Content uploaded by Brooks B. Gump
All content in this area was uploaded by Brooks B. Gump on Oct 04, 2019
Content may be subject to copyright.
Are Vacations Good for Your Health? The 9-Year Mortality Experience After the
Multiple Risk Factor Intervention Trial
BROOKS B. GUMP,PHD, MPH, AND KAREN A. MATTHEWS,PHD
Objective: The objective of this study was to determine the risk for various causes of posttrial death associated with
vacation frequency during the Multiple Risk Factor Intervention Trial (MRFIT). Methods: Middle-aged men at high
risk for coronary heart disease (CHD) were recruited for the MRFIT. As part of the questionnaires administered
during the first five annual visits, men were asked whether they had had a vacation during the past year. For trial
survivors (N⫽12,338), the frequency of these annual vacations during the trial were used in a prospective analysis
of posttrial all-cause and cause-specific mortality during the 9-year follow-up period. Results: The relative risk (RR)
associated with more annual vacations during the trial was 0.83 (95% confidence interval [CI], 0.71–0.97) for
all-cause mortality during the 9-year follow-up period. For cause of death, the RRs were 0.71 (95% CI, 0.58–0.89)
and 0.98 (95% CI, 0.78–1.23) for cardiovascular and noncardiovascular causes, respectively. The RR was 0.68 (95%
CI, 0.53–0.88) for CHD (including acute myocardial infarction). These associations remained when statistical
adjustments were made for possible confounding variables, including baseline characteristics (eg, income), MRFIT
group assignment, and occurrence of a nonfatal cardiovascular event during the trial. Conclusions: The frequency
of annual vacations by middle-aged men at high risk for CHD is associated with a reduced risk of all-cause mortality
and, more specifically, mortality attributed to CHD. Vacationing may be good for your health. Key words: coronary
heart disease, vacations, respite, restorative behaviors, Multiple Risk Factor Intervention Trial.
CHD ⫽coronary heart disease; CI ⫽confidence inter-
val; DBP ⫽diastolic blood pressure; ICD-9 ⫽Interna-
tional Classification of Diseases, ninth revision; MI ⫽
myocardial infarction; MRFIT ⫽Multiple Risk Factor
Intervention Trial; RR ⫽relative risk; SES ⫽socioeco-
nomic status; SI ⫽special intervention; UC ⫽usual
Psychological stress is positively associated with
morbidity and mortality for a variety of diseases, in-
cluding atherosclerosis (1), MI (2, 3), metabolic syn-
drome (4, 5), infectious disease (6, 7), and AIDS (8, 9).
Stress is thought to influence disease risk through a
number of pathways, including alterations in health
behaviors such as smoking (10), alcohol consumption
(11), compliance with medications (12), and emotional
states (13). More recently, it has been recognized that
normal restorative activities, such as sleep (14), exer-
cise (15), and other leisure time activities, might influ-
ence disease risk. The present study examined
whether frequent annual vacations, a common form of
respite, serves a health protective function.
The Framingham Heart Study found an association
between infrequent vacationing and increased inci-
dence of MI or death due to coronary causes during a
20-year follow-up of women participants (16). Because
coronary death was not analyzed separately in this
study (there were only 10 coronary deaths), this in-
creased risk associated with infrequent vacationing
does not necessarily represent an increased risk of
mortality. In another study, men who developed psy-
chosomatic illnesses were less likely to take vacations
than men who never developed such illnesses (17).
Although taking regular annual vacations may serve
a protective function, it is important to consider alter-
native explanations. For example, higher SES may pro-
duce both lower morbidity and mortality (18) and
more frequent annual vacations, thereby producing a
spurious association between frequent annual vaca-
tions and physical health. Along similar lines, poor
health (eg, nonfatal MIs) may produce both higher
rates of subsequent mortality (19) and prevent annual
vacations. It is important to consider these alternative
explanations when analyzing the effects of vacationing
annually on mortality.
The purpose of this study was to evaluate 9-year
posttrial mortality and cause of death among MRFIT
participants as a function of the number of annual
vacations assessed during 5 years of the trial. This
study includes statistical controls for nonfatal health-
related events and SES.
Design of the MRFIT
A total of 361,662 men aged 35 to 57 years were initially screened
at 22 clinical centers in 18 US cities for MRFIT. A total of 12,866 of
these men were enrolled in the trial. At the time of study entry, these
men were in the upper 10% to 15% of the risk score distribution,
which was derived from the Framingham Heart Study data. Exclu-
sion criteria included clinical evidence of CHD, based on history,
From the Department of Psychology (B.B.G.), State University of
New York, Oswego, NY; and the Department of Psychiatry (K.A.M.),
University of Pittsburgh, Pittsburgh, PA.
Address reprint requests to: Karen Matthews, Department of Psy-
chiatry, University of Pittsburgh, 3811 O’Hara Street, Pittsburgh, PA
15213. Address correspondence to: Brooks B. Gump, PhD, MPH,
Department of Psychology, State University of New York, Oswego,
NY 13126. Email: firstname.lastname@example.org
Received December 27, 1999; revision received March 15, 2000.
608 Psychosomatic Medicine 62:608–612 (2000)
Copyright © 2000 by the American Psychosomatic Society
physical examination, or electrocardiography (at rest); a serum cho-
lesterol level ⱖ9.05 mmol/liter (350 mg/dl); DBP readings ⱖ115
mm Hg; or a body weight at least 50% greater than the standard
weight for height (20–22). After exclusion, the remaining partici-
pants were randomly assigned to an SI (N⫽6428) or UC (N⫽6438)
group. Participants in the SI group received dietary instructions
designed to alter eating patterns to reduce the intake of saturated fats
and to reduce blood cholesterol levels, a smoking cessation program,
and stepped-care drug therapy for hypertension. Details of the mul-
tifactor intervention program are described elsewhere (23, 24). In
our analyses, we included survivors of the trial (N⫽12,338).
Measurement of Annual Vacations
Vacationing was assessed by an item on a checklist of life events
(“Within the last 12 months, have you experienced a vacation?”).
The vacation item was not included on the checklist used at baseline
but was included on the checklist used at the first, second, third,
fourth, and fifth annual examinations. This checklist did not include
an option for indicating that an event had not occurred; as a result,
a missing response on items on this questionnaire did not enable
differentiation of missing data from intentional indications that an
event had not occurred. However, the life events checklist was
administered along with a number of other questionnaires; there-
fore, we treated data as missing for participants with known missing
data for five forced-choice questions that either directly preceded or
followed (depending on year) the life events questionnaire. Missing
data increased steadily from year 1 (N⫽695, 5.4%) to year 5 (N⫽
1550, 12.0%) because of participant death or attrition. Mean vaca-
tioning was computed for all valid responses and reflects the pro-
portion of years the participant reported vacationing for those years
with valid data.
Risk Factor Assessment
The following risk factors were considered: age at study entry;
MRFIT group assignment (SI or UC); education, measured on a scale
from 1 (eighth grade or less) to 9 (graduate or professional degree);
total family income, measured on a scale from 1 (⬍$4200) to 9
(ⱖ$35,000); DBP, defined as the average of two random-zero ma-
nometer readings; serum cholesterol concentration; and cigarette
smoking, determined by self-report. To index cumulative risk factor
burden, we computed trial means for blood pressure, serum choles-
terol, and smoking. For smoking, the average corresponds to the
proportion of assessments when smoking was reported. Additional
details about these assessments are published elsewhere (25, 26).
Morbidity and Mortality Ascertainment
A nonfatal cardiovascular event during the trial was defined as
angina (Rose questionnaire), intermittent claudication (Rose ques-
tionnaire), congestive heart failure, peripheral arterial occlusive dis-
ease, stroke, left ventricular hypertrophy (electrocardiography), im-
paired renal function, accelerated hypertension, coronary artery
bypass surgery, serial electrocardiographic evidence of MI, or defi-
nite clinical MI (determined by coding criteria of electrocardio-
graphic tracings and/or physician inspection of hospital records)
Since February 1982, vital status has been ascertained by match-
ing identifying information reported by participants at the time of
enrollment with the National Death Index. The latest search of the
National Death Index was for all deaths through December 1990 and
is considered to be essentially 100% complete (28). To determine
cause of death, death certificates were collected and coded indepen-
dently by two nosologists using ICD-9 codes (29). Disagreements
between the two nosologists were adjudicated by a third nosologist.
Cause-specific mortality categories (and corresponding ICD-9 codes)
considered in the present study were selected by the MRFIT Mor-
tality and Morbidity Committee and are reported elsewhere (27).
All analyses included trial survivors and considered mortality
during the 9-year follow-up period. Posttrial mortality was analyzed
using Cox proportional hazard regression equations and 95% CIs for
the RRs associated with mean vacationing during the trial. In these
analyses, proportion of years with report of vacation was treated as
a continuous variable (ranging from 0 to 1.00). To control for possi-
ble confounding variables, additional analyses included age, MRFIT
group assignment, DBP, serum cholesterol, education, family in-
come, and cigarette smoking as covariates. Because the occurrence
of nonfatal cardiovascular events (ie, morbidity) might both interfere
with vacationing and predict future cardiovascular mortality, the
occurrence of a nonfatal cardiovascular event during the trial was
included as an additional covariate.
For analysis of differences in participants’ trial characteristics as
a function of annual vacationing, vacationing frequency was dichot-
omized into those reporting vacationing ⱕ60% of the years surveyed
(low-frequency group, N⫽6745) and those reporting vacationing for
⬎60% of the years surveyed (high-frequency group, N⫽5822) in
these analyses. This dichotomous variable (low vs. high vacationing
frequency) was used in analyses of variance for tests of continuous
variables (blood pressure, age, educational attainment, smoking, and
cholesterol) and a 2 ⫻2
analyses of group proportions for tests of
categorical variables (ie, study group and occurrence of nonfatal
cardiovascular events during the trial).
Characteristics of Participants Taking Frequent
and Infrequent Vacations
For the five assessments of vacationing in the past
year, 12.8% of participants in MRFIT reported never
taking an annual vacation, 10.1% reported 1 annual
vacation, 11.4% reported 2 annual vacations, 14.4%
reported 3 annual vacations, 18.2% reported 4 annual
vacations, and 25.6% reported 5 annual vacations. The
remaining 7.6% had missing data; therefore, frequency
of annual vacations for these participants was the pro-
portion of years in which the participant took a vaca-
tion for those years with valid data. The differences in
characteristics of participants taking frequent and in-
frequent annual vacations are shown in Table 1. The
large sample in MRFIT provides power to detect rather
small effects. Those taking frequent annual vacations
during the trial were significantly younger at study
entry, were less likely to be in the SI group, were less
likely to report cigarette smoking, were more educated,
had a higher family income, were less likely to expe-
rience a nonfatal cardiovascular event during the trial,
and had a higher serum cholesterol level during the
VACATIONS AND HEALTH
609Psychosomatic Medicine 62:608–612 (2000)
Frequency of Annual Vacations During the Trial
and Posttrial Mortality
The median length of follow-up for survivors of the
trial through December 1990 was approximately 9
years (8.8 years; range ⫽7.8 to 10.1 years). During this
time, 770 cardiovascular and 743 noncardiovascular
deaths occurred. The RRs associated with more fre-
quent annual vacations were 0.68 (95% CI, 0.59–0.79)
for all-cause mortality, 0.61 (95% CI, 0.50–0.75) for
cardiovascular mortality, and 0.77 (95% CI, 0.62–0.96)
for noncardiovascular mortality. As shown in Table 2,
after the addition of covariates to control for potential
confounding factors, the RRs associated with more
frequent annual vacations were 0.83 (95% CI, 0.71–
0.97) for all-cause mortality, 0.71 (95% CI, 0.58–0.89)
for cardiovascular mortality, and 0.98 (95% CI, 0.78–
1.23) for noncardiovascular mortality.
In addition to analyses of all cardiovascular and all
noncardiovascular deaths, specific causes of death
within each category were considered. For cardiovas-
cular mortality, the frequency of annual vacations was
associated with a significant RR, 0.68 (95% CI, 0.53–
0.88), for CHD. Within the CHD category, cause of
death was further categorized into acute MI (ICD-9
code 410) and other ischemic (coronary) heart disease
(ICD-9 codes 411–414 and 429.2). The RRs associated
with more frequent annual vacations were marginally
significant for acute MI (RR ⫽0.70, 95% CI, 0.49–1.01)
and significant for other ischemic (coronary) heart dis-
ease (RR ⫽0.66, 95% CI, 0.46–0.94). The frequency of
annual vacations was not associated with significant
RRs for any other specific cause of cardiovascular mor-
tality or noncardiovascular mortality.
The effect of morbidity on both vacationing fre-
quency and subsequent mortality is a particular con-
cern when considering the effects of the frequency of
annual vacations on subsequent mortality. Therefore,
we conducted additional analyses using only the fre-
quency of annual vacations assessed at the first, sec-
ond, and third annual examinations to predict mortal-
ity in survivors of the trial. Using this approach, a
4-year gap is placed between the assessment of the
frequency of annual vacations and the assessment of
mortality, making it unlikely that any observed asso-
ciation between vacationing frequency and mortality
would be a spurious association created by their com-
mon association with morbidity. In these analyses and
with the addition of covariates to control for potential
confounding factors, the RRs associated with more
frequent annual vacations were 0.86 (95% CI, 0.76–
1.01, p⫽.079) for all-cause mortality and 0.76 (95%
CI, 0.60–0.95, p⫽.018) for CHD mortality.
TABLE 1. Characteristics of Participants in MRFIT Taking
Infrequent and Frequent Annual Vacations During the Trial
Age at entry into study, years 46.57 46.21 .001
Study group, % in SI group 51.49 48.29 ⬍.001
Smoking, %† 48.66 43.45 ⬍.001
DBP, mm Hg
86.36 86.21 .189
Serum cholesterol, mg/dl
239.16 241.48 ⬍.001
5.14 5.70 ⬍.001
Total family income
6.21 6.85 ⬍.001
Nonfatal cardiovascular event
during trial, %
24.02 19.34 ⬍.001
Values are trial means.
Education was measured using a nine-point scale (1 ⫽eighth grade
or less, 9 ⫽graduate or professional degree).
Total family income was measured using a nine-point scale (1 ⫽
⬍$4200, 9 ⫽ⱖ$35,000).
TABLE 2. Cause of Death, Number of Deaths, and RR (with 95%
CI) for Deaths Through 9 Years of Post-trial Follow-Up
Associated With Frequency of Annual Vacations During the Trial
Cause of Death Deaths
)RR (95% CI)
All causes 1443 0.83 (0.71–0.97) .018
All cardiovascular causes 745 0.71 (0.58–0.89) .002
CHD 540 0.68 (0.53–0.88) .003
Acute MI 262 0.70 (0.49–1.01) .058
Other ischemic CHD 278 0.66 (0.46–0.94) .020
Cardiac dysrhythmias 27 0.72 (0.23–2.23) .571
Hypertensive heart disease 14
Other hypertensive 7
Cerebrovascular 61 0.86 (0.41–1.82) .699
Other cardiovascular disease 97 0.76 (0.42–1.38) .361
All noncardiovascular causes 696 0.98 (0.78–1.23) .868
Neoplastic 464 1.16 (0.88–1.53) .294
Lip, oral cavity, and pharynx 10
Digestive organs and peritoneum 120 1.25 (0.72–2.18) .415
Colorectal 47 1.00 (0.42–2.40) .996
Other gastrointestinal 73 1.45 (0.71–2.95) .305
Respiratory and intrathoracic
187 1.15 (0.74–1.77) .541
Lung 178 1.07 (0.69–1.68) .751
Other neoplasms 147 1.12 (0.68–1.82) .663
Respiratory 41 1.16 (0.45–2.95) .760
Digestive system 48 0.54 (0.23–1.25) .152
Accidents 57 0.65 (0.30–1.42) .282
Other noncardiovascular disease 86 0.71 (0.38–1.34) .292
In all Cox proportional hazard models, the following characteris-
tics were included as covariates: age, study group (SI vs. UC), edu-
cational attainment, income, occurrence of a nonfatal cardiovascular
event during the trial, smoking, DBP, and serum cholesterol (the
later three values were trial averages). Missing covariate values
resulted in a subject being dropped from the analysis.
Insufficient number of events for calculation of RR.
B. B. GUMP AND K. A. MATTHEWS
610 Psychosomatic Medicine 62:608–612 (2000)
More frequent annual vacations during the MRFIT
was associated with a significant reduction in the risk
of death during the 9-year posttrial period. The spe-
cific cause of death most strongly associated with va-
cationing frequency was CHD. This association per-
sisted with the addition of statistical controls for
Although frequent vacationing may have a direct
protective effect on health, it is important to consider
alternative explanations for the observed associations.
First, it is possible that morbidity produces both less
frequent vacationing and an increased risk of death,
thereby producing a spurious association between the
frequency of annual vacations and mortality. In other
words, those who are ill are both unable to take a
vacation and more likely to die. The continued asso-
ciation of annual vacation frequency and CHD mortal-
ity in the context of statistical controls for nonfatal
cardiovascular events and when using the frequency of
annual vacations assessed during the first 3 years of
the trial to predict mortality 4 years latter (during the
posttrial period) does not support this alternative
The positive association between SES and health is
well documented (18), and in the current sample,
lower SES was associated with less frequent vacation-
ing. Therefore, another possibility is that those of
lower SES are both unable to take a vacation and more
likely to die. A continued significant association be-
tween vacationing frequency and CHD mortality in the
context of a statistical control for participants’ educa-
tional attainment and income does not support this
There are a few possible mechanisms through
which vacationing might have direct protective effects
on health. First, vacations may reduce stress by remov-
ing ongoing stressors (eg, avoidance). The health ben-
efits of stress reduction are well documented (30, 31).
Furthermore, the current pattern of findings (ie, stron-
ger effects with CHD relative to other causes of death,
eg, cancer) is consistent with research demonstrating
stronger stress and disease associations for CHD rela-
tive to cancer (32). Second, vacations may reduce
stress by removing potential stressors and anticipated
threats, providing a period of “signaled safety” (33).
Anticipated threats are known to have adverse effects
as great as (34), if not greater than (35), the threat itself.
Finally, annual vacations may provide a unique oppor-
tunity for behaviors having restorative effects on ana-
bolic physiological processes, such as social contact
with family and friends (36–38) and physical activity
(15), in the context of reduction of stress-initiated cat-
Some limitations of the current study should be
noted. First, vacationing frequency was assessed using
a single question about “a vacation” within the past
year. Therefore, we have no information about the
quantity or length of vacations within each year nor
information about the quality of these vacations. Such
information might enable a description of the type and
pattern of vacationing that have health-protective ef-
fects. Second, the MRFIT included the question about
annual vacations as a measure of attention to the ques-
tionnaires. Therefore, it is possible that the generally
hasty or less than fully compliant participants were
both at greater risk of CHD death and reported few
annual vacations as a consequence of low adherence.
Finally, vacationing frequency may serve as a marker
of other activities or personality characteristics that
are, in turn, associated with the reduced risk of mor-
tality. For example, those taking annual vacations may
also engage in more health-promoting leisure-time ac-
tivities. Leisure-time physical activity was measured
in MRFIT using a questionnaire assessment of various
physical activities. This questionnaire has been vali-
dated against treadmill exercise performance in MR-
FIT (39). Furthermore, leisure-time physical activity
measured with this questionnaire had a modest in-
verse relationship with CHD and overall mortality dur-
ing MRFIT (15). However, the inclusion of leisure-time
physical activity as a covariate in the current analyses
did not alter the significant risk reduction in CHD
mortality associated with vacationing (RR ⫽0.69, p⫽
.006 with all covariates, including leisure-time physi-
In conclusion, more frequent annual vacations dur-
ing the MRFIT seemed to exert a direct positive effect
on mortality during the 9-year posttrial period. Al-
though the specific mechanism of this association re-
mains unknown, these findings suggest the impor-
tance of considering the health benefits of restorative
behaviors, such as vacationing. Vacations may not
only be enjoyable but also health promoting.
This work was supported by Grant HL58867 from
the National Institutes of Health. We thank James Nea-
ton, Ronald Prineas, Lewis Kuller, Greg Grandits, and
Yue-fang Chang for their consultation on analytic
strategy and manuscript preparation.
1. Rozanksi A, Blumenthal JA, Kaplan J. Impact of psychological
factors on the pathogenesis of cardiovascular disease and impli-
cations for therapy. Circulation 1999;99:2192–217.
VACATIONS AND HEALTH
611Psychosomatic Medicine 62:608–612 (2000)
2. Kamarck T, Jennings JR. Biobehavioral factors in sudden death.
Psychol Bull 1991;109:42–75.
3. Karasek RA, Theorell T, Schwartz JE, Schnall PL, Pieper CF,
Michela JL. Job characteristics in relation to the prevalence of
myocardial infarction in the US Health Examination Survey
(HES) and the Health and Nutrition Examination Survey
(HANES). Am J Public Health 1988;78:910–8.
4. Räikkönen K, Keltikangas-Järvinen L, Hautanen A, Adlercreutz
H. Neuroendocrine mechanisms in chronic perceived stress:
associations with the metabolic syndrome. Endocrinol Metab
5. Räikkönen K, Keltikangas-Järvinen L, Adlercreutz H, Hautanen
A. Psychosocial stress and the insulin resistance syndrome.
6. Cohen S, Williamson GM. Stress and infectious disease in hu-
mans. Psychol Bull 1991;109:5–24.
7. Cohen S, Tyrrell DAJ, Smith AP. Psychological stress and sus-
ceptibility to the common cold. N Engl J Med 1991;325:606–12.
8. Leserman J, Jackson ED, Petitto JM, Golden RN, Silva SG, Per-
kins DO, Cai J, Folds JD, Evans DL. Progression to AIDS: the
effects of stress, depressive symptoms, and social support. Psy-
chosom Med 1999;61:397–406.
9. Evans DL, Leserman J, Perkins DO, Stern RA, Murphy C, Zheng
B, Gettes D, Longmate JA, Silva SG, van der Horst CM, Hall CD,
Folds JD. Severe life stress as a predictor of early disease pro-
gression in HIV infection. Am J Psychiatry 1997;154:630–4.
10. Cohen S, Lichenstein E. Perceived stress, quitting smoking, and
smoking relapse. Health Psychol 1990;9:466–78.
11. Jennison KM. The impact of stressful life events and social
support on drinking among older adults: a general population
survey. Int J Aging Hum Dev 1992;35:99–123.
12. Coons SJ, Sheahan SL, Martin SS, Hendricks J, Robbins CA,
Johnson JA. Predictors of medication noncompliance in a sam-
ple of older adults. Clin Ther 1994;16:110–7.
13. Eckenrode J. Impact of chronic and acute stressors on daily
reports of mood. J Pers Soc Psychol 1984;46:907–18.
14. Kripke DF, Ancoli-Israel S, Mason W, Messin S. Sleep related
mortality and morbidity in the aged. In: Chase M, Weitzman E,
editors. Sleep disorders: basic and clinical research. New York:
SP Medical & Scientific Books; 1983. p. 415–29.
15. Leon AS, Connett J, Jacobs DR Jr, Rauramaa R. Leisure-time
physical activity levels and risk of coronary heart disease and
death: the Multiple Risk Factor Intervention Trial. JAMA 1987;
16. Eaker ED, Pinsky J, Castelli WP. Myocardial infarction and cor-
onary death among women: psychosocial predictors from a 20-
year follow-up of women in the Framingham Study. Am J Epi-
17. Vaillant GE. Natural history of male psychological health. IV.
What kinds of men do not get psychosomatic illness? Psycho-
som Med 1978;40:420–31.
18. Adler NE, Boyce T, Chesney MA, Cohen S, Folkman S, Kahn RL,
Syme SL. Socioeconomic status and health: the challenge of the
gradient. Am Psychol 1994;49:15–24.
19. The Multiple Risk Factor Intervention Trial Research Group.
Mortality after 16 years for participants randomized to the Mul-
tiple Risk Factor Intervention Trial. Circulation 1996;94:
20. Multiple Risk Factor Intervention Trial Research Group. Statis-
tical design considerations in the NHLI Multiple Risk Factor
Intervention Trial. J Chronic Dis 1977;30:261–75.
21. Sherwin R, Kaelber CT, Kezdi P, Kjelsberg MO, Thomas HE Jr.
The Multiple Risk Factor Intervention Trial (MRFIT). II. The
development of the protocol. Prev Med 1981;10:402–25.
22. Neaton JD, Grimm RH Jr, Cutler JA. Recruitment of participants
for the Multiple Risk Factor Intervention Trial (MRFIT). Con-
trolled Clin Trials 1987;8:41S–53S.
23. Benfari RC. Multiple Risk Factor Intervention Trial (MRFIT). III.
The model for intervention. Prev Med 1981;10:426–42.
24. Caggiula AW, Christakis G, Farrand M, Hulley SB, Johnson R,
Lasser NL, Stamler J, Widdowson G. Multiple Risk Factor Inter-
vention Trial (MRFIT). IV. Intervention on blood lipids. Prev
25. Multiple Risk Factor Intervention Trial Research Group. Multi-
ple Risk Factor Intervention Trial: risk factor changes and mor-
tality results. JAMA 1982:248,1465–77.
26. Multiple Risk Factor Intervention Trial Research Group. Base-
line rest electrocardiographic abnormalities, antihypertensive
treatment, and mortality in the Multiple Risk Factor Interven-
tion Trial. Am J Cardiol 1985;55:1–15.
27. Multiple Risk Factor Intervention Trial Research Group. Mortal-
ity after 16 years for participants randomized to the Multiple
Risk Factor Intervention Trial. Circulation 1996;94:946–51.
28. Wentworth D, Neaton J, Rasmussen W. An evaluation of the
Social Security Administration master beneficiary record file
and the National Death Index in the ascertainment of vital
status. Am J Public Health 1983;73:1270-4.
29. Anonymous. International Classification of Diseases, North
American Clinical Modification. Ann Arbor (MI): Edwards;
30. McCubbin JA, Wilson JF, Bruehl S, Ibarra P, Carlson CR, Norton
JA, Colclough GW. Relaxation training and opioid inhibition of
blood pressure response to stress. J Consult Clin Psychol 1996;
31. Dusseldorp E, van-Elderen T, Maes S, Meulman J, Kraaij V. A
meta-analysis of psychoeducational programs for coronary heart
disease patients. Health Psychol 1999;18:506–19.
32. Rosengren A, Tibblin G, Wilhelmsen L. Self-perceived psycho-
logical stress and incidence of coronary artery disease in mid-
dle-aged men. Am J Cardiol 1991;68:1171–5.
33. Weinberg J, Levine S. Psychobiology of coping in animals: the
effects of predictability. In: Levine S, Ursin H, editors. Coping
and health. NATO conference series III: human factors. New
York: Plenum; 1980. p. 39–60.
34. Spacapan S, Cohen S. Effects and aftereffects of stressor expec-
tations. J Pers Soc Psychol 1983;45:1243–54.
35. Nomikos MS, Opton EM, Averill JR, Lazarus RS. Surprise versus
suspense in the production of stress reaction. J Pers Soc Psychol
36. House JS, Landis KR, Umberson D. Social relationships and
health. Science 1988;241:540–4.
37. Kaplan GA, Cohn BA, Cohen RD, Guralnik J. The decline in
ischemic heart disease mortality: prospective evidence from the
Alameda County study. Am J Epidemiol 1988;127:1131–42.
38. Orth-Gomer K, Johnson JV. Social network interaction and
mortality: a six year follow-up study of a random sample of the
Swedish population. J Chronic Dis 1987;40:949–57.
39. Leon AS, Jacobs DR Jr, DeBacker G, Taylor HL. Relationship of
physical characteristics and life habits to treadmill exercise
capacity. Am J Epidemiol 1981;113:653–60.
B. B. GUMP AND K. A. MATTHEWS
612 Psychosomatic Medicine 62:608–612 (2000)