Does Dietary Health Education Reach
Only the Privileged?
The Stanford Three Community Study
STEPHEN P. FORTMANN, M.D., PAUL T. WILLIAMS, M.S., STEPHEN B. HULLEY, M.D., M.P.H.,
NATHAN MACCOBY, PH.D., AND JOHN W. FARQUHAR, M.D.
SUMMARY The relationship of selected social factors to diet, weight and plasma cholesterol was studied
in one control and two treatment towns before and after a 3-year, bilingual, mass-media health education
program. Spanish-speaking persons reported higher dietary cholesterol and saturated fat than Englishspeaking participants at baseline, and this remained true after adjusting for the confounding influence of
socioeconomic status (SES). Obesity was also more prevalent in Spanish-language and low-SES groups, but
plasma cholesterol was not related to these sociodemographic factors. Over the 3 years of the education
program, all groups reported 20-40% decreases in dietary cholesterol and saturated fat. These decreases
were as large in low-SES groups as in high-SES groups; Spanish-speaking participants reported significantly greater decreases in dietary saturated fat (p = 0.02). Weight change was not related to either SES or
language group, but change in plasma cholesterol was marginally more favorable in Spanish-speaking
subjects (p = 0.06).
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
Materials and Methods
THE STANFORD Three Community Study was a
quasi-experimental field study to determine if a community-directed health education program could reduce the risk of cardiovascular disease (CVD). The
hypothesis was that behavioral changes resulting in
reduced levels of CVD risk factors would occur if the
residents of a community were educated about CVD
and subsequently trained in specific skills to reduce
these factors. The design, educational strategy and
overall results of this study have been reported. 1-7 The
educational program was associated with improvement
in self-reported dietary habits after 2 and 3 years3 6 of
the effort. Relative weight and cholesterol changes
were consistent with reported diet change but were less
striking.6
In this report, we concentrate on the influence of
socioeconomic status (SES) and language group on
dietary habits, relative weight and plasma cholesterol.
Community education efforts have often failed to reach
the lower SES groups as effectively as they reached the
affluent.8 We therefore designed an educational program aimed at overcoming this limitation. In this paper, we test the effectiveness of this effort to reach the
entire community with dietary information by analyzing changes in self-reported diet, obesity and plasma
cholesterol in different socioeconomic and language
groups in the treatment towns.
The Stanford Three Community Study
Three northern California communities - Watsonville, Gilroy and Tracy - were selected for study.
These towns are semirural market towns with economies centered largely around agriculture and with 1970
populations of 13,000-15,000.1"3
In each community, a multistage random sample of
men and women residents 35-59 years old were invited to participate in the baseline survey during the fall
of 1972 and in subsequent annual surveys through
1975. Each survey included an assessment of CVDrelated knowledge, attitudes and behavior, and measurements of blood pressure, relative weight and blood
lipids. Plasma cholesterol was measured by the techniques of the Lipid Research Clinics.9 Relative weight
was defined as actual weight divided by ideal weight
(the mean of the weight-for-height ranges given in the
Metropolitan Life Insurance Company ideal weight
table). 10
Watsonville and Gilroy are separated by a range of
low hills, but they share television stations and so were
selected to receive the educational program. This program began in January 1973 and continued through the
summer of 1975, although during the last year the
campaign was reduced to about half of its original
intensity. The educational material was bilingual
(English and Spanish) and presented in various media,
including television and radio programs and public
service announcements, newspaper columns, billboards and other communications. Pamphlets and
cookbooks were also distributed by direct mail to the
baseline survey participants. A small portion of the
baseline survey participants in Watsonville also received personal "intensive instruction" on risk factor
reduction in small groups,7 but these participants are
excluded from the analyses in this report, with appropriate correction for this exclusion.
Educational materials of varying complexity and
From the Stanford Heart Disease Prevention Program and the Departments of Medicine and Communication, Stanford University School of
Medicine, Stanford, California.
Supported in part by grants HL-14174 and HL-21906, NHLBI.
Dr. Fortmann was supported in part by NIH training grant 5 T32
07034 and the Robert Wood Johnson Clinical Scholars Program.
Dr. Hulley's present address: Department of Epidemiology and International Health, University of California, San Francisco, California.
Address for correspondence: Stephen P. Fortmann, M.D. Stanford
Heart Disease Prevention Program, Stanford University, 730 Welch
Road, Stanford, California 94305.
Received August 28, 1981; revision accepted November 20, 1981.
Circulation 66, No. 1, 1982.
77
78
CIRCULATION
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
reading levels were prepared. The use of mass media
and direct mail ensured that the material was available
to all segments of the community. The education team
of five professionals and three students included two
media professionals who were not only bilingual but
were educated in Spanish-speaking countries. They
translated the generally complex concepts of the campaign into culturally meaningful and understandable
products. It is difficult to present an exact comparison
of the relative amounts of effort expended in the Spanish and English programs. Each member of the education team worked on all projects to various degrees.
Comparing direct costs is misleading because television productions cost more than radio productions but
may have a similar effect. Also, the effort needed to
reach Spanish-language people may not be different
from the effort needed to reach English-speakers even
though there are fewer of the former. The number of
people that a medium reaches is more important. In
this instance, the existence of Spanish-language radio
stations with a broad reach into this community made it
possible to expend less effort in the Spanish program,
with potentially equal effect. This is reflected in the
components of each program: television constituted
40%, radio 5%, print 50%, and other media 5% of the
English language effort. In Spanish, television constituted 15%, radio 35%, print 45%, and other media 5%
of the total.
Dietary Questionnaire
The dietary questionnaire was designed to characterize the general dietary behavior of participants.
Consisting of 47 precoded questions, it provided estimates of cholesterol, saturated and unsaturated fat,
refined sugar and alcohol intake, but not of total caloric intake. Trained interviewers used food models to
assist in estimating portion sizes. A computer program
later converted the questionnaire data into estimates of
the daily consumption of cholesterol, saturated and
polyunsaturated fat, and alcohol, using food composition data published by Fetcher et al."I We compared
this dietary questionnaire to a 3-day food record and
found that it gave very similar estimates of group average cholesterol and saturated fat intake but lower estimates of polyunsaturated fat.6 We therefore report only
the first two nutrients. While theoretically less reliable
than a fixed-time dietary recall, this questionnaire is
also theoretically more accurate at identifying general
dietary habits.
Analysis
We divided the participants into five SES groups
using Hollingshead's two-factor classification, 12
where group 1 has the highest income and most education and group 5 the least. Based on their response to
questions during the second survey, we also divided
the participants into three language groups. Those who
preferred to be interviewed in Spanish were designated
Spanish-speaking. English-speaking participants were
those who used English almost exclusively in their
daily lives. The bilingual group includ
inindividuals of
VOL 66, No 1, JULY 1982
Mexican-American descent who were fluent in Spanish but who also possessed sufficient knowledge of
English to converse freely. Based on these definitions,
Watsonville and Gilroy were both 8% Spanish-speaking and Tracy was 3%. Bilingual participants constituted 9.5% of our sample in Watsonville, 17.9% in
Gilroy, and 6.0% in Tracy.
We used logarithmic transformation of the variables
in the analyses presented except for alcohol consumption. 13 These transformations substantially reduced
skewness and kurtosis; distributions of the transformed
variables for all groups were symmetric and similar.
For descriptive purposes geometric means were obtained by transforming the mean logarithms back to the
original scale. Packaged computer software was used
for much of the analysis. One-way analysis of variance
and covariance, t tests, and frequency distributions
were obtained using the Statistical Package for the
Social Sciences. For regression analyses, the Biomedical
Computer Programs statistical package was used. All p
values are two-tailed.
Results
Baseline Analysis
For analyzing the baseline dietary information we
combined the results in all three communities. Table 1
shows the geometric mean intake of various nutrients,
relative weight, and plasma cholesterol stratified by
SES and language group. There is a strong association
between language group and SES. Therefore, a relationship of either of these social factors to diet or
obesity may be confounded by the other. Table 1 allows comparison of both social factors with each dietary factor and with obesity and cholesterol. Spanish
speakers reported higher levels of dietary cholesterol
and saturated fat than English-speakers, even after adjusting for SES, age and sex. Bilingual MexicanAmericans reported intermediate cholesterol intake, but
their saturated fat intake was similar to that of Englishspeaking participants. With adjustment for language
group, age and sex, dietary cholesterol is higher in the
lower SES groups (p = 0.06) but is not different
among language groups (p = 0.14).
Table 1 shows the proportion of each language and
SES group who report any amount of alcohol intake.
These data indicate that fewer low-SES participants
and Spanish speakers report alcohol use than high-SES
people or English speakers, respectively. Since the
distribution of alcohol intake is highly skewed (in each
group, the largest category is always zero intake) we
could not satisfactorily adjust for the confounding of
these two factors (the n in several cells becomes too
small). However, the trend toward less alcohol use in
lower SES groups seen for English-speaking persons
alone is also significant (p < 0.01). Alcohol use tended to decline in all three towns during the study and
was not an object of the intervention. Because of these
factors, the skewed distribution and potential confounding factors, we did not perform longitudinal
analyses of alcohol use.
STANFORD THREE COMMUNITY STUDY/Fortmann et al.
Longitudinal Analysis
79
are significantly different between male and female
participants. An analysis of covariance (adjusting for
sex, age and language group) showed no significant
differences across SES groups for any of the changes
shown in table 2. However, the largest dietary declines
often appear in the lowest SES levels. Examining the
data across language groups, controlling for SES, also
shows generally greater improvement among bilingual
and Spanish-speaking persons, with the difference in
Table 2 shows the changes in self-reported dietary
intake, relative weight and plasma cholesterol during
the educational effort in Watsonville and Gilroy. All
comparisons are between the first survey in 1972 and
the fourth survey in 1975, for men and women combined. The percent change in various nutrients, relative weight and plasma cholesterol is displayed by SES
and language group as in table 1. None of the changes
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
TABLE 1. Mean Intake of Various Nutrients, Relative Weight, and Plasma Cholesterol at Baseline in All Three
Communities
Socioeconomic status
All
5
3
4
1
2
No. of subjects
121
104
45
885
263
352
English
101
54
3
9
34
Bilingual
1
0
0
61
7
53
Spanish
46
1047
211
393
124
273
All
Dietary cholesterol
(mg/day)*
English
Bilingual
Spanish
All
Dietary saturated
fat (g/day)*
English
Bilingual
Spanish
All
Dietary alcohol (%)t
English
Bilingual
Spanish
All
Relative weight*
English
Bilingual
Spanish
All
451
450
32.3
442
508
453
451
446
p = 0.003
32.8
32.0
37.0
32.4
32.6
67
64
168
1.12
1.12
65
32.3
p = 0.76
67
21
65
p < 0.001
1.16
1.16
1.18
1.18
p
=
1.16
0.0001
471
526
672
479
480
509
631
460
515
522 i
473
31.6
31.0
40.5
33.4
32.5
31.5
41.5
32.9
58
56
44
57
53
47
39
48
61
1.28
1.31
1.30
1.29
=
0.0001
p
=
0.0001
p
=
0.001
p
=
0.0001
p
=
0.14
637)
33.1
32.0
47.1
33.2
1.20
1.30
1.41
1.22
p
50
39
59
1.19)
1.29,2
1.31
1.21
Plasma cholesterol
(mg/dl)*
English
Bilingual
Spanish
All
199
J98
209
209
204
202
204
207
209
203
207
220
210
199
2
207
209
200)
207
p = 0.06
*The p values are by analysis of covariance after adjustment for age, sex and either socioeconomic status or language
group, as appropriate.
tPercent reporting any alcohol use; p values are unadjusted (see text).
80
CIRCULATION
dietary saturated fat decline achieving a p value of
0.02. There are no differences in relative weight
change, but the changes in plasma cholesterol are congruent with the dietary changes. That is, while cholesterol tended to rise in all three communities, it rose
significantly more in the control town.6 The rise in the
treatment towns is shown in table 2. The Spanishspeaking group showed the best results, actually showing a small decline in plasma cholesterol.
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
Discussion
Public health efforts to reduce the incidence of CVD
by modifying risk factors in large population groups
are likely to be more effective if they are based on a
knowledge of the relationship of those risk factors to
basic social factors such as SES and ethnic groups.
Furthermore, the response of different social groups to
educational efforts aimed at reducing CVD risk is important. The data presented in this paper directly concern both of these issues.
VOL 66, No 1, JULY 1982
Before beginning the education program, we found
little difference in the diet of English-speaking persons
regardless of SES. However, Mexican-Americans in
these communities consume distinctly greater amounts
of cholesterol than do English-speaking Americans,
and the Spanish-speaking subgroup of MexicanAmericans reports greater levels of dietary saturated
fat as well. Although SES is an important correlate of
diet in developing countries,'4 much less is known
about this relationship within urban, industrialized
countries such as the United States. The National Center for Health Statistics has reported on selected nutrients among white and black Americans above and below the poverty level, but dietary fat and cholesterol
were not analyzed.'" Our data indicate that MexicanAmericans in the U.S. may consume a diet of somewhat higher cholesterol and saturated fat content than
do other white Americans.
The baseline data also show mutually independent
associations of SES and language group to relative
TABLE 2. Percent Change in Various Nutrients, Relative Weight, and Plasma Cholesterolfrom 1972-1975 by Socioeconomic Status and Language Group in the Two Treatment Towns*
Socioeconomic status
1
2
4
3
5
All
No. of subjects
25
92
English
192
189
57
555
0
2
Bilingual
8
28
42
80
0
0
Spanish
I
6
45
52
All
25
94
198
226
144
687
Dietary cholesterol
- 30.3)
-20.6
-29.7
English
-27.6
32.8
33.8
Bilingual
-50.8
-33.6
-40.5
-39.7
p = 0.18
Spanish
-43.7
-39.3
-39.8)
- 32.0
37.6
All
-20.6
28.9
-30.0
-33.2
0.36
p
Dietary saturated fat
- 27.11
19.7
English
-30.4
23.8
-28.4
-29.9
- 32.5
Bilingual
-46.7
-25.1
-33.8
p = 0.02
- 53.7
Spanish
-41.6
-43.25
All
19.7
-30.7
-25.1
-28.7
34.8}
-29.0
0.30
p
Relative weight
English
-0.1
-0.5
0.1
+0.4
- 1.4
-0.2)
Bilingual
+2.9
+ 1.4
+0.3
+ 1.05
0.40
p
- 1.3
Spanish
+0.9
+ 0.6J
All
0.1
-0.3
0.0
+0.4
0.2}
0.0
p = 0.47
Plasma cholesterol
+2.4
English
+0.6
+3.1
+3.3
+2.3
+ 2.6)
Bilingual
+4.2
= 0.06
+ 2.3
-0.7
+1.1
Spanish
+ 1.3
- 1.3
All
++0.8
i 2.4
+3.1
+3.1
+0.3,
+ 2.2
p = 0.37
*All p values are by analysis of covariance after adjustment for age, sex and either socioeconomic status or language
group, as appropriate.
p
+2.2
STANFORD THREE COMMUNITY STUDYIFortmann et al.
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
weight. Numerous studies have shown obesity to be
more prevalent in lower SES groups,1621 and our data
indicate that obesity is especially prevalent among
Mexican-Americans, even compared with AngloAmericans of similar economic means.
Previous studies of CVD risk factors and disease
rates among Mexican-Americans in the United States
have created an incomplete and not fully consistent
body of knowledge. Stem and Gaskill22 found comparable mortality rates for ischemic heart disease in
Spanish-surnamed and other white residents of an urban Texas county. In contrast, Buechley and colleagues report lower ischemic heart disease mortality
in Hispanic males compared with Anglo males in New
Mexico.23 In addition to disease rates, several studies
have examined risk factors by ethnic status. Stem et
al.5 reported that overall CVD risk in the MexicanAmericans in this study was similar to the white majority, but Krauss, Borhani and Franti found a somewhat
lower risk among Spanish-Americans in northern California.24 These workers also report a strikingly higher
risk of CVD in lower SES groups for all ethnic groups
except blacks. While the evidence relating CVD to
SES is complex and somewhat'contradictory,2,29 CVD
appears to be an important cause of death in all SES
groups-and may be more prevalent in low SES groups
in the U.S. Some of the difference among studies is
probably a result of regional factors or the duration of
residence of immigrant groups in this country.
We reported the data on certain of the dietary effects
of the Stanford Three Community Study.3'6IThese data
suggest that the campaign was generally successful.
However, health education efforts aimed at reducing
CVD risk cannot ignore economic, cultural or language barriers. The data in this paper show that the
Three Community Study education program crossed
these barriers for advice concerning diet and obesity.
Generally, there were rather small differences between
SES and language groups for reported dietary change,
with an unexpected tendency for greater improvement
among lower SES and Spanish-language groups.
The occurrence of dietary behavior gains among all
SES groups in this study is reassuring empirical evidence that preventive programs can be de'signed to
appeal to all social groups in a community. Research in
the social sciences generally indicates that low-SES
persons are slower than high-SES persons to adopt
innovative health practices.' Our program was carefully' designed to reach Spanish speakers and low-SES
persons by learning and using information on their
media use habits and preparing messages suited 'to
those habits which were also culturally appropriate.
This application of formative evaluation techniques
may have resulted in messages that were more effective than is typical in health education campaigns. The
credibility of a university source for health education
may also have contributed to these results, along with
the'strong local identity of the channels for the Spanish-language program. Finally, our results may reflect
some effect of the initial risk factor survey on the
receptivity to related innovations that subsequently
appear.
81
The reported differences in dietary changes could be
explained by differential response bias. Perhaps Spanish-speaking and Mexican-American persons are more
likely than English-speaking people to supply responses that they believed were desired by the interviewers.30 The validity of the reported dietary changes
is supported, however, by the'congruent changes in
plasma cholesterol (table 2). The apparently paradoxical rise in plasma cholesterol may reflect several factors, including laboratory drift and the aging of the
cohort. This is discussed more fullj in an earlier publication.6 Studies of this type would be strengthened by
the addition of unobtrusive measures of dietary behavior such as food sales.3'
We found that of the known cardiovascular risk factors, high-fat diets and obesity are prominent in'lower
SES, Spanish-speaking and bilingual persons. Our
data suggest that a properly designed health education
effort can reach such persons at least as well as it can
reach the higher socioeconomic groups. The prevention of behavior-related disease through education
need not be limited to the highly educated and economically fortunate members of society.
Acknowledgment
The authors thank Byron W. Brown, Ph.D., for reviewing the manuscript and Susan Mellen, Ann Varady, and Brenda Craft for technical
assistance.
References
1. Farquhar JW, Maccoby N, Wood PD, Beitrose H, Haskell WL,
Meyer AJ, Alexander JK, Brown BW, McAllister AL, Nash JO,
Stem MP: Community education for cardiovascular health. Lancet
1: 1192, 1977
2. Maccoby N, Farquhar JW, Wood PD, Alexander JK: Reducing the
risk of cardiovascular disease: effects of a community-based campaign on knowledge and behavior. J Community Health 3: 100,
1976
3. Stern MP, Farquhar JW, Maccoby N, Russell SH: Results of a two
year health education campaign on dietary behavior: the Stanford
Three Community Study. Circulation 54: 826, 1976
4. Williams PT, Fortmann SP, Farquhar JW, Varady A, Mellen S: A
comparison of statistical methods for evaluating risk factor changes
in community-based studies: an example from the Stanford Three
Community Study. J Chronic Dis 34: 565, 1981
5. Stern MP, Haskell WL, Wood PD, Osann WE, King AB, Farquhar
JW: Affluence and cardiovascular risk factors in Mexican-Americans and other whites in three northern California communities. J
Chronic Dis 28: 623, 1975
6. Fortmann SP, Williams PT, Hulley SB, Haskell WL, Farquhar
JW: Effect of health education on dietary behavior: the Stanford
Three Community Study. Am J Clin Nutr 34: 2030, 1981
7. Meyer AJ, Nash JD, McAlister AL, Maccoby N, Farquhar JW:
Skills training in a cardiovascular health education campaign. J
Consult Clin Psych 48: 129, 1980
8. Rogers EM, Shoemaker FF: Communication of Innovations: A
Cross-Cultural Approach, 2nd ed. New York, MacMillan, 1971
9. United States Department of Health, Education and Welfare. Lipid
Research Clinic Manual of Laboratory Operations, vol 1. Lipid and
Lipoprotein Analysis. DHEW publ. no. (NIH) 75-628. Washington DC, U.S. Govt Print Off 1974
10. Metropolitan Life Insurance Company: New weight standards for
men and women. Stat Bull Metropol Life Ins Co 40: 1, 1959
11. Fetcher ES, Foster N, Anderson JT, Grande F, Keys A: Quantitative estimation of diets to control serum cholesterol. Am J Clin
Nutr 20: 475, 1967
12. Hollingshead AB: Two Factor Index of Social Position. New Haven, Yale University Press, 1957
13. Tukey JW: Exploratory Data Analysis. New York, Addison-Wes-
82
CIRCULATION
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
ley, 1977, ch 3
14. Garcia-Palmieri MR, Tillotson J, Cordero E, Costas R, Sorlie P,
Gordon T, Kannel WB, Colon AA: Nutrient intake and serum
lipids in urban and rural Puerto Rican men. Am J Clin Nutr 30:
2092, 1977
15. United States National Center for Health Statistics: Dietary intake
findings, United States, 1971-1974. Vital and health statistics,
series 11, number 202. DHEW publ. no. (HRA) 77-1647. Washington DC, U.S. Govt Print Off 1977
16. Oken B, Hartz A, Giefer E, Rimm AA: Relation between socioeconomic status and obesity changes in 9046 women. Prev Med 6:
447, 1977
17. Goldblatt P, Moore M, Stunkard A: Social factors in obesity.
JAMA 192: 1039, 1965
18. Silverstone J: Obesity and social class. Psychother Psychosom 18:
226, 1970
19. Stunkard A, d'Aquili E, Fox S, Filion R: Influence of social class
on obesity and thinness in children. JAMA 221: 579, 1972
20. Rimm I, Rimm A: Association between socioeconomic status and
obesity in 59,556 women. Prev Med 3: 543, 1974
21. Kohrs MB, Wang LL, Eklund D, Paulsen B, O'Neal R: The association of obesity with socioeconomic factors in Missouri. Am J Clin
Nutr 32: 2120, 1979
22. Stern MsP, Gaskill ST: Secular trends in ischemic heart disease and
stroke mortality from 1970 to 1976 in Spanish-surramed and other
white individuals in Bexar County, Texas. Circulation 58: 537,
VOL 66, No 1, JULY 1982
1978
23. Buechley RW, Key CR, Morris DL, Morton WE, Morgan MV:
Altitude and ischemic heart disease in tricultural New Mexico: an
example of confounding. Am J Epidemiol 109: 663, 1979
24. Krauss JF, Borhani NO, Franti CE: Socioeconomic status, ethnicity, and risk of coronary heart disease. Am J Epidemiol 111: 407,
1980
25. Syme SL, Hyman MM, Enterline PE: Some social and cultural
factors associated with the occurrence of coronary heart disease. J
Chronic Dis 17: 227, 1967
26. Tyroler HA, Cassel J: Health consequences of cultural change. II.
The effect of urbanization on coronary heart mortality in rural
residents. J Chronic Dis 17: 167, 1964
27. Shekelle RB, Ostfeld AM', Paul 0: Social status and incidence of
coronary heart disease. J Chronic Dis 22: 381, 1969
28. Halliday ML, Anderson TW: The sex differential in ischemic heart
disease: trends by social class 1931-1971. Epidemiol Comm
Health 33: 74, 1979
29. Jenkins CD: Recent evidence supporting psychologic and social
risk factors for coronary disease. N Engl J Med 294: 987, 1976
30. Aday LA, Chiu GY, Anderson R: Methodological issues in health
care surveys of the Spanish heritage' population. Am J Public
Health 70: 367, 1980
31. Webb EJ, Campbell DT, Schwartz RD, Sechrest L: Unobtrusive
Measures: Nonreactive Research in the Social Sciences. Chicago,
Rand McNally, 1966
Does dietary health education reach only the privileged? The Stanford Three Community
Study.
S P Fortmann, P T Williams, S B Hulley, N Maccoby and J W Farquhar
Downloaded from http://circ.ahajournals.org/ by guest on June 16, 2017
Circulation. 1982;66:77-82
doi: 10.1161/01.CIR.66.1.77
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1982 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/66/1/77
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/