,. ‘ V _ V
. .A . I A
V , V . _ ,.
,. . .
_ , , . . V. .. .V.
,,_ .. . . ....
, .
. . . A ... Al
. . .. . n
. . A.
V,. . V . ., Ag
. A . ‘ ‘ Vb
, V D
.
.. .. b
. a
. ‘ A m
, ,. . A A.
n, . , . .. A n d
A A A. u
,. . .1 . fl
. .
V , . A , u.
., V . . .
A ‘ .- A
A I . . .
. ‘ , . , . .
_ ,. , 2 .
. , . , . .
_ .. u , . .

 

 

.....- #-
.

 

 

 

 

 

_ G x :V, . y . .I
q Vv . A . A r . l .
y .A. V .
_ , A 2 .
. . .A A , . .
. . u . . . . . V
7 , , I . u
.. .2 . , , . . A .
_ V .n . A
. n. . _ A . . . , . 4
A ,. . V , . . . , . .
V . , . . . ‘ . .
. . . f . . >
.A .. A . . . A . .
. A . . ... . , V .. A. V , a , I y I

 

 

 

 

 

 

STA

‘it'lliulmllllllilli‘

3129301712

This is to certify that the

thesis entitled

SOCIOECONIMIC STATUS AND MODIFABLE RISK FACTORS
FOR CORONARY HEART DISEASE FOR WOMEN
IN NORTHERN MICHIGAN

presented by

Jane M . Denay

has been accepted towards fulfillment
of the requirements for

11.5.. degree in JEIRSJNL .

gig/S14

/ Major professor

 

Date 47/2 1/73

0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY
Michigan State
UnIversity

 

 

 

PLACE IN RETURN Box
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we glamorous-p.14

SOCIOECONOMIC STATUS AND MODIFIABLE RISK FACTORS FOR
CORONARY HEART DISEASE FOR WOMEN IN NORTHERN MICHIGAN
By

Jane M. Denay

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE IN NURSING

College of Nursing

1998

ABSTRACT

SOCIOECONOMIC STATUS AND MODIFIABLE RISK FACTORS FOR
CORONARY HEART DISEASE FOR WOMEN IN NORTHERN MICHIGAN

By

Jane M. Denay

Cardiovascular disease is the leading cause of death in women in the United
States. Half of these deaths are due to coronary heart disease. Socioeconomic status has
been demonstrated to be a powerful predictor of the development of coronary heart
disease. Also, the individual risk factors which predispose women to coronary heart
disease have been found to be correlated with socioeconomic status.

This cross-sectional study of women’s coronary heart disease risk explores two
research questions: (a) what is the extent of the correlation between the individual risk
factors for coronary heart disease in women, and (b) what is the impact of
socioeconomic status on women’s individual risk factors for coronary heart disease. The
“web of causation” is the framework through which socioeconomic status will be
explored as an antecedent to the more immediate individual risk factors of CHD.

The results of this study found a higher prevalence of individual risk factors for
CHD with lower levels of education, lower income, advanced age, and being unmarried
Multiple risk factors were associated with being overweight and having hypertension.
Advanced practice nurse implications for community level interventions, directed at
ameliorating the impact of socioeconomic status on individual risk factors for coronary

heart disease, are discussed.

a
_

. .~‘
s . ..
5.;st .

aim .

 

This work is dedicated to my family

who have lovingly and faithfully shared my journey.

ACKNOWLEDGMENTS

With much appreciation I thank the individuals and organization which made
this research project possible, Manfred Stommel, chairperson of my thesis committee,
whose research expertise facilitated all phases of the project; Joan Wood, whose
community nursing expertise guided, in particular, the theoretical framework as well as
practice and research implications; and Patricia Peek, who provided the seminal idea for
utilizing the Northern Michigan Community Health Assessment for this research project,
offered thoughtful input and gave support throughout the process.

I am grateful to the Northern Michigan Hospitals-Bums Clinic Foundation, the
Director of Research Robert Sloan, and Research Assistant Minda Latham who shared the
Northern Michigan Community Health Assessment with me and assisted cheerfully at key
times.

And special thanks to my husband, Cliff, who faithfully proofed countless drafts of
this thesis without complaint, and my children, Nathaniel and Emily, who helped me keep

my sense of humor.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES viii
LIST OF FIGURES ix
CHAPTER 1
Introduction 1
Research Problem 3 .
Theoretical Framework 3
Conceptual Definitions of Variables 6
CHAPTER 2
Literature Review 10
Critical Discussion 18
CHAPTER 3
l IAAL- A- 20
Research Design 20
Sample 21
Data Collection and Recording 2]
Ethics 22
Operational Definition of Variables 2?
Data Analysis Procedures 25
CHAPTER 4
Results 27
T‘ g r‘ ' 27
Research Questions 11
Relationship of the Results to the Review of the I " ‘ 39
CHAPTER 5
I" 41
Limitations 41
Implications for Avanced Practice Nurse ........ 45
Implications for Future Research 49
Summary 50

APPENDICES
A r I A 51
B Study Approval I fitters SS

 

 

LIST OF REFERENCES 58

vii

 

LIST OF TABLES

 

 

 

 

 

 

 

 

 

 

 

 

 

Table l - I‘ g r” "‘ ' ‘ Information
About Sample Respondents 28
Table 2 - T‘ g r" "‘ ' ' Information
About Sarnp e Respondents 29
Table 3 - Frequencies for Individual Risk Factnrs ‘40
Table 4 - Pearson’s Correlations and Significance Testing of the Relationships
Between Individual Risk factors for CHI) 32
Table 5 - Logistic Regression:
Socioeconomic Predictors of High (“L ' ‘ ‘ 33
Table 6 - Logistic Regression:
Socioeconomic Predictors of Overweight ‘44
Table 7 - Logistic Regression:
Socioeconomic Predictors of Physical Inacti ': 35
Table 8 - Logistic Regression:
Socioeconomic Predictors of Diabetes Mellitus 35
Table 9 - Logistic Regression:
Socioeconomic Predictors of Hypertension 36
Table 10- Logistic Regression:
Socioeconomic Predictors of Cigarette Smoking 38
Table l l - Logistic Regression:
Socioeconomic Predictors of Depressed Mnnd 39
Table 12 - Logistic regression:
Socioeconomic Predictors of High Risk for CHD 40

 

viii

 

LIST OF FIGURES

Figure l - The “web of causation” for coronary heart disease,
demonstrating the relationship between socioeconomic
status and individual risk factors

 

Figure 2 - The “web of causation” for coronary heart disease for
women in Northern Michigan, demonstrating the relationships
between socioeconomic status and individual risk factors as
well as the relationship between individual risk factors ................................. 49

 

 

Chapter 1

Many women believe cancer is a bigger threat to their health than heart disease
(American Heart Association, 1996). However, coronary heart disease (CI-1D) is the
leading cause of death in women in the United States. Annually, 250,000 women die
from CHD, while breast cancer claims 43,100 each year, lung cancer 55,900 each year,
and all forms of cancer 246,000 each year (American Heart Association). Not
surprisingly, coronary heart disease in women has been labeled a “silent epidemic” by
the American Heart Association (1992). Beyond mortality, the disability related to CHD
affects the quality of life and independence of women in their later years.

The risk factors that predict which women are likely to develop CHD can be
categorized into those that can be modified and those which cannot. The latter category
includes advanced age and race (American Heart Association, 1996). With increasing
age, women have a greater chance of developing CHD; but there is a ten year lag in the
onset of CHD in women as compared to men. Thus, while being female is a protective
factor until menopause, this benefit diminishes with menopause as well as with the
comorbidity of diabetes (Haan, 1996). Race also influences risk of CHD. Black women
have a greater risk of CHD than white women, and Native American men and women
under 35 years of age have a heart disease death rate twice as high as all other ethnic

groups, though this risk for mortality increases less after age 44 (Harris-Hooker &

 

Sanford, 1994).

Individually based risk factors for CHI) that are, at least in principle, amenable to
modification include having high cholesterol, being overweight and physically inactive,
having diabetes mellitus and hypertension, being a cigarette smoker, and being
depressed (J udelson, 1994). Given that these factors are amenable to modification, they
need to be a focus of the Advanced Practice Nurse (APN) primary care provider in the
prevention of CHI). Prevention of CHI) can be approached at three levels: (a) primary
prevention to reduce exposure to risks of CHI); (b) secondary prevention to identify early
clinical presence of CHI) and minimize its progression; and (c) tertiary prevention to
reduce morbidity and mortality secondary to CHI). As the individual risk factors have a
cumulative effect and precede the clinical presence of CHI) by many years, the
preponderance of prevention needs to be at the primary level. To further affect primary
prevention, the APN will need to appreciate the impact of socioeconomic status (SES) on
the individual risk factors for CHI).

SES has long been recognized for its effect on health. In general, lower SES has
been linked an increased risk of disease and poorer health outcomes. However, in the
case of CHI), the relationship with regard to SES has been changing. Early studies of
CHI) conducted on men in the 1930's demonstrated that higher SES tended to increase
the likelihood of developing CHI) in developed countries; in more recent studies this
tendency has reversed, with lower SES men having higher rates of CHI) (Luepker et al.,
1993). For women, excluded from the early studies, there has always been this inverse
relationship between CHI) and SES (Kaplan & Keil, 1993). Potentially this reflects

healthier lifestyles as a result of information based on studies of men.

3

Thus for APNs, the identification of individuals and populations at high risk for
CHI) is crucial. With an increased understanding of the variation in individual risk
factors for CHD in women based on their SES, interventions, particularly community-
based interventions, should be enhanced for primary prevention.

Research Problem

In this study, the researcher examines the impact of socioeconomic status on the
individual risk factors for coronary heart disease in women. This research evaluates two
research questions: (a) what is the extent of the correlation between the individual risk
factors for coronary heart disease in women, and (b) what is the impact of
socioeconomic status on women’s individual risk factors for coronary heart disease.

Theoretical Framework

The “web of causation,” an epidemiological model, provides the framework for
this study which examines the relationship of SES to individual risk factors for CHI).
The utility of the epidemiological model to this study is twofold first, it identifies the
risk factors for deviation from health, and, second, through the resultant knowledge it
offers support for targeted interventions for primary, secondary, and tertiary prevention
of CHD. The epidemiological characteristics of a disease include its natural history, the
patterns of occurrence, and the risk factors associated with developing the disease. The
“web of causation” is a diagrammatic approach which depicts the occurance of disease,
and reflects the interrelationship between multiple factors which result in pathogenesis.
These factors include: (a) a susceptible host with intrinsic, physical, psychological, and
immunity factors; (b) a conducive environment whether social, physical, or biological;

and, (c) causative agents which can be either physical, chemical, nutrient, biological,

4

genetic, or psychological. Additionally, it has been found in diseases which are
multifactorial, such as CHI), that synergism, the combined effects of two or more agents,
increases the likelihood of onset of the disease (Shortridge & Valanis, 1992). Thus, this
model fosters a view of disease as a complex interrelationship with multiple risk factors
and not the result of isolated individual factors (Anderson & McFarlane, 1988).

The “web of causation” was first proposed by MacMahon (MacMahon & Pugh,
1970) in response to the emphasis on the oversirnplication of epidemiology reflected by
the “chain of causation”. It was his belief that the “chain” failed to take into account the
complex genealogy of disease. In this view, the metaphor of a spider’s web captured the
complex causal pathways for disease. With an epidemiological emphasis on the direct,
individually-based and biologically plausible risk factors of CHI), the utilization of the
“web of causation” recognizes conditions which precede these individual risks factors.
Thus, by contextualizing these risk factors, the “web of causation” framework can foster
an understanding of what factors predispose women to adopt lifestyles which contribute
to the development of CHI). The diagrammatic depiction of these relationships, adapted
from MacMahon and Pugh, and shown in Figure 1, provides a visual representation of the
potential relationships of the risk factors for CHI) within the “web of causation”
framework and is based on the relationships found in the review of the literature. In this
diagram, SES, the independent variable, has non-modifiable and modifiable risk factors.
Non-modifiable risk factors are gender, age, and race. Potentially modifiable risk factors
are educational achievement, employment status, marital status and household income.
The “web of causation” then maps the interplay between the modifiable SES risk factors

and the more direct causes of CHI), the modifiable individual risk factors. The individual

flog an: 333%.:
v5 83% caeeoooemoem e853 @2232“: 05 wedge—Hoe .§% :3: .0588 L8 25.5858 me 903.. RE. - _ 03!"—

 

_ gigeaflmifio L

 

 

 

 

 

 

 

QED 8.. gm

 

 

 

 

 

 

 

 

 

risk factors are the dependent variable.

Usually, social epidemiologists and medical sociologists contextualize risk
factors for CHI) by trying to understand what it is about social conditions that influence
exposure to risk factors. Using this perspective, Link and Phelan (1995) propose
that as new risk factors become apparent, people of higher SES are better situated to
know about the risks and to have the resources which allow them to engage in efforts to
avoid them. Additionally, SES influences access to resources which are used to avoid
risk and minimize consequences of disease once it occurs. Resources associated with
higher SES include money, knowledge, power, prestige, and interpersonal resources,
including social support and a social network (Link, 1996). Other, non-modifiable,
variables to be considered are age, race, and gender as they contribute to both SES (Link
& Phelan) and risk of CHI) (American Heart Association, 1996). Thus, through the “web
of causation” framework, SES will be explored as an antecedent to the more immediate
individual risk factors of CHI) in women.

Conceptual Definitions of Variables
Wm

CHI) is the reduction of oxygen and nutrients to myocardial tissue due to a
diminished coronary blood flow. CHI) is the result of coronary atherosclerosis which is
characterimd by an abnormal accumulation of lipid substances and fibrous tissue in the
vessel wall (Smeltzer & Bare, 1996).

W
Socioeconomic status refers to both the social and economic standing of an

individual and can be further categorized by whether the measure has the potential to be

7
altered or changed. As such socioeconomic status will be treated as either an non-
modifiable or modifiable risk for CHI).
11- I'EHS' ‘E'IE E3111:

The SES risk factors for CHI) which cannot be altered or changed are the
unmodifiable risk factors for CHI). These variables include gender, age, and race.
lll'filli' 'E'IE EZHE

The SES risk factors for CH1) which can be altered or changed, at least
theoretically, are the modifiable risk factors for CHI). These variables, which are
commonly used measures to evaluate SES in the epidemiological literature, include
education, income, employment status and marital status (Kaplan & Keil, 1993).

Education, Education refers to the number of years of schooling a person has
attained.

W Employment status refers to “the state of being engaged in
services for hire,” as defined by Funk and Wagnalls (Landau, 1993, p. 208).

Maxim Marital status pertains to the state of being married, which is
defined in thk and Wagnall as “a legal contract, entered into by a man and a women, to
live together as husband and wife” (Landau, 1993, p. 397).

MW Household income refers to all taxable income of all
household members in one calendar year as filed on a federal income tax return.

II Hill I l. 'l IE'IE

Additional modifiable risk factors that can be altered or changed are the

individual risk factors for CHI). The individual risk factors considered here are high

cholesterol, overweight, physical inactivity, diabetes mellitus, hypertension, cigarette

smoking, and depression.
2. S l .

The exposure to tobacco smoke through the act of smoking cigarettes.
Hymnensinn

A higher than normal blood pressure primarily due to an increase in peripheral
resistance resulting from vasoconstriction or a narrowing of peripheral blood vessels
(Thomas, 1997).
Wasted

An excessive amount of lipoproteins in the blood which infers an increased risk
of atherosclerotic plaques in the arterial lumen (American Heart Association, 1997).
I; . l l I H.

A chronic disorder of carbohydrate metabolism, marked by hyperglycemia and
glycosuria and resulting from inadequate production or use of insulin (Thomas, 1997).
: . |

According to Healthy People 2000 (U. S. Department of Health and Human
Services, 1990) overweight has been defined, for women, as a body mass index of 27.3
kilograms/meters squared or greater.
El . l I . .

Healthy People 2000 (U. S. Department of Health and Human Services, 1990)
defined physical inactivity as less than three days per week with 30 minutes of moderate

exercise.

mm

The DSM-IV (American Psychiatric Association, 1994) describes “major
depressive disorder” as “loss of interest or pleasure in nearly all activities with at least
four additional symptoms which include changes in appetite or weight, sleep, and
psychomotor activity; decreased energy; feelings of worthlessness or guilt; difficulty

thinking, concentrating, or making decisions; or recurrent thought of death or suicide.”

Chapter 2

Literature Review

Socioeconomic status has long been identified as a strong determinant of health.
Its effect is so strong that it is routinely treated as an independent variable or a control
variable in health research. Furthermore, Adler et al (1994) found an inverse, graded
relationship between health and all levels of SES fi'om lowest to highest. Thus, at every
subsequent higher level on the SES hierarchy, there is an improved health status. These
authors propose three possible explanations for this relationship. First, it may be a
spurious relationship that is, in fact, rooted in genetically based factors. However, the
evidence does not support a biologically driven relationship between SES and health. A
second explanation is called the “drifi hypothesis” which proposes that poor health
results in a lower SES. Though there are examples of this association, such as the
individual with schizophrenia whose SES has a declining trajectory as the disease
evolves, overall this phenomenon has a limited role in explaining the SES and health
relationship. The third explanation is that SES has an effect on the causative agents of
disease, which then determines health status. The latter explanation supports application
of the “web of causation” framework for explaining the interplay between SES, health,

the physical and social environment a person is exposed to, and the health behaviors

10

which are practiced

Numerous studies from the United States, Canada and Europe have linked
multiple measures of SES with individual risk factors for CHI). One of the early and
fiequently referenced studies of this relationship is the Whitehall Study by Rose and
Marmot (1981). They studied 17,530 male, civil servants in London between ages 40-64
for symptoms, signs, and risk factors related to cardiovascular disease. When men in the
lowest employment grade were compared with those in the highest grade, men in the
lower grades had more CHI) as well as CHI) risk factors. They smoked more, exercised
less, were shorter and more overweight, had higher blood pressures, and had lower
levels of glucose tolerance.

Additional studies done in the United States support the finding of the Whitehall
study. In the Hypertension and Detection Follow-up Program (1977), 158,906 black and
white adults age 30-69 in 14 commrmities in the United States were evaluated on blood
pressure, weight, and socioeconomic status. This study found education to be inversely
related to hypertension for each race and sex group. When overweight was taken into
account, the education effect was diminished. Thus, the effect of education operates in
part through such factors as weight and dietary choices. Another potential confounder in
this study is the variability in quality of education from public to private, fi'om urban to
suburban or rural, and from segregated to desegregated. The Minnesota Heart Survey
(Luepker et al., 1993) showed that overall CHI) risk was strongly related to education in
women, with predicted 10-year CHI) mortality highest in the least-educated and least-
affluent groups. When adjustments were made for effects of education on income and

income on education, the strongest association was between education and CHI) risks.

1 2
Focusing on risk factors for CHD, this study found: (a) strong inverse relationships for
both education and income with smoking and hypertension, (b) an inverse relationship
between total cholesterol and education but no significant association with income level,
and (c) an inverse relationship between education as well as income and women’s body
mass index.

In the Stanford Five-City Project, Winkleby, Jatulis, Frank, and Fortmann (1992)
examined the independent contributions of education, income, and occupation to
cardiovascular disease and its risk factors of cigarette smoking, systolic and diastolic
blood pressure, total cholesterol and high-density lipoprotein cholesterol. They found that
the relationship between the SES measures and these risk factors was strongest and most
consistent for education, showing higher risk associated with lower levels of education.
Similarly, in a rural New York cross-sectional study of 1,063 persons over age 16, which
included 541 women, those women with the least education reported the most
atherogenic risk factor profile with elevated blood pressure, total cholesterol, body mass
index, and low, high density lipids. The least educated women smoked more, got little
physical exercise, and were more angry and depressed (Gold & Franks, 1990).

Women of higher SES were found by Ford et al (1991), in an urban community
sample, to spend 26% more time than lower SES counterparts in total daily physical
activity which included leisure time, job-related, and household physical activity.
Matthews, Kelsey, Meilahn, Kuller, and Wing (1989), found that middle-aged women
who had an advanced degree expended 56% more energy per week in nonoccupational
physical activities than women with a high school education or less. Reasons for these

differences in physical activity among groups must be examined Are the differences due

13
to variation in available leisure time, discretionary income, social support, facilities, or
psychological variables? Without understanding the cause of less physical activity, it will
be difficult to provide appropriate interventions.

In Allegheny County, Pennsylvania, Matthews et al. (1989), conducted a
community study of 541 healthy, middle-aged women between the ages of 42-50. To be
eligible, women had to be premenopausal and not on medication which would influence
CHI) risk factors. The researchers found that lower educational levels were significantly
associated with an increase in glucose intolerance and higher serum insan levels. Of the
ineligible women, the number taking insulin medication was significantly higher for
those women with less education. For those with a high school education or less, 3.9%
used insulin medication as compared to 2.8% with some college, none with college
degrees, and 1.3% with an advanced degree. These findings are corroborated in data from
the National Health Interview Study (Adams & Benson, 1989), which found lower family
incomes were associated with increased rates of diabetes.

A prospective, epidemiologic, Canadian study found the rate for major depression
was 1.9%, 4.5%, and 12.4% respectively in high, average, and low SES groups (Murphy
et al, 1991). These findings were substantiated by Kaplan, Roberts, Camacho, and Coyne
(1987) who found higher rates of depressive symptoms among those with lower income
and education, as well as Matthews et al. (1989) who found those with the least
education scored highest on the Beck Depression Inventory. These studies are
contradicted by the DSM-IV (American Psychiatric Association, 1994) which states that
“the prevalence rates for Major Depressive Disorder appear to be unrelated to ethnicity,

education, income, or marital status.”

14

The Saskatchewan Heart Health Survey, a cross-sectional study conducted in
Canada, found that women with the least education, lowest household income, having a
nonskilled work classification, and who were unemployed, had the highest prevalence of
cardiovascular disease (Reeder, Lui, & Horlick, 1996). Additionally, this study compared
urban to rural environment for incidence of cardiovascular disease. It found no
significant difference with angina, but women in a rural environment were at greater risk
for possible infarction. This difference between urban and rural women could also be a
SES effect, a function of education, income, or employment which provides access to
resources and health care.

Further studies conducted in Europe also support the inverse relationship between
SES and CHI) and risk factors for CHD. A prospective study of middle-aged, married,
Swedish women found that those with low education had a significantly increased
incidence of angina (Lapidus & Bengtsson, 1986). The cross-sectional Tromso Heart
Study found higher educated women were less overweight, smoked less, were more
physically active and had less atherogenic food habits than the least educated ones.
Fmthermore, these women had a semrn total cholesterol and systolic blood pressure
which were negatively associated with educational level (Jacobsen & Thelle, 1986). A
Finnish study by Luoto, Pekkanen, Uutela, and Tuomilehto (1994) of cardiovascular risk
factors and socioeconomic status found that lower levels of education, occupation, and
income were all significantly associated with an unfavorable risk factor profile in
women.

Two additional European studies discussed the correlation of multiple risk factors

for CHI) to lower SES in their reports. The Netherlands Monitoring Project on

15

Cardiovascular Risk Factors (I-loeymans, Smit, Verkleij, & Kromhout, 1996) found that
concurrent risk factors were more prevalent in lower educated groups than in higher
educated groups. Similarly, Connolly and Kesson (1996) found a clustering of
cardiovascular risk factors for diabetics from lowest SES in Glasgow, Scotland. The
proportion of patients with no cardiac risk factors fell by 30.6% fi'om the highest SES to
the lowest SES categories, and the proportion of patients with three or more risk factors
rose fi'om 8.6% in the highest SES category to 20.2% in the lowest SES category.
Specifically, of the highest SES, 30% had a body mass index greater than
30 kg/metered squared compared with 47% in the lowest SES categories; and with regard
to smoking, 13% of the highest SES smoked, compared with 33% in the lowest SES.

The effects of the extensive cardiovascular health promotion campaigns,
implemented in the United States in the 19805 to improve knowledge on how to reduce
the risks of cardiovascular disease, demonstrate the relevance of SES. Davis, Winkleby,
and Farquhar (1995) analyzed the changes from 1980 to 1990 in knowledge of acquired
cardiovascular risk factors, knowledge of risk-reduction strategies, and interest in risk
modification by socioeconomic status using level of education as a measure of SES.
Residents of two northern California cities were studied, and participants demonstrated a
significant baseline difference in knowledge based on educational level that widened
over the 10-year study period. From 1980 to 1990, individuals with less than 12 years of
education showed only slight improvement in their knowledge of cardiovascular risk
factors whereas those with 16 years or more of education demonstrated twice the
improvement in knowledge. Similar differences were found in knowledge of

risk-reduction strategies. These findings were contrasted by a high interest in risk

16
modification at all educational levels that remained uniform across time. This indicates a
need to investigate appropriate educational interventions for populations with less than a
high school education.

Employment status was considered in two studies of women and CHI). The cross-
sectional San Antonio Heart Study found that employed women had significantly higher
levels of high density lipids, lower triglycerides, and ate a less atherogenic diet than full-
time housewives (Hazuda et al., 1986). These differences were not explained by age,
SES, or behavioral mediators (i.e., exercise, caloric imbalance, cigarette smoking,
alcohol consumption, or estrogen use). Excluded from this study were full-time students,
retired, disabled, or not working but looking for work in the last four weeks. As the
sample was selected from three sociocultrn'ally distinct neighborhoods including
Mexican-American and non-Hispanic white women, it provides important information on
the impact of employment on Hispanic women but is limited in its generalizability.
Another limitation is that education was not controlled in the analysis, though it may be
teased out in the SES measure based on their, or their spouse’s, occupational prestige.

The Rancho Bernardo study of 242 women, aged 40-59 years, found employed
women were less likely to smoke cigarettes, drank less alcohol and exercised more than
unemployed women, and, after adjustment for possible confounders (i.e., age, BM],
estrogen use, alcohol consumption, cigarette smoking, exercise, marital status, education,
and social class) employed women had significantly lower total cholesterol and fasting
plasma glucose levels than housewives (Kritz—Liverstein, Wingard, Barrett-Connor,
1992). This study is limited by lack of clarification on the variable “unemployment”. A

woman who is “not employed” implies this status was voluntary whereas “unemployed”

17

implies this status was involuntary; therefore two very different profiles of
“unemployment” may be presented under one variable. A strength of both employment
studies is that an analysis was conducted to assure a self-selection bias known as “healthy
worker effect” had not occurred

Marital status has been infrequently discussed in the studies of SES and CHI) in
women. Hu and Goldman (1990), investigating the relationship between mortality and
marital status in 16 developed countries, found that single women have higher death rates
than married women, and in the United States, widowed and divorced women have a
slightly higher risk of dying than single women. The results of this study may be
confounded by a selection bias where the healthier women are more apt to be married A
further limitation is the lack of inclusion of other SES indicators such as income,
occupation, and education which may also impact mortality rates. In another study which
considered marital status, Luoto et al (1994) discovered that in Finland marital status was
not significantly associated with a cardiovascular risk score. The risk factor score
included cholesterol, blood pressure, body mass index, cigarette smoking and leisure
time activity. In this study the other SES indicators utilized were education, family
income, and occupation. Further, women were classified according to their current or
previous occupation if they were or ever had been working (most women, 91%) thus
limiting employment comparisons with other studies which are generally based on
current employment status. The contradiction in the impact of marital status on health
between these studies indicates a need to evaluate more carefully the effect of marital
status especially in light of sociological changes in the social and economic value of

marriage as women are increasingly working outside of the home, providing them with

18
additional social roles and economic independence.
Critical Discussion

The literature, as reviewed, only partially addresses the major issues in this study
in that some of the original research on SES and risk factors for CHI) was conducted
primarily on males. Though more recent research has included women, these works can
be augmented by additional population studies. Strength is shown in the geographic
breadth of the literature in North America and Northern Europe, though it is limited in
the impact of ethnic and cultural influence, providing mixed generalizability within the
cultural diversity found in the United States. Furthermore, the comparative impact of
SES on individual risk factors for CHI) when evaluated within any nation’s health care
delivery system and economic paradigm, remain unexplored. This could be especially
important as the SES disthy between the lowest and highest levels increases in the
United States.

Another limitation in the research on risk factors related to lifestyle is the
possible bias in such data due to under-reporting related to the stigmatization of the
behavior or its outcome. This under reporting may be further confounded by SES.
Women of higher SES, who are found to be better informed about risk factors for CHI),
may be more inclined to report their risk factors conservatively. Thus, the self-report data
on smoking, weight, and physical inactivity must be viewed as conservative estimates of
the actual behavior or its outcome, especially among higher SES strata.

The preponderance of research on the relationship of SES to CHI) risk factors is
focused on four measures of SES - education, employment, occupation, and income. Of

these, education had the strongest relationship to all the individual risk factors for CHI)

19

in women. Marital status, which demonstrated an impact on mortality but failed to
impact individual risk factors, needs to be evaluated for an interaction efl‘ect; the effect
on individual risk factors may present differently when marital status is combined with
age, education, employment, and income for analysis.

Though the literature is replete in identifying the negative impact of lower SES
on individual risk factors for CHD it is inadequate in offering interventions for primary,
secondary, and tertiary interventions which go beyond the focus of the simplistic “chain

of causation” to the multifactorial “web of causation”.

Chapter 3

Methods

Wan

This study represents a non-experimental, cross-sectional, correlational design.
The goal of the study is to develop a CHI) risk profile for women with different SES in
Northern Michigan in order to facilitate the planning of preventive interventions; thus,
manipulation would be inappropriate. The correlational design is necessary since the
research has been conducted after the variation in the independent variable has occurred
without manipulation, in the natural setting. Thus, a causal relationship cannot be
definitely established using a correlational design. Rather, an interrelationship may be
identified between two variables, meaning, as one variable varies there is a tendency for
the other variable to vary.

The data is cross-sectional as it was collected at one point in time throughout the
21 counties of the Northern Lower Peninsula of Michigan. Polit and Hungler (1995)
propose that cross-sectional designs are valued for their practicality. Cross-sectional
designs are also relatively economical and easy to manage. They are well suited for
describing the status of a phenomenon or therelationship among phenomena at a fixed

point in time. However, the disadvantage of the cross-sectional design is the limited

20

2 1
ability to make causal inferences.
Sample

The Northern Michigan Community Health Assessment survey involved a
stratified random sample. The sampling design used was Survey Sampling’s equal
probability of selection method (EPSEM) (Information Transfer Systems, Inc., 1995).
This design allowed for equal probability of selection within each county, but not across
counties, and thus resulted in the completion of 300 interviews within each of the 21
counties, for a total of 6,300 interviews. The total number of interviews completed with
women was 3,746.

The sample frame was all residential telephone numbers ringing into households
in the 21 counties of interest that fell into known banks of working residential numbers.
Within each sample household, the adult respondent was randomly chosen from all
resident adults 18 years of age or older. Exclusion criteria included households occupied
by short-term vacationers (staying fewer than 3 weeks per year at the household reached).

When a residential number was successfully contacted, the household members
18 years of age or older were listed: men were listed from oldest to youngest, and women
were then listed in the same way. A random number was then generated based upon the
total number of adults living in the household, and this number was used to choose one
adult from the list to be interviewed This in effect created a second layer of stratification
within the sample.

B C H . l E I'
Data was collected from mid-July, 1995, through the beginning of September,

1995. Information Transfer Systems, Inc. of Ann Arbor, Michigan, conducted this survey

22
for the Northern Michigan Community Health Assessment Survey. Data was collected
using computerized telephone survey software and data entry. The portion of the survey
used in this study is included as Appendix A.
W

No individual identifying characteristics were included in the data coding
process. Approval for secondary data analysis for this study by this researcher was
obtained from Michigan State University’s University Committee on Research Involving
Human Subjects and from Northern Michigan Hospital-Bums Clinic Foundation before
obtaining the data set and conducting the research (see Appendix B).

3 . l I E . . E1! . l l

The operational definition of the variables in this study reflect the survey
instrument used to gather the data for the Northern Michigan Community Health
Assessment Survey and the recoding of the original data in preparation for data analysis.
1 1 I'fi ! l E . l E

The non-modifiable socioeconomic risk factors were operationalized as follows:

1. Sex was self-reported as male or female. Only female participants were
included.

2. Age was self-reported as age in years.

3. Race was self-defined and coded as: 1- White, 2 - Black or African American,
3 - Asian, 4 - Native American Indian, and 5 - Other. Because of the lopsided
distribution, race has been recoded into: 0 - Black or African American, Asian, Native

American Indian, or Other, and 1- White.

23
lll'fillS' 'E'IE

The modifiable socioeconomic risk factors were operationalized as follows:

1. Education was coded in two ways. For a frequency count, education was
coded: 1- sixth grade or less, 2 - eleventh grade or less, 3 - high school graduate or GED,
4 - some college, 5 - college graduate, 6 - some graduate school, masters, or doctorate.
For the regression analysis, educational achievement was coded as number of years of
formal education.

2. Employment status for frequency was coded: l - Employed for wages, 2 - Self-
employed, 3 - Out of work for more than one year, 4 - Out of work for less than one year,
5 - Homemaker, 6 - Student, 7 - Retired, 8 - Unable to work. For the multivariate
analysis, employment status was coded: O - not employed, not in the labor market, I -
employed for wages, self-employed.

3. Marital status was coded: 0 - divorced, widowed, separated, never married, or
member of an unmarried couple, I - married.

4. Household income was the stated annual household income or 999,997 if
greater than $999,999 from all sources before taxes in 1994.

The individual risk factors were operationalized as follows:

1. Cigarette smoking was defined as smokes cigarettes now or has smoked daily
within the last three years and is coded: 0- nonsmoker or has not smoked within the last
three years or I- smoker.

2. Hypertension was defined as having been told by a doctor, nurse, or other

health professional within the past two years that they have high blood pressure and was

24
coded: 0 - without hypertension, or 1- has hypertension.

3. High cholesterol was defined as having been told by a doctor, nurse or other
health professional within the past two years that their blood cholesterol is too high and
was coded: 0 - without high cholesterol, or 1 - high cholesterol.

4. Diabetes mellitus was defined as having been told by a doctor, nurse or other
health professional within the past three years that they have diabetes, not including
gestational diabetes, and was coded: 0 without diabetes, or 1 - diabetes.

5. Overweight, which was calculated fi'om self-reported current weight and height
and coded as: 0 - body mass index less than 27.3 kilograms/meters squared or 1 - body
mass index greater than or equal to 27 .3 kilograms/meters squared.

6. Physical inactivity was coded: 0 - gets at least 30 minutes of moderate exercise
three or more days per week, or 1- gets 30 minutes of exercise less than three days per
week

7. Depressed mood was coded: 0 - “not depressed” if a respondent answers “some
of the time or little of the time” to has been downhearted and blue and “all of the time,
most of the time, a good bit of the time, some of the time” to has been a happy person in
the past four weeks, or I- “depressed” if a respondent answers “all of the time, most of
the time, or a good bit of the time” to has been downhearted and blue and “a little of the
time or none of the time” has been a happy person in the past four weeks.

8. Risk factor profile a simple count of all identified individual risk factors (high
cholesterol, overweight, physical inactivity, diabetes mellitus, hypertension, cigarette
smoker, or depressed mood) resulting in a possible score of 0 to 7.

9. High risk for CHD based on the risk factor profile, a dichotomous variable

25
constructed by dividing the sample into high risk and low risk groups as determined by
the risk factor profile. High risk for CHD was coded: 0 - none or one individual risk
factor for CHD, or 1 - two or more individual risk factors for CHD.
W

Statistical data analysis was performed using the statistical analysis program
SPSS for Windows 7.0 (Norusis, 1996). To evaluate the first research question
concerning the extent of the correlation between the individual risk factors for coronary
heart disease in women the Pearson’s r was utilized. Pearson’s r is typically used for
interval level data or higher; however, in the case of dichotomous variables, its use is
also acceptable as it is mathematically identical with phi or the point bisectional
correlation coefficients.

To answer the second research question regarding the impact of SES on the
individual risk factors, multiple logistic regression models were employed. The
multivariate analysis teases out the effect of the SES variables on individual risk factors
and the overall risk for CHD. This offers a fuller explanation of the variation in the
occurrence of each individual risk factor (the outcome variable) and allows for the
examination of independent effects of several simultaneous independent variables. The
independent variables to be entered into the logistic regression equation are the non-
modifiable SES variables of gender, age, and race as well as the modifiable SES
variables of education, employment, marital status and income with each dependent
variable. The dependent variables are the individual risk factors for CHD ( high
cholesterol, overweight, physical inactivity, diabetes mellitus, hypertension, cigarette

smoker, and depressed mood) and the overall risk for CHD. The logistic regression

26
model allows an examination of the extent to which each SES variable acts and interacts

through the individual risk factors to affect risk of CHD.

Chapter 4

Results

Demographics

The 21 counties surveyed represent a rural area with the largest city being
Traverse City, with a population of 15,040, C. Schlueter of Northwest Michigan Council
of Governments (personal communication, February 20, 1998). Unique to this rural area
is the influence of a resort economy based on summer and winter recreational activities,
as well as the characteristics of the residents from resort communities where there can be
a significant population of retired and higher SES individuals, especially during the
summer months when the survey was conducted.
W

The sample for this study consisted of 3,746 women age 18-97; the mean age
was 51 years. The race distribution of the sample was 97.9 % white, 0.1% black, 0.3%
Asian, and 1.2% American Indian. The demographics for the SES variables included

education, employment status, marital status, and annual household income are as

I follows. The highest educational achievement of the sample respondents was: 1.0% had a
grade school education, 12.8% some high school, 42.9% a high school education, 26.4%

had some college, 10.4% a bachelors degree, and 6.6% a graduate education.

27

28

Employment status found 40.4% employed for wages, 7.1% self-employed, 2.9% out of
work, 16.6% homemakers, 1.6% students, 26.7% retired, and 4.6% unable to work.
When this variable was receded to “employed” and “not employed”, these were 47.5%
and 52.5% respectively. Of the not employed, 50.94% were retired. Marital status
showed 60.4% were married and 39.4% were unmarried. Unmarried were, not
surprisingly, heavily concentrated at either end of the age spectrum with those under age
24 and those over age 72 least likely to be married. The median annual household
income was $26,000. The reported range of income was none to greater than $999,997.

The above demographics are shown in Tables 1 and 2.

Table l - Demographics/ Socioeconomic Information About Sample Respondents

 

 

 

 

 

 

 

 

 

 

 

 

 

in" 'i..’;»’..1 1. . T. ', .. ;.1
Age l8-29years 12.0
N = 3746
30-39years 18.3
40-49 years 17.8
50-59years 15.7
60-69years 16.6
70.79years 13.7
80-89 years 4.9
90-99years 1.0
Race White 97.9
N - 3658
Black 0.1
Asian 0.3
American 1.2
Indian

 

 

 

 

 

29

Table 2 - Demographic/Socioeconomic Information About Sample Respondents

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E W . ff We.
Education Grade School 1.0
N = 3741

Some High School 12.7
High School 42.9
Some College 26.4
Bachelors Degree 10.4
Graduate Education 6.6
Employment Employed for Wages 40.4
Status
N g 3743 Self-Employed 7.1
Homanaker 16.6
Student 1.6
Retired 26.7
Out of work 2.9
Unable to work 4.6
Marital Married 60.4
Status
N = 3740 Unmarried 39.4
Annual 5 $9,999 9.8
Household
[w $10,000 - $19,999 [5.8
N = 2829
$20,000 - $29,999 14.4
$30,000 - $39,999 11.5
$40,000 - $49,999 8.0
$50,000 - $69,999 9.7
$70,000 - $89,999 3.3
2 $90,000 3.0

The frequencies for the individual risk factors are based on the valid data for each
risk factor and which include having high cholesterol, being overweight and physically

inactive, having diabetes mellitus and hypertension, being a cigarette smoker, and having

30

a depressed mood. High cholesterol was reported in 39.1% of the respondents with a
valid response; however, there was a large amount of missing data as 39.4% of the
sample did not have a valid response to the cholesterol question. The BMI ranged from
12.2 - 76.8 kg/meters squared for the sample, with the mean BMI being 25.99 kg/meters
squared; 33.6% of the sample were overweight with a BMI of 27.3 kg/meters squared or
greater. Physical inactivity was found in 17.8% of the sample who got less than 30
minutes of exercise a day fewer than 3 days of the week. Diabetes mellitus was found in
7.7% of the respondents, while 92.3% had not been told they had diabetes. Those who
had been told they had diabetes may be too few as the standards for checking glucose and
the serum glucose level for the diagnosis of diabetes have become more conservative
since this data was collected Hypertension was reported by 32.1%. Of the respondents
56.9% were current cigarette smokers or had smoked in the last three years while non-
smokers accounted for 43.1%; however, 50.6% of the sample did not respond to this
question. A depressed mood was reported by 5.6% of the sample. The frequencies for the

individual risk factors are shown in Table 3.

Table 3 - Frequencies for Individual Risk Factors

 

 

 

 

 

 

 

 

High Cholesterol 2271 39.1 60.9
Overweight 3650 33.6 66.4
Physically Inactive 3700 17.8 82.2
Diabetic 3352 7.7 92.3
Hypertensive 3448 32.1 67.9
Smoker 1852 56.9 43.1
anessedMood 3725 5.6 94.4

 

 

 

 

 

 

31
Wells

Due to the large amount of missing data on high cholesterol and cigarette
smoking, a subset of selected cases was used to answer the research questions. Of the
3,746 respondents only 996 responded to all individual risk factors. Thus a subset of 996
cases with no missing data on any one of the seven individual risk factors for CHI) was
created.

In answering the first research question, “what is the extent of the correlation
between the individual risk factors for coronary heart disease in women,” Pearson’s
correlations were computed on the subset of 996 women for whom data was available on
all risk factors. There were significant (p s, 0.01), though weak, correlations among
several individual risk factors (see Table 4). The strongest correlations were between
overweight and hypertension (r =. 172), overweight and diabetes mellitus (r =.196), and
hypertension and high cholesterol (r =.183). The greatest number of positive and
significant (p ;<_ 0.01) correlations were found with being overweight and having
hypertension. Being overweight was correlated with having high cholesterol (r = .100 ),
being physically inactive (r = .108 ), having diabetes mellitus (r = .196), having
hypertension (r = .172) and being depressed (r’ = .092). Having hypertension was
correlated with having high cholesterol (r = .183), being overweight (r = .172), having
diabetes mellitus (r = .131), and being depressed (r = .097). Still, overall, these
correlations were weak and suggest that these individual risk factors occur quite
independently from each other. At least within this population, they do not form strong

clusters.

32

§~.e§8.o§a§.&aa8£.68u .
§.~..§.o.¢fia§.&aa8§o:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

681...

.1.. 86.1.. «8.1.. 3.1.. «8.1.. 3... 1.. 26.1.. .82

8...... :3. 1. :8... 1. R... 1. as. 1. :3... 1. .68. 1. 3.2.3

.82 .1.. 8... 1.. 86.1.. 35.1.. .8. 1.. .81.. been

688.89 831. :3... 1. 2.... 1. .8. 1 c E... 1. :8...1. 68a»...
38... .1.. 86.1.. 84.1.. 68.1.. 86.1..

6386 831 :5. 1. .81 . :«2. 1 . $2.1. 83.6.5.

.1.. «8.1.. 8... 1.. .81.. 33.63.

8.55.5. 89.1. :8... 1 . :2... 1 . .68. 1. 88......

3...... .1.. .8. 1.. 68.1.. 3.8....

88...... 89.1. :8.. 1 . .68. 1. $8.2...
9.9a. . 1.. .81..

2.8.8... 8....1 a :8.. 1. .3665

.1.. 30.8.30

£865 831. .3...

3.5.3.5

398.30 8.8.2.8

.9: 828..

9.0 8..

e38... .5. .336... .6 a..... ..< case... £283.26... 6... .6 ”.96.. 856......aa .65 86,622.60 :85. - v 2.3

 

33

In testing the second research question “what is the impact of socioeconomic
status on women’s individual risk factors for CHD,” the subset used in the first research
question was again utilized and logistic regression models were run twice for each
dependent variable. The first logistic regression was computed with all socioeconomic
variables entered - age, race, education, employment status, marital status and income.
Then a second logistic regression model was established without income entered,
increasing the number of selected cases from 756 to 996. This was done because 240 of
the 996 cases in the subset had missing data on income. The results are reported in
Tables 5-11, for each dependent variable, both with and without annual household
income as an independent variable.

High cholesterol, shown in Table 5, was best predicted by age with a 2.8%
reported increase in the odds of having high cholesterol with each additional year. When
income was excluded, higher education decreased the risk of high cholesterol with the

odds declining by more than 6% for each additional year of schooling.

Table 5 - Logistic Regression: Socioeconomic Predictors of High Cholesterol

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

High Wald Significance Adjusted Wald Significance Adjusted
Choleata'ol Odrk Ratio Odds Ratio
Age 19.5346 .0000 1.0276 23.2290 .0000 1.0263
Race .2048 .6509 .7792 1.3599 .2436 .5675
Bdliclnicn 3.2837 .0700 .9358 4.8419 .0278 .9366
Employment .2460 .6199 .9120 .6501 .4201 .8772
Marital Status 1.6799 .1949 1.2492 2.3311 .1268 1.2382

N = 756 N = 983

Chi-Squl'e Significance .0000 Chi-Square Significance .0000

Predictchorrect 60.32% Predicted Carect 59.61%

34

Overweight, shown in Table 6, was significantly predicted by income with the

odds decreasing more than 18% for each additional $1000 of household income.

However, with income excluded, each additional year of schooling decreases the risk of

being overweight by 7.5%.

Table 6 - Logistic Regression: Socioeconomic Predictors of Overweight

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Overweight Wald Significance Adjusted Wald Significance Adjusted
Odds Ratio Odds Ratio
Age 1.8142 .1780 1.0083 .4836 .4868 1.0037
Race .6204 .4309 .6517 1.1605 .2814 .6018
Education .2009 .6540 .9832 6.5310 .0106 .9248
Employment .0166 .8974 1.0254 .8456 .3578 .8582
Marital 3min 2.9300 .0869 1.3579 .4627 .4963 .9084
Income 20.5751 .0000 .8149 " ' ' ‘
N = 756 N = 983

Physical inactivity, shown in Table 7, was predicted most strongly by education

Chi-Square Significance .0000
Predicted Correct 65.08%

Chi-Square Significance .0138
Predicted Correct 66.43%

with the odds of physical inactivity decreasing by 11% for each additional year of

schooling, then income, which decreases the risk of inactivity 12% for each additional

$1,000 of household income, and age, which increases the odds of inactivity by 1.6% for

each additional year. When income was excluded, physical inactivity was predicted by

education alone with lower education predicting greater physical inactivity.

 

35

Table 7 - Logistic Regression: Socioeconomic Predictors of Physical Inactivity

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Physical Wald Significance Adjusted Wald Significance Adjusted
Age 4.2601 .0390 1.0159 3.0684 .0798 1.01 18
Race .0123 .9119 1.0779 .0252 .8738 .9112
Education 6.3076 .0120 .8905 10.1560 .0014 .8874
Employment .0356 .8504 1.0464 1.0902 .2964 .8016
Marital Statm 1.0973 .2948 1.2540 .3076 .5791 .9079
Incctne 5.2459 .0220 .8824 ’ ¥ ‘ ' ‘ ' ‘

N = 756 N = 983

Chi-Square Significance .0000 Chi-Square Significance .0001

PredictedConect8l.35% PredictedConect81.89%

Diabetes mellitus, shown in Table 8, was also best predicted by advanced age
with the odds increasing 3.3% for each additional year. Income lowered the odds of
diabetes with a 15% decrease in risk for each additional 81,000 of income. When income
was excluded, the risks of diabetes mellitus decreased 11% for each additional year of

schooling.

Table 8 - Logistic Regression: Socioeconomic Predictors of Diabetes Mellitus

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Diabetes Wald Significance Adjusted Wald Significance Adjusted
Mellitus OddsRatio Odds Ratio
Age 8.5374 .0035 1.0327 6.3276 .01 19 1.0244
Race .2509 .6165 1.7078 .4517 .5015 2.0367
Education 1.2495 .2636 .9361 6.2166 .0127 .8880
Employment .1925 .6608 .8593 3.5305 .0603 .5509
Marital Status .8344 .3610 1.3023 .2359
Income 4.1185 .0424 .8512 5.. it” “ .
N - 756 N .. 983
Chi-Square Significance .0000 Chi-Square Significance .0000

Predicted Correct 90.48% Predicted Correct 90.84%

36
Hypertension, shown in Table 9, was best predicted by age with the odds of
having hypertension increasing by 3% with each additional year. When income was

excluded, both higher education and being married decreased the risk of hypertension.

Table 9 - Logistic Regression: Socioeconomic Predictors of Hypertension

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hypertension Wald Significance Adjusted Wald Significance Adjusted
Odds Ratio Odds Ratio

Age 22.4561 .0000 1.0313 33.4176 .0000 1.0338
Race .0115 .9145 1.0646 .0626 .8025 1.1411
Education .0573 .8108 1.0090 3.9658 .0464 .9404
Employment 1.6161 .2036 .7790 1.1936 .2746 .8280
Marital Status .3994 .5274 .8940 3.8345 .0502 .7543
Income 3.4275 .0641 .9273 , ,.

N = 756 N = 983

Chi-Square Significance .0000 Chi-Square Significance .0000

Predicted Correct 65.08% Predicted Correct 65.01%

Cigarette smoking, shown in Table 10, was most strongly predicted by age. The
odds of smoking declined by almost 5% for each additional year of age. Income, the
second strongest predictor of smoking status, lowered the odds of smoking by 14% for
each additional $1,000 of household income. Similarly, each additional year of education
decreased the odds of smoking by almost 12%. And, finally, marital status affects
smoking status with being married reducing the odds of smoking by 29%. In the logistic
regression without income as a predictor, age and education of the respondent remain
predictors of similar strength and magnitude. However, with income excluded, marital
status reduces the odds of smoking by 46%. Typically the household income of married
couples tends to be higher than that of single or widowed persons, thus marital status

seems to “abso ” the income effect.

37

Table 10 - Logistic Regression: Socioeconomic Predictors of Cigarette Smoking

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cigarette Wald Significance Adjusted Wald Significance Adjusted
Smoka' 0drk Ratio Odds Ratio
Age 50.2369 .0000 .9543 73.8127 .0000 .9512
Race .1795 .6718 1.2762 .2666 .6056 1.3108
Education 10.6028 .0011 .8823 32.3572 .0000 .8313
Employment 1.2488 .2638 1.2381 .9576 .3278 1.1771
Marital Status 3.8255 .0505 .7072 17.1037 .0000 .5442
lncorne 14.1664 .0002 .8607 . . i ’ '

N = 756 N = 983

Chi-Square Significance .0000 Chi-Square Significance .0000

Predicted Correct 65.08% Predicted Correct 65.51%

Depressed mood, shown in Table 11, is predicted by age when income is included with a
more than 2% decline in the odds of reported depressed mood for each year older.
However, when income is excluded, depression is predicted by education with a 14%

decrease in the odds for each additional year of schooling.

Table 11 - Logistical Regression: Socioeconomic Predictors of Depressed Mood

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Decreased Wald Significance Adjusted Wald Significance Adjusted
Mood Odds Ratio Odds Ratio
Age 4.0066 .0453 .9796 2.6790 .1017 .9847
Race .2384 .6254 1.6819 .4200 .5169 1.9789
Education 2.4927 .1144 .8971 7.6927 .0055 .8580
Employment .0073 .9319 1.0293 .0667 .7962 .9245
Marital Status .6790 .4099 .7736 3.0233 .0821 .6422
Income 3.6643 .0556 .8483

N = 756 N=983

Chi-Square Significance .0137 Chi-Square Significmee .0277

Predicted Correct 92.46% Predicted Correct 92.98%

38

Comparing the effects of socio-demographic characteristics on the seven
individual CHI) risk factors, certain patterns stand out: (a) education has a consistent
negative effect on the risk factors, i.e., respondents with more formal education
experience generally lower odds of having any of the seven CHD risk factors, (b) while
advanced age raises the odds of hypertension, high cholesterol, diabetes mellitus, and
physical inactivity, it lowers the odds of being a smoker, (c) the income effect is
consistently negative - higher income leads to lower CHD risks. The other SES
characteristics - race, employment status, and marital status, by and large, did not predict
CHD risks in this study population.

When overall CHD risk was regressed on SES, shown in Table 12, lower income
was the most significant predictor, followed by advanced age and lower education. When
income was excluded, lower education was the most significant predictor of high risk for
CHD, followed by advanced age and being rmmarried. Thus, the odds of having two or
more individual risk factors decline with younger age, higher education, and higher

household income but remain unaffected by race, employment, and marital status.

39

Table 12 - Logistic Regression: Socioeconomic Predictors of High Risk for CHD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

High Risk for Wald Significance Adjusted Wald Significance Adjusted
CHD Odds Ratio Odds Ratio
Age 5.9088 .0151 1.0151 4.9543 .0260 1.0116
Race .7882 .3747 .5739 1.3954 .2375 .5294
Education 4.0131 .0451 .9273 26.0593 .0000 .8536
Employment .0084 .9272 1.0173 .4096 .5222 .9030
Marital Status .2005 .6544 1.0814 4.0056 .0453 .7575
income 24.9004 .0000 .8242 8‘

N = 756 N = 983

Chi-Square Significance .0000 Chi-Squire Significance .0000

Predicted Correct 64.29% Predicted Correct 62.26%

EI'I'EIEI 1E°EII'
The findings of this study support most strongly the findings found in the review
of the literature in relationship to the second research question. The results of the first
research question, “to what extent are multiple risk factors for coronary heart disease in
women correlated with each other,” provided limited support for the findings reported in
the literature. Whereas the literature generally reports individual risk factors for CHD,
such as diabetes, overweight, and smoking to be correlated, the analysis found only weak
correlations among diabetes, overweight, and hypertension. Though overall the literature
was lacking in studies pertaining to the correlation of individual risk factors for CHD,
this void was supported by this analysis which found only weak correlations among
individual risk factors implying that the individual risk factors vary independently. These
results suggest that individual risk factors for CHD need to be independently assessed.
However, when comparing the literature and the results of this study on the

second research question, “impact of socioeconomic status on women’s individual risk

4O

factors for coronary heart disease,” the similarities were numerous and, generally, the
findings concurred with the literature. Education and income were the strongest
determinants of individual risk factors for CHI). Of interest in this study, employment
status did not have any significant relationship with individual risk factors for CHD. In
contrast, the review of the literature found unemployment to be related to lower high
density lipids, higher triglycerides, higher total cholesterol, more atherogenic diets, and
higher plasma glucose levels. One possible explanation for this discrepancy is the high
percentage of the not employed women who were so categorized because of their
retirement status or their being homemakers. This made it difficult to tease out the effect
of unemployment separate from age and income. Differences were also found in the
impact of marital status. The literature found marital status to be related only to
depressed mood This study found the state of being unmarried to be associated with
having hypertension, being a cigarette smoker, and having a high risk profile for CHD
when income was excluded Marital status may, in effect, he “absorbing” the effect of

income, thus the relationship with marital status is likely to be spurious.

Chapter 5

Discussion

1...

This research is subject to limitations based on the instrument, sampling methods,
missing data, data processing errors, and generalizability. And, as this study utilized _
secondary data, the researcher was required to work within the constraints of the original
research.

The limitations of the instrument included omission of important data, a potential
for bias in self-report interviews, limitations in comparability of responses, reliability and
validity of the variables measured, and dependence on the interviewee having
independently received recent and appropriate health screening. Important SES data
which was omitted from the instrument is occupation; this would have captured much
more information as a variable than simply knowing employment status.

The potential for bias in self-report is found particularly in the variables
overweight, physical inactivity, and cigarette smoker. These biases may be further
skewed according to SES group, with the higher SES groups being more prone to false
reports; this may be due to an increased awareness of social unacceptability of these

lifestyle behaviors for their SES. This bias may have affected the results in higher SES

41

42

groups by an under report of weight, refusal to respond to cigarette smoking question,
and an overestimate of physical activity.

The responses for education are somewhat limited in their valid comparability
across respondents as they account only for years of formal education and are not
sensitive to the quality of the formal education achieved; furthermore, the responses do
not account for informal education or generational differences in educational attainment.

The measurement of the variables “depressed mood” and “physical inactivity”
possibly lack validity and likely lack reliability. Depressed mood contains two questions
related to mental health from the Medical Outcomes Study 36-item short-form (MOS SF-
36): (a) Have you felt downhearted and blue? and (b) Have you been a happy person?
(Ware & Sherbourne, 1992). These two question refer to only one symptom, mood, of
the eight diagnostic symptoms for depression in the DSM-IV (American Psychiatric
Association, 1994). In addition to the potential for bias in the self-report for physical
inactivity, self-report is also subject to individual interpretation as to what constitutes
moderate activity.

Finally, the measurements of high cholesterol, diabetes mellitus, and
hypertension, were dependent on the interviewee having independently received recent
health care with appropriate screening and follow-up. This could be confounded further
by access to health care, which in turn is influenced by SES.

An additional limitation was possible data processing errors or missing data This
was especially problematic for the variables annual household income, high cholesterol,
and cigarette smoker. Under household income, for example, it was found that 1.4% of

the respondents had an income greater than or equal to $125,000 and going up to more

43

than $1,000,000. As this seemed unlikely, a cross-tabs procedure was done to compare
income and education. The researcher determined that of these high incomes, 24 cases
had a reported education level of some college or less; this level of education generally
does not correspond to very high incomes. Though indeed these incomes may have been
accurate the researcher judged that there may have been a data processing error, possibly
an additional zero, inflating the annual household income; therefore, for the purpose of
this study, these incomes were adjusted by dropping the assumed additional zero in 24
cases. Also puzzling was the lowest level of income where annual incomes as low as zero
dollars were reported As state welfare programs provide for the most needy, reports of
no income at all are improbable and were likely the result of a misunderstanding of
“earned income” vs. “welfare” with the annual household income question, possibly
indicating a need in the instrument to more carefully explain and question household
sources of income during the data collection process.

The variables of high cholesterol and cigarette smoking status contained a large
number of cases with missing data The cholesterol variable, with only 60.6% valid
responses, was cross tabulated with age. For younger respondents (aged 40 years or less)
the invalid responses consistently exceeded 50%. Similarly, for the question about
cigarette smoking, 50.6 % of the cases had missing responses. However, these missing
responses were not related to the respondents’ age. Confounding the issue of missing
data further was the potential effect of other socio-demographics on the missing data
The treatment of this missing data created a statistical analysis contmdrum. Which would
be the greater error: (a) having a sampling bias by eliminating the cases with missing

data; or (b) having a model bias by excluding the variables with large amount of missing

44

data? The researcher elected to err with sampling bias and to focus on the subset of the
sample in which only the cases with complete data on all seven individual risk factors
were accepted.

The sampling methods may have also resulted in problems with random selection
and stratification of the sample. Random selection is threatened as telephone penetration
rates vary with SES. According to the Federal Communication Commission (1997),
telephone penetration rates are lower for Blacks, Hispanics, and Native Americans,
households with a higher number of persons in the household, younger householder age,
rmemployment, and lower income. In Michigan in 1995, the year the sru'vey for this study
was conducted, 95.3% of the households had telephones; however in households with an
annual income of $9,999 or less, telephone penetration rates dropped to 87.6% (Federal
Communication Commission, 1997). Thus lower income and particularly Native
American persons may have been under represented in this sample based on telephone
surveys. Stratification of the sample occurred when selecting the interviewee based on
the ratio of women to men living in a home and when there was more than one phone in
a home.

Generalizability was limited in at least two ways by the geographical location of
this sample. First, the survey was conducted in a rural area Second, this rural area had
pockets of resort commrmities which were occupied by summer residents during the time
when the survey was completed When combined, these two factors created a unique
disparity in SE8. The sample may have had lower SES individuals based on the rural
character of the area and the lower paying seasonal jobs characteristic of a resort

economy. Yet, within this sample is a concentration of retired individuals and higher SES

45

individuals based on the attraction of living in resort communities for those who are
economically mobile.

A potential limitation of the data analysis was the management of marital status
and being a member of an unmarried couple. The definition for being married was “a
legal contact, entered into by a man and a women, to live together as husband and wife.”
As operationalized in the study a member of an umnarried relationship was placed into
the unmarried category together with divorced, widowed, separated, or never married.
However, being in an unmarried relationship - especially if it is a long-term, stable
relationship - could indeed confer the same social support as being married In this
sample 2.2% of the respondents were in an unmarried relationship thus this limitation
would have had a minimal effect on the results.

I 1' . fl 1 1 1 E . 1 1

To direct the implications for the APN the “web of causation” has been revised to
reflect the findings of this study. The significant SES predictors for individual risk
factors for CHD were advanced age, as a non-modifiable risk factor, and lower
education, being unmarried and lower household income, as the modifiable risks factors,
these relationships are shown in Figure 2. Further, the revised “web of causation”
demonstrates the relationships between the individual risk factors.

Utilizing the “web of causation” to contextualize the complex relationship
between SES and individual risk factors for the onset of CHD in women, the APN is
directed to extend her/his scope of practice beyond primary care and the individual to the
broader community, from the individual risk factors to conditions which precede and

correlate to them. The APN must use interventions which incorporate political activity

 

 

£88“ amt 13353 50253 828322 05 mm :95 8 E38“ amt 36.5%,: 93 336 032—80208 8953
338322 05 maumbmaeEou Jamie: 80582 E 8895 new 83%. :3: 3:28 8m seesaw—38 .«e no}. 2:. - N oBmE

 

gnfimgobmgfis 4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o8: 11 Beam 1 1 mean: 11 £385 .11 11 35.295
Banana 3886 8.888%: 833 3%: . 396385 .E:
m . _ m a _ . u _ . _ m _
u. r _ u . p ..
_ _ _ _ _ _ _ _
F111.F-l_lrl.llrlrl_vi.rirl 1+1 11111 trill...
sane-mix,
can;

 

 

”0%. a;

 

 

47

and teaching approaches which are sensitive to SES to lessen the link between SES and
the individual risk factors for CHD.

As a political activist for the primary prevention of CHI) at the community level,
the essential roles the APN would employ are assessor, planner, leader, educator, and
change agent (Givens & Peek, 1995). The APN must begin by assessing the community.
Community health assessments, as utilized in this study, provide core community data to
identify SES and behaviorally mediated risk factors for CHI) within the community.
Additional valuable information for community assessment could be gleaned from
analyzing the commmlity subsystems of economics, recreation, physical environment,
education, safety and transportation, politics and government, health and social services,
and communication (Anderson & McFarlane, 1988). Combined, this diverse information
will help formulate a community CHI) risk profile, identify community resources, and
direct a community intervention plan. As a planner, the APN will work in collaboration
with the community, to mutually formulate community CHI) oriented goals and identify
resources needed to promote comprehensive cardiovascular health for women.
Leadership will be required from the APN to facilitate groups in forming common goals;
thus, leadership will impact the health care and community systems which create the
climate for good cardiovascular health. Applying learning theories and learning methods,
as an educator, the APN will teach and assist the community and all appropriate groups
in identifying and meeting their health education needs. Finally, to bring positive changes
in a community’s modifiable CHI) risk factor profile, the APN will benefit from
employing change theory to systematically and deliberately approach barriers to health.

Specific APN interventions which are indicated by these roles could take place at

48

all levels of government, the public health department, the educational system, the work
site, and the community infrastructure.

The APN needs to be aware of national and state policies which effect the
provision of health care and the distribution of income. With knowledge that lower SES
results in an increase in individual risk factors, the APN needs to be aware of social
policies that increase the economic disparity in the SES hierarchy and lobby for more
egalitarian social policies.

With education, of all the SES measures, having been demonstrated as the
measure having the greatest impact on individual risk factors for CHI), the APN must
support initiatives which provide access to quality education at the primary, secondary,
and collegiate levels in the state and community. This includes supporting the provision
of alternative education programs for the students at high risk for dropping out Also,
lobbying for frmding of adult education programs which offer graduation equivalency
diplomas (GED) is necessary as well. Additionally, both public and private schools are
ideal targets for programs aimed at promoting healthy cardiovascular lifestyles as the
individual risk factors for CHI) are behaviorally mediated patterns which begin in
childhood and youth.

The APN will need to work with public health departments to reduce barriers to
health education To reach low literacy groups, utilization of videotape teaching
materials has been demonstrated to be more effective than written materials (Meade,
McKinney, & Bamas, 1994). These videotapes would be available at primary care clinics
as well as public health departments in conjuction with programs such as Women, Infant,

and Children (WIC), Maternal Support Program, Breast and Cervical Clinics, and

49

Immunization Clinics which target the lower SES population. A further community
strategy using audiovisual educational materials will require campaigns that spread the
cardiovascular health information through the mass media, especially TV and radio,
through use of public service announcements.

Worksite wellness programs have also been found to be an effective means of
promoting cardiovascular health in adults (O’Quinn, 1995). Within the worksite, the
APN can offer health promotion programs which use peer role models to demonstrate
healthy lifestyles. This would be especially effective for lower SES groups for whom role
models have been lacking.

Furthermore, the APN needs to be involved with the local government advocating
the provision of community facilities which promote cardiovascular health by providing
visible and safe places to engage in aerobic exercise year round. In Northern Michigan,
this might include paths for bicycling and walking, wilderness areas with trails for
hiking, snowshoeing and cross-country skiing, and community accommodations for
sports and recreation.

I 1' . E E E 1

The findings of this, study, as they apply to the “web of causation” as previously
shown in the revised “web” (Figure 2), reinforce the error of approaching individual risk
factors for CHI) in isolation from SES risk factors and offer potential questions for firture
research. What educational interventions are appropriate for population with less than a
high school education? What is the interaction effect of marital status on individual risk
factors for CHI) in women when combined with the variables of age, education,

employment, and income? Are there additional variables to be considered in the “web of

50

causation” that are secondary to SES but precede the individual risk factors such as
feeling of self-efficacy, perception of vulnerability, knowledge of CHI) risk factors, and
access to health care. Finally, a further study of SES and individual risk factors for CHI)
in wOmen that incorporates sampling methods which do not under represent the lowest
income and Native American persons is indicated.

Summary

Knowledge of the relationship between SES and the individual risk factors for

CHI) makes it incumbent upon the APN to go beyond individual risk factor interventions
in primary care to community interventions through political activism in order to be
effective in the primary prevention of CHD in women. This activism will take the APN
into the arena of local, state and national government, as well as the educational system,

worksite wellness programs, public health initiatives, and the mass media.

APPENDICES

APPENDIX A
Instrument

QAGE

What is your age?
_ Years

DK

REF

QGEN

GENDER OF RESPONDENT: DO NOT ASK UNLESS UNSURE
“Just to be sure, I need to ask whether you are male or female.”

1. Male

2. Female

DK

REF

Q10.C

How much of the time during the past four weeks have you felt DOWNHEARTED AND
BLUE? Would you say:

1. ALL OF THE TIME

2. MOST OF THE TIME
3.AGOODBITOFTHETIME

4. SOME OF THE TIME

5. A LITTLE OF THE TIME

6. NONE OF THE TIME/NEVER (DO NOT READ #6)
DK

REF

Q10.D

How much of the time during the past four weeks have you been a HAPPY PERSON?
Would you say:

1. ALL OF THE TIME

2. MOST OF THE TIME

3. A GOOD BIT OF THE TIME

4. SOME OF THE TIME

5. A LITTLE OF THE TIME .

6. NONE OF THE TIME/NEVER (DO NOT READ #6)
DK

REF

51

52

Q23
About how long has it been since you last had your blood pressure taken by a doctor,
nurse or other health professional?
DAYS, WEEK, MONTHS, YEARS
DK
REF

Q24

Have you ever been told by a doctor, nurse or other health professional that you have
high blood pressure?

1. YES

2. NO

DK

REF

Q25

Blood cholesterol is a fatty substance found in the blood Have you ever had your blood
cholesterol checked:

1. YES

2. NO -> Q.28

DK -> Q.28

REF

Q26

About how long has it been since you last had your blood cholesterol checked?
_DAYS, WEEKS, MONTHS, YEARS

DK

REF

Q27

Have you ever been told by a doctor, nurse or other health professional that your blood
cholesterol is too high?

1. YES

2. NO

DK

REF

Q28

Have you ever been told by a doctor, nurse or other health professional that you have
diabetes?

1. YES

2. NO

DK

REF

53

QWGT

About how much do you weight without Shoes?
_WEIGHT IN POUNDS ROUND FRACTIONS UP
DK

REF

QTALL
About how tall are you without shoes?
FEET, INCHES

QMODEX ,

In an average week, on how many days out of seven do you get 30 MINUTES or more of
AT LEAST MODERATE exercise, over the course of the entire day? Brisk walking and
moving somewhat heavy materials are examples of moderate exercise. This exercise can
take place over one 30 minute period or even over as many as four periods of 7 or 8
minutes each

___NUMBER OF DAYS PER WEEK GET AT LEAST MODERATE EXERCISE

DK

REF

Q4 1

Do you smoke cigarettes now?
1. YES

2. NO

DK

REF

Q43

How long has it been since you last smoked cigarettes regularly, that is, daily?
_DAYS, WEEKS, MONTHS, YEARS

DK

REF

QRACE

What is your race? Would you say:

1. WHITE

2. BLACK or AFRICAN AMERICAN

3. ASIAN, PACIFIC ISLANDER

4. AMERICAN INDIAN, ALASKA NATTVE or

5. Are you 8 MEMBER OF SOME OTHER RACIAL GROUP (SPECIFY)
DK

REF

54

QMARR

Are you:

1. MARRIED

2. DIVORCED

3. WIDOWED

4. SEPARATED

5. Have you NEVER BEEN MARRIED, or are you
6. A MEMBER OF AN UNMARRIED COUPLE?
DK

REF

QEDUC

What is the highest grade or year of school that your have completed?
00. NEVER ATTENDED SCHOOL, OR KINDERGARTEN ONLY
01-11. ENTER GRADES 1 THROUGH ll

12. HIGH SCHOOL GRADUATE OR GED

13 -15. ENTER SOME COLLEGE

16. COLLEGE GRADUATE

l7. SOME GRADUATE SCHOOL

1 8. MASTERS

19. DOCTORATE

DK

REF

QINC
What was your ANNUAL HOUSEHOLD income for all sources before taxes in 1994?
__ ANNUAL HOUSEHOLD INCOME IN 1994
999,997=999,997 OR MORE
DK
REF

APPENDD( B

Study Approval Letters

55

NORTHERN

MICHIGAN

HOSPITAL
o

BURNS CLINIC

INSTITUTIONAL REVIEW BOARD

Wombat» 30, 1997

360 Connable Avenue
Petoskey. Michigan 49770
616/487-4800
Fax: 616/487—7798

 

Dedicated to
Education, Research
and Technology
A
Quality
Partnership
Stewardship

Jane M. Denay, RN, BSN

Thank you for your letter of October 14, 1997 pertaining to the your thesis
entitled “Socioeconomic status and modifiable risk factors for coronary heart
disease for women in northern Michigan”.

Your letter indicates that you need to have written consent from our agency
(the Northern Michigan Hospital Burns Clinic Foundation). You also state in
your letter that you will need access to the Community Health Assessment data
set in the future.

Your request for written consent is hereby granted along with access to the data
set. .

 

Robert Sloan, MBA
Director of Research _,
Northern Michigan Hospital Burns Clinic Foundation

56

 

OFFICE OF
RESEARCH
AND
GRADUATE
STUDIES

University Committee on
lhunnmnwmfl

Human Subjects

(ucnms)

ummgmsumumwmw
246 Administration Building
East Lansing, Michigan
«madam

517/355-2180
FAX: 517/432-1171

neuummsmumwmw
urnsnwmmmome
ammemAnm

MSU is an Mirnutiwim.

an. .43--

   
 
 
 

MICHIGAN STATE

UNIVERSITY

December 8, 1997

 

To: ManfredOStommel . .
° A-230 Life SCiences Building

RE: IRBfl: 97-788
TITLE: SOCIOECONOMIC STATUS AND MODIFABLE RISK FACTORS
FOR CORONARY HEART DISEASE FOR WOMEN IN NORTHERN
MICHIGAN
REVISION REQUESTED: N/A
CATEGORY: 2-H
APPROVAL DATE: 11/25/97

The University Committee on Research Involving Human Subjects'IUCRIHS)
review of this project is complete.. I am pleased to adVise that the
rights and welfare of the human subjects appear to be adequately
rotected and methods to obtain informed consent are appropriate.
herefore, the UCRIHS approved this project and any reVisions listed
above.

RENEWAL: UCRIHS approval is valid for one calendar year, beginning with
the approval date shown above. Investigators planning to
continue a project beyond one year must use the green renewal
form (enclosed with t e original agproval letter or when a
progect is renewed) to seek u date certification. There is a
maximum of four_such expedite renewals oseible. Investigators
wishin to continue a progect beyond tha time need to submit it
again or complete reView.

REVISIONS: UCRIHS must review any changes in procedures involving human
subjeets, rior to initiation of t e change. If this is done at
theotime o renewal, please use the green renewal_form. To
reVise an approved protocol at an 0 her time during the year,
send your written request to the. CRIHS Chair, requesting reVised
approval and referenCing the prOJect's IRB # and title. Include
in ur request a description of the.change and any revised
ins ruments, consent forms or advertisements that are applicable.

PROBLEIIB/

CHANGES: Should either of the followin arise during the course of the
work, investigators must noti y UCRIHS romptly: (1) roblems
(unexpected side effects, comp aints, e c.) involving uman
eubjects.or 12) changes in the research environment or new
information indicating greater risk to the human sub'ects than
existed when the protocol was preViously reviewed an approved.

If we can be of any future help, please do not hesitate to contact us
at (517)355-2180 or FAX (517)4 2-117] .

W

E. Wright, Ph.D.
IHS Chair

W:bed

‘,£§r:;;he Denay

Sincerely,

.4

57

LIST OF REFERENCES

LIST OF REFERENCES

Adams, P. F. & Benson, V. (1990). W
W Washington, DC: National Center for Health Statistics.

Adler, N. E., Boyce, T., Chesney, M. A., Cohen, S., Folkman, S., Kahn, R. L., &
Syme, S.L. (1994). Socioeconomic status and health: The challenge of the gradient.
mmmmmzw 1 ), 15-24.

American Diabetes Association (1996). W [On-
line].Available: http://www.diabetes.org/ada/c20e.html

American Heart Association (1997). W [On-line]. Available:
http://www.ambert.orgIHeartfland_Stroke__A_Z_Guide/chollev.html

American Heart Association (1997). W [On-line]. Available:
http://www.amhrtorg/Heart__and_Stroke__A_Z_Guidc/exercise.html

American Heart Association (1996). WWW:
[On-line]. Available: http://www.amhrt.org/hs96/women.html

American Heart Association. WW

Dallas, Tex: American Heart Association; 1995.

American Heart Association. Silmtsnidemiclhemnhmammmandhm
disease, Dallas, Tex: American Heart Association; 1992.

58

59

American Psychiatric Association. (1994). W
Wm (4th ed). Washington, DC: Author.

Anderson, E. T. & McFarlane, J. M. (1988). Epidemiology, demography and
research In W (pp. 15-100). Philadelphia: Lippincott.

Booth-Kewley, S. & Friedman, H. S. (1987). Psychological predictors of heart
disease: A quantitative review. W6), 343-362.

Connolly, V. M. & Kesson, C. M. (1996). Socioeconomic status and clustering of
cardiovascular disease risk factors in diabetic patients. W5), 419-22.

Davis, S. K., Winkleby, M. A, & Farquhar, J. W., (1995) Increasing disparity in
knowledge of cardiovascular disease risk factors and risk-reduction strategies by
socioeconomic status: Implications for policymakers. W
W16), 318-323.

Federal Commtmication Commission (1997). W W
WW5 [On-line]. Available:
http://www.fcc.gov/Bureaus/Common_Carrier/Reports/FCC-Statc_Link/monitor.html

Ford, E. S., Merritt, R. K., Heath, G. W., Powell, K. E., Washbum, R. A., Kriska,

A. & Haile, B. (1991). Physical activity behaviors in lower and higher socioeconomic

Stems Populations. W02). 1246-56.
Given, B. & Peek. P. (1995). W
W2, {NUR 501 Clinical Nurse Specialist in Primary Care Course Syllabus}. East

Lansing, MI: Michigan State University.

60
Gold, M. R. & Franks, P. (1990). The social origin of cardiovascular risk: An
investigation in a rural community. WW6),
405-416.
Haan, C. K. (1996).
mm [On-line]. Available: http://www.medscape.comMedscape/womens,
health/l996/v01.n12/w69.haan/w69.haan.html

 

Harris-Hooker, S. & Sanford, G. L. (1994). Lipids, lipoproteins and coronary
heart disease in minority populations. WSW”, 883-104.

Hazuda, H. P., Hatfner, S. M., Stem, M. P. Knapp, J. A., Eifler, C. W., &
Rosenthal, M. (1986). Employment status and women’s protection against coronary heart
disease. MW“) 623-640-

Hoeymans, N., Smit, H. A., Verkleij, H., & Kromhout, D. (1996). Cardiovascular
risk factors in relation to educational level in 36 000 men and women in The
Netherlands. WM), 5 1 8-525.

Hu, Y. & Goldman, N. (1990). Mortality differential by marital status: An
international comparison. Emma), 233-250.

Hypertension Detection and Follow-up Program. (1977). Race, education and
prevalence of hypertension WW5), 351-361.

Information Transfer Systems, Inc. (1995, September). We:

\OIICIU U I"" 'I‘ .I. \0II‘JHU.I°15|JNH'5'}? Ii..' . "all “

WW Ann Arbor. MI: Author.

61

Jacobsen, B. K. & Thelle, D. S. (1986). Risk factors for coronary heart disease
and level of education. The Tromso Heart Study. WW
121(5), 923-932.

Joint National Committee on Detection, Evaluation, and Treatment of High
Blood Pressure (1993). The fifth report of the Joint National Committee on detection,
evaluation, and treatment of high blood pressure. W153,
154-183.

Judelson, D. R. (1994). Coronary heart disease in women: Risk factors and

 

' ‘ 6), 186-197.

Kaplan, G. A & Keil J. E. (1993). Socioeconomic factors and cardiovascular
disease: a review of the literature. W4), 1973-1993.

Kaplan, G. A., Roberts, R. E., Camacho, T. C., & Coyne, J. C. (1987).
Psychosocial predictors of depression. Prospective evidence from the human population
laboratory studies. WWO), 206-220.

Kritz-Silverstein, D., Wingard, D. L., & Barrett-Conner, E. (1992). Employment
status and heart disease risk factors in middle-aged women: The Rancho Bernardo Study.
WWO). 215-219.

Landau, S. 1. (Ed). (1993). Wan New
York: Harper & Row.

Lapidus, L. & Bengtsson, C. (1986). Socioeconomic factors and physical activity
in relation to cardiovascular disease and death A 12 year follow up of participants in a

population study of women in Gothenburg, Sweden. W0), 295-
301.

62
Link, B. G. (1996). Understanding sociodemographic differences in health. The

role of fundamental social cause. WW4), 471-473.

Link, B. G. & Phelan, J. (1995). Social conditions as fundamental causes of
disease. WWW (Extra 185116). 30-94-

Luepker, R. V., Rosamond, W. D., Murphy, R., Spratka, J. M., Folsom, A. R.,
McGovern, P. G., & Blackburn, H. (1993). Socioeconomic status and coronary heart
disease risk factor trend. The Minnesota Heart Survey. W5), 2172-2179.

Luoto, R., Pekkanen, J ., Uutela, A., & Tuomilehto, J. (1994). Cardiovascular risks

and socioeconomic status: Differences between men and women in Finland. W

EpideminlmmdflommuninrflealmAS. 348-354.

MacMahon. B. & Pugh. T. F. (1970). Enidmiolnsicnrincinlcsandmethods.
Boston: Little, Brown and Company.

Matthews, K. A., Kelsey, S. F., Meilahn, E. N., Kuller, L. H., & Wing, R. R.
(1989). Educational attainment and behavioral and biologic risk factors for coronary
heart disease in middle-aged women. W6), 1132-
1144.

Meade, C. D., McKinney, W. P., & Barnas, G. P. (1994). Educating patients with
limited literacy skills: The effectiveness of printed and videotaped materials about colon
cancer. WWW. 119-121.

Murphy, J. M., Olivier, D. C., Monson, R. R., Sobol, A. M., Federman, E. B., &
Leighton, A. H. (1991). Depression and anxiety in relation to social status. A prospective

epidemiologic study. WM”. 223-229.

63

Mykkanen, L., Laakso, M., & Pyorala, K. (1993). High plasma insulin level
associated with coronary heart disease in the elderly. WW
131(1 1), 1190-1202.

Nomsis. M. J. (1996). W
Chicago: SPSS Inc..

O’Quinn, J. L., 1995. Worksite wellness programs and lifestyle behaviors. ,[mal
WM). 346-360.

Pi-Sunyer, F. X. (1993). Medical hazards of obesity. WM
WU), 655-660.

Polit, D. F. & Hungler, B. P. (1995). WW
(5“ ed). Philadelphia: J .B. Lippincott Company.

Reeder, B. A., Liu, L., & Horlick, L. (1996). Sociodemographic variation in the
prevalence of cardiovascular disease in Saskatchewan: Results from the Saskatchewan
Heart Health Survey. WW6), 271-277.

Rose, 0., & Marmot, M. G. (1981). Social class and coronary heart disease.
W0). 13-19.

Rosenberg, L., Palmer, J. R., & Shapiro, S. (1990). Decline in the risk of
myocardial infarction among women who stop smoking. WW
WM), 213-217.

Shortridge, L. & Valanis, B. (1992). The epidemiological model applied in
community health nursing. In M. Stanhope & J. Lancaster (Eds), W

. (3"l ed, pp. 151-170). St Louis:

 

Mosby Year Book.

64

Siefl‘ert, G. F ., Keown, K., Moore, W. S. (1981). Pathologic effect of tobacco
smoke inhalation on arterial intima. W 333-335.

Smeltzer, S. C. & Bare, B. G. (1996). W
MW 8‘“ Edition. Philadelphia: Lippincott-Raven.

Thomas. C. L. (13d)- (1997)- WWW 18‘“
Edition. Philadelphia: F.A. Davis Company.

US. Department of Health and Human Services. Public Health Service (1990).
11'... 1 ' ‘-.1 1‘ 1111111111.; 11'.-_ 11111111 111 z 11 1111‘ 611.1111 1-11' ‘ The
author: Washington DC.

Ware, J. E. & Sherbourne, C. D. (1992). The MOS 36-Item Short-Form Survey
(SF-36). Wfi), 473-481.

Winkleby, M. A., Jatulis, F. E., Frank, E., Fortrnann, S. P. (1992). Socioeconomic
status and health: how education, income, and occupation contribute to risk factors for

cardiovascular disease. WWW, 816-820.