PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:/Prolecc&Pres/CIRC/DateDue.indd

FACTORS ASSOCIATED WITH MAMMOGRAPHY SCREENING UTILIZATION
AMONG LATINAS: A REVISION OF THE BEHAVIORAL MODEL OF HEALTH
SERVICES USE

By

Sheila F. LaHousse

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Psychology

2009

ABSTRACT
FACTORS ASSOCIATED WITH MAMMOGRAPHY SCREENING UTILIZATION

AMONG LATINAS: A REVISION OF THE BEHAVIORAL MODEL OF HEALTH
SERVICES USE

By
Sheila F. LaHousse

Breast cancer is the most common form of cancer and second leading cause of cancer
death among women in the United States. Adherence to mammography screening
guidelines is associated with a lower risk of being diagnosed with invasive breast cancer.
One reason Latinas are more likely to be diagnosed with breast cancer at advanced stages
and have lower survival rates compared to Non-Hispanic white women is because they
have lower screening rates. Decreasing these disparities can be accomplished by
increasing breast cancer screening and early detection among Latinas. This study used
theoretically-driven logistic regression analyses to determine signiﬁcant predictors of
recent mammography screening utilization among Latinas 40 years and older in San
Diego County, California. Using Andersen’s Behavioral Model of Health Services Use as
a framework, the Behavioral Model of Latino Breast Cancer Screening Use was
developed in this study and provided a reasonable ﬁ'amework for explaining and
predicting breast cancer screening behavior in this sample of 208 Latinas 40 years and
older. Results showed that-a lower language-based acculturation, higher health literacy,
higher age, a recent physician visit in the last year, and lack of distressful cognitions I
about developing breast cancer signiﬁcantly predicted recent mammography screening
use within the past two years, aﬁer adjusting for other variables. This study demonstrated

that access to care matters in the process of obtaining preventive services since no other

factor was consistently more predictive of adherence breast cancer screening guidelines
than undergoing an annual examination. In fact, having a recent physician visit in the last
year increased the odds of having a recent mammogram by 3 or 4 times in adjusted
logistic regression models. Overall, these ﬁndings suggest that improving access to care
and addressing detracting psychosocial factors is essential for promoting mammography
screening utilization among Latinas. Future research is needed to determine if the
Behavioral Model of Latino Breast Cancer Screening Use framework applies to other

Latina health behaviors and to other recent immigrant populations to the United States.

C0pyright by
SHEILA FAYE LAHOUSSE

2009

This work is dedicated to my aunt Janet Swajanen and all of the women before her who

have lost the ﬁght against breast cancer.

ACKNOWLEDGEMENTS

I would like to acknowledge all of the people who have provided me with mentoring,
support, and guidance throughout my graduate school career. I give many thanks to
Pennie F oster-F ishman, Bill Davidson, Rene Rosenbaum, and Ralph Levine for being a
dynamic dissertation committee. I would like to give credit to Georgia Robins Sadler,
Vanessa Malcame, Natasha Riley, Viridiana Conde, and Mercedes Soto, whose
collaborative effort made this study possible. I also thank the 503 women who
participated in this study. Finally, I would also like to thank my parents for their love,

encouragement, and steadfast belief in my success.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. x
LIST OF FIGURES .......................................................................................................... xi
OVERVIEW ....................................................................................................................... 1
INTRODUCTION .............................................................................................................. 5
Breast Cancer Incidence, Mortality, and Survival .................................................. 5

Breast Cancer Screening and Early Detection ......................................................... 6

The Behavioral Model of Health Services Use ...................................................... 9
Predisposing Factors ................................................................................ 14

Age .......................................................................................... 1 5

Education ................................................................................ 16

Acculturation .......................................................................... 1 7

Language-based Acculturation ................................ 17

Country of Birth ........................................................ 19

Need Factors ........................................................................................... 20

Health-Related Quality of Life (HRQOL) ............................... 20

Enabling Factors ...................................................................................... 21

Health Insurance ..................................................................... 22

Recent Physician Visit ............................................................. 23

Discretionary and Non-discretionary Utilization .................................... 24

The Behavioral Model of Latina Breast Cancer Screening Use ........................... 25
History of Breast Cancer ......................................................... 27

Health Literacy ....................................................................... 28

Cancer Fatalistic Beliefs ......................................................... 29

Religiosity ............................................................................... 30

Breast Cancer Worry .............................................................. 31

Conclusion .............................................................................................................. 32
Hypotheses ............................................................................................................. 35
RESEARCH DESIGN AND METHODS ........................................................................ 38
Procedures ...................................................................................................... 38

Setting ............................................................................................ 38

Sample ........................................................................................... 39

Data Collection .............................................................................. 40

Recruitment .................................................................... 41

Sample Size .................................................................... 41

Survey Instrument ........................................................................ 42

Breast Cancer Screening .............................................. 43

Predisposing Variables ................................................ 44

Age .............................................................. 44

Education .................................................... 44

Country of Birth ......................................... 44

vii

Language-based Acculturation .................... 44

Enabling Variables ...................................................... 45
Health Insurance ......................................... 45
Recent Physician Visit ................................ 45
Need-related Variables ................................................. 46
Health-related Quality of Life .................... 46
Context-speciﬁc Variables ........................................... 47
History of Breast Cancer ............................. 47
Health Literacy ............................................ 47
Religiosity ................................................... 48
Cancer Fatalism ........................................... 49
Breast Cancer Worry ................................... 50
Data Analysis ...................................................................................................... 51
Data Entry ........................................................................................ 51
Data Analysis Strategy .................................................................... 52
RESULTS ......................................................................................................................... 54
Logistic Regression Diagnostics .......................................................................... 54
Testing the CFA Model ........................................................................................ 54
Univariate and Descriptive Analyses .................................................................... 58
Predisposing Variables ........................................................................... 58
Enabling Variables ................................................................................. 58
Need-related Variables ........................................................................... 59
Context-speciﬁc Variables ..................................................................... 59
Results of Hypothesis Testing .............................................................................. 60
Testing the Behavioral Model of Health Services Use ............................ 60
Bivariate Analyses ...................................................................... 60
Predisposing Variables ................................................... 6O
Enabling Variables ......................................................... 61
Need-related Variables ................................................... 61
Multivariate Analyses ................................................................. 63
Summary ..................................................................................... 68
Signiﬁcance of Predisposing Variables ............................ 68
Signiﬁcance of Enabling Variables .................................. 71
Signiﬁcance of Need-related Variables. ........................... 71
Testing the Incremental Prediction of Predisposing and Need Domains .71
Predisposing and Enabling Domains .......................................... 72
Need-related and Enabling Domains ........................................... 73
Summary ..................................................................................... 75

Testing the Incremental Prediction of the Proposed Context-speciﬁc
Variables ................................................................................................... 75
Bivariate Analyses ...................................................................... 75
Multivariate Analyses ................................................................. 78
Behavioral Model of Latina Breast Cancer Screening Use ........ 79
Summary ..................................................................................... 81

viii

DISCUSSION ................................................................................................................... 82

Key Findings .......................................................................................................... 82
Behavioral Model of Health Services Use .................................................. 83

Behavioral Model of Latina Breast Cancer Screening Use ......................... 86
Predisposing, Enabling, Need, and Contextual Predictors .......................... 88

Enabling Factors ............................................................................. 88

Predisposing Factors ....................................................................... 91

Need-related Factors ....................................................................... 94

Context-speciﬁc Factors ................................................................. 94

Implications for Interventions ..................................................................... 97

Study Limitations .................................................................................................... 98
CONCLUSION ................................................................................................................ 100
APPENDICES ................................................................................................................ 102
APPENDIX A: Recruitment Strategies ....................................... ' ....................... 103
APPENDIX B: Consent Documents ................................................................... 104
APPENDIX C: Selected Pages from English and Spanish Surveys ................... 108
APPENDIX D: Descriptive and Bivariate Statistics ........................................... 125
REFERENCES ............................................................................................................... 137

ix

LIST OF TABLES

Table 1: Variable Coding for Analysis ............................................................................. 53
Table 2: Behavioral Model Factor Loadings ..................................................................... 56
Table 3: Bivariate Relationships of Behavioral Model Variables with Mammography

Screening Utilization ......................................................................................................... 62
Table 4: Predisposing Factors Predicting Recent Mammography Screening Use ............ 64
Table 5: Enabling Factors Predicting Recent Mammography Screening Use ................... 66
Table 6: Need-related Factors Predicting Recent Mammography Screening Use ............ 67

Table 7: Reduced Behavioral Model Predicting Recent Mammography Screening Use..68

Table 8: Predisposing and Enabling Domains Predicting Recent Mammography
Screening Use .................................................................................................................... 73

Table 9: Need-related and Enabling Domains Predicting Recent Mammography
Screening Use .................................................................................................................... 74

Table 10: Bivariate Associations Between Context-speciﬁc Variables with Recent

Mammography Screening Use .................................................. 77
Table 11: Reduced Behavioral Model and Selective Context-speciﬁc Variables

Predicting Recent Mammography Screening Use ............................................................ 79

APPENDIX E TABLES

Table A: Descriptive Statistics for Categorical Variables (Total Sample) ...................... 125
Table B: Comparison of Sample Demographics with County and State Data ................ 128
Table C: Descriptive Statistics of Scales and Continuous Items (Total Sample) ............ 129
Tables D-P: Psychometric Properties of Scales (English and Spanish Versions) ........... 130
Table Q: Correlation Matrix Used for Conﬁrmatory Factor Analysis ........................... 135

Table R: Correlation Matrix of Variables in the Final Behavioral Model of Latina Breast
Cancer Screening Use, Used for Conﬁrrnatory Factor Analysis ..................................... 135

Table S: Behavioral Model of Latina Breast Cancer Screening Use, With and Without
Suppressed Variables ....................................................................................................... 136

LIST OF FIGURES

Figure 1: Behavioral Model of Health Services Use (Model 1) ....................................... 14
Figure 2: Behavioral Model of Latina Breast Cancer Screening Use (Model 2) ............. 27
Figure 3: Measurement Model of the Behavioral Model of Health Services Use ............ 55
Figure 4: Modiﬁed Measurement Model ........................................................................... 57
Figure 5: Final Behavioral Model of Latina Breast Cancer Screening Use ..................... 80

Figure 6: Measurement Model of the Behavioral Model of Latina Breast Cancer
Screening Use .................................................................................................................... 80

xi

Overview

In 2004, cancer was the second leading cause of death in California, accounting
for 23 percent of all annual deaths (American Cancer Society [ACS], 2008). Of all cancer
types, breast cancer exhibits the highest incidence and the second highest mortality rate in
California for all ethnic groups (ACS, 2008; California Cancer Registry [CCR], 2008).
Although Latinas are less likely to develop breast cancer than white women, Latinas are
more likely to exhibit late stage breast cancers at time of diagnosis and have lower ﬁve-,
10- and 15-year survival rates compared to white women in California (CCR, 2008) and
nationwide (Zambrana, Breen, Fox, & Gutierrez-Mohammad, 1999).

Regular use of mammography is associated with a decreased risk of developing
invasive breast cancer (National Cancer Institute [NCI], 2007b) and reducing breast
cancer mortality by about 25% in women aged 40 years and over (Elmore, Armstrong,
Lehman, & Fletcher, 2005; Wells & Roetzheim, 2007). However, Latinas have lower
rates of mammography use than other ethnicities (Frazier, J iles, & Mayberry, 1996;
Hubbell, Chavez, Mishra, & Valdez, 1996). Thus, much work is needed to promote
screening and early detection of breast cancer among Latinas. The purpose of this study
was to contribute to this effort by examining factors associated with and that can predict
breast cancer screening utilization among Latinas.

Given that the determinants of mammography usage are complex (Fox,
Amsberger, Owens, Nussey, Zhang, Golding, et al., 2004; Mandelblatt, Yabroff, &
Kemer, 1999) a multivariate conceptual framework called the Behavioral Model of
Health Services Use guides this study. The model has existed for 45 years and was

developed to explain why individuals use health care, to deﬁne and measure equitable

access to healthcare, and to assist in developing policies that promote equitable access to
healthcare (Andersen, 1995). The model suggests that health services use is a ﬁrnction of
three factors: one’s predisposition to use health services (age, education, and
acculturation), factors that enable use (health insurance and a recent physician’s visit),
and one’s need for healthcare (perceived health status) (Andersen, 1995). Predisposing,
enabling, and need domains of the Behavioral Model were targeted in this study.

In sum, this study aimed to contribute to the current understanding of what
predicts breast cancer screening use among Latinas 40 years and older to inform research,
health promotion practice, and primary care. The following objectives were proposed to
achieve this goal:

First, this study used the Behavioral Model to examine what predisposing,
enabling, and need factors predict mammography use. In contrast to most of the studies
that have examined multiple predictors of Latina breast cancer screening, this study was
guided by a well-established theoretical framework. Studies guided by a clear well-
established theoretical ﬁ'amework can contribute substantially to the research literature,
facilitate inter-study comparisons, and lend credible evidence for the creation of policy.
A precursor to this ﬁrst step was to determine the utility of the Behavioral Model by
examining whether the model as outlined in the literature review ﬁt this particular
population and outcome variable under investigation.

Second, a main argument of this study was that enabling factors matter more than
predisposing and need factors in regards to breast cancer screening. In fact, research
shows that for the discretionary use of health services-such as mammography utilization-

predisposing (age, education, or acculturation) and enabling (health insurance or having a

recent physician’s visit) factors may matter more than need (perceived health status)
(Andersen, 1968). Although predisposing factors do predict breast cancer screening
among Latinas, the presence of enabling factors may counteract this role (Zambrana, et
al., 1999). It was necessary to determine the relative importance of such enabling factors
(e. g., insurance coverage and regular access to a primary care provider) in facilitating
access to mammography screening among this population for future research, policy and
community-based practice. Thus, this study sought to contribute to this effort by testing
the degree to which predisposing and need factors relate to breast cancer screening, after
accounting for enabling factors.

Lastly, this study proposed a modiﬁcation of the Behavioral Model. In fact, the
Behavioral Model can be viewed as a “meta-framework”, in which context-speciﬁc
variables are added to allow for greater prediction of the utilization measure in the
targeted population (Andersen, 2008). For the current study, the Behavioral Model of
Latina Breast Cancer Screening Use was created by modifying the Behavioral Model in
order to examine the signiﬁcance of variables context-speciﬁc to Latina breast cancer
screening behavior (e.g. breast cancer worry, personal/family history of breast cancer,
cancer fatalism, religiosity, and health literacy). This study sought to determine if these
variables predicted breast cancer screening, after accounting for variables in the original
Behavioral Model. Results from this last study component were hypothesized to provide
valuable insight for future research, community-based health promotion, and clinical
practice. To accomplish these objectives, this study was guided by the following research

questions:

. Does the component structure of the Behavioral Model ﬁt the Latina population
for the outcome of breast cancer screening utilization?

. What predisposing, enabling and need factors are associated with an increased
likelihood of breast cancer screening among Latinas?

. To what extent do predisposing and need factors add to the prediction of breast
cancer screening utilization by Latinas, after controlling for enabling factors?

. To what extent do context-speciﬁc variables add to the prediction of breast cancer

screening utilization by Latinas?

INTRODUCTION

Since 1971, when President Nixon signed the National Cancer Act and declared a
“War against Cancer”, there has been a decline in cancer mortality, although this progress
has not been experienced by all Americans (Freeman, 2004). Disparities exist in cancer
incidence, mortality, and survival based on ethnicity or socioeconomic status (Blackrnan
& Masi, 2006; Freeman, 2004; IOM, 1999; Sassi, Luft, & Guadagnoli, 2004; Ward et al.,
2004). Latinos, African Americans, Asian Americans, and Native Americans experience
a greater cancer burden than their white counterpartsl (N C1, 2005; Freeman, 2004; Ward,
et al., 2004). Attention to cancer disparities experienced by Latinas is critical, given that
they are members of the fastest growing ethnic and cultural group in the United States
(US. Census Bureau, 2004). Of the 299.4 million US. residents in 2006, 44.3 million
were Latino, of which 28.3 million were Mexican origin. In 2006, there were 37.5 million
foreign-born persons residing in the United States, of which 20.1 million were from Latin
America (US. Census Bureau, 2006a). Latinos in the United States are more likely to
live in poverty, have less education and lack health insurance than other ethnic groups,
which may contribute to disparities in incidence, mortality, and burden of breast cancer
affecting Latinas (Abraido-Lanza, Chao, & Gammon, 2004; DeNavas-Walt, Proctor, &
Lee, 2005; National Center for Educational Statistics [NCES], 2007).

Breast Cancer Incidence, Mortality, and Survival

Since 1988, breast cancer mortality rates in California have decreased by 27%, yet

cancer is still the second leading cause of death, accounting for 23% of all annual deaths

(ACS, 2008). Even though mortality rates have declined over the past two decades, breast

 

lCancer health disparities are “differences in incidence, prevalence, mortality and burden of cancer and
related adverse health conditions that exist among speciﬁc population groups in the US.” (NCI, 2007a)

5

cancer is still the most common cancer diagnosis and the second leading cause of cancer
mortality in California for all ethnic groups (ACS, 2008; CCR, 2008). In 2008, breast
cancer will account for 31% of all new cancer cases and 16% of cancer deaths in
California (ACS, 2008). One of the strongest predictors of survival is the degree to which
the cancer has spread through the body (ACS, 2008). The prospect of surviving ﬁve years
after diagnosis is better if breast cancer is caught early. From 1988-2004 in California,
the ﬁve—year relative survival for woman diagnosed with localized breast cancer was
97%, compared to 20% for women diagnosed with late stage breast cancer (ACS, 2008).

Although Latinas are less likely to develop breast cancer than white women,
Latinas are more likely to exhibit late stage breast cancers at time of diagnosis (Austin,
Ahmad, McNally, & Stewart, 2002), have later initiation of treatment, (Rodriguez, Ward,
& Perez-Stable, 2005), and have lower ﬁve-, 10- and 15-year survival rates compared to
white women (Zambrana, et al., 1999). For example, from 2000-2003, 54% of Latina
breast cancer cases were detected at the local stage, compared to 63% of cases of white
women (ACS, 2007), even though white women had higher age-adjusted breast cancer
incidence (140.6 per 100,000) and mortality (25.8 per 100,000) rates compared to Latinas
nationwide (89.1 per 100,000; 16.2 per 100,000 respectively); these disparities in survival
may relate to late diagnosis due to a lack of breast cancer screening (ACS, 2007).

Breast Cancer Screening andEarly Detection

Screening is a procedure whereby sections of the population most at risk for a
disease are examined for any early indications of disease. Breast cancer screening is a
secondary prevention technique, aiming for early disease detection to allow for

interventions that prevent disease progression and emergence of symptoms (Marks, 2005;

Rajararn & Rashidi, 1998). Experts agree that screening tests should only be used when
the potential beneﬁts clearly outweigh the potential harms (Brawley & Kramer, 2005).
Direct causal factors for developing breast cancer remain largely unknown; however,
screening can detect cancer at early stages, which greatly increases chances of cure,
extends life, reduces the extent of treatment needed, and improves quality of life (ACS,
2003; NCI, 2007b). Thus, breast cancer screening and early detection play an important
role in reducing the impact of this cancer (N CI, 2007b).

From 2000-2005, the median age for a diagnosis of breast cancer was 61 years old
(NCI, 2007b; 2008a). For this reason, the US. Preventive Services Task Force
recommends mammography [with or without a clinical breast examination(CBE)] every
1-2 years for women 40 years and older, since it is associated with early detection and
improved outcomes (Agency for Healthcare Research and Quality [AHRQ], 2006).
Evidence for the utility of manual CBEs and breast self-examinations to screen for breast
cancer is inconclusive (AHRQ, 2006; Elmore, etal., 2005; Humphrey, etal., 2002).

Some studies argue that mammography screening tests are less sensitive for
younger women aged 40 to 49 years than for women 50 years and older (Buist, Porter,
Lehman, Taplin, & White, 2004) and other studies have indicated that the current
evidence on the effectiveness of screening mammography does not suggest whether
women aged 40-49 should receive mammograms or not (Ringash, 2001 ). However, the
use of regular screening mammography among women 40 to 74 years has been shown in
randomized trials to be associated with a reduction in breast cancer mortality (Humphrey,

Helfand, Chan, & Woolf, 2002; Tabar, Vitak, Chen, Yen, Duffy, & Smith, 2001; Wells &

Roetzheim, 2007). Other studies have indicated that screening mammography in women
aged 40 to 70 years decreases breast cancer mortality (N CI, 2008b).

The precise age to end screening is uncertain, since no trials have enrolled women
older than 74 years of age (AHRQ, 2006). A meta-analysis of 13 studies showed that
screening mammography signiﬁcantly reduces breast cancer mortality in women aged 50
to 74 years after seven to nine years of follow-up, regardless of screening interval or
number of mammographic views per screen (Kerlikowske, Grady, Rubin, Sandrock, &
Emster, 1995). In another study that reviewed eight randomized controlled trials,
screening mammography was found to reduce breast cancer mortality by about 20-3 5%
for women aged 50 to 69 and slightly less for women aged 40-49 years 14 years after the
initial screening (Elmore, et al., 2005). A meta-analysis of over 100 randomized
controlled trials from 1965 to 2005 showed 7-23% reduction in breast cancer mortality
rates with screening mammography for women aged 40-49 (Armstrong, Moye, Williams,
Berlin, & Reynolds, 2007). Although these studies have cited the beneﬁts of
mammography screening for women aged 40 to 74, it is unclear whether these studies
have controlled for age. Given that survival post-diagnosis may be dependent on age of
diagnosis, in that younger women may have better survival rates than older women by
virtue of their age and better health status, studies are needed to examine the effect that
age at diagnosis may or may not have on survival.

Nonetheless, although the beneﬁts of screening mammography for women 40
years and older are clearly evident, mammography usage differs by ethnicity, education,
income, acculturation, and geographic region within the United States (Rodriguez, Ward,

& Perez-Stable, 2005). Latinas are less likely to utilize mammography than other ethnic

groups in California (California Health Interview Survey [CHIS], 2005) or than Aﬁican
American and white women nationally (CDC, 2007). Thus, the need to promote
screening and early detection of breast cancer among Latinas is clear. Latinas are more
likely to exhibit late stage2 breast cancers at time of diagnosis, and have lower rates of
survival and mammography usage than other ethnicities. Since breast cancer mortality is
largely due to late diagnosis, and late diagnosis is directly related to underutilization of
mammography (Borrayo & Jenkins, 2001; Jones, Caplan, & Davis, 2003), this study
focused on factors that were proposed to inﬂuence breast cancer screening3 use among
Latinas.
The Behavioral Model of Health Services Use

Theories help provide a framework for understanding the targeted health
behaviors and the context in which they occur (N CI, 2005b) and consist of both
measurement (e.g., operationalization of constructs) and structural (e. g., interrelationships
among constructs) components (N unnally & Bernstein, 1994). Interventions based on
theory are more effective than those not theoretically-based (e. g., Painter, Borba, Hynes,
Mays, & Glanz, 2008), and thus theories are important to consider in health promotion
research. However, studies targeting health behavioral change are largely atheoretical. In
a recent review of 193 health behavior studies published from 2000-2005, Painter and
colleagues found that only 37% mentioned theory and of those, most were informed by

theory (68.1%) or applied theory (18%) and few either tested (3.6%) or created theory

 

2 Degree of cancer progression (stage I-IV), can range from “in situ” (cells have not spread), “localized”
(cells have spread to local tissue or organ(s)), “regional” (cells have spread to the lymph nodes, adjacent
tissues, or organs), and “distant” (cells have spread to other regions in the body) (ACS, 2008, p. 7).

Breast cancer screening utilization in this study refers to having a recent mammogram in the previous two
years.

(9.4%) (Painter, et al., 2008). Thus, there is a clear need for greater theory integration,
development, and testing in health behavior research.

Given that determinants of breast cancer screening use are complex (Fox, et al.,
2004; Mandelblatt, Yabroff, & Kerner, 1999; Otero-Sabogal, etal., 2004), this study used
a multivariate theoretical framework called the Behavioral Model of Health Services Use
(Andersen, 1968; 1995; 2008), to help guide hypotheses, measures included, and analyses
in this study in order examine breast cancer screening utilization among Latinas. The
Behavioral Model has existed for 45 years and was developed to explain why individuals
use health care, to deﬁne and measure equitable access to healthcare, and to aid in
developing policies promoting equitable access to healthcare (Andersen, 2008).
The Behavioral Model focuses on the individual unit of analysis and suggests that
healthcare use is a function of an individual’s predisposition to use services
(predisposing), factors that enable or impede use (enabling), and their need for care
(need) (Andersen, 1995) (Figure 1).

According to Ricketts and Goldsmith (2005), the Behavioral Model and the Fit
Model by Penchanksy, Frenk, and colleagues are the two major theories focusing on
access and utilization of health services. Ricketts and Goldsmith (2005) further explained.
that the Behavioral Model is used more often to guide research and evaluation studies on
access to healthcare. While other theories of health behavior, such as the health belief
model (NCI, 2005b), may lack components thought to be important in health services
utilization research; the Behavioral Model explicitly focuses on factors that predict health

services use, which made it an ideal framework for the current study. In a recent

10

interview with Ronald Andersen, the originator of the model, the utility of such a
framework is to:

“. .. Inform the research process by guiding the development of testable questions

or hypotheses, the collection of relevant data, and the appropriate analytic

techniques. Empirical research guided by a conceptual ﬁ'amework has a better
change of advancing the ﬁeld... [and] provides a tool to link disciplinary concepts
from the social sciences and earlier applications in health services research to the
current research questions being addressed, and also helps to suggest the health

policy relevance of the analytic results” (Ashton, 2008, p. 646).

The Behavioral Model has been used as a theoretical and analytic framework in
studies for over 30 years focusing on access to and utilization of hospital, dental, and
medical care among diverse adults (Aday & Andersen, 1974; Andersen, Bozzette,
Shapiro, St Clair, Morton, Crystal, et al., 2000; Andersen & Davidson, 1997; Andersen,
Davidson, & Nakazono, 1997; Coulter, Marcus, Freed, Der-Martirosian, Cunningham,
Andersen, et al., 2000; Davidson, Rams, & Andersen, 1997; Gallagher, Andersen,
Koegel, Gelberg, 1997; Gelberg, Andersen, & Leake, 2000; Miller, Gelberg, Kwan,
Stapanian, Fink, Andersen, et al., 2008; Padgett, Struening, & Andrews, 1990; Stein,
Andersen & Gelberg, 2007; Swanson, Andersen, & Gelberg, 2003; Wolinsky & Johnson,
1991; Wolinsky, 1978).

The model was also used to examine access to care and utilization barriers
experienced by Latinos (Andersen, Giachello & Aday, 1986; Andersen, Lewis,
Giachello, Aday, & Chiu, 1981; Estrada, et al., 1990) and Latino cancer screening test

utilization (Fernandez & Morales, 2007 ; Gorin & Heck, 2005). Andersen, et al., (1981)

ll

explored predictors of health care utilization among 5,432 Hispanic households in the
southwestern region of the United States and found that barriers to accessing care were a
lower income, a lack of health insurance and lower education. Andersen, Giachello, and
Aday (1986) used the Behavioral Model to examine access to care among a sample of
4,800 Hispanic families and found that the ability to utilize health care services was
impeded by a lower income, a lack of health insurance, and a lack of ties to a primary
physician. Estrada, et al., (1990) used data from the Hispanic Health and Nutrition
Examination Survey (HHANES) from 1982-1984 to explore health care utilization
barriers among Mexican Americans and found that the following barriers prevented
individuals ﬁ'om obtaining health care: a younger age, less acculturation, lacking health
insurance, having functional limitations, and having a poorer perceived health status.
Gorin and Heck, (2005) examined factors that predicted breast cancer screening in the
past year for Latinas 40 years and older and found that some predisposing factors (i.e.,
Dominican versus other Latina subgroups, age, education, smoking status, and a personal
history of cancer), enabling factors, (i.e., health insurance and recent physician visit in the
past year), and need factors (i.e., using at least one other screening test in the past) were
signiﬁcant. Fernandez and Morales, (2007) examined factors that predicted breast cancer
screening in the past two years for Latinas aged 40-64 and found that one predisposing
factor (i.e., age), enabling factors (i.e., health insurance and a regular health care
provider), and need (i.e., self-rated health status) were signiﬁcant.

Cultural sub-groups in the United States have different backgrounds, experiences,
beliefs, and social norms (Cauce, Coronado, & Watson, 1998; Marin, 1993; Rogler,

1999) that may inﬂuence beliefs, attitudes, and knowledge around cancer. For example,

12

Latinas have been shown to have different health beliefs around breast cancer than
mainstream cultural groups (Chavez, Hubbell, McMullin, Martinez, & Mishra, 1995).
Given cross-cultural differences in beliefs and behaviors, one cannot assume that health
behavior theories work equally across cultural and linguistic groups. There is a need for
cultural sensitivity in health promotion research (e. g., Roosa & Gonzales, 1999) and an
increased need to empirically test theoretical models to determine how dependable they
are across contexts and populations (Lucas, 2003). It was thus important to consider the
cultural appropriateness of the Behavioral Model and verify its equivalence in this sample
of Latinas 40 years and over (Nunnally & Bernstein, 1994). Demonstration of
measurement equivalence can be accomplished by testing whether the theoretical
structural model is invariant across cross-cultural or language groups (V andenberg &
Lance, 2000) through conﬁrmatory factor analysis.

Thisstudy assessed the validity of the Behavioral Model of Health Services Use
in this sample of Latinas 40 years and older by conducting a conﬁrmatory factor analysis
(Nunnally & Bernstein, 1994), which addressed the study’s ﬁrst research question: “Does
the component structure of the Behavioral Model fit the Latina population for the
outcome of breast cancer screening utilization?”

Next, this study used the Behavioral Model to answer the second research
question: “ What predisposing, enabling and need factors are associated with an
increased likelihood of breast cancer screening among Latinas? ” Variables nested in the
predisposing, enabling, and need domains, according to the Behavioral Model, are

outlined in Figure 1 and discussed below.

13

Figure 1. Behavioral Model of Health Services Use (Model I)

 

 

 

 

 

 

 

 

 

 

 

Predisposing Domain
Demographics (Age) _,
Social structure (Education, language-based acculturation & Latina
country of birth) breast
cancer
N eed-related Domain 4 screening
Perceived need (Health-related quality of life) use
Enabling Domain +
Personal resources (Health insurance and recent physician visit)

 

 

 

 

 

The remainder of the literature review is written as if the sample data did ﬁt the
Behavioral Model as illustrated in Figure 1. However, if the analyses conducted to
examine research question 1 suggests an alternative framework, that revised model will
be used to examine subsequent research questions.

Predisposing Factors

Predisposing factors refer to individual characteristics that affect the likelihood of
health care utilization and include demographics, social structure, and health beliefs
(Stein, et al., 2007). People with these characteristics are more likely to use health
services (Andersen, 1968). Demographics include age, sex, marital status, and prior
illness, which relate to health care use. Because marital status has been found to not relate
to breast cancer screening in Latinas (Gorin & Heck, 2005), it was not included in this
study. Instead, age is targeted as the demographic factor. The social structure domain
reﬂects an individual’s status in society and describes one’s lifestyle, which relates to
health service use (Andersen & Newman, 1973). In the original model, the following
were included as measures of social structure: education, race, occupation, family size,
ethnicity, religion, and residential mobility (Andersen & Newman, 1973). Recent

researchers have revised this conceptualization of social structure and are now including

14

variables that more accurately ﬁt an immigrant or Latino population as targeted in this
study, such as country of birth and language acculturation (Gelberg, Andersen, & Leake,
2000). In this study education and acculturation were the targeted social structure
factors. Health beliefs are attitudes, values, and knowledge that people have about health
and health services that may inﬂuence use (Andersen & Newman, 1973). Health beliefs
deﬁned as such were not assessed. Below are descriptions of the targeted factors and the
rational for their inclusion in this study.

Age

Age is a risk factor for breast cancer; from 2000-2005, the median age for a
diagnosis of breast cancer was 61 years old (NCI, 2007b; 2008a). For this reason, federal
guidelines recommend mammography every one to two years for women 40 and older
(AHRQ, 2006) and state law requires insurance companies to cover mammograms every
two years for women 40-49 years and annually for women 50 and older (Legislative
Council of California, 20003; 2000b; 20000). Health care providers — aware of this risk
and legislative requirement -— are also more likely to recommend mammography
screening for women over 40 as part of their standard care of practice. Because of this, as
women age, they may be more aware of their risk status and thus more likely to utilize
breast cancer screening.

In fact, in studies that have used the Behavioral Model as a ﬁ'amework, Latinas
over the age of 50 were signiﬁcantly more likely to use mammography than Latinas aged
40 to 49 years old, even after adjusting for covariates (Fernandez & Morales, 2007; Gorin
& Heck, 2005). Yet, other studies found no association between age and mammography

screening in multivariate analyses (Abraido-Lanza, Chao, & Gammon, 2004). Perhaps

l5

these discrepancies are due to study design. Gorin and Heck (2005) used 2000 data from
a national sample of more than 2,000 Latinas, and Fernandez and Morales (2007) used
2000, 2002, and 2004 data ﬁom a sample of 430 Latinas in Texas; both assessed
mammography in the last 2 years and age as a dichotomous variable. Abraido-Lanza,
Chao, and Gammon (2004) used 1991 data from a national sample of 535 Latinas and
11,209 white women, assessed mammography in the last year and age as a continuous
variable, and did not run separate logistic models for Latinas and white women, which
prevented the examination of predictors solely for Latinas. Given these discrepancies in
study samples and measures used, the current study hypothesized that age will be
positively associated with breast cancer screening and that Latinas over the age of 50 will
be more likely to utilize breast cancer screening.
Education

Education is a signiﬁcant factor in breast cancer screening: several studies have
shown that Latinas with higher education levels (e.g., more than 6th grade) have greater
use of mammography (Calle, Flanders, Thun, & Martin, 1993; Carrasquillo & Pati, 2004;
Fox, Amsberger, Owens, Nussey, Zhang, Golding, et al., 2004; Jones, Caplan, & Davis,
2003; Sambamoorthi & McAlpine, 2003; Qureshi, et al., 2000; Reyes-Ortiz, et al., 2007).
In a study that used the Behavioral Model as a framework, Latinas with at least some
college were more likely to use mammography than those with less education (Gorin &
Heck, 2005). Education is likely to matter since educational level precedes and may

predict employment4 and income level5 (Muller, 2002). Women who are employed (Heck

 

4 In 2006, the unemployment rate for US. adults 25 to 64 years old with less than a high school education
was 9.6 %, compared with 2.7 % of those with a BA. degree or higher (U .8. Census Bureau, 2006d).

l6

& Makuc, 2000) and have higher income may be more likely to have access to resources
that enable them to use services more readily. Latinas with at least high school education
were hypothesized to be more likely to utilize breast cancer screening.
Acculturation

Acculturation is the cultural change process that results from continuous ﬁrst-
hand contact between two cultural groups (Berry, 198 8), where individuals (whose
primary learning has been in one culture) adopt language, attitudes, values, and behaviors
from another culture (Suarez, 1994; Wells, 1989). Acculturation among Latinos is usually
measured by the propensity to think, read, and speak in English (Marin, Sabogal, Marin,
Otero-Sabogal, & Perez-Stable, 1987; Norris, Ford & Bova, 1996). It is also measured by
country of birth or length of stay in the United States (Abraido-Lanza, Chao & Florez,
2005; Abraido-Lanza, Chao, & Gates, 2005). This study assessed acculturation by
language-based acculturation and country of birth.

Language-based Acculturation

Language is a reliable measure of the change process that occurs as immigrants
are exposed to US. culture. Language-based acculturation is the preferred language for
thinking, reading, and speaking at home and in social situations (Norris, Ford & Bova,
1996). About 27% of Latino adults in the US. have limited English proﬁciency (LEP),
deﬁned as “living in a household where a language other than English is spoke most
often and not being comfortable conversing in English” (Brach & Chevarley, 2008, p.
10); individuals with LEP are considered less acculturated. Previous research suggests

that LEP is a barrier to accessing and using health services. Individuals with LEP have

 

5 In 2006, the median annual income for US. women 25 years and older with less than a high school
education was $13,255, compared with $36,875 for those with a BA. degree (US. Census Bureau, 2006c).

l7

greater difﬁculties communicating with and understanding providers (Doty, 2003;
Salazar, 1996), are less likely to report that their doctor has discussed the importance of
cancer screening and early detection (Stein & Fox, 1990), and are less likely to use
preventive services, have health insurance, or to have had a recent physician visit (Brach
& Chevarley, 2008; Derose & Baker, 2000; Hu & Covell, 1986).

To date, the research on the role of acculturation in predicting cancer screening
has produced mixed results. In some studies, higher language acculturation (i.e., English
proﬁciency or propensity to use English over Spanish language) positively predicts
Latina mammography screening (Stein & Fox, 1990) usage, even after controlling for
other covariates (Gorin & Heck, 2005; O’Malley, Kerner, Johnson, & Mandelblatt, 1999;
Ortero-Sabogal, et al., 2003; Suarez & Pulley, 1995). Yet, other studies have shown that
after adjusting for socio-demographic factors, the relationship between Latina
mammography screening and language acculturation disappears (Fernandez & Morales,
2007; Suarez, 1994). While English proﬁciency matters, Zambrana, et al., (1999) suggest
that factors such as health insurance may matter more in predicting Latina breast cancer
screening. For this reason, this study considered the individual role of language-based
acculturation and its role in the presence of enabling factors (e. g. health insurance and
having a recent physician’s visit). This study hypothesized that language-based
acculturation will be positively associated with breast cancer screening utilization but this
effect may disappear when enabling factors are taken into consideration and incorporated

into the model being tested.

18

Country of Birth

Country of origin is related to language-based acculturation and English
proﬁciency for Latinos. Of immigrants from Latin America residing in the US. in 2006,
90% spoke a language other than English in the home (US. Census Bureau, 2006a).
Thus, those who are foreign-born or who have recently arrived to the US. can be
considered less acculturated. Sometimes being foreign-born increases or decreases the
likelihood of breast cancer screening, and sometimes it makes no difference. Studies have
shown that foreign-born Latinas were signiﬁcantly less likely to receive cancer screening
compared to U.S.-bom Latinas (Goel, et al., 2003; Otero-Sabogal, et al., 2003). Some
studies also show that after controlling for demographic factors, foreign-born Latinas
were more likely to receive mammograms than U.S.-born white women (Rodriguez,
Ward, & Perez-Stable, 2005). Others show that foreign birth did not predict
mammography in Latinas (Hubbell, et al., 1997; Valdez, et al., 2001).

These ﬁndings might have emerged due to difference socio-demographic proﬁles
between immigrant and U.S.-bom populations. Latino immigrants are more likely to have
a lower education and income, and lack health insurance and a usual source of care than
U.S.-bom individuals (Borrayo & Guamaccia, 2000; Goel, et al., 2003; Thamer, Richard,
Casebeer, & Ray, 1997). Latina immigrants are likely to report barriers to accessing
health care services (c. g. language barriers or not knowing where to go for care) (Garces,
Scarinci, & Harrison, 2006). This study sought to clarify these mixed ﬁndings by
examining whether country of birth predicts breast cancer screening and if country of
birth predicts breast cancer screening in the presence of predisposing and enabling

factors, such as education and health insurance. Given the socio-demographic proﬁle

19

discrepancies between immigrant and U.S.-bom populations, this study hypothesized that
U.S.-born Latinas will be more likely than foreign-born Latinas to utilize breast cancer
screening.
Need Factors

An individual must perceive illness or the possibility of its occurrence in order for
utilization to occur (Andersen, 1968; Andersen & Newman, 1973). Thus, need was
measured by illness level, is indicated by self-perceptions (perceived need) and objective
evaluations (evaluated need) (Gelberg, Andersen, & Leake, 2000). Perceived need
consists of perceived illness (i.e., number of disability days, a self-report of general
health status, or perceived symptoms) and evaluated need is the objective clinical
evaluation of symptoms or diagnoses (Andersen & Newman, 1973). Ideally, physical
examinations would be included to conﬁrm a subjective health status; however these are
often costly to include in most research designs (Andersen & Newman, 1973). The
current study only assessed the perceived need domain due to lack of available data on
evaluated need, which includes self-rated health status and the number of recent disability
days in the last month (recent physical and mental health, and activity limitation). The
two measures of perceived need are referred to as health-related quality of life.
Health-Related Quality of Life (HRQOL)

Health-related quality of life (HRQOL) has been deﬁned as perceived physical
and mental health over time (Ounpuu, Chambers, Patterson, Chan, & Yusuf, 2001;
Zahran, Kobau, Moriarty, Zack, Holt, & Donehoo, 2005). HRQOL measurement
dimensions include perceived general health, recent physical/mental health, and recent

activity limitation. The latter two refer to recent disability days (the number of days in the

20

last month with physical/mental distress or activity limitation due to either type of
distress) (Dominick, Ahem, Gold, & Heller, 2002). HRQOL is related to health
behaviors. For example, research has shown that the number of unhealthy days (mental or
physical) is lower for adults doing the recommended level of physical activity compared
to physically inactive adults (Brown, et al., 2003). Research has also shown that a higher
HRQOL positively predicts health care utilization (Dominick, et al., 2002). Although few
studies have assessed the relationship between HRQOL and mammography, previous
research using the Behavioral Model found that higher health status ratings predicted
breast cancer screening (Fernandez & Morales, 2007). This study hypothesized that
Latinas with higher HRQOL (higher perceived general health, less recent unhealthy days,
and less recent activity limitation) will be more likely to utilize breast cancer screening.
Enabling Factors

Enabling factors are important because even though individuals may be
predisposed to use services, they must have the means to do so. Enabling factors are
deﬁned as conditions that permit a family to act on a value or satisfy a need regarding
health service use and make services available through community or personal resources
(Andersen, 1968). First, community resources can affect the availability and use of
services and consist of ratios of health personnel and facilities to population, price of
health services, region of the country, and urban-rural character (Andersen & Newman,
1973). Due to lack of available data, community resources were not assessed in this
study. Second, an individual’s ability to secure health services is affected by personal
resources, assessed by the extent of economic resources and sources of medical care

(Andersen, 1968). These resources include income, health insurance, type of regular

21

health care source, and access to regular source of care (Andersen & Newman, 1973) and
can be understood in the context of the literature on access to health care. According to a
1993 Institute of Medicine (IOM) report, access to health care is a general concept to
describe the degree to which individuals are able to obtain health care services when
needed, and has been deﬁned as “the timely use of personal health services to achieve
best possible health outcomes” (Millman, 1993, p. 4). Access to health care is measured
by the presence of resources that facilitate health care use (e. g. insurance coverage,
having a recent physician’s visit, and adequate economic resources) (AHRQ, 2008;
Millman, 1993). Due to lack of available data, income was not assessed in this study.
Thus, enabling resources were measured by health insurance and recent physician ’s
visit.
Health Insurance

In studies that used the Behavioral Model as a framework, health insurance was
found to predict Latina breast cancer screening (Fernandez & Morales, 2007; Gorin &
Heck, 2005). Health insurance plays a powerful role in the process of using preventive
health services, as evident by the fact that state6 legislation exists to ensure that insurance
companies cover certain breast cancer screening and treatment procedures. US. health
care costs per capita are the highest in the world and given that the breast cancer
screening or treatment procedures are covered by insurance, having health insurance can _
potentially protect individuals from paying associated high medical care costs (Hoffman

& Paradise, 2007). Yet, a signiﬁcant portion of the US. population is uninsured; in 2004,

 

6 California law requires insurers to cover the costs of a baseline mammogram for women age 35-39,
mammogram screening for women 40-49 years every two years, and annually for those 50+. Since 2000,
every health care service plan contract, except one that is specialized, shall also provide coverage for
diagnosis of, and treatment for, breast cancer (Legislative Council of California, 2000a; 2000b; 2000c).

22

45.8 million Americans were without health insurance, equating to 15.7% of the total
population (DeNavas-Walt, Proctor, & Lee 2005). Latinos are the least likely to be
insured compared to other ethnic groups (DeNavas-Walt, Proctor, & Lee, 2006;
Hargraves & Hadley, 2003).

Indeed, research shows that health insurance enables utilization of mammography
screening among Latinas (De Alba, et al., 2005; Fernandez & Morales, 2007; Hubbell, et
al., 1997; Jones, Caplan, & Davis, 2003; Otero-Sabogal, et al., 2003; Qureshi, et al.,
2000; Ramirez, et al., 2000; Zambrana, et al., 1999). In multivariate models, health
insurance is the strongest predictor of having a mammogram for Latinas (Abraido-Lanza,
Chao, & Gammon, 2004; Carrasquillo & Pati, 2004; Rodriguez, et al., 2005). Having
either private or public health insurance as opposed to no insurance predicts Latina
mammography usage, even after adjusting for other factors (Gorin & Heck, 2005;
Sambamoorthi & McAlpine, 2003). In this study, Latinas with either public or private
health insurance were hypothesized to be more likely to utilize breast cancer screening.
Recent Physician Visit

In prior research that used the Behavioral Model as a ﬁamework, visiting a
physician in the past year (Gorin & Heck, 2005) or having a usual source of care
(Fernandez & Morales 2007) predicted Latina breast cancer screening, after accounting
for other variables. Other studies have shown that visiting a physician in the past year
(Frazier, Jiles, & Mayberry, 1996) or a usual source of care (De Alba, et al., 2005; Hiatt,
et al., 2001; Valdez, et al., 2001; Selvin & Bratt, 2003; Solis, et al., 1990; Zambrana, et
al., 1999) enables use of mammography among Latinas. Physicians are a reputable source

of health information that can motivate women to get a mammogram (Metsch, McCoy,

23

McCoy, Pereyra, Trapido, & Miles, 1998). Using the F oot-in-the-Door theory (Burger,

1999; Pascual, & Gueguen, 2005), once a woman has committed to undergoing an annual

exam, she may be more receptive to other recommendations, such as those for

mammography. Thus, Latinas who have visited a physician in the past year for a routine

physical exam were hypothesized to be more likely to utilize breast cancer screening.
Discretionary and Non-discretionary Utilization

Use of health services can be conceptualized as discretionary or non-discretionary
behavior (Andersen, 196 8). The placement of and speciﬁed relationships between the
predisposing, enabling, and need factors in the model varies depending on the character
of the utilization variable (Andersen & Newman, 1973). The model assumes for non-
discretionary use, the choice of using services is based on the extent of that need. For
example, if an individual is in need of ambulatory care, the decision to use services is
based on the need for care, rather than on predisposing or enabling characteristics.
Indeed, research has shown that need-related variables mattered most for recent
hospitalization in the previous two weeks (Wolinsky, 1978).

The model assumes that for discretionary use - such as breast cancer screening -
the more likely use will be based on predisposing and enabling factors rather than need
(Andersen, 1968). Thus, for discretionary use, predisposing and enabling factors matter
more than need. This study argued that enabling factors matter more than need and
predisposing factors for breast cancer screening, which relates to the third research
question: “To what extent do predisposing and need factors add to the prediction of

breast cancer screening utilization by Latinas, after controlling for enabling factors? ”

24

While need-related variables, such as HRQOL, have been linked to health care
utilization (Dominick, et al., 2002) and self-rated health status still predicts Latina breast
cancer screening after controlling for other variables (Fernandez & Morales, 2007), other
studies have shown that self-rated health status does not predict Latina breast cancer
screening once all enabling, predisposing, and need factors are taken into account
(Carrasquillo & Pati, 2004; Gorin & Heck, 2005). Although predisposing factors (e. g.
higher language acculturation) positively inﬂuence cancer screening behaviors among
Latinas (De Alba, et al., 2004; Stein & Fox, 1990), the presence of enabling factors may
counteract this role (Fernandez & Morales, 2007). Zambrana, et al., (1999) suggests that
having health insurance or a usual source of care may matter more in predicting Latina
mammography usage than English proﬁciency and country of birth; health insurance and
a usual source of health have been shown to be two of the strongest predictors of cancer
screening usage in the general population (Carrasquillo & Pati, 2004). Thus, this study
hypothesized that after adjusting for enabling factors (e.g. health insurance and a recent
physician visit) predisposing and need variables will not be associated with breast cancer
screening use.

The Behavioral Model of Latina Breast Cancer Screening Use

Recently, Andersen (2008) suggested that research on access to care and health
services utilization should stress the importance of contextual and individual
characteristics for deﬁned populations. The Behavioral Model did not include variables
context-speciﬁc to the population, the disease, or the utilization behavior under
investigation. A recent revision called the Behavioral Model for Vulnerable Populations
was designed to include domains especially relevant to understanding the health-seeking

behavior of vulnerable populations, such as the homeless population (e.g., acculturation,

25

literacy, residential history, criminal behavior and prison history, victimization, mental
illness, and substance abuse) (Gelberg, Andersen, & Leake, 2000). Thus, the Behavioral
Model can be viewed as a “meta-framewor ”, in which researchers can amend by
including context-speciﬁc variables that potentially allow for a greater prediction of use.

To answer the fourth research question, “T 0 what extent do context-specific
variables add to the prediction of breast cancer screening utilization by Latinas? ”, this
study created the Behavioral Model of Latina Breast Cancer Screening Use, in order to
target context-speciﬁc7 variables especially relevant to the Latinas, breast cancer, and
breast cancer utilization behavior (Pasick & Burke, 2008), (See Figure 2, where context-
speciﬁc variables are italicized and bolded).

First, having a history of cancer has been found to be positively related to Latina
breast cancer screening and was thus included as a predisposing variable (Gorin & Heck,
2005). Second, although literacy was included as a predisposing factor in a recent
revision of the Behavioral Model, health literacy has not previously been tested as a
predisposing factor in the Behavioral Model. Since adequate health literacy is positively
associated with mammography use in Latinas (Guerra, Krumholz, & Shea, 2005), health
literacy was included as a predisposing variable. Third, although religiosity was intended
to be a social structure predisposing variable in the original model, spirituality has been
recently included as a health belief predisposing variable in a modiﬁed version of the
Behavioral Model (Miller, et al., 2008). Since religiosity positively inﬂuences breast
cancer screening among Latinas (Otero-Sabogal, et al., 2003) it was included as a

predisposing health belief variable because of the components of religiosity assessed in

 

7 Context-specific variables are speciﬁcally related to the population, the disease, or the utilization behavior
under investigation.

26

this study, (i.e., beliefs, behavior, and prayer). Fatalistic beliefs deter breast cancer
screening use (Barroso, McMillan, Casey, Gibson, Karninski, & Meyer, 2000) and were
also included as a belief variable. Lastly, breast cancer worry was included as a perceived
need factor since it relates to perceiving the possibility for breast cancer to occur and has
been linked with breast cancer screening (Andersen, Smith, Meischke, Bowen, & Urban,
2003). In conclusion, additional proposed predisposing variables for this study included:
history of breast cancer, health literacy, cancer fatalistic belieﬁ and religiosity. This
study also measured perceived need by breast cancer worry.

Figure 2. Behavioral Model of Latino Breast Cancer Screening Use (Model 2)

 

Predisposing Domain

 

 

 

 

 

 

 

 

 

Demographics (Age, history of breast cancer) _,
Social structure (Education, language-based acculturation &
country of birth, and health literacy) Latina
Be_li§§ (Cancer fatalistic beliefs, religiosity) breast
cancer
Enabling Domain —> screening

Personal resources (Health insurance and recent physician visit) use

N eed-related Domain "
Perceived need [Health-related quality of life (Self-rated health

 

 

 

 

 

status and disability days) and breast cancer worry]

 

History of Breast Cancer

Having a personal history of cancer was found to signiﬁcantly increase breast
cancer screening in Latinas in a study that used the Behavioral Model as a framework
(Gorin & Heck, 2005). This study focused on personal and familial histories of breast
cancer as factors that may predispose Latinas to utilize mammography services. Women
with a personal history of cancer (Kim, et al., 2008) or women with a family history of
breast cancer who have an increased risk of developing the disease (Watson, Henderson,

Brett, Bankhead, & Austoker, 2005; Zakowski, Valdimarsdottir, Bovbjerg, Borgen,

27

 

Holland, Kash, et al., 1997), may perceive themselves at risk for cancer (Kim, et al.,
2008). Perceiving oneself at risk for breast cancer is associated with higher screening
rates (McCaul, Branstetter, Schroeder, & Glasgow, 1996).

In fact, Latinas with a personal or family history of cancer show greater usage of
mammography compared to those without a history (Aparicio-Ting & Ramirez, 2003;
Cohen, 2006; McCaul, et al., 1996; McCaul & Tulloch, 1999). Some studies have found
no differences in screening or perceived risk between women with and without a family
history of breast cancer (Audrain, Lerman, Rimer, Cella, Steffens, & Gomez-Caminero,
1995; West, Greene, Kratt, Pulley, Weiss, Siegﬁied, et al., 2003), which may be due to
confounding factors, such as education. Audrain, et al., (1995) found that women with a
family history who were signiﬁcantly less aware of their risk status for developing cancer
were less educated than women with a family history who were aware of their risk status.
This study hypothesized that having either a personal or family history of breast cancer
will positively predict breast cancer screening use.

Health Literacy

Health literacy has been defrned as ‘ ‘the degree to which individuals have the
capacity to obtain, process, and understand basic health information and services needed
to make appropriate health decisions” (Selden, Zorn, Ratzan, & Parker, 2000, p. v).
Functional health literacy can be measured by frequency of assistance used in reading.
hospital materials, conﬁdence in ﬁlling out medical forms, and understanding written
health information (Chew, Bradley & Boyko, 2004). Studies show that low health
literacy levels relate to a lack of knowledge about the importance of and reason for

screening mammography, which may relate to underutilization of mammography (Davis,

28

Arnold, & Berkel, 1996). Of the few studies conducted, having adequate health literacy
was associated with mammography screening in Latinas (Guerra, Krumholz, & Shea,
2005) and cervical cancer screening in multiethnic women (Garbers & Chiasson, 2004;
Lindau, Tomori, Lyons, Langseth, Bennett, & Garcia, 2002). Latinas with higher health
literacy levels were hypothesized be more likely to utilize breast cancer screening.
Cancer F atalistic Beliefs

Cancer fatalism is the belief that death is inevitable when cancer is present (Powe,
2001) and is measured by having a belief of fear, predeterrnination, pessimism, and
inevitability of death (Powe, 1995a). Fatalistic beliefs promote an external locus of
control, causing individuals to believe that they have no control over their health, which
can deters mammography utilization (Barroso, et al., 2000; Borrayo & Guamaccia, 2000;
Niederdeppe & Levy, 2007). Latinas have a tendency to have fatalistic beliefs and
misconceptions about cancer, such as divine predetermination as a cause for cancer
(Borrayo & Jenkins, 2001; Chavez, et al., 1997; Florez, et al., 2008; Hubbell, et al., 1996;
Salazar, 1996; Schettino, Hemandez-Valero, Moguel, Hajek, & Jones, 2006; Suarez,
Nichols, Pulley, Brady, & McAlister, 1993). Few studies have assessed whether these
beliefs relate to cancer screening among Latinas (Abraido-Lanza, et a1, 2007), yet cancer
fatalistic beliefs have been found to impede Latina’s use of cancer screening exams
(Austin, et al., 2002; Chavez, et al., 1997; Hubbell, etal., 1997; Perez-Stable, Sabogal,
Otero-Sabogal, Hiatt, & McPhee, 1992; Otero-Sabogal, et al., 2003). Latinas with less

fatalistic beliefs were hypothesized to be more likely to utilize breast cancer screening.

29

Religiosity

Religiosity involves beliefs, practices, and personal devotion and is measured by
organizational, (e.g., religious services attendance), nonorganizational (e.g., prayer or
religious study), and intrinsic (e. g., experiencing the presence of the Divine) religiosity
(Storch, et al., 2004). Religiosity is a key facet of Latino culture that promotes positive
health behaviors (Magaﬁa & Clark, 1995) and church attendance among Latinas is
positively related with a healthier dietary and physical activity (Arredondo, et al., 2005).

Different facets of religiosity relate to health in different ways (Powell, Shahabi,
& Thoresen, 2003; van Olphen, Schulz, Israel, Chatters, Klem, Parker, et al., 2003).
Religious attendance is associated with health behaviors, good mental health, increased .
social relationships, marital stability, and use of preventive health services (Felix Aaron,
Levine, & Burstin, 2003; Hill, Burdette, Ellison, & Musick, 2006; Levin, Markides, &
Ray, 1996; Strawbridge, Shema, Cohen, & Kaplan, 2001). Prayer is associated with less
depressive symptoms, perhaps by encouraging emotions such as hope and forgiveness
(van Olphen, al., 2003) and prayer can help one cope with stress, pain, or illness
(Abraido-Lanza, Vasquez, Echeverria, 2004; Park, & Gaffey, 2007; Tarakeshwar,
Vanderwerker, Paulk, Pearce, Kasl, & Prigerson, 2006). Spirituality or faith is also linked
with health (Figueroa, Davis, Baker, & Bunch, 2006) and may promote positive
behaviors (van Olphen, al., 2003). ‘

Of studies conducted on religious beliefs and breast cancer screening, most has
focused on African American women (Gullate, 2006; Kinney, Emery, Dudley, & Croyle,
2002; Mitchell, Lannin, Mathews, & Swanson, 2002; Underwood & Powell, 2006). A

few studies have shown that church attendance is associated with breast and cervical

30

cancer screening among diverse women (Benjarnins, 2006) and breast cancer screening
among Latinas (Otero-Sabogal, et al., 2003). This study hypothesized that religiosity —as
measured by intrinsic, organizational, and nonorganizational religiosity- will positively
predict Latina breast cancer screening.
Breast Cancer Worry

Worry can be characterized by emotions such as fear and is sometimes considered
the cognitive aspect of the anxiety and may ﬁrnction as an adaptive coping mechanism in
attempting to psychologically prepare for an external threat (Hay, Buckley, & Ostroff,
2005; Kubzansky, Kawachi, Spiro, Weiss, Vokonas, & Sparrow, 1997). Pathological
aspects of worry, (e.g., excessive worry-related thoughts), are characteristic of
generalized anxiety disorder (Starcevic, Berle, Milicevic, Harman, Larnplugh, & Eslick,
2007). Breast cancer worry has been deﬁned as an emotional reaction to the threat of
cancer and consists of both cognitive and affective elements (Hay, Buckley, & Ostroff,
2005). Measurement targets cognitive elements of worry, (e. g., how often one thinks
about their risk of developing breast cancer) (Andersen, et al., 2003). Worrying could
stem from a fear of developing cancer. Women consider cancer to be the most feared
disease compared to men (Murray & McMillan, 1993) and fear is a barriers to cancer
screening in Latinas (Austin, et al., 2002; Garbers, Jessop, Foti, Uribelarrea, & Chiasson,
2003)

No consensus exists on whether worry promotes or inhibits screening (Hay,
Buckley, Ostroff, 2005). Moderate levels of worry have been shown to relate to higher
mammography rates (Andersen, et al., 2003; Diefenbach, Miller & Daly, 1999; McCaul,

Branstetter, O'Donnell, Jacobson & Quinlan, 1998; McCaul, Branstetter, Schroeder &

31

Glasgow, 1996; McCaul, Schroeder & Reid, 1996) and extremely high levels of worry
and distress are associated with a decreased likelihood of mammography (Andersen, et
al., 2003; Schwartz, Taylor & Willard, 2003). Other studies found that higher levels of
cancer worry relates to over-perforrnance of breast self-exarns (Cohen, 2006). Due to the
lack of research on Latinas, it was unclear whether higher levels of worry promote or
inhibit screening. Given the state of the literature, this study explored the hypothesis that
Latinas with extremely high amounts of breast cancer worry and distress will be less
likely to utilize breast cancer screening.
Conclusion

Because of the link between early breast cancer detection through screening and a
decreased risk of breast cancer mortality (AHRQ, 2006), it is important to determine
what promotes mammography usage. This study aimed to contribute to the current
understanding of what predicts breast cancer screening use among Latinas 40 years and
older to inform research, health promotion practice, and primary care. Given that the
determinants of breast cancer screening use are complex, studies examining such
determinants should be guided by a multivariate theoretical ﬁamework in order to
contribute to the research literature, facilitate inter-study comparisons, and lend credible
evidence for the creation of policy. Of the prior studies that have examined multiple
predictors of Latina breast cancer screening utilization, only a few were overtly
theoretically guided; theories included the T ranstheoretical Model (Otero-Sabogal, et al.,
2003) and the Behavioral Model of Health Services Use (Fernandez & Morales, 2007;
Gorin & Heck, 2004). This study was theoretically guided by the Behavioral Model of

Health Services Use to determine what predisposing (age, education, or acculturation),

32

enabling (health insurance or having a recent physician’s visit), and need (perceived
health status) factors relate to breast cancer screening use among Latinas.

For the discretionary use of health services, the Behavioral Model suggests that
enabling (health insurance or having a recent physician’s visit) and predisposing (age,
education, or acculturation) factors may matter more than need (perceived health status)
(Andersen, 1968). This study sought to contribute to the research literature on the
Behavioral Model by arguing that for the discretionary utilization of breast cancer
screening, enabling factors matter more than need and predisposing factors. The rationale
behind this argument was that although predisposing factors -such as age, education, and
acculturation— have been shown to be important predictors relating to breast cancer
screening in Latinas, once enabling factors-such as health insurance- are taken into
account, predisposing factors may be rendered insigniﬁcant. Therefore, this study argued
that enabling factors matter most when considering breast cancer screening behavior and
sought to determine if predisposing and need factors predict breast cancer screening, after
accounting for enabling factors.

Lastly, this study proposed a modiﬁcation of the Behavioral Model, called the
Behavioral Model of Latina Breast Cancer Screening Use, since adding context-speciﬁc
variables to the Behavioral Model can allow for a greater prediction of the utilization
measure in the targeted population (Andersen, 2008). In contrast to most of the studies
that have examined multiple predictors of Latina breast cancer screening in the past, this
study proposed to target socio-demographic (e.g., predisposing), health status (e. g., need),

and access to care (e. g., enabling) predictors, in addition to variables context-speciﬁc to

33

breast cancer screening behavior among Latinas (e.g., breast cancer worry, personal/
family history of breast cancer, cancer fatalism, religiosity, and health literacy).

Indeed, prior studies focusing on Latinas have lacked a full range of predictors
relevant to the context of breast cancer screening. A review of studies focusing on Latina
breast cancer screening published from 1990 to 2007 revealed that most examined
knowledge, socio-dernographic, or access to care variables, yet lacked variables context-
speciﬁc to breast cancer screening, such as breast cancer worry (Abraido-Lanza, Chao, &
Gammon, 2004; Aldridge, et al., 2006; Carrasquillo & Pati, 2004; Coughlin, et al., 2003;
De Alba, et a1, 2005; Hubbell, et al., 1997; Jones, etal., 2003; Qureshi, et al., 2000;
Ramirez, et al., 2000; Rodriguez, Ward, & Perez-Stable, 2005; Sambamoorthi &
McAlpine, 2003; Selvin & Brett, 2003; Stein & Fox, 1990; Suarez, 1994; Valdez, et al.,
2001; Zambrana, et al., 1999). Other studies included socio-demographic and access to
care variables in addition to one or two context-speciﬁc variables, such as literacy
(Guerra, et al., 2005), personal/family history of breast cancer (Aparcio-Tin & Ramirez,
2003; Shah, et al., 2007), or religiosity and fatalism (Otero-Sabogal, et al., 2003). The
two studies that used the Behavioral Model as a theoretical ﬁamework did not examine
the same predictor variables, produced inconsistent results, and lacked many of the
context-speciﬁc variables proposed to be included in the current study (Fernandez &
Morales, 2007; Gorin & Heck, 2004). This study sought to determine if variables context-
speciﬁc to Latina breast cancer screening behavior would allow for greater prediction of
breast cancer screening than variables in the original Behavioral Model alone, which may

provide valuable insight for future research.

34

Hypotheses

The goal of this study was to contribute to the current understanding of what

predicts breast cancer screening use among Latinas 40 years and older. To achieve this

goal, the following proposed hypotheses and sub-hypotheses were tested:

1.

Latinas with higher rates of predisposing factors will be more likely to utilize breast

cancer screening services.

a.

Latinas over the age of 50 years old compared to those aged 40 to 49 years will be
more likely to utilize breast cancer screening.

Latinas with at least a high school education will be more likely to utilize breast
cancer screening than Latinas with less than a high school education.

Latinas who are more acculturated will be signiﬁcantly more likely to use breast
cancer screening exams than Latinas with lower levels of acculturation.

U.S.-born Latinas will be signiﬁcantly (more likely to utilize breast cancer

screening exams than foreign-born Latinas.

Latinas with higher rates of enabling factors will be more likely to utilize breast

cancer screening services.

a. Latinas with either public or private health insurance will signiﬁcantly be more

likely to utilize breast cancer screening exams than Latinas without health
insurance.

Latinas that have visited a physician in the past year for a routine physical exam
will be more likely to have utilized breast cancer screening services compared to

those who have visited a physician in more than a year.

35

3. Latinas with higher rates of need factors will be more likely to utilize breast cancer

screening services.

a.

Latinas with higher levels of reported health status will be more likely to utilize
breast cancer screening than Latinas with lower levels of reported health status.
Latinas with a lower number of reported unhealthy days due to mental or physical
distress will be more likely to utilize breast cancer screening than Latinas with
higher levels of reported unhealthy days.

Latinas with fewer reported days of physical activity limitation levels due to
either mental or physical distress will be more likely to utilize breast cancer

screening than Latinas that report more days of physical activity limitation.

4. After adjusting for enabling factors in the original model, predisposing and need

variables will not be signiﬁcantly associated with breast cancer screening utilization.

5. After adjusting for variables in the Behavioral Model, context speciﬁc variables will

be signiﬁcantly associated with breast cancer screening utilization which will

improve the predictability of the Behavioral Model.

a.

Latinas with a personal or family history of breast cancer will be more likely to
utilize breast cancer screening than those without a history.

Latinas with higher health literacy levels will be more likely to utilize breast
cancer screening than Latinas with lower levels of health literacy.

Latinas with lower levels of cancer fatalistic beliefs will be more likely to utilize
breast cancer screening services than to Latinas with higher levels of cancer

fatalistic beliefs.

36

. Latinas with higher levels of intrinsic religiosity will be more likely than Latinas
with low levels of intrinsic religiosity to utilize breast cancer screening services.
Latinas with higher levels of organizational religiosity will be more likely than
Latinas with low levels of organizational religiosity to utilize breast cancer
screening services.

Latinas with higher levels of nonorganizational religiosity will be more likely
than Latinas with low levels of nonorganizational religiosity to utilize breast
cancer screening services.

Latinas with extremely high amounts of breast cancer worry and distress will be
less likely to utilize breast cancer screening services than Latinas with lower

amounts of breast cancer worry and distress.

37

RESEARCH DESIGN AND METHODS

In order to investigate the proposed hypotheses, this study used secondary data
that were previously collected from a cross-sectional survey of 208 Latinas 40 years and
older in San Diego County, California that was originally approved by the University of
California, San Diego (UCSD) Institutional Review Board (IRB). Data were collected as
part of a community-campus partnership between Vista Community Clinic (V CC) and
the UCSD Cancer Center, funded through a California Breast Cancer Research Program
(CBCRP) community research collaborative (CRC) pilot award. Approval for secondary
data analysis was also obtained from the Michigan State University IRB.

Procedures
Setting

This study focused on San Diego County, Califomia’s most southern county
located adjacent to Baja California Norte, Mexico. The County spans 4,200 square miles
(http://www.co.san-diego.ca.us/dpw/organization/historyl .htrnl). California is home to
the largest US. Latino population (Guzman, 2001). In 2006, about 8.1 percent (2.9
million) of Californians resided in San Diego County and 30.1% of San Diego residents
were Latino, of which most were Mexican-origin (87%) (U .S. Census Bureau, 2006b;
20060). In 2008, in San Diego, breast cancer was projected to be the leading type of
cancer incidence for all women, and in 2004, 63% of Latina breast cancer cases were
diagnosed at early stage compared to 71% for white women (ACS, 2008). Thus, San

Diego was an ideal setting for recruiting a large sample of Latinas.

38

Sample

Participants included in the study were drawn ﬁom a community sample of 503
Latinas. Eligibility criteria for participation were: being an adult woman, who self-
identiﬁed as Hispanic American and who reported her preferred language as English or
Spanish. Given that the outcome of interest is breast cancer screening utilization, only
those women 40 years and older (n = 208) were included for analysis. Participants’ ages
ranged from 40 to 80 years (M = 50.98; SD = 8.1). Among these women, 47.3% (n = 98)
had a high school education or greater, 67.6% (n = 140) had health insurance, 76.3% (n =
151) were Mexican-born, 23.7%, (n = 47) were U.S.-bom, and 56.3% (n = 117) chose to
be interviewed using a Sparrish-language survey and 43.8% (n = 91) chose English.

To determine the extent to which this study sample was representative of the
County and State, data from Latinas 40 years and older in the study sample were
compared to 2007 data from 121,000 Latinas 40 years and older in San Diego County and
1,861,000 Latinas 40 years and older across California (http://www.chis.ucla.edu/); 2006
data ﬁ'om 363 Latinas 40 years and older from California were also used for
comparisons8 (http://apps.nccd.cdc.gov/BRFSS). Overall, it appears that the study sample
was representative of Latinas 40 years and older state-wide on some key factors, such as
mammography screening, age, education, health status, and of Latinas 40 years and older
in San Diego county for age and mammography screening; but the sample was different
on health insurance and country of birth compared to the state and county, and education
and health status compared to the county. These differences were most likely due to the

sampling procedures utilized to collect sample, county and state-level data.

 

8 Chi-square analyses were run using an online chi-square calculator:
http://home.ubalt.edu/ntsbarsh/Business-stat/otherapplets/Catego.htrn.

39

Latinas in the sample were signiﬁcantly more likely to have a high school
education or higher (47.3%, n = 98) than those in San Diego (39.8%, n = 40,000) [x2 (1,
101,207) = 5.17, p = .023]. Latinas in the sample were signiﬁcantly less likely to have
health insurance (67.6%, n =140) than those in San Diego (77.5%, n = 93,000) [x2 (1,
120,207) = 11.53, p = .001] and those in California (79.1%, n = 1,471,000) [752 (1 ,
1,860,207) = 16.41, p = .000]. When comparing only US. or Mexican born groups,
Latinas in the sample were signiﬁcantly more likely to be Mexican born (76.3%, n = 151)
than those in San Diego (62.5%, n = 70,000) [x2 (1, 112,198) = 15.98, p = .000] and those
in California (53.8%, n = 836,000) [x2 (1, 1,555,198) = 39.72, p = .000]. Finally, Latinas
in the sample were less likely to report their health as excellent (8.3%, n = 17) compared
to those in San Diego (16.8%, n = 20,000) [x2 (4, 120,205) = 16.18, p=.003]. There were
no signiﬁcant differences between sample, county, and State data for age and
mammography screening utilization rates (Table B, Appendix D). _

Although participants for this study were drawn from a convenient sample, these
data suggest that the sample was representative of Latinas 40 years and older in San
Diego County in terms of their mammography screening utilization and age; and the
sample was representative of Latinas 40 years and older across the state in terms of their
mammography screening utilization, age, educational attainment, and health status.

Data Collection
' Data were collected from October, 2007 through February, 2008 by seven bi-
lingual interviewers representing Vista Community Clinic and UCSD Cancer Center
staff, (four Latinas, one Aﬁican American woman, and two white women). Interviewers

were IRB- and HIPAA-certiﬁed through UCSD’s IRB and trained in survey data

40

collection methods by the research team (e.g., consistent data collection protocols,
eligibility screening, informed consent, conﬁdentiality, checking surveys for complete
data, and how to handle standardized scales when reading surveys to participants).
Recruitment

Interviewers recruited and consented eligible women to participate in the study -
through face-to-face meetings at community-based sites, and by snowballing and word-
of-mouth strategies. Interviewers were instructed to keep a record of the recruitment
method used, number of eligible women, refusal rates, reasons for reﬁrsals, and potential
contacts. Interviewers logged a total of 10 women who refused to participate in the face-
to-face interviews, with lack of time being the most ﬁequent reason for refusal. To invite
participants, interviewers explained that the study would involve completing surveys
ﬁequently used in health and research settings to learn if the surveys worked equally well
for Hispanic women. Participants were given the option of having the surveys and
consent forms read to them in English or Spanish or they could ﬁll out the forms
themselves. All participants consented to participate in the study and were given a $20.00
incentive for their time. Interviews took around one hour to complete. For those who
were unable to read, it took interviewers two hours to read the survey to the participant.
Interviews took place at a convenient location, (e. g., homes, place of work, place of
worship, community center, or at the community clinic). Most (63.5%, n = 132) were
recruited by VCC staff and 36.5% (n = 76) were recruited by UCSD (Appendix A).
Sample Size

The power calculations for this study planned a sample size that would allow for

at least 80 percent complete data for each scale, while still being able to demonstrate

41

statistical signiﬁcance at a p < .05 level (Comrey, & Lee, 1992; Gorsuch, 1983; Nunnally
& Bernstein, 1994). While programs can calculate sample size and power estimates for
linear regression, ANOVA and related statistical tests (Cohen, 1988; Paul, Erdfelder,
Lang, & Buchner, 2007; Lenth, 2006), it was difﬁcult to determine sample size estimates
for multivariate logistic regression. Using a widely accepted participant to variable ratio
of 10:1, a total sample of 208 still allowed for incomplete response rate of 20% and
enough power to detect a moderate effect.
Survey Instrument

Survey administration procedures were piloted in Spanish and English with
community members and clinic staff. Participant fatigue and literacy level of scales
caused the research team to shorten the length of the survey considerably, from 24 to 18
pages. To help minimize response bias, three versions of the surveys were developed,
which only differed in order of scale placement. Measures that only had English versions
were made into content equivalent Spanish versions by the research team, including:
History of Breast Cancer items, Health Literacy, and the Cancer Worry Scale. These were
ﬁrst translated ﬁom English to Spanish by a professional translator and then back-
translated by one of the researchers, who created a list of discrepancies between the two
versions. Two additional researchers reconciled the problems to ensure proper meaning
was not lost. The process was ﬁnalized once four bilingual researchers reviewed the ﬁnal
translations alongside the original versions and veriﬁed their equivalence (Appendix C).

Given that some of the scales were either modiﬁed or translated from existing
measures, or have not been validated with samples of Latinas, factor analysis was

performed to determine the internal consistency and underlying structure of all scales and

42

to determine if the Spanish and English versions had equivalent psychometric properties
(Aﬁﬁ, Clark, & May, 2004; Nunnally & Bernstein, 1994; Vandenberg & Lance, 2000).
Item means and standard deviations, item-total correlations, and patterns of missing or
omitted responses were considered. Internal consistency for total and subscale scores of
all instruments was computed (using Cronbach’s alpha “a”) (Nunnally & Bernstein,
1994). Reliabilities were computed for the total sample and separately in English and
Spanish for: language-based acculturation (BASH), religiosity (DUREL), cancer fatalistic
beliefs (PF 1), and breast cancer worry (CWS). Scores are presented as unstandardized
values for ease of interpretation (Appendix D).
Breast Cancer Screening

Breast cancer screening was assessed in a manner to allow for comparison to
national trends, and thus items were derived ﬁom the 2007 CDC Behavioral Risk Factor
Surveillance System (BRFSS) Questionnaire, available for public use in English and
Spanish (CDC, 2006). Participants were asked: “A mammogram is an x-ray of each
breast to look for breast cancer. Have you ever had a mammogram?” If “yes”, they were
asked: “How long has it been since you had your last mammogram?” Participants could
then select from the following: within the past year, within the past 2 years (1 year but
less than 2 years ago), within the past 5 years (2 years but less than 5 years ago), 5 or
more years ago, don’t know, and never. A dichotomous outcome variable was created to
compare those compliant with federal screening guidelines9 (most recent mammogram
within the past two years) to those non-compliant with federal guidelines (never screened

or most recent mammogram more than two years ago) (Table 1 and Appendix D).

 

9The US. Preventive Services Task Force recommends screening mammography every 1-2 years for
women 40+ (AHRQ, 2006).

43

Predisposing Variables

Age was assessed by asking “What is your date of birth?”, an item derived from
the 2007 CDC BRF SS Questionnaire (CDC, 2006). By subtracting the participant’s date
of birth from the survey date, the exact age was calculated as a continuous variable in
years through syntax in SPSS. Age categories (40-49 and 50+) were created to compare
mammography utilization across two age groups (Appendix D).

Education was assessed by asking, “What is the highest level of school you
completed?”, an item derived from the 2007 CDC BRFSS Questionnaire (CDC, 2006).
Participants could respond to this question by selecting different categories, such as
“elementary school or primary school.” Participant responses were broken down into one
binary variable (less than high school versus high school degree or more) (Appendix D).

Country of Birth was the ﬁrst proxy for acculturation. Participants were asked the
open-ended question: “In what country were you born?” Responses were transformed
into a dichotomous categorical variable (i.e., United States and Mexico) (Appendix D).

Language-based Acculturation was the second proxy for acculturation and was
assessed by an established language acculturation scale. The Brief Acculturation Measure
for Hispanics (BASH) is a 4 item scale that treats language use and preference as
indicator for level of acculturation (Norris, Ford & Bova, 1996). The BASH is a modiﬁed
version of the 12 item Short Acculturation Scale for Hispanics (SASH) (Marin, et al.,
1987), which consists of 3 factors: language use, ethnic social relations, and media. The 5
item language subscale was shown to be reliable (or = .90) and similar to the 12-item total
scale (Marin, et al., 1987), which suggested that it could be used as a shorter measure of .

acculturation. The BASH used in this study includes 4 of these 5 items; excluding “What

was the language(s) you used as a child?” The 4-item BASH has adequate reliability ((1 =
80-92) (Norris, Ford & Bova, 1996). Reponses are rated on a 5-point scale ranging ﬁom
“only Spanish” to “only English” and a mean score can be created from the 4 items.

Psychometric analyses were conducted to explore the internal consistency and
underlying structure of the BASH scale and to determine if the Spanish and English
versions had equivalent psychometric properties. Specifying 1 factor and performing an
orthogonal varimax rotation using the principle components as the method of factor
extraction yielded a 1 factor solution as expected for both language versions. Factor I
loaded highly on all 4 items. The 4-item BASH had an alpha of .96 for the total sample
(English a = .95, Spanish (1: .84). A single mean score was created (ranging from 1-5)
with higher scores indicating greater language-based acculturation (Appendix D).
Enabling Variables

Health Insurance was assessed by asking: “Do you have any kind of health care
coverage, including health insurance, prepaid plans such as HMO’s, or government plans
such as Medicare/Medi-Cal?” If “yes”, they were asked what type of by selecting from
thefollowing: Medi-Cal/Medicaid, Medicare, HMO, PPO, Employer-provided, and
other; items were derived from the 2007 CDC BRFSS Questionnaire (CDC, 2006). A
categorical variable was created for insurance type [private (HMO, PPO, military); public
(Medicaid and Medicare); and none]. A dichotomous variable was created to compare
those with any insurance to none; two dummy variables were created to compare private,
public, and no insurance groups (Appendix D).

Recent Physician Visit was assessed by asking: “About how long has it been since

you last visited a doctor for a routine checkup? A routine checkup is a general physical

45

exam, not an exam for a speciﬁc injury, illness, or condition”, an item derived from the
2007 CDC BRFSS Questionnaire (CDC, 2006). Participants could select: within the past
year, within the past 2 years (1 year but less than 2 years ago), within the past 5 years (2
years but less than 5 years ago), 5 or more years ago, don’t know, and never. A
dichotomous variable was created (visit in the past year versus otherwise) (Appendix D).
Need-related Variables

Health-related Quality of Life (HRQOL) has been deﬁned as perceived physical
and mental health over time and is measured by the CDC HRQOL-4, known as the
“Healthy Days Measures.” The CDC HRQOL-4 assesses: self-rated health, number of
recent days with physical or mental health was impaired, and number of days of activity
limitation due to poor health (Zahran, et al., 2005) and is available in English and Spanish
(http://www.cdc.goV/hrqol). The CDC HRQOL-4 has demonstrated reliability and
validity for population use (Andersen, et al., 2003). First, participants are asked “Would
you say that in general your health is” and can choose from a range of “excellent” to
“poor.” Second, participants are asked: “Now thinking about your physical health, which
includes physical ﬁtness and injury, for how many days during the past 30 days was your
physical health not good?” “Now thinking about your mental health, which includes
stress, depression, and problems with emotions, for how many days during the past 30
days was your mental health not good?” Lastly, participants are asked: “During the past
30 days, for about how many days did poor physical or mental health keep you from
doing your usual activities, such as self-care, work, or recreation?”

Self-rated general health status (item 1) ranges from 1-5 where greater scores

reﬂected a worse self-rated health status. Second, the surrrrnary index of overall unhealthy

46

days ranges from 0-30 and was created by combining responses to items 2 and 3, with a
logical maximum of 30 days. Third, item 4 ranges from 0-30 and is the number of days of
activity limitation due to unhealthy days in the last 30 days. Item 4 was only valid if
items 2 or 3 > 0; higher scores for the latter 2 indicated more unhealthy days and more
activity limitation days due to unhealthy days (Appendix D).
Context-Specific Variables

History of Breast Cancer was assessed by two items created by modifying an item
from the 2007 CDC BRFSS questionnaire (e.g., “Have you ever been told by a doctor,
nurse, or other health professional that you had prostate cancer?” (CDC, 2006).
Participants were asked: “Has a family member ever had breast cancer?”; if “yes” they
were asked to speciﬁc which family member by selecting from the following categories:
mother, grandmother, sister, aunt, cousin, or other. Participants were then asked: “Have
you ever had breast cancer?” A single (dichotomous variable was created (women with a
positive personal or family history of breast cancer versus otherwise) (Appendix D).

Health Literacy (HL) can be measured by frequency of assistance used in reading
hospital materials, conﬁdence in ﬁlling out medical forms, and understanding written
health information and was assessed by 2 items: “How conﬁdent are you ﬁlling out
medical forms by yourself?” and “How often do you have someone help you read
hospital materials?” (Chew, Bradley & Boyko, 2004). The HL was translated into
Spanish for this study. The two items were signiﬁcantly correlated in the total sample (r
= .24) and in English (r = .35) (p S .01); and were not correlated in Spanish. The two HL .
items were thus kept separate. Item 1 was reverse-scored and both items ranged frOm 1-5;

higher scores indicated higher HL (more conﬁdence and less help needed) (Appendix D).

47

Religiosity was assessed by the Duke University Religion Index (DUREL), a 5-
item scale that assesses organizational, nonorganizational, and intrinsic religiousness
(Storch, et al., 2004). Organizational religiosity refers to the ﬁequency of attending
formal religious services, nonorganizational refers to the amount of time spent in private
religious activities such as prayer, and intrinsic refers to the degree to which one
integrates religiosity into their life (Storch, et al., 2004). The DUREL was translated into
Spanish for this study. Responses to organizational and nonorganizational items are rated
on a 6-point scale ranging from “never” to “several times a week” or “several times a
day.” Reponses to the 3 intrinsic items are rated on a 5-point scale ranging from
“deﬁnitely not true” to “deﬁnitely true.” Items are reverse scored and a total score and an
intrinsic mean score can be created (Sherman, Plante, Simonton, Adams, Harbison, &
Burris, 2000). The 5-item DUREL (a = .91) (Sherman, etal., 2000; Storch, et al., 2004)
and the intrinsic subscale (a = .94) (Sherman, et al., 2000) showed high internal
consistency in previous studies.

Psychometric analyses were conducted to explore the internal consistency and
underlying structure of the DUREL and to determine if the language versions were
equivalent in this study. First, a 1 factor model was speciﬁed and an orthogonal varimax
rotation using the principle components method of extraction did not yield a 1 factor
solution across the two versions. Items 3-5 loaded highly on the 1 factor, where items 1
and 2 did not. Using the same methods, a 3 factor model was speciﬁed and as expected,
items 3-5 (intrinsic) loaded highly on factor. 1, item 1 (organizational) loaded highly on
factor 2, and item 2 (non-organizational) loaded highly on factor 3 for both versions. All

5 DUREL items were reverse coded. An intrinsic mean score was created from items 3-5,

48

with higher scores reﬂecting greater intrinsic religiousness; this 3-item scale had an alpha
of .75 for the total sample (English a = .80, Spanish (1 = .74). For items 1 and 2, higher
scores reﬂected greater organizational and nonorganizational religiosity (Appendix D).

Cancer F atalism is deﬁned as the belief that death is inevitable when cancer is
present and is measured in this study by the Powe F atalism Inventory (PFI) (Powe,
1995a) and the Spanish Powe F atalism Inventory (SPFI) (Lopez-McKee, et al., 2003),
which measure four aspects of cancer fatalism: inevitability of death, fears, pessimism,
and predeterrnination. The SPFI was translated from the English PF 1; both have similar
reliabilities (SPFI a = .81, PF I a = .80) (Lopez-McKee, etal., 2003). The SPFI/PF I
consist of 15 items with a “yes/no” response format. Sample items ﬁ'om each domain
include: “I think if someone has cancer and gets treatment for it, they will probably still
die from the cancer” (inevitability of death items 7, 11, 12, and 15), “I think getting
checked for cancer makes people scared that they may really have cancer” (fear items 8
and 10), “I think if someone is meant to get cancer, they will get it no matter what they
do” (predeterrnination items 1, 3-5, 9, 13, and 14), and “I think if someone gets cancer,
their time to die is soon” (pessimism items 2 and 6). Items are intended to be summed
into a total score, with higher values indicating stronger fatalistic beliefs (Powe, 1995b).

Although the 15-item PFI has been presented as a one factor scale (Powe, 1995b),
since there are 4 subscales, psychometric analyses were conducted to explore the internal
consistency and underlying structure of the PF I/SPFI scale and to determine if the two
versions had equivalent psychometric properties. First, a 4 factor model was speciﬁed
representing the 4 theoretical factors of the PFI/SPFI (inevitability of death,

predeterrnination, pessimism, and fear). The orthogonal varimax rotation using the

49

 

 

 

principle components extraction method showed that items 2, 3, 6, 7, 13 and 14 did not
load as expected, these items produced inconsistent factor loadings across the PFI and
SPFI, and no clear or consistent factor representing pessimism was found. Thus, these six
items were omitted. The 9 remaining items representing a reduced version of the
predeterrnination, inevitability of death, and fear scales were used for a second factor
analysis and all items loaded as expected on 3 factors consistently across the SPFI/PFI.
In this study the modiﬁed version of the SPFI /PFI consisted of the inevitability of
death (items 11, 12, and 15), predeterrnination (items 1, 4, 5 and 9), and fear (items 8 and
10) scales. Mean scores were created and range from 0-1, higher scores indicate more
cancer fatalistic beliefs. No total score was created. The 3-item inevitability of death
scale had an alpha of .78 (SPFI a = .82, PFI a = .74), the 4-item predeterrnination scale
had an alpha of .72 for the total sample (SPFI a = .69, PFI or = .76), and the 2 fear items
had a signiﬁcant correlation of r = .32 (SPFI r = .34, PF I r = .32) (p S .01) (Appendix D).
Breast Cancer Worry is deﬁned as an emotional reaction to the threat of cancer
and consists of both cognitive and affective elements (Hay, Buckley, & Ostroff, 2005).
The 5-item Cancer Worry Scale (CWS) was used to assess cognitive elements of recent
worry about developing breast cancer (Andersen, et al., 2003). Items 1-4 are rated on a 4-
point scale (e.g., “During the past month, how often have you worried about your own
chances of developing breast cancer?”) and a sum score is created representing recent
worry about developing breast cancer (Gramling, et al., 2007). Item 5, rated on a 4-point
scale, asked women to describe feelings of distress about their risk of developing breast

cancer. The CWS was translated into Spanish in this study.

50

Psychometric analyses were conducted to explore the internal consistency and
underlying structure of the CWS and to determine if the two language versions were
equivalent. A 2 factor model was speciﬁed representing breast cancer worry and distress
and the orthogonal varimax rotation using the principle components extraction method
showed that factor loading were as expected for the Spanish CWS, where items 1-4
(worry) loaded on factor 1 and item 5 (distress) loaded on factor 2. Since item 4 did not
load as expected in the English CWS, it was dropped from the analysis. Using the
remaining 4 items a second factor analysis was conducted and all items loaded as
expected on the 2 factors across both language versions. In this study the modiﬁed
version of the CWS consisted of breast cancer worry (items 1, 2, and 3) and breast cancer
distress (item 5). A mean score was created for the breast cancer worry 3-item scale
(ranging from 1-4) and higher scores reﬂected more recent worry; this scale had an alpha
of .86 for the total sample (English a = .86, Spanish (1 = .87). Item 5 ranged from 1-4;
higher scores reﬂect more recent distress about developing breast cancer (Appendix D).

Data Analysis
Data Entry

Survey data were entered into a database in SPSS version 14.0 in English by six
research assistants trained in data entry by the research team. Data were entered and
checked by two independent data enterers and a third person double-checked
discrepancies between the ﬁrst and second data entry, and resolved entry errors. Data
enterers were given a master survey with numeric codes to enter for each item. Data
enterers also computed the percentage of missing data per scale for each participant (e.g.,

two items answered equates to 50% missing data for a scale of four total items).

51

Data Analysis Strategy

A series of chi-square tests, t-tests, and logistic regression analyses were used to
conduct analyses for this study and test the proposed hypotheses. Table 1 includes an
illustration of variable coding for analysis. Prior to hypothesis testing, a conﬁrmatory
factor analysis was conducted to determine how well the three-factor Behavioral Model
ﬁt this particular sample and regression diagnostics techniques were used to detect the
presence of collinearity among the IV s and to determine if there were any outliers in the
dataset (Kleinbaum, etal., 2007).

Logistic regression was used as the primary method of analysis given the
dichotomous nature of the outcome, consisting of: compliant with breast cancer screening
recommendations (e.g., screening received in the last 2 years) versus non-complaint
(never screened or last screened 2 or more years ago). Overall ﬁt of the logistic
regression models was assessed by the Hosmer and Lemeshow Goodness of Fit (GOF)
Test and the model chi-square (12). In SPSS, if the p-value was > .10 for the GOF Test,
then the model was a good ﬁt to the data (Lemeshow & Hosmer, 1982). If the p-value
was < .05 for the model chi-square, then the model predicted the DV better than a model
containing only the intercept (Cohen, et al., 2003; Menard, 2002). The Likelihood Ratio
78 Test was used to compare the signiﬁcance difference between reduced and full models
(i.e., hypothesis 4 and 5). If the p-value for this x2 Test was < .05, then the added set IVs
were signiﬁcantly related to the DV, after adjusting for IVs in the reduced model. To
convey the magnitude and direction of the signiﬁcance of each IV, odds ratios, 95% CIs,

and p-values were reported for all logistic regression analyses (Kleinbaum, et al., 2007).

52

Table 1. Variable Coding for Data Analysis

 

Variables

Coding

 

Age
Age (2 categories)
History of Breast Cancer

Education

Language-based Acculturation
Country of Birth (COB)

Health Literacy] (Conﬁdence)
Health Literacy 2 (Needs help reading)
Cancer F atalism l (Predeterrnination)
Cancer Fatalism 2 (Inevitable death)
Cancer Fatalism 3 (Fear)

Intrinsic Religiosity

Organizational Religiosity
Nonorganizational Religiosity

Health Insurance

Health Insurance Type

(3 categories; 2 dummy variables)

Recent Physician Visit

Breast Cancer Worry

Distress about Breast Cancer

Health Status (HRQOLI)

Unhealthy days (HRQOLZ)

Activity limitation (HRQOL3)

Activity limitation transformed (HRQOL3T)
Recent Mammogram (< 2 years ago)

53

Continuous, in years

1: 50 years and older 0: 40 to 49 years old
1: Personal/family history of breast cancer
0: otherwise

1: 2 high school degree 0: < high school
Continuous mean score, (range: 1 -5)

1: U.S.-bom O: Mexican-bom
Continuous score, (range: 1-5)
Continuous score, (range: 1-5)
Continuous mean score, (range: 0-1)
Continuous mean score, (range: 0-1)
Continuous mean score, (range: 0-1)
Continuous mean score, (range: 1-5)
Continuous score, (range: 1-6)
Continuous score, (range: 1-6)

lzyes 0: otherwise

Insl =0 and InsZ=0: No insurance
Insl=l and In32=0z Private insurance
Insl=O and In32=lz Public insurance

1: < 1 year 0: otherwise

Continuous mean score, (range: 1 -4)
Continuous score (range: 1-4)
Continuous score (range: 1-5)
Continuous, in days (range: 0-30)
Continuous, in days (range: 0-30)
Continuous, (range: 1-1.60)

l: < 2 years 0: otherwise

RESULTS
Logistic Regression Diagnostics

Three separate linear regressions were conducted to obtain tolerance statistics for
all IV 5: all 12 IVS from the Behavioral Model, all 11 context-speciﬁc IVS, and all 23
IVS; each set was regressed on an arbitrary continuous DV. No presence of collinearity
existed in the dataset. Potential outliers were examined by plotting the standardized
residuals and leverage values of all 12 IVS from the Behavioral Model and all 11
context-speciﬁc IVS against the predicted probabilities of the DV (mammogram < 2
years). Four cases were identiﬁed as outliers by plotting the standardized residuals
against the predicted probabilities of the DV. All four cases had never had a mammogram
or had their last mammogram more than 2 years ago. Since 76.2% (n = 154) of the
sample reported having a mammogram in the past 2 years and only 23.8% (n = 48)
reportednever having a mammogram or having their last mammogram more than 2 years
ago, these cases were kept in the sample. Although removing these four cases did change
the results slightly (see Tables 3 - 11), these cases were kept in the sample for analysis
because their removal would have decreased the variability of the non-compliant group.

Testing the CFA Model

To answer the ﬁrst research question, “Does the component structure of the
Behavioral Model fit the Latino population for the outcome of breast cancer screening
utilization? ” a conﬁrmatory factor analysis (CF A) taking a cluster analytic approach to
the measurement model was conducted to determine how well the three-factor Behavioral
Model ﬁt this particular sample. The model includes three latent factors and 9 observed
variables (Figure 3). The LISREL (J oreskog & Sorbom, 2007) structural equation

modeling program was used to conduct the CFA.

54

Multiple goodness of ﬁt indices were used to assess how well the model
accounted for the data, including the xz/df ratio, the non-normed ﬁt index (NNFI), the
comparative ﬁt index (CFI), the root mean square error of approximation (RMSEA), and
the standardized root mean squared residual (SRMR) (Anderson & Gerbing, 1988; Kline,
1998). The CFA showed that the model ﬁt the data [x2 = 81.23 (df= 24, p < .01): x2/ df
ratio = 3.50; CFI = 0.87, NNFI = 0.81; RMSEA = 0.11; and SRMR = 0.10]. The overall
model was a good ﬁt, except for age (age had a standardized factor loading of .03 on the
latent predisposing factor), and thus an alternative model was analyzed.

Figure 3. Measurement Model of the Behavioral Model of Health Services Use”

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.00 -U Age
K .03
.62 Education <— 61 Pre-
. disposing
.50 -F COB " .71 >\
A’ 1.03
-.05 BASH .28
-.O3
.96 -u Phys Visit " .20 '\
-.05
.86 _. HRQOL] x
.37
24 menu
.87
.60 rumors II’ .63

 

As a ﬁrst step towards identifying an alternative model, an exploratory factor
analysis (EFA) was conducted in SPSS to determine how the variables loaded on an

unspeciﬁed number of factors (Aﬁﬁ, Clark, & May, 2004; Nunnally & Bernstein, 1994).

 

'0 Note: Since the latent factors are correlated in this measurement model, the factor loadings of the
indicators on the latent factors are not correlations, but regression coefficients, and are thus able to range
beyond one. The standardized factor loadings for acculturation (BASH) and health insurance (ins) were
greater than one, which does not present a problem (J oreskog, 1999). In Figure 3, only the factor loadings
for education, acculturation (BASH), country of birth (COB), health status (HRQOLI), unhealthy days
(HRQOL2), and activity limitation (HRQOL3) were statistically signiﬁcant at the .05 level.

55

The orthogonal varimax rotation using the principle components extraction yielded a
three factor solution, explaining 64.5% of the variance. Using .4 as cut-off criteria (e. g.,
Malcame, et al., 2005), language-based acculturation, country of birth, and education
loaded highly on factor 1, health status, unhealthy days, and activity limitation loaded
highly on factor 2, and a recent physician’s visit and age loaded highly on factor 3.
Health insurance loaded both on factor 1 and 3 (Table 2).

Table 2. Behavioral Model Factor Loadings

Predigmsing Need Enabling
Language-based acculturation .91
Country of birth .80
Education .77
Health Status .83
Unhealthy Days .79
Activity Limitation .59
Age .62
Physician Visit .80
Health Insurance .49

 

 

Although age has been previously conceived of as part of the predisposing
domain, age facilitates or enables use of screening because mammography screening is
an age-dependent preventive service. An increased age is a risk factor for developing
breast cancer (N CI, 2007b; 2008a), and California state law requires insurance companies
to cover mammograms every two years for women 40-49 years and annually for women
50 and older (Legislative Council of California, 2000a; 2000b; 2000c). Thus, older
women are enabled to have a greater frequency of screening than younger women.

Health insurance, as with a recent physician visit, acts to “enable” or “facilitate”
access to preventive care, such as mammography screening (Andersen, 1995). Both have
been considered enabling factors in the Behavioral Model since its inception (Andersen

& Newman, 1973), both have been found be two of the strongest predictors of general

56

cancer screening utilization (Carrasquillo & Pati, 2004)., and are both considered key
resources facilitate health care use (AHRQ, 2008; Millman, 1993). Thus, health insurance
was still considered an enabling variable in the second CF A.

A second CFA model, with age as an enabling variable, was tested to empirically
validate this EFA structure (Figure 4). The CF A showed that the model ﬁt the data ()8 =
79.74 (df= 24, p < .01): x2/ dfratio = 3.32; CFI = 0.88, NNFI = .83; RMSEA = 0.11; and
SRMR = 0.099). This suggested that the slightly modiﬁed 9-variable 3-factor structure of
the Behavioral Model shown in Figure 4 was supported in the current sample. Thus,
exploratory and conﬁrmatory analyses suggested that for this sample of Latinas, a
modiﬁed Behavioral Model — where age is treated as an indicator of the enabling domain
rather than the predisposing domain — was a better ﬁt for the data. Subsequent logistic
regression analyses for the hypothesis testing were conducted based on the new
measurement model, with age as an enabling, rather than a predisposing variable.

Figure 4. Modified Measurement Model H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.62 "U Education K .61
.50 -.L COB *— ,71 Pre'
disposing

-,05 -> BASH 4’ 1.03 V\
.98 ‘P A e .

8 K .13 31
-,71 —> Health Ins <— 1.31 Enabling /

-.03
.95 -* Phys Visit A/ '22 '\
-.06
.86 ‘5 HRQOL] % 37
+

~23 HRQOLZ ]" .88
-61 HRO0L3 I" '63

 

 

 

” In Figure 4, factor loadings for education, acculturation (BASH), country of birth (COB), insurance,
physician visit, health status (HRQOLI ), unhealthy days (HRQOL2), and activity limitation (HRQOL3)
were statistically signiﬁcant at the .05 level.

57

Univariate and Descriptive Analyses

Means, standard deviations, skewness, and kurtosis of continuous IVs were
examined in order to ensure that each was normally distributed within acceptable limits.
The IV ‘activity limitation’ (hrqol3) that fell outside of acceptable ranges for kurtosis (i
2) was transformed (Tabachnick & Fidell, 2001) and all reported analyses were computed
using the transformed IV, [new hrqol3 = loglO (hrqol3 + 10)]. Analyses were also run
using an untransformed IV and did not change the results. An overall correlation matrix
is located in Table R in Appendix D. Descriptive statistics of all IV s and DVs are
reported in Tables A-D in Appendix D and are described below.

Predisposing Variables. Over half (52.7%, n = 109) reported not graduating high
school and 47.3% (n = 98) reported having at least a high school education. Almost three-
fourths [73.7% (n = 151)] of participants reported being Mexican-born and 23.7% (n =
47) reported U.S.-bom. Language-based acculturation ranged ﬁ'om 1 to 5, with higher
scores indicating a greater English language acculturation. The average score for the
entire sample was M = 2.35 (SD = 1.34, n = 200); those interviewed in English had a
signiﬁcantly higher average scores (M = 3.53, SD = 1.14, n = 87) compared to those
interviewed in Spanish (M = 1.44, SD = 0.61, n = 113) [T = 16.675 (df = 198, p = .000)].

Enabling Variables. Participants’ ages ranged from 40 to 80 (M = 50.8; SD =
8.1); 49.5% (n = 103) were 40 to 49 years old and 50.5% (n = 105) were 50 years or
older. Several participants (67.6 %, n = 140) indicated that they had health insurance, and
for those who speciﬁed type, 77.6% (n = 104) reported private (HMO, PPO, or military)
and 22.4% (n = 30) reported public (Medicare or Medicaid). Many (67.1%, n = 139)

reported visiting their physician in the last year for a routine checkup.

58

Need-related Variables. When asked about their general health status, 27.3% (n =
56) rated their health as excellent or very good. Participants reported an average of 9.57
unhealthy days (SD = 11.00, n = 201) in the past 30 where their physical health and/or
mental health was perceived as not good. Of those that reported at least one unhealthy
day in the last 30 days, participants reported an average of 5.37 days (SD = 9.23, n = 147)
of activity limitation in the last 30 days where poor physical or mental health keep them
from doing their usual activities, such as self-care, work, or recreation.

Context-specific Variables. Few [2.9% (n = 6)] reported having a personal history
of breast cancer and 26.4% (n = 55) reported that a family member had breast cancer;
28.4% (n = 59) reported having a family or personal history of breast cancer. The two
health literacy items ranged from 1 to 5, with higher scores indicating higher health
literacy. Participants rated themselves as having an average conﬁdence level for ﬁlling
out medical forms by themselves (M = 3.96, SD = 1.05, n = 208); and reported asking
someone (like a family member, friend, hospital/clinic worker, or caregiver) occasionally
to help them read hospital materials (M = 3.75, SD = 1.37, n = 207).

The three cancer fatalistic belief subscales ranged from 0 to 1, with higher scores
indicating a higher endorsement of cancer fatalistic beliefs. Participants reported a less
than average level of beliefs about one’s predeterrnination to get cancer (M = 0.36, SD =
0.36, n = 198), were slightly more likely to endorse the belief about being fearful of
getting cancer (M = 0.65, SD = 0.31, n = 206), and less likely to believe that death is
inevitability once cancer is present (M = 0.17, SD = 0.31, n = 204). Intrinsic religiosity
ranged from 1 to 5 and organizational/non-organizational religiosity ranged from 1-6,

with higher scores indicating higher endorsement of religious beliefs, practices, or

59

behavior. Participants were most likely to endorse intrinsic beliefs, such as having their
religious beliefs lie behind their whole approach to life (M = 4.44, SD = 0.80, n = 204)
reported attending church or other religious meetings more than a few times per year (M
= 3.24, SD = 1.27, n = 194), and reported spending a few times per week in private
religious activities, such as prayer (M = 3.12, SD = 1.39, n = 162).

Breast cancer worry and distress ranged from 1 to 4, with higher scores indicating
a greater prevalence of recent worry and distress about developing breast cancer (e.g.,
frequency of thinking and worrying about one’s chance of getting breast cancer).
Participants reported sometimes worrying about breast cancer (M = 1.59, SD = 0.71, n =
206) and were occasionally distressed about developing breast cancer in the past month
(M = 1.68, SD = 0.67, n = 205) (Tables A-D, Appendix D).

Results of Hypothesis Testing
Testing the Behavioral Model of Health Services Use

To address the question, “ What predisposing, enabling and need factors are
associated with an increased likelihood of breast cancer screening among Latinas?
several hypothesized relationships between predisposing, enabling, and need-related
factors and screening were explored through bivariate and multivariate analyses.
Bivariate Analyses

Bivariate chi-square (x2) and t-tests analyses were conducted to determine
whether individual variables in each of the three domains were signiﬁcantly related to
having a mammogram in the past two years (sub-hypotheses la-ld, 2a-2b, and 33-30).

Predisposing Variables. Results ﬁom the'bivariate analyses did not provide

support for the hypothesized relationships between predisposing factors and

60

mammography screening. Hypotheses lb-ld were not supported: mammography
screening rates did not differ by education [)8 (1, 201) = 2.412 (p = .120)], language-
based acculturation level (T = -0.102, df = 193, p = .919), or country of birth [x2 (1, 193)
= 0.047 (p = .828)] (Table 3).

Enabling Variables. Signiﬁcant bivariate relationships existed between breast
cancer screening and all enabling factors. Although screening rates did not differ when
comparing those 50 and older to those 40.49 years old [x2 (1, 202) = 2.733 (p = 098)],
hypothesis 1a was partially supported since those who reported a mammogram in the last
two years were signiﬁcantly older (M = 51.99 years) than those who did not have their
mammogram in the last two years (M = 47.86 years) (T = 2.846, df = 200, p = .005).

Hypotheses 2a and 2b were supported: Latinas with either public or private health
insurance were signiﬁcantly more likely to have received a mammogram in the last two
years (83.8%, n = 114), compared to Latinas without health insurance (60.0%, n = 39) [x
2'(1, 201) = 13.732 (p = .000)]. Latinas that had visited a physician in the past year for a
routine physical exam were signiﬁcantly more likely to have received a mammogram in
the last two years (85.8%, n = 115) compared to Latinas who did not visit a physician in
the last year (56.7%, n = 38) [x2 (1, 201) = 20.814 (p = .000)] (Table 3).

Need-related Variables. Results from bivariate analyses did not provide support
for the hypothesized relationships between need-related factors and breast cancer
screening. Hypotheses 3a, 3b, and 3c were not supported: screening rates among Latinas
40 years and older did not differ signiﬁcantly by health status (T = -1.823, df = 197, p =
.070), recent unhealthy days (T = -1.805, df = 193, p = .073), or recent physical activity

limitation (T = 0.143, df= 141, p = .806) (Table 3).

61

Table 3. Bivariate Relationships of Behavioral Model Variables with Mammography

Screening Utilization

 

Chi-Square Analyses

 

Education
Less than high school
High school or greater

Country of birth
United States
Mexico

Age
40-49 yrs
50 +

Physician visit < 1 yr
Yes
No

Health insurance
Yes
No

Health Insurance Groups
Private

Marmnography (< 2 years)
% (n) Sig.
Yes No
71.7 (76) 28.3 (30) x2 (1, 201) = 2.412
81.1 (77) 18.9(18) (p=.120)
74.5 (35) 25.5 (12) x2 (1, 193) = .047
76.0 (111) 24.0 (35) (p = .828)

71.3 (72) 28.7 (29) x2 (1, 202) = 2.733
81.2 (82) 18.8 (19) (p = .098)“

85.8 (115) 14.2 (19) x2 (1, 201) = 20.814
56.7 (38) 43.3 (29) (p = .000)*

83.8 (114) 16.2 (22) f(1,201)=13.732
60.0 (39) 40.0 (26) (p = .000)*

84.0 (84) 16.0 (16) x2 (2, 195) = 13.476

 

 

Public 83.3 (25) 16.7 (5) (p = .001)*
None 60.0 (39) 40.0 (26)
Mammography (< 2 yrs)
Mean Comparisons M (SD) Sig.
Yes No
Language-based acculturation a 2.35 (1.34) 2.43 (1.49) t = -0.102
(n=151) (n=44) (df=193,p=.919)
Age (years) 51.99 (9.25) 47.86 (6.96) t = 2.846
(n = 154) (n = 48) (df = 200, p = .005)*
Health status b 2.99 (1.04) 3.30 (0.97) t = -1.823
(n = 152) (n = 47) (df = 197, p = .070)A
Unhealthy days b 8.99 (10.66) 12.32 (12.11) = -1.805
(n = 148) (n = 47) (df = 193, p = .073)“

62

Table 3. Cont’d

 

Activity limitationb 5.55 (9.43) 5.29 (9.13) t = 0.143
(n=108) (n= 35) (df= 141,p= .806)

Activity limitation transformed “C 1.14 (0.20) 1.13 (0.20) t = 0.109
(n=108) (n=35) (df= 141,p= .914)

 

Note: Chi-Square tests and independent samples t-tests (equal variances assumed) were used to compare
yariables across mammography adherence groups. Incomplete data are due to participant non-response.

Approaching signiﬁcance at the .05 level (.05 > p < .10); * P S .01; ** P S .05
a Possible range 1- 5; higher scores denote higher acculturation to English language.
b Health status ranges from 1-5; unhealthy days and activity limitation range from 0-30; higher scores
denote worse health status; cActivity limitation transformed.

d Analyses were conducted with removing the 4 outlier cases and the only change was that dichotomous
age became signiﬁcant (women 50 + were more likely to have a mammogram in the last 2 years (83.7 %,

n=82) than women 4049 (72.0 %, n=72) [x2 (1, 198) = 3.902 p = .048)].

In sum, bivariate analyses revealed that only enabling variables were related to
mammography screening (Table 3). Although predisposing and need-related factors were
not associated with screening, hypotheses 1, 2, and 3 postulated signiﬁcant predictive
relationships between the three domains and screening. These relationships were
examined in three multivariate logistic regression models. Results from these analyses
are presented below.

Multivariate Analyses

To answer hypotheses about the signiﬁcance of predisposing, enabling, and need
domains in predicting breast cancer screening among Latinas, three logistic regression
models were examined to determine the predictive signiﬁcance of each domain.

HYPOTHESIS 1: Latinas with higher rates of predisposing factors will be more likely to
utilize breast cancer screening services.

Table 4 displays the adjusted odds ratios (OR) and corresponding 95% conﬁdence
intervals (CI) for all predisposing variables predicting likelihood of breast cancer
screening as a group. The Hosmer and Lemeshow Goodness of Fit (GOF) Test showed

that the model was a good ﬁt to the data (12 = 2.317, df = 6, p = .888), indicating that

63

there was no difference between observed and expected values of the dependent variable
(Menard, 2002). The model chi-square was insigniﬁcant (x2 = 6.467, df= 3, p = .091),
indicating that at least one of the logistic regression coefﬁcients equaled zero.

Having a high school degree or greater predicted screening (OR = 2.999),
adjusting for other factors (p S .05): the odds of having a mammogram in the past two
years were almost 3 times higher for those with at least a high school degree compared to
those with less than a high school education. Language-based acculturation was inversely
related to screening (OR = 0.641), adjusting for education and country of birth (p S .05).
In other words, for every one unit decrease in acculturation, the odds of having a
mammogram in the past two years increased almost two fold [OR = 1.560 (1 / 0.641)].
Adjusting for other factors, country of birth did not relate to screening.

Table 4. Predisposing Factors Predicting Recent Mammography Screening Use

 

 

Predisposing Recent Mammography (< 2 years)“ c
Factors ,
OR 95% CI. P-value

Education

Less than high school b 1.00

High school or greater 2.999 (1.154, 7.793) 0.024**
Language-based acculturation 0.641 (0.414, 0.995) 0.047**
Country of Birth

Mexico b 1.00

United States 2.042 (0,590, 7.060) 0.259

 

Note: C .1. = conﬁdence interval; Acculturation no longer predicted screening when 4 outlier cases were
excluded (OR=.656, CI= (.413 - 1.044), p=.075)

Approaching signiﬁcance at the .05 level (.05 > p < .10); * P S .01; ** P S .05;
a n = 185; Reference category
cModel chi-square: 6.467 (df= 3, p = .091); GOF x2 = 2.317, (df = 6, p = .888).

In sum, although the model ﬁt the data at an acceptable level, the model chi-

square was not signiﬁcant, indicating that at least one of the predictors equaled zero.

64

However, in the adjusted model, language-based acculturation and education
signiﬁcantly predicted recent mammography screening; country of birth was not
signiﬁcant. Thus, hypothesis 1 was supported: Latinas with certain predisposing factors
(a greater education and lower language-based acculturation) were more likely to utilize
mammography screening services in the past two years.

HYPOTHESIS 2: Latinas with higher rates of enabling factors will be more likely to
utilize breast cancer screening services.

Table 5 displays the adjusted ORs and corresponding 95% C18 for all enabling
IVs predicting likelihood of breast cancer screening as a group. The model was a good ﬁt
to the data (x2 = 4.568, df = 8, p = .803), indicating that there was no difference between
observed and expected values of the dependent variable (Menard, 2002). The model chi-
square showed that this model predicted the likelihood of mammography screening better
than a null model (x2 = 29.355, df = 4, p = .000), indicating that at least one of the
predictors was signiﬁcantly related to the dependent variable (Menard, 2002).

Age predicted screening (OR = 1.062), adjusting for health insurance and a recent
physician visit (p S .05); for every one unit increase in age, the odds of having a recent
mammogram increased 1.06 times. A physician visit in the past year signiﬁcantly
predicted screening (OR = 3.336), adjusting for insurance and age (p S .01); the odds of
having a recent mammogram were 3 times higher for those who went to see a physician
for a physical exam in the past year‘compared to those who had not. Private insurance
was signiﬁcant (OR = 2.391), adjusting for age and a recent physician visit (p S .05); the
odds of a recent mammogram were 2 times higher for those with private insurance
compared to those with none. Adjusting for other factors, public insurance was not

signiﬁcant (Table 5).

65

Table 5. Enabling Factors Predicting Recent Mammography Screening Use

 

 

Enabling Recent Mammography (< 2 years) a‘ c
Factors
OR 95% CI. P-value
Age 1.062 (1.006, 1.120) 0029““
Health insurance
None b 1.00
Public 1.746 (0.538, 5.667) 0.354
Private 2.391 (1.093, 5.228) 0.029**
Physician visit < 1 yr
No b 1.00
Yes 3.336 (1.603, 6.944) 0001*

 

Note: CI. = conﬁdence interval; an analysis without 4 outlier cases did not change the results. Insurance
remained a signiﬁcant predictor (OR = 2.180, CI= (1048-4538), p= .037) when substituting the three
insurance categories with a dichotomous insurance variable (l=yes, 0=no).

* P S .01; ** P S .05;

a n = 195; Reference category

° Mode1 chi-square: 29.355 (df= 4, p = .000); GOF x2 = 4.568, (df= 8, p = .803).

In sum, the model ﬁt the data at an acceptable level and the model chi-square was
signiﬁcant, indicated that at least one of the predictors was signiﬁcant. In particular, a
greater age, a recent physician visit, and private insurance predicted having a recent
mammogram. Hypothesis 2 was supported: Latinas with certain enabling resources were
more likely to utilize mammography screening services in the previous two years.

HYPOTHESIS 3: Latinas with higher rates of need factors will be more likely to utilize
breast cancer screening services.

Table 6 displays the adjusted ORs and corresponding 95% C1s for all need-related
IVs predicting likelihood of breast cancer screening as a group. The model was a good
ﬁt to the data (x2 = 4.155, df = 8, p = .843), which means that there was no difference
between observed and expected values of the dependent variable (Menard, 2002).

However, the model chi-square was not signiﬁcant ()6 = 3.819, df = 3, p = .282),

66

indicating that this partial regression model did not predict likelihood of mammography
screening better than a model containing only the intercept term, which means that at
least one of the logistic regression coefﬁcients equaled zero (Menard, 2002). Indeed, in
the adjusted model, screening rates did not differ by health status, unhealthy days, or by
activity limitation. Hypothesis 3 was not supported: There was no difference in likelihood
of utilizing breast cancer screening based on the level of need among Latinas.

Table 6. Need-related Factors Predicting Recent Mammography Screening Use

 

 

Need Recent Mammography (< 2 years) a‘ C
Factors
OR 95% CI. P-value
Health status 0.877 (0.580, 1.325) 0.534
Unhealthy days 0.962 (0.922, 1.004) 0078"
Activity limitation b 4.932 (0.411, 59.2 L4) 0.208

 

Note: CI. = conﬁdence interval; an analysis substituting the transformed activity limitation variable with
an untransformed variable did not change results. Removing 4 outlier cases did not change results.

A Approaching signiﬁcance at the .05 level (.05 > p < .10); a n = 141; bActivity limitation is transformed
c Model chi-square: 3.819 (df= 3, p = .282); GOF x2 = 4.155, (df= 8, p = .843)

The three partial logistic regression models showed that all of the enabling
variables, and all of the predisposing variables except country of birth, signiﬁcantly
predicted the likelihood of having a mammogram in the previous two years. These
variables were included in a full regression model (Table 7). The model was a good ﬁt to
the data (x2 = 6.198, df = 8, p = .625), indicating that the model explained a signiﬁcant
proportion of variance in the criterion. The model chi-square was signiﬁcant (x2 = 35.533,
df = 6, p = .000); indicating at least one of the predictors was signiﬁcant (Menard, 2002).
After adjusting for other factors, having a high school degree (OR = 3.528), a lower
language-based acculturation (OR = 0.538), a greater age (OR = 1.067), and a physician
visit in the last year (OR = 3.350) signiﬁcantly predicted screening (p S .05); private or

public insurance were not signiﬁcant (Table 7).

67

Table 7. Reduced Behavioral Model Predicting Recent Mammogrgrhy Screening Use

 

 

Factors Recent Mammography (< 2 years)3
OR 95% CI P-value
Education
Less than high school 1’ 1.00
High school or greater 3.528 (1.076, 11.564) 0.037**
Language-based Acculturation 0.538 (0.386, 0.881) 0010'”
Age 1.067 (1.010, 1.126) 0.021"
Health insurance
None b 1.00
Public 1.793 (0.495, 6.493) 0.374
Private 2.589 (0.917, 7.314) 0073"
Physician visit < 1 yr
No b 1.00
Yes 3.350 (1.519, 7.386) 0003*

 

Note: C .I. = conﬁdence interval. Education no longer predicted screening when 4 outlier cases were
excluded (OR= 3.544, CI: (.962, 13.060), p= .057). An analysis with dichotomous insurance (yes/no) was

not signiﬁcant. Approaching signiﬁcance at the .05 level (.05 > p < .10); * P S .01; ** P S .05
a n = 194; Reference category
° Model chi-square: 35.533 (df= 6, p = .000); cor x2 = 6.198, (df= 8, p = .625)

Summary

Significance of Predisposing Variables. None of the predisposing variables were
associated with screening in bivariate analyses (Table 3). However, results ﬁom the
adjusted logistic regression analysis revealed that Latinas with a greater education and
lower language-based acculturation were more likely to obtain a recent mammogram
(Table 4), which partially supports the hypothesis that predisposing variables increases
the likelihood of mammography screening. A closer examination of these analyses
revealed that language-based acculturation and education went from non-signiﬁcant in

bivariate analyses to signiﬁcant in the logistic regression model containing three

68

predisposing variables, which suggests that a suppression12 situation was present (Collins
& Schmidt, 1997; Conger, 1974; Conger & Jackson, 1972; Lynn, 2003; MacKinnon,
Krull, & Lockwood, 2000; Rosenberg, 1973; Tzelgov & Henik, 1991). To determine
which variable(s) acted as a suppressor, additional analyses were examined with the three
predictors (language-based acculturation, education, and country of birth).

T-tests analyses showed that those with high school or greater education had a
higher language-based acculturation (M = 3.25, SD = 1.22, n = 95) than those with less
than a high school education (M= 1.53, SD = .89, n = 104), [T = 11.467 (df= 197) p =
.000]; and U.S.-born Latinas had a higher language-based acculturation (M = 4.15, SD =
.95, n = 44) than Mexican-born Latinas (M = 1.80, SD = .94, n = 146) [T = 14.457 (df =
188) p = .000]. A chi-square analyses showed that U.S.-bom Latinas were more likely to
have a high school education or greater (85.1%, n = 40) compared Mexican-born Latinas
(35.3%, n = 53) [x2 (1, 197) = 35.574, p = .000]. Thus, a higher education was positively
related to higher English-language acculturation and being U.S.-bom; and being U.S.
born was positively related to a higher English-language acculturation level.

Simple unadjusted logistic regression analyses revealed that education (p = .122),
language-based acculturation (p = .918), and country of birth (p = .829) were not
signiﬁcant predictors of recent mammography screening. Then, three regression models
were conducted with two predictors entered simultaneously (i.e., country of birth and

acculturation, country of birth and education, and education and acculturation). Country

 

‘2 A suppressor is a non-redundant variable that increases the predictive validity of another variable (or set
of variables) by its inclusion in the regression equation by removing irrelevant predictive variance in other
predictor(s) (Conger & Jackson, 1972). Usually, the variable being suppressed was uncorrelated with the
dependent variable in bivariate analyses and became signiﬁcant in multivariate analyses (Conger, 1974;
Tzelgov & Henik, 1991).

69

of birth (p = .344) and acculturation (p = .392) were not signiﬁcant when entered
simultaneously and country of birth (p = .349) and education (p = .094) were not
signiﬁcant when entered simultaneously. However, when education and acculturation
were entered simultaneously, acculturation was non-signiﬁcant (p = .116) and education
became signiﬁcant (OR = 2.910, p = .026). When all three variables were entered
simultaneously (e. g., Table 4) country of birth was not signiﬁcant (p = .259), yet a higher
education (OR = 2.999, p = .024) and lower acculturation (OR = .641 , p = .047)
signiﬁcantly predicted screening. No interaction or confounding relationships existed
among the three variables (Lynn, 2003; MacKinnon, Krull, & Lockwood, 2000).

Since education became signiﬁcant only when acculturation was entered
simultaneously, this suggested that acculturation suppressed some of the irrelevant
variance in the relationship between education and screening (Rosenberg, 1973). Since
acculturation only became signiﬁcant once both education and country of birth were
entered into the model, this suggested that country of birth and education together
suppressed some of the irrelevant variance in the relationship between acculturation and
screening (Rosenberg, 1973). Thus, education was suppressed by acculturation and
acculturation was suppressed by education and country of birth. These two variables were
not excluded from further analyses because their inclusion improves the predictability of
the outcome and provides information about the relationship among the predictors and
the outcome that needs to be explored further (Collins & Schmidt, 1997; Conger &

Jackson, 1972; Rosenberg, 1973; Tzelgov & Henik, 1991)”.

 

'3 For example, if A suppressed some of the irrelevant variance in the relationship between B and the
criterion, and A was correlated with B, this means that the uniqueness of B unrelated to A is associated
with the criterion. The unique part of B would need to be explored in future studies (Rosenberg, 1973).

70

Signiﬁcance of Enabling Variables. As hypothesized, all of the enabling variables
were signiﬁcantly associated with screening in bivariate analyses (Table 3). Latinas with
enabling characteristics, (greater age, private insurance, and a physician visit in the past
year), were more likely to utilize screening, after adjusting for all other enabling
variables. Thus, this dataset supported the hypothesis that enabling variables increased
the likelihood of having a recent mammogram in the past two years.

Signiﬁcance of Need-related Variables. Need-related variables were not related to
screening in bivariate analyses (Table 3) or in the partial regression model containing all
three need-related variables (Table 6). Thus, these data failed to support the hypothesis
that the need-related domain increases the likelihood of having a recent mammogram.

Signiﬁcant variables from the three partial logistic regression models, (i.e., all
enabling variables and education and language-based acculturation predisposing
variables) were included in a full regression model. Results showed that the model
explained a signiﬁcant portion of the variance in screening and after adjusting for all
other factors, having a high school degree, a lower language-based acculturation, a
greater age, and a recent physician visit signiﬁcantly predicted screening; and private or
public insurance were not signiﬁcant (Table 7). This model was further modiﬁed by
adding contextual variables (Table 11). First, attention is drawn to analyses conducted to
determine the incremental signiﬁcance of predisposing and need-related domains.

Testing the Incremental Prediction of Predisposing and Need Domains

To address the third research question, “T 0 what extent do predisposing and need

factors add to the prediction of breast cancer screening utilization by Latinas, after

controlling for enablingfactors? two hierarchical logistic regression analyses were

71

conducted to determine the incremental prediction of predisposing and need-related
domains. The ﬁrst model placed all enabling domain variables in Step 1 and included
predisposing variables in Step 2 (Table 8). The second model placed enabling variables in
Step 1 and included need-related variables in Step 2 (Table 9).
HYPOTHESIS 4: After adjusting for enabling factors in the original model, predisposing
and need variables will not be signiﬁcantly associated with breast cancer screening
utilization.
Predisposing and Enabling Domains

Table 8 displays the adjusted ORs and corresponding 95% C1s for all
predisposing and enabling variables examined in a step-wise fashion to predict likelihood
of breast cancer screening; Model 1, contained all enabling variables and Model 2
contained all enabling plus predisposing variables. Model 2 was not a good ﬁt to the data
()8 = 17.459, df = 8, p = .026), indicating that there was a difference between observed
and expected values of the dependent variable (Menard, 2002). However, the model chi-
square was signiﬁcant, indicating that at least one of the predictors was signiﬁcantly
related to the dependent variable (x2 = 36.062, df= 7, p=.000), (Menard, 2002).

A Likelihood Ratio x2 Test was used to compare the hierarchical models; results
of which indicated that the three predisposing variables signiﬁcantly added to Model 1
(x2 = 8.686, df = 3, p S .05) (Table 9). Although the likelihood ratio test showed that after
accounting for enabling factors, predisposing factors signiﬁcantly added to the model, the
overall model including enabling and predisposing factors did not account for a
signiﬁcant proportion of variance in mammography screening. Thus, hypothesis 4 was

supported.

72

Table 8. Predisposing and Enabling Domains Predicting Recent Mammography
Screening Use

 

 

 

All Behavioral Model Model 1 (Enabling Only) Model 2 (Enabling + Predisposing)
Independent Variables Recent Mammography Recent Mammo phy
(<2 yet-rs)" (<2 years)
OR 95% CL P- OR 95% CI. P-
value value
Enabling Factors
Age 1.052 (.995, 1.111) .072/\ 1.059 (1.00, 1.121) .048“
Health insurance
None c 1.00 1.00
Public 1.448 (.430, 4.882) .550 1.870 (.482, 7.249) .365
Private 2.029 (.876, 4.698) .099/\ 3.172 (1.043, 9.644) .042**
Physician visit < 1 yr
No c 1.00 1.00
Yes 4.097 (1.876, 8.948) .000* 3.411 (1.530, 7.736) 003*
Predisposing Factors
Education
Less than high school C “ “ " 1-00
High school or greater -- -- -- 3.053 (1.019, 10.048) 066‘
Language-based acculturation -- -- -- .462 (.263, .814) .007*
Country of birth
Mexico C " -- -- 1.00
United States -- -- -- 2.712 (.596, 12.345) .197
-2 log likelihood 167.132 d 158.446 °

 

Note: CI. = conﬁdence interval. Analyses substituting insurance dummy variables with a dichotomous
variable and substituting continuous age with a dichotomous age only changed the results slightly: Age
became non-signiﬁcant in model 2 (OR=] .538, CI=.699, 3.380, p=.284). Age became signiﬁcant in Model
1 when 4 outlier cases were excluded (OR= 1.084, CI= 1.016, 1.156, p= .014).

*P S .01; ** P S .05; a n =178;b n =178;c Reference category;
d Model chi-square: 27.377 (df= 4, p = .000); GOF x2 = 6.676, (df= 8, p = .572)
° Model chi-square: 36.062 (df= 7, p = .000); cor x2 = 17.459, (df= 8, p = .026). A Likelihood Ratio x2

Test indicated that the three additional predisposing variables signiﬁcantly added to the model 1 containing
enabling factors [x23= 8.686 > critical value (x23 (0 = _05) = 7.815), p S .05].
Need-related and Enabling Domains

Table 9 displays the adjusted ORs and corresponding 95% CIs for need and

enabling domains examined in a step-wise fashion to predict likelihood of screening;

73

Model 1 contained the enabling domain and Model 2 contained the enabling plus need-
related variables. Model 2 was a good ﬁt to the data ()8 = 7.253, df = 8, p = .510),
indicating there was no difference between observed and expected values of the outcome.
The model chi-square was signiﬁcant (x2 = 19.695, df = 7, p = .006), showing that at least
one of the predictors was signiﬁcant (Menard, 2002). Yet, a Likelihood Ratio x2 Test
showed that the need domain did not add to the model containing enabling factors (x2 =
2.641 , df = 3, p > .05); thus, need did not matter after accounting for enabling factors.

Table 9. Need-related and Enabling Domains Predicting Recent Mammography
Screening Use

 

 

 

All Behavioral Model Model 1 (Enabling Only) Model 2 (Enabling + Need)
Independent Variables Recent Mammography Recent Mammo aphy
(< 2 years)a (< 2 years)
OR 95% C .I. P-value OR 95% CI. P-value
Enabling Factors
Age 1.036 (.980, 1.096) .215 1.044 (.983, 1.108) .162
Health insurance
None C 1.00 1.00
Public 2.063 (.533, 7.993) .294 2.264 (.547, 9.379) .260
Private 2.161 , (.871, 5.362) .096" 1.770 (.667, 4.698) .252
Physician visit < 1 yr
No c 1.00 1.00
Yes 3.361 (1.436, 7.866) .005* 3.369 (1.409, 8.056) .006*
Need Factors
Health status -- -- -- .927 (.572, 1.501) .758
Unhealthy days -- -- -- .965 (.920, 1.011) .136
Activity limitation -- -- -- 1.737 (.108, 28.032) .697
-2 log likelihood 135.323 d 132.682 °

 

Note: Substituting the transformed activity limitation variable with an untransformed variable, insurance
with a dichotomous variable and substituting the age with a dichotomous age did not change results. Age
was signiﬁcant in Model 2 when 4 outliers were excluded. (OR= 1.075, CI= 1.001, 1.156, p= .048).* P S

.01; ** P S .05; an = 135; bn = 135; cReference category; d Model chi-square: 17.055 (df= 4, p = .002);
GOF x2 = 11.718, (df= 8, p = .164); ° Model chi-square: 19.695 (df= 7, p = .006); GOF x2 = 7.253, (df= 8,
p = .510). A Likelihood Ratio x2 Test indicated that the three need-related variables did not signiﬁcantly
add to model 1 containing enabling factors [x23= 2.641 < critical value (783 (a = .05) = 7.815), p S .05).

74

Summary

In sum, the Likelihood Ratio )6 Test showed that after adjusting for enabling
factors, predisposing variables as a group were signiﬁcantly related to breast cancer
screening. However, the model was not a good ﬁt to the data, meaning that the overall
model including enabling and predisposing factors failed to account for a signiﬁcant
proportion of variance in the dependent variable (Table 8). After adjusting for enabling
variables, the need-related variables were not related to screening (Table 9). These results
demonstrated that predisposing and need domains did not matter in the prediction of
recent mammography screening among Latinas 40 years and older, after accounting for
enabling variables, which supports this study’s hypotheses.

Testing the Incremental Prediction of the Proposed Context-Specific Variables

To address the question, “To what extent do context-speciﬁc variables add to the
prediction of breast cancer screening utilization by Latinas? ”, several hypotheses about
the relationship between contextual predisposing and need-related factors and breast
cancer screening were explored through bivariate and multivariate analyses .
Bivariate Analyses

Bivariate chi-square (x2) and t-tests analyses were conducted to determine
whether the individual context-speciﬁc variables were signiﬁcantly related to having a
mammogram in the past two years (sub-hypotheses 5a-5 g).

Breast cancer screening rates did not differ by a history of breast cancer in 3 Chi-
square analysis [x 2 (1, 202) = .424, (p = .515)]. Hypothesis 5a was not supported,
suggesting that a personal or family history of breast cancer was not related to breast

cancer screening in this sample of Latinas.

75

Latinas who reported having a mammogram in the past two years (M = 4.03, SD
= 1.00, n = 154) had a greater conﬁdence in ﬁlling out medical forms compared to those
who reported having a mammogram more than 2 years ago (M = 3.63, SD = 1.16, n = 48)
(T = 2.360, df = 200, p = .019). Hypothesis 5b was partially supported: screening rates
differed by level of conﬁdence in ﬁlling out medical forms. Assistance needed in reading
hospital materials was not associated with screening (T = .224, df =199, p = .824).

Overall, hypotheses 5c -1 f were not supported, suggesting that cancer fatalistic
beliefs were not related breast cancer screening in this sample of Latinas 40 years and
older. Screening rates did not differ by beliefs about a predeterrnination of getting cancer
(T = .330, df = 190, p = .742), an inevitability of death once cancer is present (T = .777,
df= 190, p = .438), or a fear of getting cancer (T = -0.289, df =198, p = .773) (Table 10).

Hypotheses 5d-5f were not supported, suggesting that religiosity was not related
breast cancer screening in this sample of Latinas 40 years and older. Mammography
screening rates did not differ when comparing different levels of intrinsic religiosity (T =
0.411, df= 196, p = .517), organizational religiosity (T = 0.649, df = 187, p = .517), or
non-organizational religiosity (T = -0.513, df = 156, p = .609) (Table 10).

Overall, hypothesis 5 g was partially supported. While breast cancer worry was
not related to screening, (T = 0.407, df = 197, p = .684), Latinas who had a mammogram
in the two years had a lower level of recent distress about developing breast cancer (M =
1.50, SD =.62, n = 153) when compared to those who had a mammogram more than two

years ago (M= 1.85, so = .93, n = 47) (T = -2.961, df= 198, p = .003) (Table 10).

76

Table 10. Bivariate Associations between Context-speciﬁc Variables with Recent

 

 

 

 

Mammography Screening Use
Mammography (< 2 years)
Chi-Square Analyses % (n) Sig.
Yes No
History of breast cancer
Yes 79.3 (46) 20.7 (12) x2 (1, 202) = .424
No 75.0 (108) 25.0 (36) (p = .515)

Mammography (< 2 yrs)

Mean Comparisons M (SD) Sig.
Yes No

Cancer fatalistic beliefs: 0.46 (0.37) 0.44 (0.34) t = 0.330
predetennination a (n = 146) (n = 46) (df = 190, p = .742)
Cancer fatalistic beliefs: 0.17 (0.31) 0.13 (0.28) t = 0.777
inevitability of death a (n = 150) (n = 48) (df = 196, p = .438)
Cancer fatalistic beliefs: 0.65 (0.39) 0.67 (0.40) t = -0.289
fear a (n = 152) (n = 48) (df= 198, p = .773)
Intrinsic religiosity b 4.45 (0.80) 4.40 (0.81) t = 0.411

(n=151) (n=47) (df=196,p=.681)
Organizational religiosity c 3.26 (1.21) 3.11 (1.43) t = 0.649

(n=145) (n=44) (df=187,p=.517)
Non-organizational 3.08 (1.37) 3.22 (1.44) t = -0.513
religiosityc (n = 121) (n = 37) (df = 156, p = .609)
Health literacy: conﬁdence 4.03 (1.00) 3.63 (1.16) t = 2.360
in ﬁlling out medical formsd (n = 154) (n = 48) (df= 200, p = .019)**
Health literacy: assistance 3.75 (1.32) 3.70 (1.52) t = 0.224
needed in reading hospital (n = 154) (n = 47) (df = 199, p = .823)
materials d
Breast cancer worry 8 1.68 (0.66) 1.64 (0.71) t = 0.407

(n = 152) (n = 47) (df = 197, p = .684)
Breast cancer distress e 1.50 (0.62) 1.85 (0.93) = -2.961

(n = 153) (n = 47) (df= 198, p = .003)*

 

Note: Chi-Square tests and independent samples t-tests (equal variances assumed) were used to compare
variables across mammography adherence groups. An analysis without 4 outliers did not change the results.

A Approaching signiﬁcance at the .05 level (.05 > p < .10); * P S .01; ** P S .05; a Range from 0-1; higher
scores denote higher levels of fatalistic beliefs. b Ranges from 1-5 and c Ranges from 1- 6; higher scores
denote greater religiosity. d Ranges ﬁom 1-5, higher scores denote more conﬁdence in ﬁlling out forms and
less assistance needed. 6 Range from 1-4; higher scores denote higher levels of worry and distress.

Multivariate Analyses

Contextual factors associated with screening in bivariate analyses included health
literacy (conﬁdence in ﬁlling out medical forms) and breast cancer distress. To determine
the incremental prediction of these variables, a hierarchical logistic regression analysis
was performed with Step 1 including Behavioral Model variables (i.e., education,
language-based acculturation, age, health insurance, and recent physician visit) and Step
2 included conﬁdence in ﬁlling out medical forms and breast cancer distress.
HYPOTHESIS 5: After adjusting for variables in the Behavioral Model, context speciﬁc
variables will be signiﬁcantly associated with breast cancer screening utilization which
will improve the predictability of the Behavioral Model.

Table 11 displays the adjusted ORs and corresponding 95% C1s for the two
models. Model 1 contained selective Behavioral Model variables and Model 2 contained
the selective Behavioral Mode1 variables plus the two context-speciﬁc variables. Model 2
was a good ﬁt to the data (x2 = 6.392, df = 8, p =.603), meaning there was no difference
between observed and expected values of the outcome variable. The model chi-square
was signiﬁcant (x2 = 47.657, df = 8, p = .000), indicating that at least one of the
predictors was signiﬁcant (Menard, 2002). A Likelihood Ratio )8 Test indicated that the
two contextual variables signiﬁcantly added to Model 1 (x2 = 12.571, df = 2, p S .05).
After controlling for the reduced Behavioral Model variables, a lower breast cancer
distress (OR = .582) and a greater conﬁdence in ﬁlling out medical forms (OR = 1.701)
predicted recent mammography screening (p S .05). Thus, the addition of the two
context-speciﬁc variables improved the prediction of the reduced Behavioral Mode1,
(containing all enabling and predisposing variables except for country of birth), together

these constitute the ﬁnal Behavioral Model of Latina Breast Cancer Screening Use.

78

Table 11. Reduced Behavioral Model and Selective Context-speciﬁc Variables
Predicting Recent Mammography Screening Use

 

 

Factors Model 1 Model 2
Mammography (< 2 yearsf1 Mammography (< 2 years)b
OR 95% CI. P-value OR 95% CI. P-value
Education
Less than high school c 1-00 1-00

High school or greater 3.523 (1.076, 11.532) .037** 2.937 (.854, 10.103) .087"

 

Language-based Ace. .584 (.386, .882) .011** .509 (.331, .782) .002*
Age 1.066 (1.009, 1.126) .022" 1.087 (1.025, 1.152) .005*
Health insurance

None c 1.00 1.00

Public 1.776 (.490, 6.433) .382 2.038 (.525, 7.914) .304

Private 2.586 (.916, 7.300) .073A 2.541 (.865, 7.468) .090"
Physician visit < 1 yr

No c 1.00 1.00

Yes 3.342 (1.516, 7.367) .003* 4.170 (1.793, 9.698) .001*
Contextual Factors
Breast cancer distress -- -- -- .582 (.347, .976) .040"
Conﬁdence in ﬁlling out -- -- -- 1.701 (1.134, 2.552) .010*
medical forms
-2 log likelihood 166.055 d 153.484 6

 

Note: Substituting insurance with a dichotomous variable and the continuous age with a dichotomous
variable changed the results slightly. Education (OR=2.430, CI=.837-7.052, p=.103) and age (OR=1.652,
CI=.766-3.560, p=.200) became non-signiﬁcant in Model 1 and age became non-signiﬁcant in Model 2
(OR=2.07 3, CI=.919-4.687, p=.079). An analysis without 4 outlier cases did not change the results.

A Approaching signiﬁcance at the .05 level (.05 > p < .10); * P s .01; ** P s .05

an = 186; n = 186; cReference category;

d Model 1 chi-square: 35.086 (df= 6, p = .000); GOF x2 = 7.009, (df= 8, p = .536)

6 Model 2 chi-square: 47.657 (df=8, p = .000); GOF x2 = 6.392, (df=8, p= .603). Model 2 was a better ﬁt. A
Likelihood Ratio x2 Test indicated that the two context-speciﬁc variables signiﬁcantly added to the reduced
model 1 [122 = 12.571 > critical value (sz (a = .01) = 9.210), p s .01].

Behavioral Model of Latina Breast Cancer Screening Use
A ﬁnal step in this analysis was to determine how well the structure of the ﬁnal
three-factor 7 observed variable model ﬁt this particular sample by running a

conﬁrmatory factor analysis (Figure 5). Since the latent factor labeled “need” only

79

contained one indicator, breast cancer worry was included as an additional indicator of
the “need” latent factor to examine all three latent factors simultaneously in the CFA
(Figure 6). Multiple goodness of ﬁt indices showed that overall, the model ﬁt the data
well (x2=36.51 (df = 17): x2/ df ratio = 2.15; CFI= 0.93, NNFI= 0.88; RMSEA= 0.074;
and SRMR: 0.062). The conﬁrmatory analyses suggested that for this sample of Latinas,
the eight-variable three latent factor modiﬁed Behavioral Model depicted in Figure 6 was
a good ﬁt to the data; suggesting that the Behavioral Model of Latina Breast Cancer
Screening Use (Figure 5) was also supported in this sample of Latinas and was a better ﬁt
than the original Behavioral Model (see pages 55 and 57).

Figure 5. Final Behavioral Model of Latina Breast Cancer Screening Use
Predisposing Domain
H

 

 

 

 

M (Education, language-based acculturation, health literacy) Latina
breast
Enabling Domain cancer
Personalﬁresources (age, health insurance and recent physician visit) + screening
use

 

 

 

Need Domain _,
Perceived need (breast cancer distress)

 

 

 

 

 

Figure 6. Measurement Model of the Behavioral Model of Latina Breast Cancer

Screening Use”
.31 .4 Health Lit \

 

 

 

 

 

.44
.37 -> Education <— 30 pm-
K .80 disposing

.36 1 BASH

‘\
.36
Enabling /

 

 

.98 "P Age K .13

 

 

-.98 Health Ins <— 1-41

 

 

'96 " ths Visit

-.36
'\
—.05
.43 4 worry $ .75
.75 Distress 4* .50

14 In Figure 6, factor loadings for health literacy, education, acculturation (BASH), country of birth (COB),
insurance, breast cancer worry, and breast cancer distress were statistically signiﬁcant at the .05 level.

80

 

 

 

 

 

 

 

 

Summary

Religiosity (intrinsic, organizational, and non-organizational), history of breast
cancer, cancer fatalistic beliefs (fear, predeterrnination, and inevitability of death),
assistance needed in ﬁlling out medical forms, and breast cancer worry were not
associated with recent mammography screening. The Likelihood Ratio 38 Tests showed
that a higher conﬁdence in ﬁlling out medical forms and a lower breast cancer distress
were signiﬁcantly related to screening, after adjusting for age, education, language-based
acculturation, health insurance, and a recent physician visit, which supported the
hypothesis that some contextual variables increased the prediction of having a recent
mammogram in the past two years. These seven variables together were used to create
the Behavioral Model of [ﬁtina Breast Cancer Screening Use, which accounted for a
signiﬁcant proportion of variance in recent mammography screening utilization among
Latinas. Conﬁrrnatory analyses suggested that this modiﬁed seven-indicator three-factor
model was a good ﬁt to the data, which can be appropriately used to predict the

likelihood of mammography screening among Latinas 40 years and older.

81

DISCUSSION

This study sought to contribute to the current understanding of what inﬂuences
Latina’s use of breast cancer screening by testing a theoretical model, the Behavioral
Model of Health Services Use (Behavioral Model), which explains why individuals use
health care (Andersen, 2008) and includes three domains: predisposing, enabling, and
need-related. Predisposing factors refer to individual characteristics that increase the
likelihood of health care use (Stein, etal., 2007); measured in this study by education,
language-based acculturation, and country of birth. Enabling factors are conditions that
permit access to health care and make services available (Andersen, 1968), measured in
this study by age, health insurance, and recent physician’s visit. Need-related factors are
measured by illness level, indicated by self-perceptions and objective evaluations
(Gelberg, Andersen, & Leake, 2000), measured in this study by health-related quality of
life (health status, unhealthy days, and activity limitation).

Because there is a need for a greater focus on the importance of contextual and
individual characteristics for deﬁned populations in studies that examine access to care
and health services utilization, a proposed modiﬁcation of the Behavioral Model, called
Behavioral Model of Latina Breast Cancer Screening Use, was created and tested in this
study. The model was modiﬁed by including context-speciﬁc variables theorized to be
especially relevant to the Latinas, breast cancer, and breast cancer utilization behavior
(Pasick & Burke, 2008) that could allow for a greater prediction of screening.

Key Findings

Overall, ﬁndings suggest that the Behavioral Model was a reasonably good ﬁt for

examining breast cancer screening in this sample of Latinas 40 years and older, with the

slight modiﬁcation of treating age as an enabling variable. The suppression effect that

82

emerged in the partial regression model of predisposing variables of the Behavioral
Model, illustrated a complex relationship among predictors and the criterion that needs to
be explored further (Rosenberg, 1973). Results showed the enabling domain mattered
more than the need and predisposing domains for predicting screening (Tables 8-9). Yet,
once country of birth was omitted, (Table 7), enabling and predisposing variables
together predicted screening, thus supporting the claim that predisposing and enabling
factors matter more than need for the discretionary use of health care (Andersen, 1968).

Findings from the logistic regression analysis of the ﬁnal Behavioral Model of
Latinas Breast Cancer Screening Use showed that an older age, being unacculturated, a
recent physician visit in the past year, a lower recent level of cognitive distress about
developing breast cancer, and a higher conﬁdence in ﬁlling out medical forms were key
to ensuring compliance with current mammography screening guidelines among Latinas.
A recent physician visit was the most signiﬁcant predictor of screening in this study,
suggesting that above all, access and entry into the health care system is critical to receipt
of mammography screening among Latinas (e. g., Zambrana, et al., 1999).

A discussion of this study’s results in regards to the Behavioral Model and
Behavioral Model of Latina Breast Cancer Screening Use are outlined below. Discussion
of speciﬁc ﬁndings related to this study’s research questions, implications for future
research, and implications for interventions are included.

Behavioral Model of Health Services Use

Exploratory and conﬁrmatory analyses suggested that a modiﬁed Behavioral

Model — where age was treated as an enabling not a predisposing variable — provided a

reasonable framework for examining breast cancer screening behavior for Latina women

83

40 years and older in this sample. These ﬁndings suggest that within different disease
contexts, various factors may play unique roles in utilization behaviors. Future studies
should empirically examine the structure of the Behavioral Model within the context of
different diseases, health care utilization types, and ethnic populations (Nunnally &
Bernstein, 1994; Lucas, 2003; Vandenberg & Lance, 2000). Contrary to previous
research, (Andersen, 1995; Gelberg, Andersen, & Leake, 2000) age was treated as an
enabling variable in this study. Because an older age is a risk factor for breast cancer
incidence (N CI, 2007b; 2008a) and federal screening recommendations and state
insurance coverage laws are age-speciﬁc, these factors together may cause older women
to be enabled to have a greater frequency of mammography screening use than younger
women. To conﬁrm this enabling effect of age, future studies are needed to examine age
in relation to use of other preventive health services, such as colorectal cancer or
cholesterol screening.

The second goal of this study was to determine what predisposing, enabling and
need factors were associated with an increased likelihood of breast cancer screening
among Latinas. Overall, predisposing and need-related variables were not associated with
screening in bivariate analyses; however all the enabling variables had signiﬁcant
bivariate relationships with screening (Table 3). In the adjusted regression model of need-
. related variables, none of the predictors were signiﬁcantly related to screening (Table 6),
contrary to what was hypothesized. In the adjusted regression model of the enabling
variables, age (OR = 1.062), private health insurance (OR = 2.391), and a recent

physician visit (OR = 3.336) poSitively predicted the likelihood of mammography

84

screening in the last two years, which did support this study’s hypotheses. Public health
insurance did not predict screening, once adjusting for other enabling variables.

A closer examination of the predisposing variables revealed that language-based
acculturation and education went ﬁom non-signiﬁcant in bivariate analyses (Table 3) to
signiﬁcant in the partial logistic regression model [acculturation OR = 1.560 (1/0.641);
education OR= 2.999] (Table 4). In fact, ﬁrrther analyses revealed that education was
suppressed by language-based acculturation and language-based acculturation was
suppressed by education and country of birth. Suppressor variables are more common
than researchers tend to believe (Rosenberg, 1973) and should not necessarily be
excluded once discovered because they may provide key insight for future research.
Suppressors can improve the predictability of the outcome variable (Collins & Schmidt,
1997; Conger & Jackson, 1972; Tzelgov & Henik, 1991) and may contribute to theory
development (Collins & Schmidt, 1997). Omitting suppressors could potentially increase
the risk of a Type 11 error of rejecting a true hypothesis as false (Rosenberg, 1973) which
could lead to a misinterpretation of results (Collins & Schmidt, 1997). Suppressors can
also lend interesting information about the relationship among predictors and the criterion
that needs to be explored further (Rosenberg, 1973) in order to conﬁrm that the
suppression effect was not due to chance (Collins & Schmidt, 1997).

While this study found that enabling factors mattered more than need and
predisposing factors (Tables 8 and 9), once the non-signiﬁcant country of birth
predisposing variable was omitted, (Table 7), the combination of predisposing and
enabling variables explained a signiﬁcant proportion of the variance in breast cancer

screening. Findings from this enabling/predisposing model that excluded country of birth

85

showed that Latinas who had a high school education or greater (OR = 3.528), a lower
language-based acculturation [OR = 1.859 (1 / 0.538)], a greater age, (OR = 1.076) and a
recent physician visit in the last year (OR = 3.3 50), were more likely to obtain a recent
mammogram in the past two years than Latinas without these characteristics, adjusting
for other variables (Table 7); and private and public insurance were not signiﬁcant. This
collection of variables suggest that older Spanish-speaking Latinas who have graduated
high school and who have had a recent doctor visit are also likely to have received a
mammogram within the last two years, irrespective of insurance coverage.

These results also imply that enabling and certain predisposing variables (i.e.,
education and language-based acculturation) contribute signiﬁcantly to the prediction of
screening. These results support the original assumptions of the Behavioral Model,
suggesting that for the discretionary use of health care - such as breast cancer screening -
the more likely use will be based on predisposing and enabling factors rather than need
(Andersen, 1968). Since research suggests that enabling factors, such as health insurance
and a usual source of care matter more for breast cancer screening than predisposing
factors such as education and acculturation (Zambrana, et al., 1999), studies are needed to
explore these relationships further. Studies are also needed to determine the extent to
which predisposing and need-related domains matter in the face of enabling domains for
other disease contexts, types of health care use, and ethnic populations.

Behavioral Model of Latina Breast Cancer Screening Use

The conﬁrmatory analyses suggested that for this sample of Latinas, the

Behavioral Model of Latina Breast Cancer Screening Use created in this study provided

a reasonable framework for examining breast cancer screening behavior (Figure 6).

86

Future studies are needed to empirically test this model in order to determine how
dependable it is across disease contexts, health care utilization types, and ethnic
populations (Lucas, 2003; Nunnally & Bernstein, 1994). Findings ﬁ'om the logistic
regression analysis of the Behavioral Model of Latinos Breast Cancer screening Use
showed that these two context-speciﬁc variables signiﬁcantly added to the prediction of
mammography screening use. In fact, a lower language-based acculturation [OR = 1.965
(1 / 0.509)], a greater age (OR = 1.087), a recent physician visit in the past year (OR =
4.170), a greater conﬁdence in ﬁlling out medical forms (OR= 1.701), and a lower level
of cognitive distress about developing breast cancer (OR = 1.718 (1/ 0.582) were
important facilitators to breast cancer screening use. Education and private insurance
were marginally insigniﬁcant (p < .10), adjusting for the preceding variables (Table 11).
Given that education and health literacy were related - those with a high school degree or
more had a higher conﬁdence in ﬁlling out medical forms (M = 4.39, SD = 0.83, n = 98)
than those with less than a high school education (M = 3.58, SD = 1.09, n = 109) (T =
5.951 (df = 205, p = .000) - perhaps health literacy took away some of the variance in the
relationship between education and screening, suggesting that education should be
excluded. However, a post-hoe analysis showed that the two variables that were being
suppressed (education and language-based acculturation) added signiﬁcantly to the
Behavioral Model of Latina Breast Cancer Screening Utilization ’s prediction of
screening (Table S, Appendix D), thus supporting their inclusion in the ﬁnal model.
Some of these ﬁndings support prior research showing that a higher health
literacy (Guerra, Krumholz, & Shea, 2005), visiting a physician in the past year (Frazier,

J iles, & Mayberry, 1996; Gorin & Heck, 2005) and a higher age (Fernandez & Morales,

87

2007) promote mammography screening use among Latinas and how extremely high
levels of worry and distress create a barrier to mammography screening (Andersen, et al.,
2003; Schwartz, Taylor, & Willard, 2003). However, other study ﬁndings contradicted
prior research which has showed that health insurance is the leading predictor of
mammography screening among Latinas, (Carrasquillo & Pati, 2004; Rodriguez, et al.,
2005), and a higher language-based acculturation (Gorin & Heck, 2005; O’Malley,
Kerner, Johnson, & Mandelblatt, 1999) and higher education predict screening after
controlling for other factors (Gorin & Heck, 2005; Jones, Caplan, & Davis, 2003).Thus,
ﬁndings from this San Diego sample did not fully replicate what other researchers have
found when examining predictors of breast cancer screening among Latinas.
Predisposing, Enabling, Need, and Contextual Predictors

Enabling Factors

Although this study failed to replicate previous study ﬁndings showing that
Latinas over the age of 50 are more likely to use mammography than those aged 40 to 49,
(Fernandez & Morales, 2007; Gorin & Heck, 2005), this study did ﬁnd a positive
relationship between age and recent mammography screening, in adjusted and unadjusted
analyses. Since this study included age as an enabling variable, studies are needed to
determine what component of age enables screening, such as age-speciﬁc physician
referral systems or personal motivations to obtain a mammogram, stemming from cancer
risk awareness or knowledge of age-appropriate cancer screenings exams. Post-hoe
analyses revealed there were no age differences between those who have been to see a
physician in the past year and those who have not (T = 1.668 (df = 205) p = .097), but

those who had insurance were signiﬁcantly older (M = 52.04, SD = 9.45, n = 140) than

88

those without insurance (M = 48.89, SD = 6.89, n = 67) (T = 2.433 (df = 208) p = .016).
However, since Latinas with health insurance and who were employed15 were
signiﬁcantly younger (M = 49.81, SD = 6.83, n = 97) than Latinas with health insurance
and unemployed (M= 57.30, SD = 12.55, n = 41) (T = - 4.517 (df= 136), p = .000), there
is little evidence to suggest that employment status explained the link between a higher
age and likelihood of insurance coverage. Thus, studies should examine the relationship
between age and health insurance in relation to other enabling resources, such as income.

A recent physician visit remained the most signiﬁcant predictor of recent
mammography screening adjusting for other enabling variables, (OR = 3.336) and
adjusting for variables in the Behavioral Model of Latinos Breast Cancer Screening Use
(OR = 4.170) (Tables 5 and 11). These results substantiate prior research showing that
visiting a physician in the past year (Gorin & Heck, 2005) predicted Latina breast cancer
screening, adjusting for other variables. This suggests that above all, access and entry into
the health care system is critical to receipt of mammography screening among Latinas
(e. g., Zambrana, et al., 1999). In order to resolve why a physician visit may matter so
much, research is needed to determine if an increased exposure to the health care system,
somewhat like a dosage effect, promotes screening among Latinas. Research is also
needed to understand the role that physicians as trusted sources of information plays in
patients’ motivation to obtain mammograms (Metsch, et al., 1998).

This study failed to replicate previous research showing that health insurance was

the strongest predictor of having a mammogram for Latinas (Abraido-Lanza, Chao, &

 

'5 In this sample, 60.6% (n = 123) were currently employed (n = 110 employed for wages and n = 13 self-
employed) and 39.4% (n = 80) were unemployed (n = 5 retired, n = 47 homemaker, n = 6 on social
security/SS1, n = 10 unable to work, n = 11 out of work, and n = 1 student).

89

Gammon, 2004; Carrasquillo & Pati, 2004; Rodriguez, et al., 2005). Although private
insurance was related to screening adjusting for enabling variables (Table 5), private and
public insurance did not signiﬁcantly predict screening adjusting for variables in the
Behavioral Model of Latina Breast Cancer Screening Use (Table 11). Since previous
studies show that either private or public insurance predict mammography screening
among Latinas (Gorin & Heck, 2005), the role that insurance plays in the process of
obtaining a mammogram needs to be explored further.

Perhaps health insurance was not found to be as nearly important as previous
studies because of a context-speciﬁc phenomenon unique to this study. One plausible
explanation is the availability and utilization of federally-qualiﬁed community health
centers (CHCs) throughout San Diego County that offer bilingual breast cancer screening
services for the underinsured and uninsured”. In fact, most (63.5%, n = 132) of the
sample of Latinas 40 years and older were recruited by staff of one such federally-
qualiﬁed CHC located in northern San Diego County and 36.5% (n = 76) were recruited
by UCSD staff. Although CHC staff were instructed not to recruit from the CHC patient
population, it is possible that patients were recruited at community-based events in the
catchment area of the CHC. Recruitment data from the entire sample of 503 Latinas 18

years and older, showed that CHC staff recruited 42.5% (n = 149) through community-

 

'6 Across the United States, there are around 1,200 federally-funded community health centers (CHCs), that
provide comprehensive health care services, such as mammography screening at free or reduced rates, for
around 15 million underserved patients annually; underserved patients include the under-insured and
uninsured, and other disadvantaged groups, such as the elderly, homeless, migrant farmworkers, and public
housing residents (National Association of Community Health Centers (N ACHC), 2008a; 2009). Many
CHCs have bilingual staff; in 2007, of the 260 CHCs surveyed, 84% had Spanish-speaking staff and 68%
had more than 10% of their patients who were Spanish-speaking (NACHC, 2008c). In California, 110
federally-funded CHCs with 796 health care delivery sites served 2.3 million patients in 2007 (NACHC,
2008b). In 2007, 45% of California CHC patients were uninsured compared to 19% of the state; 76% of
patients were at or below 100% of poverty compared to 19% of the state; and 62% of patients were Latino
compared to 36% of the state (NACHC, 2008b). In San Diego, there were 92 delivery sites in 2009 (Health
Resources and Services Administration, 2009).

90

based events hosted by the CHC, (e. g., health fair) (Appendix A). Given that CHC staff
did not inquire potential participants if they were CHC patients, it is entirely possible that
a signiﬁcant proportion of women recruited by the CHC were indeed clinic patients.

Post-hoe analyses showed that there was no signiﬁcant difference in likelihood of
recent mammography screening [)8 (1, 202) = 2.722, p = .099] and a recent physician
visit [7(2 (1 , 207) = 2.325, p = .127] based on CHC versus UCSD recruitment. However,
those recruited by UCSD were more likely to have insurance (81.6%, n = 62) than those
recruited by the CHC (59.5%, n = 78) [x2 (1, 207) = 10.670, p = .001]; and of those with
health insurance, those recruited by the CHC were more likely to have public as opposed
to private (33.8%, n = 25) versus those recruited ﬁ'om UCSD (8.3%, n = 5) [x2 (1, 74) =
12.352, p = .000]. Even though CHC-recruited participants were more likely to be
uninsured or have public health insurance, they did not differ in their mammography
screening rates. This might suggest that communities can link to available local resources
and opportunities that override the negative consequences of being uninsured, which may
help explain why insurance was not an important predictor of screening relative to other
factors in the current study. Research is needed to examine these speciﬁc community and
healthcare contextual pathways that positively inﬂuence screening.
Predisposing Factors

Although prior research has found that a having at least a high school degree
(Jones, Caplan, & Davis, 2003), or at least a college degree (Gorin & Heck, 2005)
predicts breast cancer screening use among Latinas in multivariate analyses; in this study,
the relationship between education and screening was concealed until acculturation was

accounted for. In fact, language-based acculturation suppressed some of the irrelevant

91

variance of the relationship between education and screening (Rosenberg, 1973),
meaning, the uniqueness of education unrelated to acculturation is what predicted
screening. Since some studies show that education does not relate to Latina breast cancer
screening use after adjusting for other variables (Fernandez & Morales, 2007), research is
needed to conﬁrm the suppression effect was not due to chance (Collins & Schmidt,
1997) and to clarify the causal mechanisms by which education leads to screening.

The original Behavioral Model assumed that education leads towards a greater
knowledge of health, diseases, and available health care services (Andersen, 1968).
Education may relate to employment and higher income (Muller, 2002; US. Census
Bureau, 2006d), which increase access to resources that enable utilization of preventive
health care services (Andersen, 1968; Heck & Makuc, 2000). In fact, Latinas in this
sample who were employed were more likely to have insurance (79.5%, n = 97) than
those not employed (51.3%, n = 41) [x2 (1, 202) = 17.825, p = .000], however,
employment status was not associated with recent mammography screening [x2 (1, 198) =
0.827, p = .363]. Future studies should examine how education, employment status and
other resources such as income relate to use of mammography screening among Latinas.

In this study, the relationship between language-based acculturation and screening
was concealed until education and country of birth were accounted for. In fact, these two
variables suppressed some of the irrelevant variance of the relationship between
language-based acculturation and screening (Rosenberg, 1973), meaning, the uniqueness
of acculturation unrelated to these two variables is what predicted screening. Since
acculturation has been deﬁned as the cultural change process of adopting language,

attitudes, values, and behaviors ﬁ'om another culture (Suarez, 1994; Wells, 1989), future

92

studies should explore aspects of the acculturation process other than language that could
be relevant to screening, irrespective of education level and country of birth.

Studies show that a limited English proﬁciency is a barrier to accessing and using
health services because of the difﬁculty of communicating with and understanding
providers (Doty, 2003; Salazar, 1996). Spanish-language has been shown to be a barrier
to mammography screening among Latinas (Gorin & Heck, 2005; O’Malley, et al., 1999;
Ortero-Sabogal, et al., 2003; Suarez & Pulley, 1995); however sometimes it makes no
difference, after considering other demographic factors (Fernandez & Morales, 2007;
Suarez, 1994). Results from this study contrast the literature that deems Spanish-language
usage as a barrier to mammography screening use, which leads one to question the
unique characteristics of the Southern California context that allowed predominately
Spanish-speaking Latinas in this study to obtain mammography screening, irrespective of
their education level and country of birth. Around 30% of San Diego County residents
are Latino, (US. Census Bureau, 2006b), of which a signiﬁcant portion is unacculturated
(e.g., in 2007 18.6% of Latinos spoke only Spanish at home) (http://www.chis.ucla.edu)
and a signiﬁcant portion of community health centers (CHCs) and private sector health
care systems, such as Sharp Healthcare (http://www.sharp.com), throughout the county
offer bilingual health care services. Perhaps in this context, English-language proﬁciency
was not found to be as nearly important as in previous studies.

Studies show that foreign-born Latinas are more (Rodriguez, Ward, & Perez-
Stable, 2005) or less likely (Goel, et al., 2003; Otero-Sabogal, et al., 2003) to receive
mammography screening, and other studies show that Latina foreign birth did not predict

mammography screening (Valdez, et al., 2001); and this study sought to clarify these

93

mixed ﬁndings. In fact, study results showed no relationship between nativity and
screening, even after controlling for other factors. Even though research shows that
Latina immigrants are likely to report barriers to accessing health care services (Garces,
Scarinci, & Harrison, 2006), US. and Mexican-born Latina participants did not differ in
their likelihood of having a recent physician visit [7(2 (1, 197) = .946, p = .331]. Although
U.S.-born Latinas were more likely to be insured and have a higher education than
Mexican-born Latinas”, country of birth did not predict screening in the presence of
factors, such as education and insurance (Tables 4 and 8). Studies should assess other
aspects of nativity, such as citizenship status or time since immigration, that may create
barriers to knowing how and being able to navigate the local health care system.
Need-related Factors

Health-related quality of life, (including health status, recent unhealthy days, and
recent activity limitation), was not signiﬁcantly related to screening in this study. Given
that prior research has shown that a higher health-related quality of life positively
predicts health care utilization (Dominick, et al., 2002) and a higher health status
predicted breast cancer screening in Latinas, even after adjusting for other factors
(Fernandez & Morales, 2007), research is needed to fully examine the role of mental and
physical health status in health behaviors, such as breast cancer screening.
Context-specific Factors

Having a personal or family history was not related to screening, which contrasts

prior research showing that a personal or family history of cancer increased the likelihood

 

l7U.S.-born Latinas (85.1%, n = 40) were more likely than Mexican-born Latinas (35.3%, n = 53) to have a
high school education or greater [x (1, 197) = 35.574, p = .000] and U.S.-born Latinas (85.1%, n = 40)
were more likely that Mexican-born Latinas (60.7%, n = 91) to be insured [x2 (1, 197) = 9.595, p = .002].

94

of breast cancer screening among Latinas (Aparicio-Ting & Ramirez, 2003; Cohen, 2006;
Gorin & Heck, 2005). Since a personal/family history of breast cancer has been shown to
increase one’s perceived risk of developing cancer (Kim, et al., 2008), future studies are
needed to examine the relationship among a history of breast cancer, perceived cancer
risk and breast cancer screening among Latinas.

Cancer fatalistic beliefs were not related to screening in this study, which
contrasts prior research showing that cancer fatalistic beliefs impede Latina’s use of
breast cancer screening exams (Hubbell, et al., 1997). Fatalistic beliefs are thought to
promote an external locus of control, causing individuals to believe that they have no
control over their health, which can deter mammography utilization (Barroso, et al.,
2000; Borrayo & Guarnaccia, 2000; Niederdeppe & Levy, 2007). Latinas in this study
tended to have moderate levels of fatalistic beliefs, which supports prior research
(Borrayo & Jenkins, 2001; Chavez, etal., 1997; Florez, et al., 2008; Hubbell, et al.,
1996); perhaps these beliefs were not severe enough to impede screening. Research is
needed to examine the relationship between health locus of control, severity of fatalistic
beliefs, and mammography screening use.

Religiosity were not related to screening, which contrasts prior research showing
that church attendance was associated with breast cancer screening among Latinas
(Otero-Sabogal, et al., 2003). Given that religiosity is a key facet of Latino culture that
can promote positive health behaviors (Magaﬁa & Clark, 1995), studies are needed to
ﬁrrther examine the role that different components of religiosity among Latinas play in

health behaviors such as mammography screening.

95

Frequency of assistance used in reading hospital materials was not associated with
screening. In bivariate and multivariate analyses, conﬁdence in ﬁlling out medical forms
was a signiﬁcant predictor of screening, which supports previous research showing that
an adequate level of health literacy was associated with mammography screening in
Latinas (Guerra, Krumholz, & Shea, 2005). Since a low health literacy level relates to a
lack of knowledge about the importance of and reason for screening mammography,
which may relate to underutilization of mammography (Davis, Arnold, & Berkel, 1996),
research is needed to examine the interplay of health literacy and other predisposing and
enabling variables, such as education and income, in predicting use of screening among
Latinas. Studies are also needed to understand additional components of health literacy,
such as knowledge of navigating the health care system and informed decision making
(Selden, Zorn, Ratzan, & Parker, 2000), and how these relate to patient’s perceptions of
provider cultural competency, patient satisfaction with patient-provider communication
(DeVoe, Wallace, & Fryer, 2009; Lloyd, Ammary, Epstein, Johnson, & Rhee, 2006), and
use of mammography screening services.

Breast cancer worry was not associated with screening in this sample. However,
in bivariate and multivariate analyses, lower levels of breast cancer distress were
associated with an increased likelihood of screening, which supported ﬁndings in
previous studies where extremely high levels of worry and distress decreased the
likelihood of mammography among Latinas (Andersen, et al., 2003; Schwartz, Taylor &
Willard, 2003). Worry is sometimes considered the cognitive aspect of the anxiety and
may function as an adaptive coping mechanism in attempting to psychologically prepare

for an external threat (Hay, Buckley, & Ostroff, 2005; Kubzansky, et al., 1997). Future

96

studies are needed to examine how aspects of extreme levels worry, distress, and
cognitive aspects of anxiety impede health behaviors among Latinas and how these
thoughts may be overcome.

Implications for Interventions

The results of this study yield several implications that may inform future
research and the development of breast cancer screening promotion interventions for
Latinas 40 years and older. Breast cancer screening utilization in this sample differed by
level of education, age, acculturation, health literacy, cognitive distress about developing
breast cancer, and whether women have visited a physician recently. Thus, interventions
need to be tailored to the context of women’s lives and may vary depending on the
predisposing, enabling, and need-related characteristics of the target population.

Research shows that interventions that were culturally-tailored, incorporated the
health belief model, or included physician recommendations signiﬁcantly increased
breast cancer screening adherence in diverse samples (Sohl & Moyer, 2007). Indeed,
results from this study imply that individual behavioral change interventions to increase
breast cancer screening adherence among Latinas should be culturally and linguistically
tailored based on the level of health literacy, language of choice, and extent of cognitive
distress about developing breast cancer experienced. System-level change interventions
should focus on incorporating concepts of increasing access to care, such as age-based
targeting of patients or promoting mammography referrals by general practice physicians
(Legler, Meissner, Coyne, Breen, Chollette, & Rimer, 2002; Sohl & Moyers, 2007).
Since no other factor was consistently more predictive of adherence to breast cancer
screening guidelines, the encouragement to have an annual physical examination appears

to be the single most important message health educators need to convey.

97

Study Limitations

This study has several limitations. One limitation is that the secondary data for
this study were derived from a cross-sectional self-report survey, which although can be
considered a snapshot of participants’ reality, limited the ability to develop causal
inferences about the relationships among variables examined in this study. Inherent in the
cross-sectional design was the focus on retrospective accounts of the amount of time
since their last mammogram. This limitation could be overcome by examining patterns of
mammography utilization over time. Perhaps future studies can also include independent
observations of screening, obtained ﬁ'om health care providers or medical records, to
verify the reliability of self-reported mammography screening. Another limitation of this
study was the single geographic focus from which the convenient sample of study
participants was drawn. All participants were recruited throughout San Diego County,
California and most of the sample was recruited by community clinic staff, thus,
generalizations to other areas in the United States must be made with caution. Not all of
the measures had reliable estimates since several variables were single items (e. g., age,
health insurance, and recent physician visit) rather than aggregate reliable scale scores;
thus, unreliability of selective independent and dependent variables cannot be completely
ruled out as a potential inﬂuence on this study’s results. In addition, the assumptions of
the Andersen Behavioral Model are limited in the sense that the model does not take into
account signiﬁcant contextual features (e. g., time, changes in health insurance coverage,
and changes in recommended standards of utilization) that may affect the predisposing,
enabling, and need factors and the utilization outcome of interest. There is also a need for

longitudinal studies that can examine the sequencing of events that trigger obtainment of

98

mammography screening services to determine if an annual examination leads to
subsequent receipt of mammography screening utilization. Despite these limitations,
results from this study have the potential to lend insight to future research, community-
based health promotion, and primary care practice in relation to increasing breast cancer

screening adherence among Latinas in Southern California.

99

CONCLUSION

Increasing breast cancer screening usage is one key tactic in comprehensive
cancer control by minimizing the proportion of late-stage diagnoses (True, Kean, Nolan,
Haviland, & Hohman, 2005). This study used theoretically-driven analyses to determine
signiﬁcant predictors of breast cancer screening utilization among Latinas 40 years and
older in San Diego County, California. Using Andersen’s Behavioral Model of Health
Services Use (Andersen, 1995) as a framework, the Behavioral Model of Latina Breast
Cancer Screening Use was developed in this study and provided a reasonable framework
for explaining and predicting breast cancer screening behavior in this sample of 208
Latinas 40 years and older.

Healthcare policy-makers and community-based practitioners working with
Latinas should recognize the importance of individual characteristics and behaviors, such
as acculturation, health literacy, age, access to healthcare, and distressﬁrl cognitions
about developing breast cancer, as crucial facilitators or impediments to mammography
screening use among this population. Public health researchers should incorporate the
Behavioral Model of Latina Breast Cancer Screening Use framework into the process of
designing culturally appropriate evidence-based cancer control interventions. Future
research is needed to determine if this framework applies to other Latina health behaviors
and to other recent immigrant populations to the United States.

This study showed that access to care matters in the process of obtaining
preventive services since no other factor was consistently more predictive of adherence
breast cancer screening guidelines than undergoing an annual examination. In fact,

having a recent physician visit in the last year increased the odds of having a recent

100

mammogram by 3 or 4 times in adjusted logistic regression models. Future research
should explore the role that culturally competent primary care providers as trusted
sources of health information play in motivating Latinas to obtain preventive health care
services (Metsch, et al., 1998). Future research is also needed to explore the application
of the F oot-in-the-Door theory (Burger, 1999; Pascual & Gueguen, 2005), by which
access to the health care system through non-invasive means leads to an increased
receptivity to recommendations for mammography screening and other more invasive

procedures.

101

APPENDICES

102

APPENDIX A

 

 

Recruitment Strategies

Recruitment Location Strategy Interviewer

(# of Surveys
Collected)
Community-based Word-of-mouth, VCC (n=149)
Clinic event phone, and
intemet bulletin
Hospital/clinic staff Word-of-mouth UCSD (n=15)
University staff Word-of-mouth UCSD (n=41)
(main campus and
medical centers)

Participant’s Home Phone, email, UCSD (n= 38)
and word-of-mouth VCC (n=114)

Participant’s Workplace Phone, ﬂyer, UCSD (n=9)

and word-of-mouth VCC (n=29)

Latino Organization Phone, email, UCSD (n=18)

(community-based, and word-of-mouth VCC (n=22)

professional, or
educational)

Church Word-of-mouth VCC (n=3 7)
Community Center Word-of-mouth UCSD (n=19)
Recruitment Totals UCSD (n=140)

VCC (n=351)

Total (N =491)

 

103

Note: UCSD= Staff ﬁom the Moores UCSD Cancer Center Outreach Department,
University of California, San Diego (located in La J olla, California), who recruited
participants throughout Central and Southern San Diego County, CA.

Note: VCC= Staff from the Health Promotion Center, Vista Community Clinic (located
in Vista, CA), who recruited participants throughout North San Diego County, CA.
Note: Total sample size was n=503 (n=144 (28.6%) participants were recruited and had
surveys administered by UCSD and n=359 (71.4%) participants were recruited and had
surveys administered by VCC.

APPENDIX B
Consent Documents
English Consent Document
UCSD Informed Consent Document

Dr. Georgia Robins Sadler, Associate Director for Outreach at the Moores UCSD Cancer
Center and Ms. Natasha Riley, MA Program Manager for Vista Community Clinic are
conducting a research study to evaluate the effectiveness of a breast cancer clinical trials
education program. The program has been speciﬁcally developed to help adult Aﬁican
American and Hispanic American women to learn about clinical trials. You have been
asked to participate in the study because you are believed to be a member of one of these
two communities.

There are three parts to this study. You are being asked to participate in study checked
below.

1:1 Study 1 Many surveys have been developed to evaluate people’s health and well-
being. Few have been proven to work when given to Hispanic Americans. This study is
inviting 250 English-speaking and 250 Spanish-speaking women to complete some of the
surveys that are frequently used in health and research settings. Those responses will
allow us to learn if the surveys work equally well for Hispanic women. The Vista
Community Clinic is inviting 500 Hispanic American women to complete those surveys.
It will take about 10 minutes to complete the informed consent process and another 40
minutes to complete the surveys. There will also be time afterward for discussion or
additional questions once the surveys have been completed. You will also be asked if you
would like to help invite other women to join this study. You will be offered light
reﬁeshments while you complete the surveys and $20 for your time.

1:1 Study 2 The National Cancer Institute created a clinical trials education program and
then modiﬁed it so it would be culturally appropriate for Hispanics. After testing, the
program was not proven to be effective and no program was ever created speciﬁcally for
the Aﬁican American community. We have modiﬁed the original program in a way that
we think will make them useﬁrl for both communities. Now we would like to ask you to
watch them and give us your opinion of how useful and acceptable these programswill
be to your community. The program is approximately 45 nrinutes plus time for discussion
and questions and answers. You will be offered light reﬁ'eshments and $35 for your time.
We are asking 80 African American and Hispanic American women to help with these
focus groups.

:1 Study 3 This is a pilot study that will begin to test the clinical trials education
programs that are developed in the second study. First you will complete the surveys that
were found to be useful in Study 1 and answer some general knowledge questions. Then
half of the group will receive the program developed in Study 2 and the other half will

104

 

participate in a different health program. After you complete the program, you will be
asked to answer some questionnaires again. You will be compensated $35 for your time.
We will invite 20 Aﬁican American, 20 English-speaking Hispanic American, and 20
Spanish-speaking Hispanic American women to join this study.

Participation in this study will help raise awareness about health issues related to breast
cancer clinical trials education. The information you learn may beneﬁt you or your family
members or other members of your community in the future. Collectively, what we learn
from you should help us offer more effective clinical trials education in the future and
that will potentially beneﬁt the entire community.

All personal information that you provide (such as your name, address, e-mail address,
phone number) will remain conﬁdential. The only known risk to participating in this
study is that an unauthorized person may illegally gain access to university ﬁles in spite
of the University’s and research staﬂ members ’ best eﬂorts to avoid this possibility. To
reduce this risk, personal identiﬁable information is kept separately ﬁ'om other study
information.

Participation is voluntary. You may discontinue participation at any time without penalty
or loss of beneﬁts. If you have study related questions or wish to report a concern please
contact Dr. Sadler at (858) 534-7611. If you have other research related questions or are
injured as a result of this study, call the Human Research Protection Program at (858)
455-5050.

You have received a copy of this consent document and a copy of the Experimental
Subject’s Bill of Rights for your records.

Dr. Sadler, Ms. Riley, or one of their assistants has
explained this study to you. By signing below you are agreeing to take part in this study.

 

 

 

Participant Signature Date

 

 

Witness Signature Date

105

Spanish Consent Document
UCSD Documento de Consentimiento Informado

Dra. Georgia Robins Sadler, Directora Asociada de ‘Outreach’ del Centro para Cancer
Moores de UCSD y Natasha Riley, Gerente del Prograrna MA de la Clinica de la
Comunidad de Vista, dirigen la investigacion y los estudios para evaluar la efectividad de
un programa de educacion sobre las pruebas clinicas del cancer de seno. El programa es
desarrollado especiﬁcamente para ayudar alas mujeres Hispanas y Aﬁo Americanas
adultas, para dar a conocer las pruebas clinicas. Le hemos pedido que participe en el
estudio porque usted es crel’do ser un miembro de una de estas comunidades.

La investigacion consiste de tres estudios. Le pedimos que participe en el marcado a
continuacion:

1:1 Estudio 1. Se han desarrollado muchas encuestas para evaluar la salud y el bienestar,
sin embargo, muchas fallan cuando se les dan a 10s Hispanos Americanos. Este estudio
consiste en invitar a 250 mujeres Hispanas que hablan Espaﬁol y 250 que hablan Inglés a
responder parte de las encuestas que son usadas comunmente en lugares de salud e
investigacion. Las respuestas nos permitirén aprender si dichas encuestas tienen la rrrisma
efectividad con la mujer Hispana. La Clinica de la Comunidad de Vista ha invitado a 500
mujeres Hispano Americanas a completar las encuestas. Le tomara unos 10 minutos para
el procedimiento para su consentimiento y otros 40 minutos para contestar las encuestas.
Ademas, tendra tiernpo extra para hacer preguntas adicionales, después de completar la
encuesta. También 1e pediremos si gusta invitar a otras muj eres para participar en este
estudio. Le ofrecemos $20 dolares y bebidas ligeras por su tiernpo.

1:1 Estudio 2. El Instituto Nacional del Cancer creo el programa de educacion para
pruebas clinicas y después 10 modiﬁed para hacerlo culturalmente adecuado para los
Hispanos Americanos. Después de las pruebas se encontro que el programa no es efectivo
y no se creo ningr’rn programa especiﬁco para la mujer Afro Americana. Hemos
modiﬁcado el programa original de forma que creernos sera util para ambas
comunidades. Hoy, 1e pedimos que lo vea y nos diga, en su opinion gQué utiles y
aceptados seran estos programas en su comunidad? Su duracién es de 45 minutos, mas
tiernpo extra para preguntas y respuestas, se le ofrecen bebidas ligeras y $35 délares por
su tiernpo. Se 1e pedira a 80 mujeres Hispano Americanas y Afro Americanas que nos
ayuden con estos grupos.

1:1 Estudio 3. Este es un estudio piloto para empezar a probar los prograrnas de educacion
para las pruebas clinicas, que seran desarrollados en el Estudio 2. Primero deberé
contestar a1 gunas preguntas de conocimiento general y las encuestas que se encontraron
l'rtiles en el Estudio 1. La mitad del grupo recibira el programa desarrollado en el Estudio
2 y la otra mitad participara en un programa de salud diferente. Después de completar el
programa se le pedira que conteste otras encuestas. Se 1e ofrecen bebidas ligeras y $35
dolares por su tiernpo. Se le pedira participacion a 20 Hispano Americanas y 20 Afro
Americanas que hablan Inglés y 20 Hispano Americanas que solo hablan Espafiol.

106

Su participacion en estos estudios nos ayuda a concientizar sobre los problemas del
cancer de seno y las pruebas clinicas para educacion y salud. La inforrnacién que usted
aprenda le puede beneﬁciar en el ﬁrturo, asi como a su farnilia y a otros miembros de la
comunidad. Lo que aprendemos de usted, potencialmente debera ayudamos a dar mejor
educacion sobre las pruebas clinicas que puede beneﬁciar a la comunidad entera.

T oda la informacio'n que usted nos dé (nombre, domicilio, tele'fono y correo electro'nico)
es conﬁdencial. El unico riesgo posible es si alguna persona no autorizada, a pesar de
los mejores esfuerzos del personal universitario para evitar esta posibilidad, pueda
obtener acceso a los archivos de la universidad. Para reducir este riesgo, la informacio'n
identzﬁcable se mantiene separada del resto del estudio.

Su participacion es voluntaria y puede suspenderla en cualquier momento, sin penalidad
o pérdida de sus beneﬁcios. Si tiene alguna pregunta o quiere reportar alguna
preocupacion, por favor llame a la Dra. Sadler al (858) 534-7611. Si tiene otras preguntas
relacionadas al estudio 0 se ha lastirnado como resultado de participar, llame por favor a1
Programa de Proteccién para Investigaciones Hurnanas a1 (85 8) 455-5050.

Para su informacién y registros, se le ha entregado una copia de este Documento de
Consentimiento Informado y una copia de Las Leyes y Derechos de los Sujetos a
Experimentacion.

La Dra. Sadler, Sra. Riley, 0 uno de sus asistentes 1e ha
explicado este estudio.

 

Con su ﬁrma abajo usted acepta participar en este estudio.

 

 

Firrna del Participante Fecha

 

 

Firrna del Testi go Fecha

107

APPENDIX C
Selected Pages ﬁ'om English and Spanish Surveys
PERSONAL HEALTH SURVEY

Please answer the questions the best you can. You may skip questions that make you
uncomfortable.

BACKGROUND

1. What is your date of birth (mm/dd/yyyy)? _ _/_ _/_ _ _ _
2. What is your current employment status?
Employed for wages
Self-employed - '1
Out of work for more than one year
Out of work for less than one year
Homemaker
Student
Retired
Unable to work
SSI

 

000000000

3. What is your marital status?
Married

Divorced

Widowed

Separated

Never been married
Living with a partner

000000

4. Do you have children?
0 No 0 Don’t know 0 Yes -—>

a. IF YES: How many?

 

5. What is the highest level of school you completed?

Never attended school or only attended nursery school/kindergarten
Elementary school or primary school

Middle school or junior high school

Some high school but no diploma

High school graduate (high school diploma or equivalent, such as
GED, foreign equivalent)

0 Vocational or trade school graduate

00000

108

Some college, but no degree

Associate degree in college

Bachelor’s degree

Master’s degree

Professional school or doctorate degree (MD, DDS, JD, DVM, Ph.D.,
Ed.D., etc.)

0 Don’t know

00000

6. What is your religious preference?
Christian

Catholic

Jewish
Muslim/Islamic
Buddhist
Unitarian/Universalist
Hindu

Native American

None
Other

OOOOOOOOOO

 

7. In what country were yo_u_ born?

 

HEALTH CARE ACCESS

8. Do you have any kind of health care coverage, including health insurance,
prepaid plans such as HMO’s, or government plans such as Medicare/MediCal?

O No 0 Don’t know 0 Yes—>

. IF YES: What type?
Medi-Cal/Medicaid
Medicare

HMO

PPO
Employer-provided
Other

000000

 

9. Was there a time in the past 12 months when you needed to see a doctor but
couldn’t?

O No 0 Don’t know 0 Yes—r

a. IF YES: Why couldn’t you go?
0 Did not know where to go
0 Language barriers (Didn’t understand; no Hispanic/bilingual
staff)

109

0 Don’t have a regular doctor

0 No health insurance

0 Cost too much (for co-pay or sliding scale fee)

0 Lack of doctors who provide services to Medicaid patients
0 Lack of transportation (or lack of money for gas)

0 Lack of time (had to work)

O Fear of being reported to the immigration service

0 Not sick

0 Other

 

10. About how long has it been since you last visited a doctor for a routine checkup?
A routine checkup is a general physical exam, not an exam for a speciﬁc injury,
illness, or condition.

Within past year (anytime less than 12 months ago)

Within past 2 years (1 year but less than 2 years ago)

Within past 5 years (2 years but less than 5 years ago)

5 or more years ago

Don’t know

Never

000000

11. Have you ever been to a traditional healer like a curandero?

O No 0 Don’t know . 0 Yes —>
a. If yes, how long ago did you see the healer? (months)
b. Where did you go see the healer? (country)

 

c. What type of healer was it?
0 Curandero (faith healer)
O Huesero (bone-setter)
O Espiritualista (Spiritualist)
O Sobador (massage therapist)
O Yerbero (herbalist)

d. What problem did you go to the healer for?

 

e. Why did you go to the healer instead of the clinic or doctor?
0 Some illnesses cannot be treated by clinical medicine.
0 Less expensive than clinical medicine
0 Have tried clinical care, and it did not work
0 Someone advised me to see a traditional healer
O Prefer natural medicines
O Other
0 Don’t know

 

110

12. A mammogram is an x-ray of each breast to look for breast cancer. Have you
ever had a mammogram?

O No 0 Don’t know 0 Yes —>
a. IF YES: How long has it been since you had your last
mammogram?
0 Within past year (anytime less than 12 months ago)
0 Within past 2 years (1 year but less than 2 years ago)
0 Within past 5 years (2 years but less than 5 years ago)
0 5 or more years ago
0 Don’t know
0 Never

13. A clinical breast exam is when a doctor, nurse, or other health professional feels
the breast for lumps. Have you ever had a clinical breast exam?

ONo

0 Don’t know 0 Yes—>

a. IF YES: How long has it been since your last breast exam?

OOOOOO

Within past year (anytime less than 12 months ago)
Within past 2 years (1 year but less than 2 years ago)
Within past 5 years (2 years but less than 5 years ago)
5 or more years ago

Don’t know

Never

14. Has a family member ever had breast cancer?

ONo

0 Don’t know 0 Yes—>

a. IF YES: Which family member?

OOOOOO

Mother
Grandmother
Sister

Aunt

Cousin
Other

 

15. Have you ever had breast cancer?

ONo

0 Don’t know 0 Yes

111

ENCUESTA SOBRE LA SALUD PERSONAL Y F AMILIAR

Por favor responda las preguntas lo mejor que pueda. Se puede pasar las preguntas que lo
hagan sentir incomodo.

ANTECEDENTES

1. LSu fecha de nacimiento (mes/dia/aﬁo)? _ _/_ _/_ _ _ _
2. aCual es su posicién de empleo actual?
Empleo a salariado
Auto-empleado

Sin trabajo por mas de un aﬁo
Sin trabajo por menos de un aﬁo
Ama de casa

Estudiante

Retirada

Incapacitada para trabajar
Seguro Social (SSI)

000000000

3. 5Cual es su estado civil?
Casada
Divorciada
Viuda
‘ Separada
Soltera
Viviendo con mi pareja

000000

4. LTiene niiios?
O No 0 Si -—>

a. Si si: gCuantos?

 

5. éCual es su nivel mas alto de escolaridad completado?
O Nunca asisti a la escuela, o solo a] kinder
O Primaria
O Secundaria
O Preparatoria, pero sin diploma 0 certiﬁcado
O Graduada de preparatoria (diploma, certiﬁcado o equivalente
extranjero, ejemplo GED)
Graduada de escuela vocacional o técnica »
Universidad, pero sin diploma 0 certiﬁcado
Certiﬁcado Universitario Asociado
Diploma o certiﬁcado de licenciatura
Diploma o certiﬁcado de Maestria

00000

112

O Doctorado de Escuela Profesional Universitaria (Doctor, Dentista,
Abogado, Doctorado, etce’tera)
O No sé

6. Lean es su preferencia religiosa?
Cristiana

Catélica

Judia

Musulmana / Islarnica
Budista

Unitaria / Universalista
Hindu

Nativa Americana
Ninguna

Otra

OOOOOOOOOO

 

7. (En qué pais nacio usted?

 

ACCESO A CUIDADOS DE SALUD

8. aTiene cualquier tipo de cobertura de salud? Incluyendo: Seguro médico, planes
HMO, 0 del gobierno como Medicare 0 Medi-Cal?

O No ONosé O Si—->

a. Si si: gDe qué tipo?

Medi-Cal / Medicaid

Medicare

HMO

PPO

Proporcionado por su trabajo (empresa)
Otro

OOOOOO

 

9. LEn los ultimos 12 meses, necesito ver a un doctor pero no pudo?
O No 0 No sé O Si —>

a. Si si: gPor qué no pudo verlo?

O No sabia dénde acudir

O Problemas de idioma (N o entendia; no tenian personal
Hispano/bilingiie en espafiol)

O No tengo un doctor regular

0 No tengo seguro de salud (médico)

O Muy caro (de co-pay o pago escalonado)

O F alta de doctores para los pacientes de Medicaid

O Falta de transporte (o dinero para gasolina)

113

O Falta de tiernpo (tenia que trabajar)

O Ternor de ser reportada a los servicios de migracién (migra)
O No he estado enferma

O Otra razén

 

10. gMas o menos hace cuanto que le hicieron un examen médico de rutina? Un
examen de rutina en general es una revision fisica, no por causa de herida,
enfermedad o condicién especial.

En este aﬁo (cualquier vez en los ultimos 12 meses)

Ultimos 2 aﬁos (mas de un aﬁo, pero menos de dos)

En los ﬁltimos 5 aﬁos (mas de dos, pero menos de cinco)

5 o mas aﬁos

No sé

Nunca

000000

11. LAlguna vez ha consultado a un curandero tradicional?

 

O No 0 No sé O Si ——+
a. Si si ghace cuanto tiempo que vio al curandero? (meses)
b. LDénde vio al curandero? (pais)

 

c. aQue’ tipo de curandero?
O Curandero

O Huesero

O Espiritualista

O Sobador

O Yerbero

d. 5Por qué problema fue al curandero?

 

e. aPor qué fue al curandero en lugar de ir a un doctor 0 a una
clinica?

O Porque a1 gunas enfermedades no pueden ser curadas por la medicina
Clinica

O Mas barato que la medicina Clinica

O He tratado con la medicina Clinica pero no sirvic')

O Alguien me recomendé a1 curandero tradicional

O Preﬁero 1a medicina naturista

O Otro

O No sé

 

12. Una mamografia son rayos—X en cada seno para buscar cancer. 5Alguna vez le

han hecho una mamografia?
O No 0 No sé O Si —+

114

OOCDOOO?D

Si si: (,Hace cuanto tiempo 1e hicieron su ultima mamografia?
En este aﬁo (cualquier vez en los ultimos 12 meses)

Ultimos 2 aﬁos (mas de un aﬁo, pero menos de dos)

En los ultimos 5 aﬁos (mas de dos, pero menos de cinco)

5 o mas aﬁos

No sé

Nunca

13. Un examen clinico de los senos es cuando un doctor, enfermera, u otro
profesional de salud palpa su seno buscando bultos. (,Alguna vez le han hecho un
examen clinico de seno?

O

000000?

No O Nosé ‘ O Si—’

Si si: (,Cuanto hace de su ultimo examen clinico de seno?
En este aﬁo (cualquier vez en los ultimos 12 meses)
Ultimos 2 aﬁos (mas de un aﬁo, pero menos de dos)

En los ﬁltimos 5 afios (mas de dos, pero menos de cinco)

5 o mas aﬁos

No sé

Nunca

14. (Ha tenido un miembro de la familia el cancer de seno?

O

a.
O
O
O
O
O
O

No 0 Nosé O Si—>

Si si: 5Cual miembro familiar?

Mama
Abuela
Hermana
Tia
Prima
Otro

 

15. LHa tenido usted el cancer de seno?

O

No 0 No sé O Si

115

1.

Health Literacy (English)
How conﬁdent are you ﬁlling out medical forms by yourself?

(1) Extremely (2) Quite a bit (3) Somewhat (4) A little bit (5) Not at all

How often do you have someone (like a family member, ﬁ-iend, hospital/clinic
worker, or caregiver) help you read hospital materials?

(1) Always (2) Often (3) Sometimes (4) Occasionally (5) Never
Health Literacy (Spanish)

LCuanta conﬁanza siente usted para llenar las formas médicas?

(1) Mucha (2) Suﬁciente (3) Regular (4) Poca (5) Nada

 

2. (Que tan frecuente 1e ayuda alguien a leer los materiales del hospital? (familiar,
amiga, trabajador del hospital / clinica, cuidador)
(1) Siempre (2) Frecuente (3) A veces (4) Ocasional (5) Nunca
Language-based Acculturation (English)
BASH
The next questions are about your use of English and Spanish.
Spanish English
Only more Both more Only
Spanish than Equally than English
English Spanish
(1) (2) (3) (4) (5)

 

1. In general, what language do
on read and speak?

 

2. What language do you usually
speak at home?

 

3. In what language do you
usually think?

 

 

4. What language do you usually
speak with your friends?

 

 

 

 

 

 

116

 

Language-based Acculturation (Spanish)
BASH- Espaiiol

Las pro'ximas preguntas son acerca de su uso de inglés y espaﬁol.

 

Espafiol Los Inglés

Solo mas Dos mas Solo
Espaﬁol que Por que Inglés
Inglés Igual Espaﬁol
(1) 12L (3) (4) (5)

 

1. En general, 5Qué lenguaje Ud.
lee y habla?

 

2. (Que idioma habla usualmente
en su hogar?

 

3. (En qué idioma piensa
usualmente?

 

 

4. (En qué idioma habla
usualmente con sus amigos?

 

 

 

 

 

 

DUREL: Duke University Religion Index (English)

Directions: Please answer the following questions about your religious beliefs and/or
involvement. Please indicate your answer with a checkmark.

(1) How often do you attend church or other religious meetings?
1. More than once/wk
2. Once a week
3. A few times a month
4. A few times a year
5. Once a year or less
6 Never

(2) How often do you spend time in private religious activities, such as prayer, meditation or
Bible study?

More than once a day

Daily

Two or more times/week

Once a week

A few times a month

Rarely or never

QMPP’NZ‘

117

 

The following section contains 3 statements about religious belief or experience. Please
mark the extent to which each statement is true or not true for you.

(3) In my life, I experience the presence of the Divine (i. e. ,God).
.Deﬁnitely true of me

Tends to be true

Unsure

Tends not to be true

Deﬁnitely not true

9:59P)!"

(4) My religious beliefs are what really lie behind my whole approach to life.
1. Deﬁnitely true of me
2. Tends to be true
3. Unsure
4. Tends not to be true
5. Deﬁnitely not true

(5) I try hard to carry my religion over into all other dealings in life.
Deﬁnitely true of me

Tends to be true

Unsure

Tends not to be true

Deﬁnitely not true

.U‘PP’NT"

DUREL: Duke University Religion Index (Spanish)
Para las preguntas 1 y 2, rodee con un circulo e1 numero que sea mas adecuado:

1. [,COI] cuanta frecuencia atiende usted a la iglesia u otros encuentros religiosos?
1. Nunca
2. Una vez a1 aﬁo o menos
3. Unas pocas veces a1 aﬁo
4. Unas pocas veces a1 mes
5. Una vez a la semana
6. Mas de una vez a la semana

2. bCon cuanta frecuencia dedica usted tiernpo a actividades religiosas privadas, como
por ejemplo rezar, meditar, o estudiar 1a Biblia?

l. Mas de una vez a1 dia

2. Diariamente

3. Dos o mas de dos veces a1 dia

4. Una vez a la semana

5. Unas pocas veces a] mes

6. Raras veces o nunca

118

Lea las siguientesfrases y rodee con un circulo el numero que sea mas adecuado:

3. En mi Vida, yo siento la presencia de lo Divino (por ejemplo, Dios)
1. Deﬁnitivamente no es cierto
2. Tiende a no ser cierto
3. No estoy segura
4. Tiende a ser cierto
5. Deﬁnitivamente cierto para mi

4. M13 creencias religiosas son lo que realrnente esta detras de mi enfoque hacia la Vida.
1. Deﬁnitivamente no es cierto
2. Tiende a no ser cierto
3. No estoy segura
4. Tiende a ser cierto
5. Deﬁnitivamente cierto para mi

5. Trato de llevar mis ﬁmdamentos religiosos a todos los demas aspectos de mi Vida.
1. Deﬁnitivamente no es cierto
2. Tiende a no ser cierto
3. No estoy segura
4. Tiende a ser cierto
5. Deﬁnitivamente cierto para mi

HEALTH RELATED QUALITY OF LIFE (HRQOL-4) (English)
1. Self-Perceived Health

Would you say that in general your health is:

O Excellent
0 Very good
0 Good

0 Fair

0 Poor

2. Recent Physical Health

Now thinking about your physical health, which includes physical illness and injury,
for how many days during the past 30 days was your physical health not good?

 

119

3. Recent Mental Health

New thinking about your mental health, which includes stress, depression, and
problems with emotions, for how many days during the past 30 days was your mental
health not good?

 

4. Recent Activity Limitation

During the past 30 days, for about how many days did poor physical or mental health
keep you from doing your usual activities, such as self-care, work, or recreation?

 

Calidad de la Vida en Relacion a La Salud (HRQOL-4) (Spanish)
1. Estado de Salud
LDiria usted que su estado de salud general es:
0 Excelente
O Muy bueno
O Bueno
0 Regular
0 Malo

2. Salud Fisica Reciente

' Con respecto a su estado de salud ﬁsica, lo que incluye tanto enfermedades como
lesiones ﬁsicas, Lcuantos dias durante los ultimos 3O dias tuvo problemas de salud?

 

3. Salud Mental Reciente

Con respecto a su estado de salud mental, lo que incluye estrés, depresion y
problemas emocionales, (gcuantos dias durante los ultimos 30 dias sintié que su estado
de salud mental no era bueno?

 

4. Limitacion Reciente de Actividad

Durante los ultimos 30 dias, (gen cuantos dias sintio que los problemas relacionados
con su salud mental o fisica 1e impidieron realizar sus actividades habituales, tales
como cuidados personales, trabajo o recreacién?

 

120

Cancer Worry Scale (English)

During the past month, how often have you thought about your own chances of
deve10ping breast cancer?

Not at all, or rarely Sometimes Often Almost all the time
D l D 2 D 3 D 4

During the past month, how often have you worried about your own chances of
developing breast cancer?

Not at all, or rarely Sometimes Often Almost all the time
[:1 1 D 2 D 3 [:1 4

During the past month, how often have thoughts about your chances of getting breast
cancer affected your mood?

Not at all, or rarely Sometimes Often Almost all the time
[:1 1 D 2 D 3 D 4

During the past month, how often have thoughts about your chances of getting breast .
cancer affected your ability to perform daily activities?

Not at all, or rarely Sometimes Often Almost all the time
D 1 D 2 D 3 D 4

The next question asks about any feelings of distress you may have about your risk of
breast cancer. During the past four weeks, which one of the following best describes you?
I] 1 Generally happy and ﬂee from worry
[I 2 Occasionally distressed
D 3 Often distressed
E] 4 Almost always or extremely distressed

If yes to 4, would you say you have been so extremely distressed that you require
hospitalization?

D Yes

D No

121

Cancer Worry Scale (Spanish)

Escala de Preocupacion del Cancer

Durante el mes pasado (gQue tanto ha pensado usted en su propia posibilidad de
desarrollar cancer de seno?

Nunca/raramente A veces Frecuentemente Casi todo el tiempo
D l D 2 D 3 D 4

Durante e1 mes pasado 5Qué tanto se ha preocupado usted en la posibilidad de desarrollar
cancer de seno?

Nunca/raramente A veces Frecuentemente Casi todo el tiempo
D l D 2 D 3 D 4

Durante e1 mes pasado LQué tanto 1e ha afectado 1a posibilidad de tener cancer de seno en
su estado de animo?

Nunca/raramente A veces Frecuentemente Casi todo el tiempo
Ur Dz U3 D4

Durante e1 mes pasado 5Qué tanto 1e ha afectado la posibilidad de tener cancer de seno en
su habilidad para hacer sus actividades diarias?

Nunca/rararnente A veces Frecuentemente Casi todo el tiempo
Dr D2 Us D4

La siguiente pregunta es acerca de sus sentimientos de estrés, por la posibilidad del riesgo
del cancer de seno. Durante las ultimas cuatro semanas 5Cua1 descripcién siguiente 1a
describe mejor?

El 1 Generalmente feliz y libre de preocupacién

E] 2 Estresada ocasionalmente

C13 Estresada frecuentemente

D 4 Casi Siempre o estresada extremadamente

Si responde 81 a la anterior (4), gDiria que ha estado tan estresada que
requeriria hospitalizacion?

C] Si

E] No

122

Powe F atalism Inventory (English)

Directions: Please answer the following questions.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. I think if someone is meant to have cancer, it doesn't matter what YES NO
kinds of food they eat, they will get cancer anyway.

2. I think if someone has cancer, it is already too late to get treated for YES NO
it.

3. I think someone can eat fatty foods all their life, and if they are not YES NO
meant to get cancer, they won't get it.

4. I think if someone is meant to get cancer, they will get it no matter YES NO
what they do.

5. I think if someone gets cancer, it was meant to be. YES NO

6. I think if someone gets cancer, their time to die is soon. YES NO

7. I think if someone gets cancer, that's the way they were meant to YES NO
die.

8. I think getting checked for cancer makes people scared that they YES NO
may really have cancer.

9. I think if someone is meant to have cancer, they will have cancer. YES NO

10. I think some people don't want to know if they have cancer because YES NO
they don't want to know they may be dying from it.

11. I think if someone gets cancer, it doesn't matter whether they ﬁnd it YES NO
early or late, they will still die from it.

12. I think if someone has cancer and gets treatment for it, they will YES NO
probably still die from the cancer.

13. I think if someone was meant to have cancer, it doesn't matter what YES NO
doctors and nurses tell them to do, they will get cancer anyway.

14. I think if someone is meant to have cancer, it doesn't matter if they YES NO
eat healthy foods, they will still get cancer.

15. I think cancer will kill you no matter when it is found and how it is YES NO

treated.

 

 

 

123

 

 

Powe Fatalism Inventory (Spanish)

Tabla de Fatalismo de Powe

Instrucciones: Por favor conteste las siguientes preguntas. Marque sus respuestas con un
circulo en la contestacién correcta.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Yo pienso que si a una persona 1e va a dar cancer, no importa qué tipo SI N0
de comidas coma, de todos modos le va a dar cancer.

2. Yo pienso que si a1 guien tiene cancer, ya es demasiado tarde para tratar SI N0
de buscar tratamiento.

3. Yo pienso que una persona puede comer comida con grasa toda su Vida, SI N0
pero si no le toca que le de cancer, no le va a dar cancer.

4. Yo pienso que si a una persona 1e va a dar cancer, 1e va a dar no importa SI N0
lo que haga.

5. Yo pienso que si a una persona 1e da cancer, asi 1e tocaba. SI N0

6. Yo pienso que si a una persona 16 da cancer, esa persona se va a morir SI N0
pronto.

7. Yo pienso que si le da cancer a una persona, ese es el modo en cual le SI N0

tocaba morirse a esa persona.

8. Yo pienso que a la gente le da miedo examinarse para el cancer porque SI N0
les da miedo que de veras vayan a tener cancer.

9. Yo pienso que si a una persona le toca que le de cancer, 1e va a dar SI N0
cancer.

10. Yo pienso que algunas personas no quieren saber si tienen cancer, SI N0
porque no quieren saber si ya se estan muriendo de esa enfermedad.

11. Yo pienso que si alguien tiene cancer, no importa si se lo encuentran SI N0
temprano o tarde, porque de todos modos va a morir de cancer.

12. Yo pienso que 31 a1 guien tiene cancer y recibe tratamiento para curarse, SI N0
de todas maneras se va a morir de esta enfermedad.

13. Yo pienso que si a una persona 1e toca que le de cancer, no importa qué SI N0
le digan los doctores y enfermeras que haga, de todos modos 1e va a dar
cancer.

14. Yo pienso que si a una persona 1e toca que 1e de cancer, no importa si SI N0
come comidas saludables, pues de todos modos le va a dar cancer.

15.Yo pienso que el cancer matara a una persona no importa cuando 10 SI N0

encuentren o como lo curen.

 

 

 

124

 

APPENDIX D
Descriptive and Bivariate Statistics

Table A. Descriptive Statistics for Categorical Variables (Total Sample)

 

 

Variable Response % (n)
Chosen language English 43.8 (92)
for survey Spanish 56.3 (117)
administration Total 100.0 (208)
Age categories 40-49 years old 49.5 (103)
50+ years old 50.5 (105)
Total 100.0 (208)
What is the Never attended or only nursery school/
highest level of kindergarten 3.9 (8)
school you Elementary or primary school 25.1 (52)
completed? Middle or junior high school 16.9 (35)
Some high school but no diploma 6.8 (14)
High school graduate 17.4 (36)
Vocational or trade school graduate 5.8 (12)
Some college but no degree 14.0 (29)
Associate degree in college 3.9 (8)
Bachelor's degree 3.4 (7)
Master's degree 2.9 (6)
Total 100.0 (207)
Education Did not graduate high school 52.7 (109)
Categories High school graduate or higher 47.3 (98)
Total 100.0 (207)
Do you have any Yes 67.6 (140)
kind of health Otherwise 32.4 (67)
care coverage? Total 100.0 (207)
Health insurance Private (HMO, PPO, or military) 51.7 (104)
type categories Public (Medicaid or Medicare) 14.9 (30)
No health insurance 33.3 (67)
Total 100.0 (201)
About how long <1 year ago 67.1 (139)
has it been since 1 year to < 2 years ago 15.0 (32)
you last visited a 2 + years ago 11.6 (24)
doctor for a Don't know 1.4 (3)
routine checkup? Never 4.8 (10)

Total 100.0 (207)

 

125

 

 

Variable Response % (n)
Recent physician < 1 year ago 67.1 (139)
visit categories Otherwise 32.9 (68)
Total 100.0 (207)
In what country Mexico 76.3 (151)
were you born? United States 23.7 (47)
Total 100.0 (205)
Has a family Yes 26.4 (55)
member ever had No 68.3 (142)
breast cancer? Don’t know 5.3 (11)
Total 100.0 (208)
If “yes”, who? 1St degree relative (mother, sister, daughter) 32.7 (18)
Other family (aunt, niece, grandmother, in-law) 67.3 (3 7)
Total 100.0 (55)
Have you ever Yes 2.9 (6)
had breast No 93.3 (194)
cancer? Don’t know 3.8 (8)
Total 100.0 (208)
History of breast Personal or family history of breast cancer 28.4 (59)
cancer categories Otherwise 71.6 (149)
Total 100.0 (208)
Have you ever Yes 84.6 (176)
had a No 14.9 (31)
mammogram? Don’t know .5 (1)
(40+) Total 100.0 (208)
If yes, how long < 1 year ago 57.4 (116)
has it been since 1 year to < 2 years ago 18.8 (38)
you had your last 2 + years ago 7.9 (16)
mammogram? Don’t know .5 (1)
(40+) Never 15.3 (31)
Total 100.0 (202)
Time since last < 1 year ago 57.4 (116)
mammogram Otherwise 42.6 (86)
(40+) Total 100.0 (202)

 

126

 

 

Variable Response % (n)
Time since last < 1 year ago 63.4 (64)
mammogram Otherwise 36.6 (37)
(50+) Total 100.0 (101)
Time since last < 2 years ago 76.2 (154)
mammogram Otherwise 23.8 (48)
(40+) Total 100.0 (202)
Time since last < 2 years ago 71.3 (72)
mammogram Otherwise 28.7 (29)
(40—49) Total 100.0 (103)

 

127

Table B. Comparison of Sample Demographics with County and State Data

 

 

Variable Response Sample San Diego California
County
(Latinas, 40+) (Latinas, 40+) (Latinas, 40+)
% (n) % (n) % (n)
Agea 40-49 years old 49.5 (103) 46.1 (56,000) 44.7 (832,000)
50+ years old 50.5 (105) 55.9 (65,000) 55.3 (1,028,000)
Total 100.0 (208) 100.0 (121,000) 100.0 (1,860,000)
Educationa Did not graduate high school 52.7 (109) 60.2 (61,000) 54.1(858,000)
High school or higher 47.3 (98) 39.8 (40,000) 45.9 (729,000)
Total 100.0 (207) 100.0 (101,000)" 100.0 (1,587,000)
Has health Yes 67 .6 (140) 77.5 (93,000) 79.1 (1,471,000)
insurancea Otherwise 32.4 (67) 22.5 (27,000) 20.9 (389,000)
Total 100.0 (207) 100.0 (120,000)* 100.0 (1,860,000)*
Health Private 51.7 (104) 52.4 (63,000) 45.3 (842,000)
insurance Public 14.9 (30) 25.1 (30,000) 33.8 (629,000)
categories 3 None 33.3 (67) 22.5 (27 ,000) 20.9 (389,000)
Total 100.0 (201) 100.0 (120,000)* 100.0 (1,860,000)*
Country of Mexico 76.3 (151) 62.5 (70,000) 53.8 (836,000)
birtha United States 23.7 (47) 37.5 (42,000) 45.9 (714,000)
Total 100.0 (198) 100.0 (112,000)* 100.0 (1,555,000)*
Ever had a Yes 84.6 (176) 85.7 (104,000) 88.0 (1,637,000)
mammograma No 15.4 (32) 14.3 (17,000) 12.0 (224,000)
Total 100.0 (208) 100.0 (121,000) 100.0 (1,861,000)
Mammogram Yes 76.2 (154) n/a 76.4 (274)
within the past No 23.8 (48) 23.6 (89)
2 years (40 +)b Total . 100.0 (202) 100.0 (363)
Mammogram Yes 81.2 (82) n/a 84.8 (168)
within the past No 18.8 (19) 15.2 (35)
2 years (50+)b Total 100.0 (101) 100.0 (203)
Health statusa Excellent 8.3 (17) 16.8 (20,000) 10.2 (190,000)
Very Good 19.0 (39) 17.7 (21,000) 20.0 (372,000)
Good 37.1 (76) 27.4 (33,000) 30.7 (571,000)
Fair 29.8 (61) 32.3 (39,000) 31.0 (576,000)
Poor 5.9 (12) 5.8 (7,000) 8.1 (151,000)
Total 100.0 (205) 100.0 (120,000)* 100.0 (1,860,000)

 

a Source: 2007 California Health Interview Survey (http://www.chis.ucla.edu/; San Diego County,
California sample only includes Latinas 40 years and older.

b Source: 2006 Behavioral Risk Factor Surveillance Survey Data (http://apps.nccd.cdc.gov/BRFSS/);
California sample only includes Latinas 40 years and older.
Note: Chi-square tests were used to compare state versus sample and county versus sample data.
*Difference from sample data statistically signiﬁcant (p S .01)

"Difference from sample data statistically signiﬁcant (p S .05)

128

Table C. Descriptive Statistics of Scales and Continuous Items (Total Sample)

 

 

Range
Dimensions Scale and M (SD) of Possible Cronbach Skew Kurtosis
Item # (N= 208) Scores Range Alpha
Language-based BASH 235(134)
acculturationl #1-4 (n=200) 1-5 1-5 .96 .65 -.91
Intrinsic DUREL 4.44 (.80)
religiosity #3_5 (11:2 04) 1-5 1-5 .75 -1.61 2.05
Organizational DUREL 3.24 (1.27) -125
religiosity #1 (n=l94) 1'5 1'6 n/a "12
Nonorganizational DUREL 3.12 (1.39)
religiosity #2 (n=162) 1-5 1-6 n/a -.41 -1.26
Predetermination PFI .45 (.36)
#1,4,5,9 (n= 198) 0-1 0-1 .76 .14 -1.34
Inevitability of death PFI .17 (.31)
#11,]2,15 (n=204) 0-1 0-1 .78 1.69 1.51
Fear PFI .65 (.39)
#8,10 (n=206) 0-1 0-1 .32* -.58 -l .11
Breast cancer worry CWS 1.68 (.67)
#1_3 (11:205) 1-4 1-4 .86 .99 .66
Distress about breast CWS 1.59 (.71)
cancer #5 (n=206) 1-4 1-4 n/a 1.20 1.50
Self-rated health status HRQOL - 3.06 (1.03)
#1 (11:20 5) 1-5 1-5 n/a -.26 -.45
Summary index for HRQOL 9, 57 (11 ()0)
' - - . 4 -.6
unhealthy days2 #2'3 (n=201) O 30 0 30 n/a 9 6
Days of activity HRQOL 5.37 (923)
_ - 1. .
1imitation3 #4 (”=1“) 0 30 0 30 n/a 86 2 18
Days of activity HRQOL
. . . 4 #4 1.13 (.20) _ _
l1m1tat10n(New) (11:14.7) 1 1.6 1 1.6 n/a 1.44 .70
Conﬁdence in ﬁlling HL #1 3.96 (1.05)
out medical forms (n=208) 1.5 1-5 “874 '170
Assistance needed for HL#2 3.75 (1.37) .24*
reading (n=207) 1-5 1-5 -.688 -.761
Age continuous Age #1 50.98 (8.81.) 40.01-
(n=208) 40+ 80.35 n/a 1.33 1.77

 

*Correlation is signiﬁcant at the .01 level;

lHigher scores denote higher acculturation to English language usage, greater religiosity, higher levels of
breast cancer worry and distress, higher levels of fatalistic beliefs, higher health literacy (more conﬁdence
in ﬁlling out forms and less assistance needed reading), higher age; and lower self-rated health status

Sum of physical and mental distress in the last 30 days

3Activity limitation due to physical/ mental distress in the last 30 days (valid only if previous item

“summary index for unhealthy days” is > 0)
4Transformed variable [new activity limitation= loglO (activity limitation + 10)]

129

Tables D-P. Psychometric Properties of Scales (English and Spanish Versions)

Table D. Brief Acculturation Scale for Hispanics (BASH): Language-based Acculturation
Language-based Acculturation English (n=87) Spanish (n=113)
Mean Item-Total Mean Item-Total

 

 

 

Items (SD) Correlations (SD) Correlations
.mw
2 2:231]: 33%;?) do you usually (1:421?) .88 1.39 (. 70) .75
3 xihnitganguage do you usually (3.33) .91 1.41 (.78) .59
“tang“..-m 3:32. mam

Version Mean (SD) Rggffegf Possible Range (21:11:31?
SpaniSh 1.44 (.61) 1-3 1-5 '84
English 3.53 (1.14) 1-5 1-5 .95

 

Note: Scale ranges from 1- 5, with higher scores indicating greater acculturation, (1= only Spanish, 2=
Spanish more than English, 3=both equally, 4: English more than Spanish, and 5=only English).

Table E. Duke University Religion InventoﬂDUREL): Organizational Religiosity
Item 1 How often do you attend church or other religious meetings?

 

 

Version Mean (SD) Range of Possible Range Cronbach
Scores Alpha

Spamsh 3.17(1.22) 1_5 1-6 n/a

English 3.32 (1.34) 1-5 1-6 n/a

 

Note: Scale ranges from 1- 6, with higher scores indicating more religiosity, (l= never, 2= once a year or
less, 3=a few times a year, 4: a few times a month, 5=once a week, and 6= more than once a week).

Table F. Duke University Religion Inventory (DUREL): Non-Organizational Religiosity

Item 2 How often do you spend time in private religious activities, such as prayer,
mediation or Bible study?

 

 

Version Mean (SD) Range Of Possible Range Cronbach
Scores Alpha

Spanlsh 3.16(1.36) 1_5 1-6 ,1),

English 3.07 (1.44) f 1-5 1-6 n/a

 

Note: Scale ranges from 1- 6, with higher scores indicating more religiosity, (l= rarely or never, 2= a few
times a month, 3=once a week, 4= two or more times a week, 5: daily, and 6= more than once a day).

130

Table G. Duke University Religion Inventory (DUREL): Intrinsic Religiosity Subscale

 

 

 

 

 

Intrinsic Religiosity English (n=89) Spanish (n=115)
Items Mean Item—Total Mean Item—Total
(SD) Correlations (SD) Correlations
3 In my life, I experience the 4.47 65 4.67 .41
presence of the Divine (i.e., God). (.94) ' (.81)
4 My religious beliefs are what 4 30 4 58 .64
really lie behind my whole ( 88) .69 ( '93)
approach to life. ' '
5 I try hard to carry my religion over 4.08 63 4.41 .66
into all other dealings in life. (1.13) ' (1 .00)
Version Mean (SD) RS232? Possible Range C123) 1:23}!
SPamSh 4.55 (.76) 1-5 1-5 '74
English 4.28( .84) 1-5 1-5 .80

 

Note: Scale ranges {tom 1- 5, with higher scores indicating more religiosity, (I: deﬁnitely not true of me,
2= tends to be true of me, 3=unsure, 4= tends to be true, and 5=deﬁm'tely true of me).

Table H. Powe F atalism Scale (PFI): Predetermination Subscale

 

 

 

 

Predetermination English (n=86) Spanish (n=112)
Items Mean Item-Total Mean Item-Total
(SD) Correlations (SD) Correlations
1 I think if someone is meant to
have cancer, it doesn't matter .48 .40
what kinds of food they eat, (.50) '56 '58 ('50)
they will get cancer anyway
I think if someone is meant to .49
get cancer, they will get it no .55 (.50) .66 .51 (.50)
matter what they do
I th1nk if someone gets cancer, it .34 (.48) .58 .39 (.49) .536
was meant to be
I think if someone is meant to .43
have cancer, they will have .33 (.47) .43 .43 (.50)
cancer
. Range of Possible Cronbach
Versron Mean (SD) Scores Range Alpha
Spanish .48(.37) 0-1 0-1 '69
English .42 (.37) 0-1 0-1 .76

 

131

Note: Scale ranges ﬁ'om 0-1, with higher scores indicating more cancer fatalistic beliefs, (0= no, 1=yes).

Table I. Powe Fatalism Scale (PFI): Inevitability of Death Subscale
Inevitability of Death Enﬂsh (n=88) Spanish (n=116)
Mean Inter Item Mean Item-Total

Items (SD) Correlations (SD) Correlations

11 I think if someone gets

 

 

cancer, it doesn't matter
whether they ﬁnd it early or '11 ('32) '66 '14 ('35) .65
late, they will still die from it
12 I think if someone has cancer
and gets treatment for it, they .69
will probably still die ﬁ'om '20 ('41) '52 '16 ('36)
the cancer
15 I think cancer will kill you no
matter when it is found and .10 (.30) .54 .28 (.45) .69 .
how it is treated 51.1
. Range of Possible Cronbach
Versron Mean (SD) Scores Range Alpha
Spanish .l9(.33) 0-1 0—1 '82
English .14 (.28) 0-1 0-1 .71

 

Note: Scale ranges from 0-1, with higher scores indicating more cancer fatalistic beliefs, (0= no, 1=yes).

Table J. Powe F atalism Scale (PFI): Fear Subscale

 

 

 

 

 

Fear English (n=90) Spanish (n=116)
Items Mean Inter-item Mean Inter-item
(SD) Correlation (SD) Correlation
8 I think getting checked for
cancer makes people scared
that they may really have '61 ('49) '66 ('48)
cancer
10 I think some people don't .32* .34*
want to know if they have
cancer because they don't .74 (.44) .60 (.49)
want to know they may be
dying from it
Version Mean (SD) Range of Scores Possible Range
Spanish .63 (.40) 0-1 0-1
English .68 (.38) 0-1 0-1

 

Note: Scale ranges from 0-1, with higher scores indicating more cancer fatalistic beliefs, (0= no, 1=yes).

132

Table K. Powe F atalism Scale (PFI): Deleted items based on factor analysis and
reliability testing of English and Spanish versions

 

Items originally part of the Predetermination Scale

3 I think someone can eat fatty foods all their life, and if they are not meant to get
cancer, they won't get it

13 I think if someone was meant to have cancer, it doesn't matter what doctors and
nurses tell them to do, they will get cancer anyway

14 I think if someone is meant to have cancer, it doesn't matter if they eat healthy
foods, they will still get cancer

 

Item originally part of the Inevitability of Death Scale
7 I think if someone gets cancer, that's the way they were meant to die

 

Items originally part of the Pessimism Scale
2 I think if someone has cancer, it is already too late to get treated for it
6 I think if someone gets cancer, their time to die is soon

 

Table L. Health Literacy: Conﬁdence in filling out medical forms
Item 1 How conﬁdent are you ﬁlling out medical forms by yourself?

 

. Mean Range of . ' Cronbach
Verswn (SD) Scores Possrble Range Alpha
Spanrsh 3.78 (.93) 1_5 1_5 n/a
English 4.20 (1.16) 1-5 1-5 n/a

 

Note: Scale ranges from 1- 5, with higher scores indicating more health literacy or greater conﬁdence,
(1= not at all, 2= a little bit, 3=somewhat, 4= quite a bit, and 5=extremely).

Table M. Health Literacy: Assistance in reading hospital materials

 

Item 2 How often do you have someone (like a family member, ﬁiend, hospital/clinic
worker, or caregiver) help you read hospital materials?

 

 

. Mean Range of . Cronbach
Versron (SD) Scores Possrble Range Alpha
Spamsh 3.22 (1.36) 1_5 1_5 n/a
English 4.43 (1.03) 1-5 1-5 n/a

 

Note: Scale ranges from 1- 5, with higher scores indicating more health literacy or less help needed in
ﬁlling out medical forms, (1= always, 2= often 3=sometimes, 4= occasionally, and 5=never).

133

Table N. Cancer Worry Scale: Breast Cancer Worry
Breast Cancer Worry English Spanish
(n=89) (n=116)
Mean Item-Total Mean Item-Total

Items (SD) Correlations (SD) Correlations

1 During the past month, how often

 

 

have you thought about your own 1.70 78 1.76 73
chances of developing breast (.73) ° (.78) °
cancer?
2 During the past month, how often
have you worried about your own 1.58 85 1.91 84
chances of developing breast (.67) ' (.85) '
cancer?
3 During the past month, how often
have thoughts about your chances 1.29 58 1.72 70
of getting breast cancer affected (.59) ' (.79) '
your mood?
Version Mean (SD) R3252? Possible Range (31:11:11?h
SpamSh 1.80 (.72) 1-4 1-4 '87
English 1.52 (.59) 1-3.33 1-4 .86

 

Note: Scale ranges from 1- 4, with higher scores indicating more worry, (1= not at all, or rarely, 2=
sometimes, 3=oiten, and 4= almost all the time).

Table O. Cancer Worry Scale (C WS): Deleted item based on factor analysis and
reliability testing of English and Spanish versions:

Item originally part of the breast cancer worry scale
4 During the past month, how often have thoughts about your chances of getting
breast cancer affected your ability to perform daily activities?

 

Table P. Cancer Worry Scale: Breast Cancer Distress

Item 5 The next question asks about any feelings of distress you may have about your
risk of breast cancer. During the past 4 weeks, which of the following best describes you?

 

 

. Mean Range of . Cronbach
Versron (SD) Scores Possrble Range Alpha
Spamsh 1.72 (.76) 1_4 1_4 n/a
English 1.42 (.62) 1-4 1-4 n/a

 

Note: Scale ranges from 1- 4, with higher scores indicating more distress, (l= generally happy and free
from worry, 2= occasionally distressed, 3=often distressed, and 4= almost always or extremely distressed).

134

Table Q. Correlation Matrix Used for Conﬁrmatory Factor Analysis

 

 

1 2 3 4 5 6 7 8 9
1. EDU -—

2. BASH .633** --

3. COB .425** .726* --

4. INS .383* .412* .221* -—

5. PHYS .131 .065 .069 286* --

6. AGE -.148** .030 .109 .168** .116

7. HRQOL] -.313* -.256* -.162**-.241* -.106 .154** --

8. HRQOL2 -.066 .001 .018 -.O68 -.018 .155** .318* -—

9. HRQOL3 -.005 -.029 .029 .079 .079 .225* .236* .552* --
Note: * P _<_ .01; ** P S .05.
1) Education (dichotomous), 2) Language-based acculturation, 3) Country of birth (dichotomous), 4)
Insurance (dichotomous), 5) Recent physician visit (dichotomous), 6) Age (continuous), 7) Health status, 8)
Unhealthy days, and 9) Activity limitation.
Higher scores denote greater education, higher language-based acculturation, being U.S.-born (1= US, 0=
Mexico), having insurance, a recent physician visit < 1 year, a greater age, a lower health status, less
amount of unhealthy days, and a less amount of activity limitation.
Activity limitation is not transformed.

 

Table R. Correlation Matrix of Variables in Final Behavioral Model of Latina Breast
Cancer Screening Use, Used for Conﬁrmatory Factor Analysis

 

 

1 . 2 3 4 5 6 7 8
1. HL __

2. EDU .384* -—

3. BASH .318* .633* --

4. AGE -.195* -.148** .030 -—

5. INS .174** .383** .412* .168** --

6. PHYS -.013 .131 .065 .116 .286* --

7. WORRY -.082 -.195* -.236* .043 -.014 .099 --

8. DISTRESS -.188* -.108 -.177** -.008 -.089 -.042 .378* --

 

Note: * P S .01; ** P S .05.

1) Health literacy: conﬁdence in ﬁlling out medical forms (continuous), 2) Education (dichotomous), 3)
Language-based acculturation (continuous),.4) Age (continuous), 5) Insurance (dichotomous), 6) Recent
physician visit (dichotomous), 7) Recent breast cancer worry (continuous), and 8) Recent breast cancer
distress (continuous)

Note: Higher scores denote higher health literacy, greater education, higher language-based acculturation, a
greater age being having insurance (l=yes, 0=no), a recent physician visit < 1 year (l=yes, 0=no), and
greater recent worry and distress about one’s own risk of developing breast cancer.

135

Table S. Behavioral Model of Latina Breast Cancer Screening Use, With and Without

 

 

Suppressed Variables
Factors Model 1 Model 2
Mammography (< 2 years)a Mammography (< 2 years)b
OR 95% CI. P-value OR 95% CI. P-value
Education
Less than high school c " “ “ 1-00
High school or greater -- -- -- 2.937 (.854, 10.103) 0.087"
Language-based Acc. -- -- -- 0.509 (.331, .782) 0002*
Age 1.068 (1.009, 1.130) 0.023" 1.087 (1.025, 1.152) 0005*
Health insurance
None c 1.00 1.00
Public 1.484 (0.431, 5.102) 0.531 2.038 (0.525, 7.914) 0.304
Private 1.632 (0.694, 3.837) 0.262 2.541 (0.865, 7.468) 0090"
Physician visit < 1 yr
No C 1.00 1.00
Yes 4.610 (2.055, 10.344) 0.000* 4.170 (1.793, 9.698) .001*
Contextual Factors
Breast cancer distress 0.614 (0.372, 1.014) 0.056" .582 (.347, .976) .040**

Conﬁdence in ﬁlling out 1.517 (1.039, 2.215) 0.031“ 1.701 (1.134, 2.552) .010*
medical forms

 

-2 log likelihood 163.946 " 153.484 °’ f
Note: C .I. = conﬁdence interval. Analyses substituting the insurance dummy variables (private, public, and
none) with a dichotomous variable (yes/no) and the continuous age with a dichotomous (40-49 versus 50+)
variable changed the results slightly. Age became non-signiﬁcant in Model 1(OR=1.745, CI=.801-3.802,
p=.16l) and Model 2 (OR=2.073, CI=.919-4.687, p=.079). An analysis without 4 outlier cases did not
change the results.

AApproaching signiﬁcance at the .05 level (.05 > p < .10); * P S .01; ** P S .05
a n = 186; n = 186; c Reference category;

d Model chi-square: 37.195 (df= 6, p = .000); GOF x2 = 5.383, (df= 8, p = .716)

6 Model chi-square: 47.657 (df = 8, p = .000); GOF [2 = 6.392, (df= 8, p= .603)

fA Likelihood Ratio x2 Test indicated that the two additional predisposing variables signiﬁcantly added to
the reduced model 1 [38(2) = 10.462 > critical value (382 (n.0, , = 9.210), p s .01].

136

 

 

 

REFERENCES

137

REFERENCES

Abraido-Lanza, A. E., Viladrich, A., Florez, K. R., Cespedes, A., Aguirre, A. N., & De
La Cruz, A. A. (2007). Commentary: Fatalismo reconsidered: A cautionary note
for health-related research and practice with Latino populations. Ethn Dis, 1 7(1),
153-158.

Abraido-Lanza, A. F., Chao, M. T., & F lorez, K. R. (2005). Do healthy behaviors decline
with greater acculturation? Implications for the Latino mortality paradox. Soc Sci
Med, 61(6), 1243-1255.

Abraido-Lanza, A. F., Chao, M. T., & Gammon, M. D. (2004). Breast and cervical cancer
screening among Latinas and non-Latina whites. Am J Public Health, 94(8),
1393-1398.

Abraido-Lanza, A. F., Chao, M. T., & Gates, C. Y. (2005). Acculturation and cancer
screening among Latinas: results from the National Health Interview Survey. Ann
Behav Med, 29(1), 22-28.

Abraido-Lanza, A. F., Vasquez, E., & Echeverria, S. E. (2004). En las manos de Dies [in
God's hands]: Religious and other forms of coping among Latinos with arthritis. J
Consult Clin Psychol, 72(1), 91-102.

Aday, L. A., & Andersen, R. (1974). A framework for the study of access to medical
care. Health Serv Res, 9(3), 208-220.

Aﬁﬁ, A., Clark, V. A., & May, S. (2004). Computer-aided Multivariate Analysis (4th ed).
Boca Raton, FL: Chapman & Hall/ CRC.

Agency for Healthcare Research and Quality (AHRQ). (2004). National Healthcare
Disparities Report, 2004—Full Report [Electronic Version], from
http://www.ahrq.gov/qual/nhdr04/ﬁrllreport/

AHRQ. (2006). The Guide to Clinical Preventive Services: Recommendations of the US.
Preventive Services Task Force. Retrieved May 1, 2007, from http://pda.ahrq.gov

AHRQ. (2008). National Healthcare Disparities Report, 2007 [Electronic Version],
AHRQ Publication No. 08-0040, Rockville, MD, Retrieved December 1, 2008,
ﬁ'om http://www.ahrq.gov/qual/nhdr07/nhdr07.pdf

Aldridge, M. L., Daniels, 1. L., & Jukic, A. M. (2006). Mammograms and healthcare

access among US Hispanic and non-Hispanic women 40 years and older. F am
Community Health, 29(2), 80-88.

138

American Cancer Society (ACS). (2003). Cancer Facts & Figures for Hispanics/Latinos
2003-2005. Atlanta, GA: American Cancer Society.

ACS. (2007). Breast Cancer Facts & Figures, 2007-2008. Retrieved February 1, 2008,
from http://www.cancer.org/downloads/STT/BCF F -F ina1.pdf

ACS. (2008). California Cancer Facts & Figures, 2008. Retrieved February 14, 2008,
from http://www.ccrca1.org/PDF/ACS2008.pdf

Andersen, M. R., Smith, R., Meischke, H., Bowen, D., & Urban, N. (2003). Breast cancer
worry and mammography use by women with and without a family history in a
population-based sample. Cancer Epidemiol Biomarkers Prev, 12(4), 314-320.

Andersen, R., Bozzette, S., Shapiro, M., St Clair, P., Morton, S., Crystal, 8., et a1. (2000).
Access of vulnerable groups to antiretroviral therapy among persons in care for
HIV disease in the United States. HCSUS Consortium. HIV Cost and Services
Utilization Study. Health Serv Res, 35(2), 389-416.

Andersen, R., Lewis, S. Z., Giachello, A. L., Aday, L. A., & Chiu, G. (1981). Access to
medical care among the Hispanic population of the southwestern United States. J
Health Soc Behav, 22(1), 78-89.

Andersen, R., & Newman, J. F. (1973). Societal and individual determinants of medical
care utilization in the United States. Milbank Mem Fund Q Health Soc, 51(1), 95-
124.

Andersen, R. M. (1968). A Behavioral Model of Farnilies' Use of Health Services,
Research Series No. 25. Center for Health Administration Studies, University of
Chicago. Retrieved August 1, 2008, from
http://www.chas.uchicago.edu/documents/Publications/RS/RSZS.PDF

Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care:
does it matter? J Health Soc Behav, 36(1), 1-10.

Andersen, R. M. (2008). National health surveys and the behavioral model of health
services use. Medical Care, 46(7), 647-653.

Andersen, R. M., & Davidson, P. L. (1997). Ethnicity, aging, and oral health outcomes: a
conceptual framework. Adv Dent Res, 1 1(2), 203 -209.

Andersen, R. M., Davidson, P. L., & Ganz, P. A. (1994). Symbiotic relationships of
quality of life, health services research and other health research. Qua] Life Res,
3(5), 365-371.

Andersen, R. M., Davidson, P. L., & Nakazono, T. T. (1997). Oral health policy and
programmatic implications: lessons from ICS-II. Adv Dent Res, 11(2), 291 -303.

139

'11]

 

Andersen, R. M., Giachello, A. L., & Aday, L. A. (1986). Access of Hispanics to health
care and cuts in services: a state-of-the-art overview. Public Health Rep, 101(3),
23 8-252.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A
review and recommended two-step approach. Psychological Bulletin, 103(3),
411-423.

Aparicio-Ting, F., & Ramirez, A. G. (2003). Breast and cervical cancer knowledge,
attitudes, and screening practices of Hispanic women diagnosed with cancer. J
Cancer Educ, 18(4), 230-236.

Armstrong, K., Moye, E., Williams, 8., Berlin, J. A., & Reynolds, E. E. (2007).
Screening mammography in women 40 to 49 years of age: a systematic review for
the American College of Physicians. Ann Intern Med, 146(7), 516-526.

Arredondo, E. M., Elder, J. P., Ayala, G. X., & Campbell, N. R. (2005). Is church
attendance associated with Latinas' health practices and self-reported health? Am J
Health Behav, 29(6), 502-511.

Ashton, C. M. (2008). An interview with Ronald Andersen, PhD. Medical Care, 46(7),
645-646.

Audrain, J ., Lerman, C., Rimer, B., Cella, D., Steffens, R., & Gomez-Caminero, A.
(1995). Awareness of heightened breast cancer risk among ﬁrst-degree relatives
of recently diagnosed breast cancer patients. The High Risk Breast Cancer
Consortium. Cancer Epidemiol Biomarkers Prev, 4(5), 561-565.

Austin, L. T., Ahmad, F., McNally, M. J ., & Stewart, D. E. (2002). Breast and cervical
cancer screening in Hispanic women: a literature review using the health belief
model. Womens Health Issues, 12(3), 122-128.

Baker, D. W., Parker, R. M., Williams, M. V., Clark, W. S., & Nurss, J. (1997). The
relationship of patient reading ability to self-reported health and use of health
services. Am J Public Health, 87(6), 1027-1030.

Barroso, J., McMillan, S., Casey, L., Gibson, W., Kaminski, G., & Meyer, J. (2000).
Comparison between Aﬁ'ican-American and white women in their beliefs about
breast cancer and their health locus of control. Cancer Nurs., 23(4), 268-276.

Benjamins, M. R. (2006). Religious inﬂuences on preventive health care use in a
nationally representative sample of middle-age women. J Behav Med, 29(1), 1-16.

140

Bentley, J. R., Delﬁno, R. J., Taylor, T. H., Howe, S., & Anton-Culver, H. (1998).
Differences in breast cancer stage at diagnosis between non-Hispanic white and
Hispanic populations, San Diego County 1988-1993. Breast Cancer Res Treat,
50(1), 1-9. '

Berestein, L. (2008). Decline in border crossings crimps economy: Waits, security issues
hitting tourism hard. Retrieved March 3, 2009, from
http://www.signonsandiego.com/news/mexico/tijuana/200803 1 1-9999-
lnl 1cross.html

Berry, J. W. (1988). Acculturation and psychological adaptation: a conceptual overview.
In J. W. Berry & R. C. Annis (Eds), Ethnic psychology: research and practice
with immigrants, refugees, native peoples, ethnic groups and sojourners.
Berwyn: Swets North America.

Blackrnan, D. J., & Masi, C. M. (2006). Racial and ethnic disparities in breast cancer
mortality: are we doing enough to address the root causes? J Clin Oncol, 24(14),
2170-217 8.

Borrayo, E. A., & Guarnaccia, C. A. (2000). Differences in Mexican-born and U.S.-bom
women of Mexican descent regarding factors related to breast cancer screening
behaviors. Health Care Women Int, 21(7), 599-613.

Borrayo, E. A., & Jenkins, S. R. (2001). Feeling indecent: Breast cancer screening
resistance of Mexican descent women. Journal of Health Psychology, 6, 53 7-550.

Brach, C., & Chevarley, F. M. (2008). Demographics and Health Care Access and
Utilization of Limited-English-Proﬁcient and English-Proﬁcient Hispanics.
Research Findings No. 28. [Electronic Version]. Agency for Healthcare Research
and Quality, Rockville, MD. Retrieved May 10, 2008, ﬁom
http://www.meps.ahrq.gov/papers/r128/rf28.pdf

Brawley, O. W., & Kramer, B. S. (2005). Cancer screening in theory and in practice. J
Clin Oncol, 23(2), 293-300.

Brody, H., & Hunt, L. M. (2005). Moving beyond cultural stereotypes in end-of-life
decision making. Am Fam Physician, 71(3), 429-430.

Brown, D. W., Balluz, L. S., Heath, G. W., Moriarty, D. G., Ford, E. S., Giles, W. H., et
al. (2003). Associations between recommended levels of physical activity and
health-related quality of life. Findings from the 2001 Behavioral Risk Factor
Surveillance System (BRFSS) survey. Prev Med, 37(5), 520-528.

Buist, D. S., Porter, P. L., Lehman, C., Taplin, S. H., & White, B. (2004). Factors

contributing to mammography failure in women aged 40-49 years. J Natl Cancer
Inst, 96(19), 1432-1440.

141

Burger, J. M. (1999). The foot-in-the-door compliance procedure: a multiple-process
analysis and review. Pers Soc Psychol Rev, 3(4), 303-325.

California Cancer Registry (CCR). (2008). Public Use Data Set: California Cancer
Incidence and Mortality Rates Plus Interactive Maps Retrieved August 1, 2008,
from http://www.ccrcal.org/dataquery.htrnl

California Health Interview Survey (CHIS). (2005). Mammography Screening History.
Retrieved August 1, 2008, from http://www.chis.ucla.edu

Calle, E. E., Flanders, W. D., Thun, M. J ., & Martin, L. M. (1993). Demographic
predictors of marmnography and Pap smear screening in US women. Am J Public
Health, 83(1), 53-60.

Carrasquillo, O., & Pati, S. (2004). The role of health insurance on Pap smear and
mammography utilization by immigrants living in the United States. Prev Med,
39(5), 943-950.

Cauce, A. M., Coronado, N., & Watson, J. (1998). Conceptual, methodological, and
statistical issues in culturally competent research. In M. Hernandez & M. Isaacs
(Eds), Promoting Cultural Competence in Children’s Mental Health Services
(pp. 305—329). Baltimore, MD: Paul H. Brookes.

Centers for Disease Control and Prevention (CDC). (2006). 2007 Behavioral Risk Factor
Surveillance System Questionnaire. Retrieved May 15, 2008, ﬁom
http://www.cdc.gov/brfss/questionnaires/pdf-ques/2007brfss.pdf

CDC. (2007). Prevalence Data, Nationwide (States and DC) - 2006, Women's Health.
Retrieved May 10, 2008, from
http://apps.nccd.cdc. gov/brfss/race.asp?cat=WH&yr=2006&qkey=4421 &state=U
B

CDC. (2008). National Breast and Cervical Cancer Early Detection Program, About the
Program. Retrieved May 10, 2008, ﬁ'om
http://www.cdc.gov/cancer/NBCCEDP/about.htm

Chavez, L. R., Hubbell, F. A., McMullin, J. M., Martinez, R. G., & Mishra, S. I. (1995).
Understanding knowledge and attitudes about breast cancer. A cultural analysis.
Arch Fam Med, 4(2), 145-152.

 

Chavez, L. R., Hubbell, F. A., Mishra, S. I., & Valdez, R. B. (1997). The inﬂuence of
fatalism on self-reported use of Papanicolaou smears. Am J Prev Med, 13(6), 418—
424.

Chew, L. D., Bradley, K. A., & Boyko, E. J. (2004). Brief questions to identify patients
with inadequate health literacy. Fam Med, 36(8), 588-594.

142

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.).
Hillsdale, NJ: Lawrence Erlbaum.

Cohen, J ., Cohen, R, West, S. G., & Aiken, L. S. (2003). Applied Multiple Regression/
Correlation Analysis for the Behavioral Sciences (3rd ed.). Mahwah: NJ .:
Lawrence Erlbaum Associates.

Cohen, M. (2006). Breast cancer early detection, health beliefs, and cancer worries in
randomly selected women with and without a farnily history of breast cancer.
Psychooncology, 15(10), 873-883.

Collins, J. M., & Schmidt, F. L. (1997). Can suppressor variables enhance criterion-
related validity in the personality domain? Educational and Psychological
Measurement, 5 7(6), 924-936.

Comrey, A. L., & Lee, H. B. (1992). A First Course in Factor Analysis (2nd ed.).
Hillsdale, NJ: Lawrence Erlbaum Associates.

Conger, A. J. (1974). A revised deﬁnition for suppressor variables: A guide to their
identiﬁcation and interpretation. Educational and Psychological Measurement,
34, 35-46.

Conger, A. J ., & Jackson, D. N. (1972). Suppressor variables, prediction, and the
interpretation of psychological relationships. Educational and Psychological
Measurement, 32, 579-599.

Coulter, I. D., Marcus, M., Freed, J. R., Der-Martirosian, C., Cunningham, W. B.,
Andersen, R. M., et a1. (2000). Use of dental care by HIV-infected medical
patients. J Dent Res, 79(6), 1356-1361.

Davidson, P. L., & Andersen, R. M. (1997). Determinants of dental care utilization for
diverse ethnic and age groups. Adv Dent Res, 11(2), 254-262.

Davidson, P. L., Rams, T. E., & Andersen, R. M. (1997). Socio-behavioral determinants
of oral hygiene practices among USA ethnic and age groups. Adv Dent Res, 11(2),
245-253.

Davis, T. C., Arnold, G, Berkel, H. J ., Nandy, 1., Jackson, R. H., & Glass, J. (1996).
Knowledge and attitude on screening mammography among low-literate, low-
income women. Cancer, 78(9), 1912-1920.

De Alba, 1., Hubbell, F. A., McMullin, J. M., Sweningson, J. M., & Saitz, R. (2005).

Impact of US. citizenship status on cancer screening among immigrant women. J
Gen Intern Med, 20(3), 290-296.

143

De Alba, 1., Sweningson, J. M., Chandy, C., & Hubbell, F. A. (2004). Impact of English
language proﬁciency on receipt of pap smears among Hispanics. J Gen Intern
Med, 19(9), 967-970.

Demidenko, E. (2008). Sample size and optimal design for logistic regression with binary
interaction. Stat Med, 27(1), 36-46.

DeNavas-Walt, C., Proctor, B. D., & Lee, C. H. (2005). Income, Poverty, and Health
Insurance Coverage in the United States: 2004. Current Population Reports, P60-
229, U. S. Governmental Printing Oﬂice, Washington, D. C. Retrieved January
16, 2008, from http://www.census.gov/prod/2005pubs/p60-229.pdf

Derose, K. P., & Baker, D. W. (2000). Limited English proﬁciency and Latinos' use of
physician services. Med Care Res Rev, 57(1), 76-91.

DeVellis, R. F. (2003). Scale Development: Theory and Applications. (2nd ed.). Applied
Social Research Methods Series, Vol 26. Thousand Oaks, CA: Sage Publications.

DeVoe, J. B., Wallace, L. S., & Fryer, G. E., Jr. (2009). Measuring patients' perceptions
of communication with healthcare providers: do differences in demographic and
socioeconomic characteristics matter? Health Expect, 12(1), 70-80.

Diefenbach, M. A., Miller, S. M., & Daly, M. B. (1999). Speciﬁc worry about breast
cancer predicts mammography use in women at risk for breast and ovarian cancer.
Health Psycho], 18(5), 532-536.

Dominick, K. L., Ahern, F. M., Gold, C. H., & Heller, D. A. (2002). Relationship of
health-related quality of life to health care utilization and mortality among older
adults. Aging Clin Exp Res, 14(6), 499-508.

Doty, M. M. (2003). Hispanic Patients' Double Burden: Lack of Health Insurance and
Limited English. The Commonwealth Fund. [Electronic Version]. Retrieved May
10, 2008, from
http://www.commonwealthfund.org/usr_doc/doty_hispanicdoubleburden_592.pdf
?section=4039

Elmore, J. G., Armstrong, K., Lehman, C. D., & Fletcher, S. W. (2005). Screening for
breast cancer. JAMA, 293(10), 1245-1256.

Estrada, A. L., Trevino, F. M., & Ray, L. A. (1990). Health care utilization barriers
among Mexican Americans: evidence from HHANES 1982-84. Am J Public
Health, 80 Suppl, 27-31.

Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A ﬂexible

statistical power analysis program for the social, behavioral, and biomedical
sciences. Behavior Research Methods, 39(175-191).

144

Felix Aaron, K., Levine, D., & Burstin, H. R. (2003). Aﬁican American church
participation and health care practices. J Gen Intern Med, 18(11), 908-913.

Fernandez, L. B., & Morales, A. (2007). Language and use of cancer screening services
among border and non-border Hispanic Texas women. Ethn Health, 12(3), 245-
263.

Figueroa, L. R., Davis, B., Baker, S., & Bunch, J. B. (2006). The inﬂuence of spirituality
on health care-seeking behaviors among Aﬁican Americans. ABNF .1, 1 7(2), 82-
88.

Florez, K. R., Aguirre, A. N., Viladrich, A., Cespedes, A., De La Cruz, A. A., & Abraido-
Lanza, A. F. (2008). Fatalism or Destiny? A Qualitative Study and Interpretative
Framework on Dominican Women's Breast Cancer Beliefs. J Immigr Minor
Health.

Fox, P., Amsberger, P., Owens, D., Nussey, B., Zhang, X., Golding, J. M., et a1. (2004).
Patient and clinical site factors associated with rescreening behavior among older
multiethnic, low-income women. Gerontologist, 44(1), 76-84.

Frazier, E. L., Jiles, R. B., & Mayberry, R. (1996). Use of screening mammography and
clinical breast examinations among black, Hispanic, and white women. Prev Med,
25(2), 118-125.

Freeman, H. P. (2004). Poverty, culture, and social injustice: determinants of cancer
disparities. CA CancerJ Clin, 54(2), 72-77.

Gallagher, T. C., Andersen, R. M., Koegel, P., & Gelberg, L. (1997). Determinants of
regular source of care among homeless adults in Los Angeles. Med Care, 35(8),
814-830.

Garbers, S., & Chiasson, M. A. (2004). Inadequate functional health literacy in Spanish
as a barrier to cervical cancer screening among immigrant Latinas in New York
City. Prev Chronic Dis, 1(4), A07.

Garbers, S., J essop, D. J ., Foti, H., Uribelarrea, M., & Chiasson, M. A. (2003). Barriers to
breast cancer screening for low-income Mexican and Dominican women in New
York City. J Urban Health, 80(1), 81-91.

Garces, I. C., Scarinci, I. C., & Harrison, L. (2006). An examination of sociocultural
factors associated with health and health care seeking among Latina immigrants. J
Immigr Minor Health, 8(4), 377-385.

Gelberg, L., Andersen, R. M., & Leake, B. D. (2000). The Behavioral Model for

Vulnerable Populations: application to medical care use and outcomes for
homeless people. Health Serv Res, 34(6), 1273-1302.

145

Goel, M. S., Wee, C. C., McCarthy, E. R, Davis, R. B., Ngo-Metzger, Q., & Phillips, R.
S. (2003). Racial and ethnic disparities in cancer screening: the importance of
foreign birth as a barrier to care. J Gen Intern Med, 18(12), 1028-1035.

Gorin, S. S., & Heck, J. E. (2005). Cancer screening among Latino subgroups in the
United States. Prev Med, 40(5), 515-526.

Gorsuch, R. L. (1983). Factor Analysis (2nd ed.). Hillsdale, NJ.: Lawrence Erlbaum
Associates.

Gramling, R., Anthony, D., Frierson, G., & Bowen, D. (2007). The cancer worry chart: a
single-item screening measure of worry about developing breast cancer.
Psychooncology, 16(6), 593-597.

Guerra, C. E., Krumholz, M., & Shea, J. A. (2005). Literacy and knowledge, attitudes and
behavior about mammography in Latinas. J Health Care Poor Underserved,
16(1), 152-166.

Gullate, M. (2006). The inﬂuence of spirituality and religiosity on breast cancer

screening delay in Aﬁican American women: application of the Theory of
Reasoned Action and Planned Behavior (TRA/TPB). ABNF J, 1 7(2), 89-94.

Guzman, B. (2001). The Hispanic Population: Census 2000- Brief (02kbr01-3).
Retrieved June 10, 2008, from http://www.census. gov/prod/2001pubs/02kbr01-
3.pdf

Hargraves, J. L., & Hadley, J. (2003). The contribution of insurance coverage and
community resources to reducing racial/ethnic disparities in access to care. Health
Serv Res, 38(3), 809-829.

Hay, J. L., Buckley, T. R., & Ostroff, J. S. (2005). The role of cancer worry in cancer
screening: a theoretical and empirical review of the literature. Psychooncology,
14(7), 517-534.

Health Resources and Services Administration (HRSA). (2009). HRSA- Find a Health
Center- Search by County or State. Retrieved March 11, 2009, from
http://ﬁndahealthcenter.hrsa.gov/SearchByCounty.aspx

Helm, H. M., Hays, J. C., Flint, E. P., Koenig, H. G., & Blazer, D. G. (2000). Does
private religious activity prolong survival? A six-year follow-up study of 3,851
older adults. J Gerontol A Biol Sci Med Sci, 55(7), M400—405.

Hiatt, R. A., Pasick, R. J ., Stewart, S., Bloom, J ., Davis, R, Gardiner, P., et a1. (2001).
Community-based cancer screening for underserved women: design and baseline
ﬁndings from the Breast and Cervical Cancer Intervention Study. Prev Med,
33(3), 190-203.

146

Hill, T. D., Burdette, A. M., Ellison, C. G., & Musick, M. A. (2006). Religious
attendance and the health behaviors of Texas adults. Prev Med, 42(4), 309-312.

Hof’ﬁnan, C. B., & Paradise, J. (2007). Health Insurance and Access to Health Care in the
United States. Ann N Y Acad Sci.

Horn, J. L., & McArdle, J. J. (1992). A practical and theoretical guide to measurement
invariance in aging research. Exp Aging Res, 18(3-4), 117-144.

Hoyle, R. H., & Smith, G. T. (1994). Formulating clinical research hypotheses as
structural equation models: a conceptual overview. J Consult Clin Psychol, 62(3),
429-440.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for ﬁt indices in covariance structure
analysis: Conventional criteria, versus new alternatives. Structural Equation
Modeling, 6, 1-55.

Hubbell, F. A., Chavez, L. R., Mishra, S. I., & Valdez, R. B. (1996). Differing beliefs
about breast cancer among Latinas and Anglo women. West J Med, 164(5), 405-
409.

Hubbell, F. A., Mishra, S. 1., Chavez, L. R., & Valdez, R. B. (1997). The inﬂuence of
knowledge and attitudes about breast cancer on mammography use among Latinas
and Anglo women. J Gen Intern Med, 12(8), 505-508.

Huerta, E. E. (2003). Cancer statistics for Hispanics, 2003: good news, bad news, and the
need for a health system paradigm change. CA Cancer J Clin, 53(4), 205-207.

Humphrey, L. L., Helfand, M., Chan, B. K., & Woolf, S. H. (2002). Breast cancer
screening: a summary of the evidence for the US. Preventive Services Task
Force. Ann Intern Med, 137(5 Part 1), 347-360.

Institute of Medicine (IOM). (1999). The Unequal Burden of Cancer: An Assessment of
NIH Research and Programs for Ethnic Minorities and the Medically
Underserved. Washington, D. C.: National Academy Press.

Jones, A. R., Caplan, L. S., 81. Davis, M. K. (2003). Racial/ethnic differences in the self-
reported use of screening mammography. J Community Health, 28(5), 303-316.

J oreskog, K. G. (1999). How Large Can a Standardized Coefﬁcient be? Retrieved
March 1, 2009, from
http://www.ssicentral.com/1isrel/techdocs/HowLargeCanaStandardizedCoefﬁcient

be.pdf

J oreskog, K. G., & Sorbom, D. (2007). LISREL 8.80 Student Edition. Chicago: Scientiﬁc
Software International, Inc.

147

Kerlikowske, K., Grady, D., Rubin, S. M., Sandrock, C., & Ernster, V. L. (1995).
Efﬁcacy of screening mammography. A meta-analysis. JAMA, 273(2), 149-154.

Kim, S. E., Perez-Stable, E. J ., Wong, S., Gregoriclr, S., Sawaya, G. F ., Walsh, J. M., et
a1. (2008). Association between cancer risk perception and screening behavior
among diverse women. Arch Intern Med, 168(7), 728-734.

Kinney, A. Y., Emery, G., Dudley, W. N., & Croyle, R. T. (2002). Screening behaviors
among African American women at high risk for breast cancer: do beliefs about
god matter? Oncol Nurs Forum, 29(5), 835-843.

Kleinbaurn, D. G., Kupper, L. L., Nizarn, A., & Muller, K. E. (2007). Applied Regression
Analysis and Multivariable Methods (4th ed.). Paciﬁc Grove, CA: Duxbury Press.

Kline, R. B. (1998). Principles and Practice of Structural Equation Modeling. New York,
London: The Guilford Press.

Kubzansky, L. D., Kawachi, I., Spiro, A., 3rd, Weiss, S. T., Vokonas, P. S., & Sparrow,
D. (1997). Is worrying bad for your heart? A prospective study of worry and
coronary heart disease in the Normative Aging Study. Circulation, 95(4), 818-
824.

Legislative Council of California (2000a). Califomia Health & Safety Code, § 1367.65.
Retrieved August 1, 2008, ﬁ'om: http://www.leginfo.ca.gov/cgi-
bin/displaycode?section=hsc&group=01001-02000&ﬁle=1 367-1 3 74. 1 9

Legislative Council of California (2000b). California Welfare & Institutions Code, §
14132.16. Retrieved August 1, 2008, from: http://www.leginfo.ca.gov/cgi-
bin/displaycode?section=wic&group=14001-1 5000&ﬁle=1413 1-141 38

Legislative Council of California (20000). California Insurance Code, § 10123.81.
Retrieved August 1, 2008 from: http://www.leginfo.ca.gov/cgi-
bin/displaycode?section=ins&group=10001 -1 1000&ﬁle=1 01 10-1 0127.1 8

Legler, J ., Meissner, H. I., Coyne, C., Breen, N., Chollette, V., & Rimer, B. K. (2002).
The effectiveness of interventions to promote mammography among women with
historically lower rates of screening. Cancer Epidemiol Biomarkers Prev, 11, 59-
71.

Lenth, R. V. (2006). Java Applets for Power and Sample Size [Computer software].
Retrieved September 1, 2008, ﬁ'om http://www.stat.uiowa.edu/~rlenth/Power.

Levin, J. S., Markides, K. S., & Ray, L. A. (1996). Religious attendance and

psychological well-being in Mexican Americans: a panel analysis of three-
generations data. Gerontologist, 36(4), 454-463.

148

Lindau, S. T., Tomori, C., Lyons, T., Langseth, L., Bennett, C. L., & Garcia, P. (2002).
The association of health literacy with cervical cancer prevention knowledge and
health behaviors in a multiethnic cohort of women. Am J Obstet Gynecol, 186(5),
93 8-943.

Lloyd, L. L., Ammary, N. J ., Epstein, L. G., Johnson, R., & Rhee, K. (2006). A
transdisciplinary approach to improve health literacy and reduce disparities.
Health Promot Pract, 7(3), 331-335.

Lopez-McKee, G., McNeil], J. A., Eriksen, L. R., & Ortiz, M. (2007). Spanish translation
and cultural adaptation of the Powe F atalism Inventory. J Nurs Scholarsh, 39(1),
68-70.

Lucas, J. W. (2003). Theory-testing, generalization, and the problem of external validity.
Sociological Theory, 21(3), 236-253.

Lynn, H. S. (2003). Suppression and confounding in action. The American Statistician,
57(1), 58-61.

MacKinnon, D. P., Krull, J. L., & Lockwood, C. M. (2000). Equivalence of the
mediation, confounding and suppression effect. Prevention Science, 1(4), 173-
181.

Magana, A., & Clark, N. M. (1995). Examining a paradox: does religiosity contribute to
positive birth outcomes in Mexican American populations? Health Educ Q, 22(1),
96-109.

Malcame, V. L., Fernandez, S., & Flores, L. (2005). Factorial validity of the
multidimensional health locus of control scales for three American ethnic groups.
J Health Psycho], 10(5), 657-667.

Mandelblatt, J. S., Yabroff, K. R., & Kerner, J. F. (1999). Equitable access to cancer
services: A review of barriers to quality care. Cancer, 86(11), 2378-2390.

Marin, G. (1993). Deﬁning culturally appropriate community interventions: Hispanics as
a case study. Journal of Community Psychology, 21(2), 149—161.

Marin, G., Sabogal, R., Marin, B.V., Otero-Sabogal, R., & Perez-Stable, E. (1987).
Development of a short acculturation scale for Hispanics. Hispanic Journal of
Behavioral Sciences, 9, 183-205.

Marks, D. F., Murray, M., Evans, B., Willig, C., Woodall, C., & Sykes, C. M. (2005).

Health Psychology: Theory, Research & Practice (2nd ed.). London, Thousand
Oaks, New Delhi: Sage Publications.

149

McCaul, K. D., Branstetter, A. D., O'Donnell, S. M., Jacobson, K., & Quinlan, K. B.
(1998). A descriptive study of breast cancer worry. J Behav Med, 21(6), 565-579.

McCaul, K. D., Branstetter, A. D., Schroeder, D. M., & Glasgow, R. E. (1996). What is
the relationship between breast cancer risk and mammography screening? A
meta-analytic review. Health Psycho], 15(6), 423-429.

McCaul, K. D., Schroeder, D. M., & Reid, P. A. (1996). Breast cancer worry and
screening: some prospective data. Health Psycho], 15(6), 430-433.

McCaul, K. D., & Tulloch, H. E. (1999). Cancer screening decisions. J Natl Cancer Inst
Monogr(25), 52-58.

Mechanic, D. (1979). Correlates of physician utilization: why do major multivariate
studies of physician utilization ﬁnd trivial psychosocial and organizational
effects? J Health Soc Behav, 20(4), 387-396.

Metsch, L. R., McCoy, C. B., McCoy, H. V., Pereyra, M., Trapido, B., & Miles, C.
(1998). The role of the physician as an information source on mammography.
Cancer Pract, 6(4), 229-236.

Miller, D. C., Gelberg, L., Kwan, L., Stepanian, S., Fink, A., Andersen, R. M., et a1.
(2008). Racial Disparities in Access to Care for Men in a Public Assistance
Program for Prostate Cancer. J Community Health, 33 (5), 318-335.

Millman, M. (Ed.). (1993). Access to Health Care in America. Washington, DC: Institute 9
of Medicine (IOM), Committee on Monitoring Access to Personal Health Care
Services.

Mitchell, J ., Lannin, D. R., Mathews, H. F., & Swanson, M. S. (2002). Religious beliefs
and breast cancer screening. J Womens Health (Larchmt), 11(10), 907-915.

Muller, A. (2002). Education, income inequality, and mortality: a multiple regression
analysis. BMJ, 324(7328), 23-25.

Murray, M., & McMillan, C. L. (1993). Gender differences in perceptions of cancer. J
Cancer Educ, 8(1), 53-62.

National Association of Community Health Centers (NACHC). (2008a). National
Association of Community Health Centers Annual Report 2007. Retrieved March
11, 2009, from http://www.nachc.com/client/documents/2008NACHCAdef

NACHC (2008b). California Health Center Fact Sheet, 2007. Retrieved March 11, 2009,

ﬁom http://www.nachc.com/client/docmnents/research/Z008-State-Fact-
Sheets/CAstatefactsheet08.pdf

150

NACHC. (20080). Serving Patients with Limited English Proﬁciency: Results of a
Community Health Center Survey. Retrieved March 11, 2009, ﬁ'om
http://www.nachc.com/client/documents/LEP_report.pdf

NACHC. (2009). About Our Health Centers. Retrieved March 11, 2009, from
http://www.nachc.com/about-our—health-centers.cfm

National Cancer Institute (N CI). (2005a). Cancer Health Disparities: A F act Sheet
[Electronic Version]. BenchMarks, 5 (6). Retrieved December 1, 2007, from
http://www.cancer.gov/newscenter/benchmarks-vol5-issue6/page2

NCI. (2005b). Theory at a Glance: A Guide for Health Promotion Practice (second
edition). National Institutes of Health (NIH) Publication No. 05-3896, Printed
September 2005. Retrieved June 1, 2008, from
http://www.cancer.gov/PDF/481f5d53-63df-41bcbfaf5aa48eelda4d/TAAG3.pdf.

NCI. (2007a). Health Disparities Deﬁned. Retrieved December 6, 2007, ﬁ'om
http://crchd.cancer.gov/deﬁnitions/deﬁned.html

NCI. (2007b). Breast Cancer Prevention (PDQ®). Retrieved December 7, 2007, from
http://www.cancer.gov/cancertopics/pdq/prevention/breast/Patient/

NCI. (2007c). Screening Marnmograrns: Questions and Answers. Retrieved August 1,
2008, from http://www.cancer.gov/cancertopics/factsheet/detection/screening-
mammograms

NCI. (2008a). Surveillance Epidemiology and End Results (SEER) Cancer Statistics
Review 1975-2005: Table I-11, Median Age of Cancer Patients at Diagnosis,
2000-2005. Retrieved August 1, 2008, from
http://seer.cancer. gov/csr/ l 975_2005/resu1ts_single/sect_0 1 _table.1 1_2pgs.pdf

NCI. (2008b). Breast Cancer Screening (PDQ®). Retrieved January 15, 2009, ﬁom:
http://www.cancer.gov/cancertopics/pdq/screening/breast/healthprofessional/allpa
ges

National Center for Educational Statistics (N CES). (2007). Status and Trends in the
Education of Racial and Ethnic Minorities. NCES 200 7-039, from _
http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2007039

Niederdeppe, J., & Levy, A. G. (2007). Fatalistic beliefs about cancer prevention and
three prevention behaviors. Cancer Epidemiol Biomarkers Prev, 16(5), 998-1003.

Norris, A. B., Ford, K., & Bova, C. A. (1996). Psychometrics of a brief acculturation

scale for Hispanics in a probability sample of urban Hispanic adolescents and
young adults. Hispanic Journal of Behavioral Sciences, 18(1), 29-3 8.

151

Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). New York. :
McGraw-Hill.

O'Malley, A. S., Kerner, J ., Johnson, A. B., & Mandelblatt, J. (1999). Acculturation and
breast cancer screening among Hispanic women in New York City. Am J Public
Health, 89(2), 219-227.

Otero-Sabogal, R., Owens, D., Canchola, J ., Golding, J. M., Tabnak, F., & Fox, P.
(2004). Mammography rescreening among women of diverse ethnicities: patient,
provider, and health care system factors. J Health Care Poor Underserved, 15(3),
390-412.

Otero-Sabogal, R., Stewart, S., Sabogal, F., Brown, B. A., & Perez-Stable, E. J. (2003).
Access and attitudinal factors related to breast and cervical cancer rescreening:
Why are Latinas still underscreened? Health Educ Behav, 30(3), 337-359.

Ounpuu, S., Chambers, L. W., Patterson, c., Chan, D., & Yusuf, s. (2001). Validity of
the US Behavioral Risk Factor Surveillance System's Health Related Quality of
Life Survey Tool in a Group of Older Canadians [Electronic Version]. Chronic
Diseases in Canada, 22. Retrieved March 10, 2008, ﬁ'om http://www.phac-
aspc.gc.ca/publicat/cdic-mcc/22-3/c_e.html

Padgett, D., Struening, E. L., & Andrews, H. (1990). Factors affecting the use of medical,
mental health, alcohol, and drug treatment services by homeless adults. Med
Care, 28(9), 805-821.

Painter, J. E., Borba, C. P., Hynes, M., Mays, D., & Glanz, K. (2008). The use of theory
in health behavior research from 2000 to 2005: a systematic review. Ann Behav
Med, 35(3), 358-362.

Park, C. L., & Gaffey, A. E. (2007). Relationships between psychosocial factors and
health behavior change in cancer survivors: an integrative review. Ann Behav
Med, 34(2), 115-134.

Pascual, A., & Gueguen, N. (2005). Foot-in-the—door and door-in-the—face: a comparative
meta-analytic study. Psychol Rep, 96(1), 122-128.

Pasick, R. J ., & Burke, N. J. (2007). A Critical Review of Theory in Breast Cancer
Screening Promotion across Cultures. Annu Rev Public Health.

Perez-Stable, E. J ., Sabogal, F., Otero-Sabogal, R., Hiatt, R. A., & McPhee, S. J. (1992).
Misconceptions about cancer among Latinos and Anglos. JAMA, 268(22), 3219-
3223.

152

Powe, B. (2001). Cancer fatalism among elderly African American women: Predictors of
the intensity of the perceptions. Journal of Psychosocial Oncology, 19(3/4), 85-
95.

Powe, B. D. (1995a). Fatalism among elderly Aﬁican Americans. Effects on colorectal
cancer screening. Cancer Nurs, 18(5), 385-392.

Powe, B. D. (1995b). Cancer fatalism among elderly Caucasians and Aﬁican Americans.
Oncol Nurs Forum, 22(9), 1355-1359.

Powe, B. D., & Finrrie, R. (2003). Cancer fatalism: the state of the science. Cancer Nurs,
26(6), 454-465; quiz 466-457.

Powell, L. H., Shahabi, L., & Thoresen, C. E. (2003). Religion and spirituality. Linkages
to physical health. Am Psycho], 58(1), 36-52.

Qureshi, M., Thacker, H. L., Litaker, D. G., & Kippes, C. (2000). Differences in breast
cancer screening rates: an issue of ethnicity or socioeconomics? J Womens Health
Gend Based Med, 9(9), 1025-1031.

Rajaram, S. S., & Rashidi, A. (1998). Minority women and breast cancer screening: the
role of cultural explanatory models. Prev Med, 27(5 Pt 1), 757-764.

Ramirez, A. G., Talavera, G. A., Villarreal, R., Suarez, L., McAlister, A., Trapido, E., et
a1. (2000). Breast cancer screening in regional Hispanic populations. Health Educ
Res, 15(5), 559-568.

Reyes-Ortiz, C. A., Camacho, M. E., Amador, L. F ., Velez, L. F., Ottenbacher, K. J ., &
Markides, K. S. (2007). The impact of education and literacy levels on cancer
screening among older Latin American and Caribbean adults. Cancer Control,
14(4), 388-395.

Ricketts, T. C., & Goldsmith, L. J. (2005). Access in health services research: the battle
of the frameworks. Nurs Outlook, 53(6), 274-280.

Ringash, J. (2001). Preventive health care, 2001 update: screening mammography among
women aged 40-49 years at average risk of breast cancer. CAMJ, 164(4), 469-476.

Rodriguez, M. A., Ward, L. M., & Perez-Stable, E. J. (2005). Breast and cervical cancer
screening: impact of health insurance status, ethnicity, and nativity of Latinas.
Ann Fam Med, 3(3), 235-241.

Rogler, L. H. (1999). Methodological sources of cultural insensitivity in mental health
research. Am Psycho], 54(6), 424-433.

153

Roosa, M. W., & Gonzales, N. A. (2000). Minority issues in prevention: introduction to
the special issue. Am J Community Psycho], 28(2), 145-148.

Rosenberg, M. (1973). The logical status of suppressor variables. The Public Opinion
Quarterly, 37(3), 359-372.

Salazar, M. K. (1996). Hispanic women's beliefs about breast cancer and mammography.
Cancer Nurs, 19(6), 43 7-446.

Sambamoorthi, U., & McAlpine, D. D. (2003). Racial, ethnic, socioeconomic, and access
disparities in the use of preventive services among women. Prev Med, 3 7(5), 475-
484.

Sassi, F., Luft, H. S., & Guadagnoli, E. (2006). Reducing racial/ethnic disparities in
female breast cancer: screening rates and stage at diagnosis. Am J Public Health,
96(12), 2165-2172.

Schettino, M. R., Hernandez-Valero, M. A., Moguel, R., Hajek, R. A., & Jones, L. A.
(2006). Assessing breast cancer knowledge, beliefs, and misconceptions among
Latinas in Houston, Texas. J Cancer Educ, 21 (1 Suppl), S42-46.

Schwartz, M. D., Taylor, K. L., & Willard, K. S. (2003). Prospective association between
distress and mammography utilization among women with a family history of
breast cancer. J Behav Med, 26(2), 105-117.

Selden, C. R., Zorn, M., Ratzan, S., & Parker, R. M. (2000). Health Literacy. Current
Bibliographies in Medicine. No 2000-1. Bethesda, MD: National Library of
Medicine. Retrieved May 21, 2008, from:
http://www.n1m.nih. gov/ archivel/20061 214/pubs/cbm/hliteracy.htrnl

Shah, M., Zhu, K., Palmer, R. C., Jatoi, I., Shriver, C., & Wu, H. (2007). Breast,
colorectal, and skin cancer screening practices and family history of cancer in
US. women. J Womens Health (Larchmt), 16(4), 526-534.

Sherman, A. C., Plante, T. G., Simonton, S., Adams, D. C., Harbison, C., & Burris, S. K.
(2000). A multidimensional measure of religious involvement for cancer patients:
the Duke Religious Index. Support Care Cancer, 8(2), 102-109.

Sohl, S. J., & Moyer, A. (2007). Tailored interventions to promote mammography
screening: a meta-analytic review. Prev Med, 45(4), 252-261.

Solis, J. M., Marks, G., Garcia, M., & Shelton, D. (1990). Acculturation, access to care,

and use of preventive services by Hispanics: ﬁndings from HHANES 1982-84.
Am J Public Health, 80 Suppl, 1 1-19.

154

SPSS for Windows [computer program]. (2005). Version 14.0.0. 2005. Chicago: SPSS
Inc. '

Starcevic, V., Berle, D., Milicevic, D., Harman, A., Lamplugh, C., & Eslick, G. D.
(2007). Pathological worry, anxiety disorders and the impact of co-occurrence
with depressive and other anxiety disorders. J Anxiety Disord, 21(8), 1016-1027.

Stein, J. A., Andersen, R., & Gelberg, L. (2007). Applying the Gelberg-Andersen
behavioral model for vulnerable populations to health services utilization in
homeless women. J Health Psycho], 12(5), 791-804.

Stein, J. A., & F ox, S. A. (1990). Language preference as an indicator of mammography
use among Hispanic women. J Natl Cancer Inst, 82(21), 1715-1716.

Storch, E. A., Roberti, J. W., Heidgerken, A. D., Storch, J. B., Lewin, A. B., Killany, E.
M., et a1. (2004). The Duke Religion Index: A Psychometric Investigation.
Pastoral Psychology, 53(2), 175-182.

Strawbridge, W. J ., Shema, S. J ., Cohen, R. D., & Kaplan, G. A. (2001). Religious
attendance increases survival by improving and maintaining good health
behaviors, mental health, and social relationships. Ann Behav Med, 23(1), 68-74.

Suarez, L. (1994). Pap smear and mammogram screening in Mexican-American women:
the effects of acculturation. Am J Public Health, 84(5), 742-746.

Suarez, L., & Pulley, L. (1995). Comparing acculturation scales and their relationship to
cancer screening among older Mexican-American women. J Natl Cancer Inst
Monogr(18), 41-47.

Swanson, K. A., Andersen, R., & Gelberg, L. (2003). Patient satisfaction for homeless
women. J Womens Health (Larchmt), 12(7), 675-686.

Tabachnick, B., & Fidell, L. (2001). Using Multivariate Statistics (4th ed.). Boston: Allyn
& Bacon.

Tabar, L., Vitak, B., Chen, H. H., Yen, M. F., Duffy, S. W., & Smith, R. A. (2001).
Beyond randomized controlled trials: organized mammographic screening
substantially reduces breast carcinoma mortality. Cancer, 91(9), 1724-1731

Tarakeshwar, N., Vanderwerker, L. C., Paulk, E., Pearce, M. J ., Kasl, S. V., & Prigerson,
H. G. (2006). Religious coping is associated with the quality of life of patients
with advanced cancer. J Palliat Med, 9(3), 646-657.

True, S., Kean, T., Nolan, P. A., Haviland, E. S., & Hohman, K. (2005). In conclusion:

the promise of comprehensive cancer control. Cancer Causes Control, 16 Suppl 1,
79-88.

155

Tzelgov, J., & Henik, A. (1991). Suppression situations in psychological research:
Deﬁnitions, implications, and applications. Psychological Bulletin, 109(3), 524-
536.

US. Census Bureau. (2004). US. Interim Projections by Age, Sex, Race, and Hispanic
Origin: 2000-2050. Retrieved May 15, 2008, from
http://www.census.gov/ipc/www/usinterimproj/

US. Census Bureau. (2006a). Selected Characteristics of the Foreign-Bom Population by
Region of Birth: Latin America. Retrieved April 20, 2008, ﬁom:
http://factﬁnder.census.gov/servlet/STTable?_bm=y&-
qr_name=AC S_2006__EST_GOO_S0506&-geo_id=01 000US&-
ds_name=ACS_2006_EST_G00_&-_lang=en&-formaF&-CONTEXT=st

US. Census Bureau. (2006b). California, Selected Characteristics of the Foreign-bom
Population by Region of Birth: Latin America. Retrieved April 20, 2008, from:
http://factﬁnder.census.gov/servlet/STTable?_bm=y&-context=st&-
qr_name=ACS_2006_EST_G00_SOS06&-ds_narne=ACS_2006_EST_G00_&-
CONTEXT=st&-tree_id=306&-redoLog=fa1se&-_caller=geoselect&-
geo_id=04000USO6&-format=&-_lang=en

US. Census Bureau. (2006c). San Diego County, Califomia- General Demographic
Characteristics. Retrieved April 20, 2008, from:
http://factﬁnder.census.gov/servlet/QTTable?-ds_name=PEP_2006_EST&-
qr_name=PEP_2006_EST_DP1 &-geo_id=05000US06073

US. Census Bureau. (2006d). Subject Tables, American FactFinder: United States
Employment Status. Retrieved August 1, 2008, from:
http:// factﬁnder. census. gov/ servlet/ STTable?_bm=y&- geo_id=0 1 000US&-
qr_name=AC S_2006_EST_G00_SZ3 01 &-ds_name=AC S_2006_EST_G00_&-
redoLog=false

US. Census Bureau. (2006c). Detailed Tables, American FactFinder: Median earnings in
the past 12 months (in 2006 inﬂation-adjusted dollars) by sex by educational
attainment for the population 25 years and over. Retrieved April 20, 2008, from:
http://factﬁnder.census. gov/servlet/DTTable?_bm=y&-geo_id=01 000US&-
ds_name=ACS_2006_EST_G00_&-mt_name=ACS_2006_EST_G2000_B20004

Underwood, S. M., & Powell, R. L. (2006). Religion and spirituality: inﬂuence on
health/risk behavior and cancer screening behavior of Aﬁican Americans. ABNF
.1, 17(1), 20-31.

Valdez, A., Banerjee, K., Ackerson, L., Fernandez, M., Otero-Sabogal, R., & Somkin, C.

P. (2001). Correlates of breast cancer screening among low-income, low-
education Latinas. Prev Med, 33(5), 495-502.

156

van Olphen, J ., Schulz, A., Israel, B., Chatters, L., Klem, L., Parker, R, et a1. (2003).
Religious involvement, social support, and health among African-American
women on the east side of Detroit. J Gen Intern Med, 18(7), 549-557.

Vandenberg, R. J. & Lance, C. E. (2000). A review and synthesis of the measurement
invariance literature: Suggestions, practices, and recommendations for
organizational research. Organizational Research Methods, 3(1), 4-70.

Ward, E., J emal, A., Cokkinides, V., Singh, G. K., Cardinez, C., Ghafoor, A., et a1.
(2004). Cancer disparities by race/ethnicity and socioeconomic status. CA Cancer
J Clin, 54(2), 78-93.

Watson, E. K., Henderson, B. J ., Brett, J ., Bankhead, C., & Austoker, J. (2005). The
psychological impact of mammographic screening on women with a farnily
history of breast cancer--a systematic review. Psychooncology, 14(11), 939-948.

Weathers, A., Minkovitz, C., O'Carnpo, P., & Diener-West, M. (2003). Health services
use by children of migratory agricultural workers: exploring the role of need for
care. Pediatrics, 111 (5 Pt 1), 956-963.

Wells, K. B., Golding, J. M., Hough, R. L., Bumam, M. A., & Karno, M. (1989).
Acculturation and the probability of use of health services by Mexican
Americans. Health Serv Res, 24(2), 237-257.

Wells, K. J ., & Roetzheim, R. G. (2007). Health disparities in receipt of screening
mammography in Latinas: a critical review of recent literature. Cancer Control,
14(4), 369-379.

West, D. S., Greene, P. G., Kratt, P. P., Pulley, L., Weiss, H. L., Siegﬁied, N., et a1.
(2003). The impact of a family history of breast cancer on screening practices and

attitudes in low-income, rural, African American women. J Womens Health
(Larchmt), 12(8), 779-787.

Wolinsky, F. D. (1978). Assessing the effects of predisposing, enabling, and illness-
morbidity characteristics on health service utilization. J Health Soc Behav, 19(4),
384-396. .

Wolinsky, F. D., & Johnson, R. J. (1991). The use of health services by older adults. J
Gerontol, 46(6), 8345-357.

Young, J. L., Jr., Ries, L. G., & Pollack, E. S. (1984). Cancer patient survival among
ethnic groups in the United States. J Natl Cancer Inst, 73(2), 341-3 52.

Zahran, H. S., Kobau, R., Moriarty, D. G., Zack, M. M., Holt, J ., & Donehoo, R. (2005).

Health-related quality of life surveillance--United States, 1993-2002. MM WR
Surveill Summ, 54(4), 1-35.

157

Zakowski, S. G., Valdimarsdottir, H. B., Bovbjerg, D. H., Borgen, P., Holland, J ., Kash,
K., et a1. (1997). Predictors of intrusive thoughts and avoidance in women with
family histories of breast cancer. Ann Behav Med, 19(4), 362-369.

Zambrana, R. E., Breen, N., Fox, S. A., & Gutierrez-Mohamed, M. L. (1999). Use of

cancer screening practices by Hispanic women: Analyses by subgroup. Prev Med,
29(6 Pt 1), 466-477.

158

ml W Ill ll]! llvllllllllllllllllﬂ
3 1293 03062 5614