” (␬⫽0.2– 0.4), and “slight or poor
agreement” (␬⫽0.0 – 0.2).28 A weighted kappa statistic was
used to investigate the concordance between the number of
self-reported anthrax vaccinations and the number of vaccinations recorded in the electronic database, including only
participants who were identified as vaccinated by both
sources.
Multivariable logistic regression was performed to evaluate
associations between hospitalization and vaccination concordance status. Participants were classified as hospitalized if they
were hospitalized for any cause, except for child-bearing
reasons, in the 12 months prior to completing the survey.
Analysis of covariance was performed to evaluate the association between functional status (SF-36V) and vaccination
concordance. Functional status scales were scored with increasing score reflecting better functional health.23–26 Imputed values were used for missing questions if the respondent answered at least half of the questions in a scale. Values
were imputed based on the mean of the score for the
complete portion of that scale. If more than one half of the
questions in a scale were missing the participant was not
included in the analysis for that scale. Both logistic regression
and analysis of covariance modeling adjusted for all demographic and military-specific variables. All data analyses were
completed using SAS software (version 9.1, SAS Institute,
Inc., Cary NC, 2004).

Results
Of the 67,796 Millennium Cohort participants that
were not deployed to the first Gulf War, 67,018 (98.9%)
completed the anthrax vaccination survey question.
Most self-reported they had never received an anthrax
vaccination (n⫽46,945; 70.0%). This agreed with electronic vaccination records for 46,612 (69.6%) participants. Participants who reported receiving an anthrax
vaccination (n⫽20,073; 30%) and also had an electronic record indicating vaccination included 15,041
(22.4%) participants. The remaining 5365 (8.0%) participants had discordant results. Most discordance was
defined by self-reported receipt of vaccination, whereas
electronic records indicated no vaccination (n⫽5032;
7.5%). Finally, 333 (0.5%) participants reported not
receiving a vaccination despite the indication of one or
more anthrax vaccinations in the electronic records.
The demographic characteristics of these four vaccination groups were compared (Table 1). A greater
proportion of those who had an electronically confirmed
self-reported vaccination (concordant vaccinated) were
male; previously deployed to Southwest Asia, Bosnia, or
Kosovo; active duty; Air Force or Marine service members;
and worked as combat or electrical/mechanical repair
specialists than those whose self-report of not receiving a
vaccination was confirmed by absence of an electronic
record of vaccination (concordant unvaccinated). When
compared with concordant vaccinated participants,
proportionately more participants who self-reported
being vaccinated without electronic confirmation were
April 2007

in transitional occupations (relative %⫽2.1), in the
Navy or Coast Guard (2.0), older (1.6), less educated
(1.4), and Reserve/Guard members (1.4).
Of all participants with electronic documentation of
anthrax vaccinations, 98% self-reported being vaccinated, and of all participants with no electronic record
of vaccination, 90% self-reported not receiving a vaccination (Table 2). The kappa statistic indicated greater
than substantial agreement (␬⫽0.80). Of those whose
self-report and electronic record reflected anthrax vaccination (concordant vaccinated) there was moderate
agreement (weighted ␬⫽0.53) for the number of anthrax vaccinations received.
Among active-duty service members, hospitalization
rates for any cause in the year prior to survey participation were not significantly different between the
vaccination groups, after adjustment for gender, age,
education, marital status, race/ethnicity, previous deployment experience, pay grade, service component,
service branch, and occupation (Table 3).
Adjusted SF-36V mean scores ranged from 62 to 95
for all participants (Table 3). Means were adjusted for
gender, age, education, marital status, race/ethnicity,
previous deployment experience, pay grade, service
component, service branch, and occupation. Although
SF-36V adjusted mean scores were relatively high for all
functional health components in each vaccination
group, significantly lower scores for all eight components were found among participants whose electronic
records reflected no vaccination but who self-reported
receiving the anthrax vaccination when compared to
groups with concordant status. Additionally, the adjusted mean score for role limitations because of
physical problems was significantly higher among the
concordant vaccinated group compared with the
concordant unvaccinated group.

Discussion
Previous studies have used self-reported anthrax vaccination status to compare vaccinated and unvaccinated
participants on subjective and objective health measures.6,10,12,13 However, the validity of self-reported
anthrax vaccination has not yet been well established.
Overall, self-report and electronic records of vaccination were very consistent, with 92% of participants
having identical responses from the two sources. Agreement levels28 suggest that the overall kappa demonstrates greater than substantial agreement between
electronic and self-reported anthrax vaccination
(␬⫽0.80). Participants with electronic documentation
of anthrax vaccinations also reported vaccination 98%
of the time, which is higher than in a similar study
conducted by Mahan et al.,10 where sensitivity of selfreported anthrax vaccination among Gulf War veterans
was found to be 74%. The lower agreement in vaccination status, however, found by Mahan et al.,10 may be
Am J Prev Med 2007;32(4)

349

Table 1. Characteristics of Millennium Cohort Study participants by anthrax vaccination status
Anthrax vaccination statusa

Study sample
nⴝ67,018
%b

Concordant
unvaccinated
nⴝ46,612
n (%)

Discordantc
self-report
vaccinated
nⴝ5032
n (%)

71.5
28.5

31,517 (67.6)
15,095 (32.4)

4041 (80.3)
991 (19.7)

254 (76.3)
79 (23.7)

12,110 (80.5)
2,931 (19.5)

21.4
35.7
30.1
12.8
0.0

9,634 (20.7)
15,586 (33.4)
14,412 (30.9)
6,962 (14.9)
18 (0.0)

1230 (24.4)
1611 (32.0)
1622 (32.2)
565 (11.2)
4 (0.1)

107 (32.1)
144 (43.2)
71 (21.3)
11 (3.3)
0 (0.0)

3,364 (22.4)
6,564 (43.6)
4,073 (27.1)
1,038 (6.9)
2 (0.0)

48.6
25.7
16.8
8.8
0.0

22,988 (49.3)
10,848 (23.3)
8,260 (17.7)
4,511 (9.7)
5 (0.0)

3015 (59.9)
1065 (21.2)
630 (12.5)
322 (6.4)
0 (0.0)

201 (60.4)
79 (23.7)
37 (11.1)
16 (4.8)
0 (0.0)

6,390 (42.5)
5,253 (34.9)
2,344 (15.6)
1,054 (7.0)
0 (0.0)

39.2
60.9

18,271 (39.2)
28,341 (60.8)

2091 (41.6)
2941 (58.5)

164 (49.3)
169 (50.8)

5,710 (38.0)
9,331 (62.0)

70.1
13.2
16.4
0.4

32,702 (70.2)
6,167 (13.2)
7,548 (16.2)
195 (0.4)

3573 (71.0)
615 (12.2)
818 (16.3)
26 (0.5)

191 (57.4)
76 (22.8)
64 (19.2)
2 (0.6)

10,477 (69.7)
1,972 (13.1)
2,564 (17.1)
28 (0.2)

29.1
70.9

8,893 (19.1)
37,719 (80.9)

2140 (42.5)
2892 (57.5)

129 (38.7)
204 (61.3)

8,336 (55.4)
6,705 (44.6)

77.3
22.7

35,344 (75.8)
11,268 (24.2)

4190 (83.3)
842 (16.7)

287 (86.2)
46 (13.8)

11,965 (79.6)
3,076 (20.5)

53.4
46.7

24,867 (53.4)
21,745 (46.7)

1420 (28.2)
3612 (71.8)

80 (24.0)
253 (76.0)

3,058 (20.3)
11,983 (79.7)

46.7
29.9
18.6
4.9

25,573 (54.9)
12,118 (26.0)
7,359 (15.8)
1,562 (3.4)

1563 (31.1)
1033 (20.5)
2023 (40.2)
413 (8.2)

115 (34.5)
95 (28.5)
90 (27.0)
33 (9.9)

4,030 (26.8)
6,768 (45.0)
2,957 (19.7)
1,286 (8.6)

19.6
8.9
7.1
10.6
2.6
20.5
14.4
3.1
8.3
4.8
0.0

8,546 (18.3)
3,656 (7.8)
3,267 (7.0)
5,863 (12.6)
1,163 (2.5)
10,341 (22.2)
5,480 (11.8)
1,379 (3.0)
4,088 (8.8)
2,813 (6.0)
16 (0.0)

1034 (20.6)
57 (11.4)
380 (7.6)
303 (6.0)
137 (2.7)
852 (16.9)
1061 (21.1)
157 (3.1)
363 (7.2)
172 (3.4)
2 (0.0)

78 (23.4)
24 (7.2)
18 (5.4)
21 (6.3)
8 (2.4)
77 (23.1)
60 (18.0)
6 (1.8)
29 (8.7)
12 (3.6)
0 (0.0)

3,473 (23.1)
1,705 (11.3)
1,100 (7.3)
906 (6.0)
440 (2.9)
2,463 (16.4)
3,074 (20.4)
541 (3.6)
1,104 (7.3)
234 (1.6)
1 (0.0)

c

Characteristic
Gender
Male
Female
Age (years)
17–24
25–34
35–44
⬎44
Unknown
Education
High school or less
Some college
Bachelor’s degree
Master’s/PhD
Unknown
Marital status
Not married
Married
Race/ethnicity
White non-Hispanic
Black non-Hispanic
Other
Unknown
Recent deployment experience
Deployment experience
No deployment
Military pay grade
Enlisted
Officer
Service component
Reserve/Guard
Active duty
Branch of service
Army
Air Force
Navy/Coast Guard
Marines
Occupational category
Combat specialists
Electronic repair
Communications/intel
Healthcare specialists
Other technical
Functional support
Electrical/mechanic
Craft workers
Service support
Trainees, others
Unknown

Discordantc
electronic
vaccinated
nⴝ333
n (%)

Concordant†c
vaccinated
nⴝ15,041
n (%)

a

All unadjusted associations between anthrax vaccination status and individual characteristics were statistically significant (p⬍0.01).
Percents may not sum to 100 due to rounding.
Concordant unvaccinated: both self-report and electronic database reflect no vaccination; discordant self-report vaccinated: self-reported
vaccination but electronic database reflects no vaccination; discordant electronic vaccinated: self-reported no vaccination but electronic database
reflects vaccination; concordant vaccinated: both self-report and electronic database reflect vaccination.

b
c

350

American Journal of Preventive Medicine, Volume 32, Number 4

www.ajpm-online.net

For all eight functional status components, participants whose self-reported vaccination was not confirmed by electronic record had lower adjusted mean
scores, indicating worse health. These participants also
had lower unadjusted means compared with national
norms for ages 18 – 64 years for four of the eight
scales.34 The other three vaccination status groups had
higher unadjusted means compared with the national
norms for all components, except the concordant
unvaccinated group, which had a lower bodily pain
score. Although there has been little research to date
linking anthrax vaccination concordance to functional
health, at least one other previous study has indicated
that individuals who otherwise fit our definition
of “vaccination concordant” self-report significantly
higher levels of functional impairment, healthcare utilization, and selected medical conditions.10 In contrast,
those in our study whose self-report of vaccination was
confirmed in the electronic database did not have
significantly lower functional status scores, compared
with those who self-reported and were confirmed unvaccinated. These results may indicate there are a small
percentage of participants with lower functional status
who reported being vaccinated that either did not
actually receive an anthrax vaccine or received one that
failed to show up in the electronic database. Although
the likelihood exists that not all anthrax vaccinations
are being documented in the military electronic data,
these results may also suggest a small amount of report-

Table 2. Agreementa between self-reported anthrax
vaccination and electronic records
Electronic record
Self-report

Vaccine

No vaccine

Vaccine
No vaccine

15,041
333

5,032
46,612

a

Agreement measures for self-report are as follows: kappa⫽0.80; 98%
of participants with electronic documentation of anthrax vaccination
self-reported being vaccinated, 90% of participants with no electronic
record of vaccination self-reported not receiving a vaccination.

because of a longer period between vaccination and
recall in addition to differences in sample size and
population.
Individuals whose electronic and self-reported anthrax vaccination status did not agree were more likely
to be in transitional occupations, Navy/Coast Guard
service members, older, less educated, and Reserve/
Guard members when compared with participants
whose electronic and self-reported vaccination status
were in agreement. Previous studies have found lower
education to be associated with lower rates of response,29 higher rates of refusal,30,31 and lower rates of
reliability in self-reported data,32,33 similar to the current study. Service-specific differences and differences
found between Reserve/Guard and active-duty subgroups may be because of variations in procedures for
reporting of vaccinations.

Table 3. Adjusted odds of hospitalization and adjusted means of SF-36V health scores among Millennium Cohort
participantsa by anthrax vaccination status
Anthrax vaccination status

Health outcome
d

Any cause hospitalization
e

SF-36V

Physical functioning
Role physical
Bodily pain
General health
Social functioning
Role emotional
Mental health
Vitality

Concordantb
unvaccinated

Discordantb
self-report
vaccinated

Discordantb
electronic
vaccinated

Concordantb
vaccinated

OR (95% CI)c

OR (95% CI)

OR (95% CI)

OR (95% CI)

1.00

1.08 (0.88–1.32)

1.61 (0.89–2.90)

0.95 (0.82–1.09)

Mean

Mean

Mean

Mean

1

92.1
93.91
78.51
80.31
88.81
94.31
80.21
64.41

2

89.6
91.12
74.52
76.92
85.22
92.32
78.02
62.12

1,2

92.1
94.21,3
79.61
81.51
89.61
93.71,2
80.71
66.61

92.51
94.63
78.81
80.61
89.01
94.71
80.31
64.51

a

Participants with missing covariate data (n⫽294) or whose SF-36V component could not be scored due to insufficient questionnaire responses
were removed from analyses (n varies by component, maximum removed 347).
Concordant unvaccinated: both self-report and electronic database reflect no vaccination (reference group); discordant self-report vaccinated:
self-reported vaccination but electronic database reflects no vaccination; discordant electronic vaccinated: self-reported no vaccination but
electronic database reflects vaccination; concordant vaccinated: both self-report and electronic database reflect vaccination.
c
Odds ratio (OR), 95% confidence interval (CI), and mean are adjusted for gender, age, education, marital status, race/ethnicity, deployment
status, pay grade, service component (SF-36V only), service branch, and occupation.
d
Hospitalization analysis includes active-duty service members only, and reflects odds of any-cause admission in year prior to survey response.
e
SF-36V, Medical Outcomes Study 36-Item Short Form Health Survey for Veterans.
1,2,3
Different numbers indicate vaccination status scores that are significantly different from each other (p⬍0.05) using Scheffe’s adjustment for
multiple comparisons.
b

April 2007

Am J Prev Med 2007;32(4)

351

ing and/or recall bias with this question in the Millennium Cohort questionnaire.
There are notable limitations to these analyses that
should be mentioned. The study population consists of
a sample of responders to the Millennium Cohort
questionnaire, and may not be representative of the
U.S. military population in general. Those who might
be severely ill may not choose to respond or have the
capacity to do so. The kappa statistic is dependent on
the true prevalence of the variable being examined,
with the kappa statistic tending toward zero as the true
prevalence approaches 0 or 1.35 However, because a
considerable percentage of U.S. military service members were receiving anthrax vaccines from the late
1990s to the early 2000s, this dependence on prevalence should not substantially affect these findings. It
should be noted that the electronic vaccination database includes anthrax vaccinations only from 1998
onward. However, those that were most likely to have
received the anthrax vaccine before 1998, namely those
deployed to the first Gulf War, were excluded from this
study. Additionally, quantifying the level of incomplete
documentation in the DoD electronic vaccination database for anthrax vaccination remains a challenge. Medical records, if available, may have provided additional
vaccination history. Recent research, however, shows a
strong level of agreement between electronically
maintained anthrax vaccination records and anthrax
vaccination data abstracted from medical charts.36
Despite these limitations, this study has several
strengths. The large sample size allowed for the robust
comparison of self-report and objective measures of
anthrax vaccination. Furthermore, this study was able
to link both objective and subjective measures of morbidity to self-reported and electronically recorded vaccination status. This has not been previously done in a
single population.
In summary, military service members represent a
unique population that relies on advanced protection
during deployment and combat situations. Vaccines
against biologic agents are an important component of
protection. It is of further importance to the healthcare
community to be able to document the use of vaccines,
such as anthrax, and investigate potential associations
between vaccination and morbidity.7,37–39 This analysis
found that most people accurately recall their vaccination history. In addition, no differences in health, as
defined objectively by morbidity severe enough to
require hospitalization, were found by vaccine history.
However, service members who self-reported receiving
the anthrax vaccine with no objective evidence of
vaccination reported more health difficulties. Further
investigation into this subset will add important insight
into those who may indiscriminately report more exposures as well as more health challenges.
352

We thank Scott L. Seggerman from the Management Information Division, Defense Manpower Data Center, Seaside,
California. We also are very grateful for the help of other
team members, including Laura Chu, Gia Gumbs, Isabel
Jacobson, Travis Leleu, Steven Spiegel, Christina Spooner,
Linda Wang, James Whitmer, and Dr. Sylvia Young, from the
Department of Defense Center for Deployment Health Research, Naval Health Research Center, San Diego, California.
Additionally we thank Mary McDonell from the VA Puget
Sound Health Care System. We appreciate the support of the
Henry M. Jackson Foundation for the Advancement of Military Medicine, Rockville, Maryland.
This represents report 06-02, supported by the DoD, under
work unit no. 60002. The views expressed in this article are
those of the authors and do not reflect the official policy or
position of the Department of the Navy, Department of the
Army, Department of the Air Force, Department of Defense,
Department of Veterans Affairs, or the U.S. Government.
This research has been conducted in compliance with all
applicable federal regulations governing the protection of
human subjects in research (Protocol NHRC.2000.007).
No financial conflict of interest was reported by the authors
of this paper.

References
1. Institute of Medicine: Committee to assess the safety and efficacy of the
anthrax vaccine MF-UA. The anthrax vaccine: is it safe? Does it work?
Washington DC: National Academy Press; 2002.
2. Department of Defense and Department of Veterans Affairs. Combined
analysis of the VA and DoD Gulf War Clinical Evaluation Programs: a study
of the clinical findings from the systematic medical examinations of
100,339 U.S. Gulf War veterans. Washington DC: Author; 2002.
3. United States General Accounting Office. Chemical and biological
defense: observation on DoD’s plans to protect U.S. forces (Report No.
GAO/T-NSIAD-98-83). Washington DC: Author; 1998.
4. United States General Accounting Office. Medical readiness: DoD continues to face challenges in implementing its anthrax vaccine immunization
program (Report No. GAO/T-NSIAD-00-157). Washington DC: Author;
2000.
5. Rostker B. Information paper: military medical recordkeeping during and
after the Gulf War, 1999. Available from: http://www.gulflink.osd.mil/
mrk/.
6. Schumm WR, Reppert EJ, Jurich AP, et al. Self-reported changes in
subjective health and anthrax vaccination as reported by more than 900
Persian Gulf War era veterans. Psychol Rep 2002;90:639 –53.
7. Sato PA, Reed RJ, Smith TC, Wang L. Monitoring anthrax vaccine safety in
U.S. military service members on active duty: surveillance of 1998 hospitalizations in temporal association with anthrax immunization. Vaccine
2002;20:2369 –74.
8. Mazzuchi JF, Claypool RG, Hyams KC, et al. Protecting the health of U.S.
military forces: a national obligation. Aviat Space Environ Med 2000;
71:260 –5.
9. Murphy D, Dandeker C, Horn O, et al. UK armed forces responses to an
informed consent policy for anthrax vaccination: a paradoxical effect?
Vaccine 2006;24:3109 –14.
10. Mahan CM, Kang HK, Dalager NA, Heller JM. Anthrax vaccination and
self-reported symptoms, functional status, and medical conditions in the
National Health Survey of Gulf War Era Veterans and Their Families. Ann
Epidemiol 2004;14:81– 8.
11. Cherry N, Creed F, Silman A, et al. Health and exposures of United
Kingdom Gulf war veterans. Part II: the relation of health to exposure.
Occup Environ Med 2001;58:299 –306.
12. Unwin C, Blatchley N, Coker W, et al. Health of UK servicemen who served
in Persian Gulf War. Lancet 1999;353:169 –78.
13. Steele L. Prevalence and patterns of Gulf War illness in Kansas veterans:
association of symptoms with characteristics of person, place, and time of
military service. Am J Epidemiol 2000;152:992–1002.

American Journal of Preventive Medicine, Volume 32, Number 4

www.ajpm-online.net

14. Hotopf M, David A, Hull L, Ismail K, Unwin C, Wessely S. Role of
vaccinations as risk factors for ill health in veterans of the Gulf War:
cross-sectional study [see comment]. BMJ 2000;320:1363–7.
15. Mac Donald R, Baken L, Nelson A, Nichol KL. Validation of self-report of
influenza and pneumococcal vaccination status in elderly outpatients. Am J
Prev Med 1999;16:173–7.
16. Stange KC, Zyzanski SJ, Smith TF, et al. How valid are medical records and
patient questionnaires for physician profiling and health services research?
A comparison with direct observation of patients visits. Med Care 1998;
36:851– 67.
17. Shenson D, Dimartino D, Bolen J, Campbell M, Lu PJ, Singleton JA.
Validation of self-reported pneumococcal vaccination in behavioral risk
factor surveillance surveys: experience from the sickness prevention
achieved through regional collaboration (SPARC) program. Vaccine
2005;23:1015–20.
18. Nichol KL, Korn JE, Baum P. Estimation of outpatient risk characteristics
and influenza vaccination status: validation of a self-administered questionnaire. Am J Prev Med 1991;7:199 –203.
19. Andrews RM. Assessment of vaccine coverage following the introduction
of a publicly funded pneumococcal vaccine program for the elderly in
Victoria, Australia. Vaccine 2005;23:2756 – 61.
20. Zimmerman RK, Raymund M, Janosky JE, Nowalk MP, Fine MJ. Sensitivity
and specificity of patient self-report of influenza and pneumococcal
polysaccharide vaccinations among elderly outpatients in diverse patient
care strata. Vaccine 2003;21:1486 –91.
21. Gray GC, Chesbrough KB, Ryan MA, et al. The Millennium Cohort Study:
a 21-year prospective cohort study of 140,000 military personnel. Mil Med
2002;167:483– 8.
22. Ryan MAK, Smith TC, Smith B, et al. Millennium cohort: enrollment
begins a 21-year contribution to understanding the impact of military
service. J Clin Epidemiol 2007;60:181–91.
23. Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey
(SF-36). I. Conceptual framework and item selection. Med Care 1992;
30:473– 83.
24. Kazis LE, Lee A, Spiro A 3rd, et al. Measurement comparisons of the
medical outcomes study and veterans SF-36 health survey. Health Care
Financ Rev 2004;25:43–58.

April 2007

25. Kazis LE, Miller DR, Clark JA, et al. Improving the response choices on the
veterans SF-36 health survey role functioning scales: results from the
Veterans Health Study. J Ambul Care Manage 2004;27:263– 80.
26. Kazis LE, Miller DR, Skinner KM, et al. Patient-reported measures of
health: the Veterans Health Study. J Ambul Care Manage 2004;27:70 – 83.
27. Cohen JA. A coefficient of agreement for nominal scales. Educ Psychol
Meas 1960;20:37– 46.
28. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159 –74.
29. Madigan MP, Troisi R, Potischman N, et al. Characteristics of respondents
and non-respondents from a case– control study of breast cancer in
younger women. Int J Epidemiol 2000;29:793– 8.
30. Voigt LF, Koepsell TD, Daling JR. Characteristics of telephone survey
respondents according to willingness to participate. Am J Epidemiol 2003;
157:66 –73.
31. Holt VL, Martin DP, LoGerfo JP. Correlates and effect of non-response in
a postpartum survey of obstetrical care quality. J Clin Epidemiol 1997;
50:1117–22.
32. Johnson TP, Mott JA. The reliability of self-reported age of onset of
tobacco, alcohol and illicit drug use. Addiction 2001;96:1187–98.
33. Huerta M, Chodick G, Balicer RD, Davidovitch N, Grotto I. Reliability of
self-reported smoking history and age at initial tobacco use. Prev Med
2005;41:646 –50.
34. Ware J, Kosinski M, Gandek B. SF-36 Health Survey: manual and interpretation guide. Lincoln, NE: QualityMetric Incorporated; 2002.
35. Thompson WD, Walter SD. A reappraisal of the kappa coefficient. J Clin
Epidemiol 1988;41:949 –58.
36. Payne DC, Rose CE, Aranas A, et al. Assessment of anthrax vaccination data
in the Defense Medical Surveillance System, 1998 –2004. Pharmacoepidemiol Drug Saf 2007; In press.
37. Grabenstein JD. Anthrax vaccine: a review. Immunol Allergy Clin North
Am 2003;23:713–30.
38. Lange JL, Lesikar SE, Rubertone MV, Brundage JF. Comprehensive
systematic surveillance for adverse effects of anthrax vaccine adsorbed, U.S.
Armed Forces, 1998 –2000. Vaccine 2003;21:1620 – 8.
39. Wasserman GM, Grabenstein JD, Pittman PR, et al. Analysis of adverse
events after anthrax immunization in U.S. Army medical personnel. J
Occup Environ Med 2003;45:222–33.

Am J Prev Med 2007;32(4)

353