Citation
Predictors of Attrition in Latino Alzheimer's Disease Caregivers in the REACH Trial: An Archival Investigation

Material Information

Title:
Predictors of Attrition in Latino Alzheimer's Disease Caregivers in the REACH Trial: An Archival Investigation
Creator:
CACHO, IVETTE MAGALY FREYTES
Copyright Date:
2008

Subjects

Subjects / Keywords:
Alzheimers disease ( jstor )
Caregivers ( jstor )
Dementia ( jstor )
Diseases ( jstor )
Health status ( jstor )
Hispanics ( jstor )
Logistic regression ( jstor )
Modeling ( jstor )
Research studies ( jstor )
Statistical models ( jstor )
City of Miami ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Ivette Magaly Freytes Cacho. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/12/2007
Resource Identifier:
659860308 ( OCLC )

Downloads

This item has the following downloads:

freytescacho_i ( .pdf )

freytescacho_i_Page_082.txt

freytescacho_i_Page_055.txt

freytescacho_i_Page_083.txt

freytescacho_i_Page_092.txt

freytescacho_i_Page_104.txt

freytescacho_i_Page_046.txt

freytescacho_i_Page_094.txt

freytescacho_i_Page_026.txt

freytescacho_i_Page_062.txt

freytescacho_i_Page_071.txt

freytescacho_i_Page_110.txt

freytescacho_i_Page_028.txt

freytescacho_i_Page_072.txt

freytescacho_i_Page_064.txt

freytescacho_i_Page_066.txt

freytescacho_i_Page_063.txt

freytescacho_i_Page_029.txt

freytescacho_i_Page_043.txt

freytescacho_i_Page_042.txt

freytescacho_i_Page_039.txt

freytescacho_i_Page_096.txt

freytescacho_i_Page_033.txt

freytescacho_i_Page_021.txt

freytescacho_i_Page_003.txt

freytescacho_i_Page_019.txt

freytescacho_i_Page_010.txt

freytescacho_i_Page_095.txt

freytescacho_i_Page_108.txt

freytescacho_i_Page_101.txt

freytescacho_i_pdf.txt

freytescacho_i_Page_052.txt

freytescacho_i_Page_087.txt

freytescacho_i_Page_006.txt

freytescacho_i_Page_113.txt

freytescacho_i_Page_085.txt

freytescacho_i_Page_084.txt

freytescacho_i_Page_035.txt

freytescacho_i_Page_061.txt

freytescacho_i_Page_080.txt

freytescacho_i_Page_051.txt

freytescacho_i_Page_045.txt

freytescacho_i_Page_102.txt

freytescacho_i_Page_099.txt

freytescacho_i_Page_013.txt

freytescacho_i_Page_106.txt

freytescacho_i_Page_076.txt

freytescacho_i_Page_074.txt

freytescacho_i_Page_070.txt

freytescacho_i_Page_048.txt

freytescacho_i_Page_112.txt

freytescacho_i_Page_040.txt

freytescacho_i_Page_041.txt

freytescacho_i_Page_034.txt

freytescacho_i_Page_091.txt

freytescacho_i_Page_098.txt

freytescacho_i_Page_030.txt

freytescacho_i_Page_105.txt

freytescacho_i_Page_054.txt

freytescacho_i_Page_017.txt

freytescacho_i_Page_049.txt

freytescacho_i_Page_097.txt

freytescacho_i_Page_022.txt

freytescacho_i_Page_075.txt

freytescacho_i_Page_088.txt

freytescacho_i_Page_027.txt

freytescacho_i_Page_011.txt

freytescacho_i_Page_038.txt

freytescacho_i_Page_050.txt

freytescacho_i_Page_016.txt

freytescacho_i_Page_079.txt

freytescacho_i_Page_001.txt

freytescacho_i_Page_073.txt

freytescacho_i_Page_036.txt

freytescacho_i_Page_009.txt

freytescacho_i_Page_060.txt

freytescacho_i_Page_031.txt

freytescacho_i_Page_023.txt

freytescacho_i_Page_090.txt

freytescacho_i_Page_002.txt

freytescacho_i_Page_107.txt

freytescacho_i_Page_086.txt

freytescacho_i_Page_058.txt

freytescacho_i_Page_037.txt

freytescacho_i_Page_056.txt

freytescacho_i_Page_025.txt

freytescacho_i_Page_069.txt

freytescacho_i_Page_014.txt

freytescacho_i_Page_004.txt

freytescacho_i_Page_100.txt

freytescacho_i_Page_024.txt

freytescacho_i_Page_057.txt

freytescacho_i_Page_047.txt

freytescacho_i_Page_114.txt

freytescacho_i_Page_005.txt

freytescacho_i_Page_032.txt

freytescacho_i_Page_081.txt

freytescacho_i_Page_089.txt

freytescacho_i_Page_077.txt

freytescacho_i_Page_103.txt

freytescacho_i_Page_007.txt

freytescacho_i_Page_008.txt

freytescacho_i_Page_012.txt

freytescacho_i_Page_109.txt

freytescacho_i_Page_067.txt

freytescacho_i_Page_018.txt

freytescacho_i_Page_059.txt

freytescacho_i_Page_020.txt

freytescacho_i_Page_093.txt

freytescacho_i_Page_111.txt

freytescacho_i_Page_068.txt

freytescacho_i_Page_053.txt

freytescacho_i_Page_078.txt

freytescacho_i_Page_015.txt

freytescacho_i_Page_044.txt

freytescacho_i_Page_065.txt


Full Text





PREDICTORS OF ATTRITION IN LATINO ALZHEIMER' S DISEASE CAREGIVERS IN
THE REACH TRIAL: AN ARCHIVAL INVESTIGATION





















By

IVETTE MAGALY FREYTES CACHO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2007

































Copyright 2007 Ivette Magaly Freytes Cacho





























To my parents, mami & papi, with love










ACKNOWLEDGMENTS

I would like to thank Dr. Edil Torres-Rivera and Dr. Sondra Smith, for chairing my

committee and for their support and guidance throughout this study. I also want to thank Dr.

Mary Fukuyama and Dr. Leilani Doty for their helpful feedback and suggestions and Dr. Mirka

Koro-Ljungberg for graciously agreeing, at the last minute, to fi11 an unexpected vacancy on my

committee. I am indebted to the original investigators of the Resources for Enhancing

Alzheimer' s Caregiver Health (REACH) study for the contribution of the data and their expertise

to this study. I would also like to thank Jocelyn Lee and Judy Callahan for their input regarding

the data analysis. I am especially grateful to Betsy Pearman for her encouragement and her

priceless contribution and assistance with the daunting task of data analysis throughout the

process.

I would like to express my deepest appreciation to Dr. Timothy Ketterson, who mentored

me throughout the years and whose ongoing support, guidance, and encouragement were

essential in the undertaking and completion of this dissertation. He spent countless hours

helping me through this journey and always believed in my ability to Einish this task.

I would also like to thank the many friends who offered their support throughout all these

years. Specifically, I want to thank Jim and Patty Probert for their ongoing support and love, and

the countless and delicious meals they provided throughout this process. I want to thank Dana

Myers for giving me a social outlet when I was burned out and for consistently showering me

with her "good vibes." I would like to thank Alice Martin and Debbie Gipple for always lending

a listening ear and for helping me to find the humor in this process.









Last but not least, I want to extend my deepest appreciation to my parents for believing in

me even when I doubted myself and encouraging me when I wanted to give up. Without their

unconditional love and support, and their relentless faith in me, I never would have come this far.













TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES ........._..__..... ._._ ...............8....


AB S TRAC T ......_ ................. ............_........9


CHAPTER


1 INTRODUCTION ................. ...............11.......... ......


Scope of the Problem ................. ...............12................
Statement of the Problem ................. ...............14................
Need for the Study ................. ...............16.......... ....
Purpose of the Study ................. ...............17.......... .....
Theoretical Fram ew ork................... .... .... .. ............... ........1
Overview of the Resources for Enhancing Alzheimer' s Caregiver Health Study .................. 19
Hypotheses............... ...............2
Definition of Terms .............. ......... ...............21
Overview of the Remainder of Study ................. ...............21..............

2 REVIEW OF RELEVANT LITERATURE ................. ...............22........... ...


Attrition Rates and Latinos ................. ...............24................
Latino Caregivers' Age............... .... ...............27
Latino Caregivers' Level of Education ................. ...............29...............
Latino Caregivers' Income Level .............. ...............30....
Depression in Latino Caregivers .............. ...............31....
Social Support and Latino Care givers .............. ...............32....
Latino Caregivers' Health Status ................. ...............34........... ...
Summary and Conclusions .............. ...............36....

3 M ETHODS .............. ...............38....


Resources for Enhancing Alzheimer' s Caregiver Health ................. ......... ................38
Sample Characteristics............. .....................3
Variables of Interest and Operational Definitions ................ ...............40...............
Null Hypotheses............... ...............4
Research Procedures............... ...............4
M measures ................. ...............43.................
Data Analysis................ ...............4
Methodological Limitations............... ...............4












4 RE SULT S .............. ...............47....


Preliminary Analyses ................. ...............47.................
Palo Alto Correlations .............. ...............49....
M iami Correlations............... ..............5

Hypothesis Testing .............. ...............50....
Data Analysis............... ...............51
Palo Alto Sample ................. ...............52........... ....
M iami Sample .............. ...............54....
Summary of Results............... ...............5


5 CONCLU SION................ ..............7


Conclusions............... ..............7
Palo Alto ............. ...... ._ ...............75...
M iam i .............. ...............75....

Im plications .............. ...............76....
Limitations ............. ...... ._ ...............78...
Recommendations............... ............8


APPENDIX


A CAREGIVER SOCIODEMOGRAPHIC INFORMATION ................. ................ ...._.84


B CENTER FOR EPIDEMIOLOGIC STUDIES DEPRESSION SCALE (CES-D) ................89


C SOCIAL SUPPORT .............. ...............94....


D CAREGIVER HEALTH AND HEALTH BEHAVIOR ......____ ........._ ................101


REFERENCE LIST .............. ...............105....


BIOGRAPHICAL SKETCH ................. ...............114......... ......












LIST OF TABLES

Table page

4-1 Variable means for total group and by location (Palo Alto and Miami) ...........................57

4-2 Variable means for total group and dropouts and completers .............. .....................5

4-3 Income level group for completion status and location ........................... ...............59

4-4 ANOVA test between sites (Palo Alto and Miami) ................. .............................60

4-5 ANOVA test between treatment dropouts and treatment completers ............... .... ........._..61

4-6 Palo Alto correlations .............. ...............62....

4-7 Miami correlations ................. ...............63........... ....

4-8 Palo Alto logistic regression, model one .............. ...............64....

4-9 Palo Alto logistic regression, model two .................._____ ......... ...........6

4-10 Palo Alto logistic regression, model three ................ ...............66........... ..

4-11 Palo Alto logistic regression, model four .............. ...............67....

4-12 Palo Alto logistic regression, model five ................. ...............68........... ..

4-13 Miami logistic regression, model one ................. ...............69...............

4-14 Miami logistic regression, model two............... ...............70..

4-15 Miami logistic regression, model three ................. ...............71...............

4-16 Miami logistic regression, model four ................. ...............72...............

4-17 Miami logistic regression, model five .............. ...............73....









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PREDICTORS OF ATTRITION IN LATINO ALZHEIMER' S DISEASE CAREGIVERS IN
THE REACH TRIAL: AN ARCHIVAL INVESTIGATION

By

Ivette Magaly Freytes Cacho

May 2007

Chair: Edil Torres-Rivera
Cochair: Sondra Smith
Major: Mental Health Counseling

Progressive memory disorders are becoming a national epidemic. It is estimated that

75% of persons with memory disorders are cared for informally, at home, by a friend or family

member. Caring for someone with cognitive impairment is a demanding and challenging

responsibility that places tremendous strain on families. Due to the changing demographics in

the composition of the U. S. population, it is expected that a significant proportion of elders with

dementia and their caregivers will be of diverse backgrounds.

Latinos, the largest and fastest growing segment of the U. S. population, are expected to

account for 25% of the dementia-affected cohort; however, little is known about Latino dementia

caregivers. Recruiting and retaining Latinos into research studies continues to be a challenge for

investigators. This study was a secondary analysis of the Resources for Enhancing Alzheimer' s

Caregiver Health (REACH) trial, a multi-site randomized trial that developed and tested the

effectiveness of multiple interventions for family caregivers of persons with Alzheimer' s disease

and related disorders. The focus of this study was to identify variables associated with attrition

from the REACH trial in Latino participants at the Palo Alto and Miami sites. Six variables (i.e.,

age, educational level, income, depression score, social support, and health status) believed to

predict attrition (i.e., drop-out status) of Latino dementia caregivers from the REACH trial were









examined. None of these variables were predictive of attrition for these two samples. Although

the outcomes of this study did not contribute new information concerning which socioeconomic,

demographic, and client factors are associated with the attrition of Latino dementia caregivers

from the REACH trial, these findings indicate some trends and thus contribute to the existing

literature on Latino dementia caregivers and serve to inform future research studies.

The understanding of the variables examined in this study are embedded in the cultural

context, thus, it is recommended that future inquiries into Latino dementia caregivers allow for

the measurement of cultural considerations. Due to the archival nature of this inquiry, the

hypotheses in this study were not powered sufficiently to test the relationship of the variables of

inquiry and data were not purposefully collected.









CHAPTER 1
INTTRODUCTION

Progressive memory disorders, such as Alzheimer's disease (AD), cardiovascular

dementias, and many other related disorders, are becoming a national epidemic. Progressive

memory disorders refer to a cluster of conditions that worsens over time, negatively impacts the

brain, and impairs the functioning of individuals. Currently, it is estimated that there are more

than 5 million people living with Alzheimer' s disease in the United States (Alzheimer' s

Association, 2007; USA Today, 2007) and this number is expected to more than double in the

coming years. In Florida, over 479,875 individuals are believed to suffer from Alzheimer' s

disease (Department of Elder Affairs, 2006). As Alzheimer's disease and related disorders

continue to affect increasing numbers of people, it is estimated that a larger proportion of those

will be culturally or ethnically diverse. Many of these individuals will be of Latino origin.

The terms Hispanics and Latino are frequently used interchangeably and refer to a

heterogeneous group that share Spanish as a common language. This study used the term Latino

rather than the term Hispanic, as Latino is an inclusive term that has been used by people of

Latin/Hispanic descent to empower themselves in political arenas, humanities, and literature

(Oboler, 1995; Shorris, 1992; Torres, 2004; Torres-Rivera, Wilbur, Roberts-Wilbur, & Phan,

1999).

Perhaps the biggest impact of Alzheimer' s disease (AD) is on the family and other

caregivers. Due to the debilitating nature of progressive memory disorders (e.g., AD),

individuals affected by these types of ailments require increasing assistance with daily living

skills and close supervision. The majority of individuals suffering from Alzheimer's disease

receive care at home. The Family Caregiving Alliance (2001) reports that 34 million adults, or

16 percent of the population, provide care to adults 50 years of age or older. This type of unpaid










caregivers is referred to as informal or family caregivers. Statistical data provided by the

Alzheimer' s Association (2005) indicate that seven out of ten people with Alzheimer' s disease

live at home in the care of family or friends. According to Schulz (2000), this translates into

over 5 million households in the U.S. providing care for a loved one with dementia.

A review of the caregiving literature shows that the terms caregivers and caregiving are

broadly defined and refer to providing a wide range of care and assistance. These range from

providing care with activities of daily living such as personal hygiene, toileting, eating, assisting

with other routines such as managing money or assets, to providing total care to individuals who

are not able to function independently (Gallagher-Thompson, Coon, Solano, Amber, Rabinowitz,

& Thompson, 2003; National Alliance for Caregiving & American Association of Retired

Persons, 2005; Family Caregiver Alliance, 2001; Shultz, 2000).

The demands associated with providing care to a person with chronic medical conditions

such as Alzheimer' s disease and related disorders may place a significant burden on caregivers.

Research findings show that due to the complex deficits of a person with dementia (i.e., declines

in intellectual functioning and self-care behaviors, personality changes, and problematic

behaviors), caregivers of persons with Alzheimer' s disease and related disorders experience

higher levels of stress and a number of additional negative outcomes such as poor health,

depression, and emotional strain (Argimon, Limon, Vila, & Cabezas, 2004; Family Caregiver

Alliance, 2001; Haley, Levine, Brown, & Bartolucci, 1987; Gallagher-Thompson, Coon, Solano,

Ambler, Rabinowitz, & Thompson, 2003; Gallagher-Thompson et al., 2003; Mitrani et al., 2006;

Schulz & Beach, 1999).

Scope of the Problem

The Alzheimer' s Association (2004) reports that Latinos have a greater risk of

developing Alzheimer's disease and related disorders. In Florida, the Department of Elder









Affairs estimates that there are currently over 34,000 Latinos diagnosed with Alzheimer's

disease (electronic communication, September 29, 2006). By 2050 the Alzheimer' s Association

(2004) estimates that over one million Latinos nationwide will be afflicted with this devastating

illness. Alzheimer's disease, the most common progressive memory disorder, is an irreversible

condition that adversely affects the intellectual functioning of the brain and eventually results in

maj or thinking and physical impairments as well as changes in personality. As the number of

persons affected with this devastating illness increases, so does the number of informal

caregivers helping a loved one with Alzheimer' s disease.

Two demographic trends contribute to the growing numbers of people with Alzheimer' s

disease in the Latino population. First, the natural chronological process of aging appears to be a

strong risk factor for Alzheimer' s disease. There has been a remarkable increase in life

expectancy over the last century. In addition, the number of older individuals is steadily rising.

By 201 1, the first wave of adults in the baby boomer generation (born between 1946 -1964) will

turn 65 years of age (Center for Health Communication Harvard School of Public Health &

MetLife Foundation, 2004). As the U.S. population ages, the Alzheimer' s Association (2004)

estimates that by age 65, one out of every ten persons will develop Alzheimer' s disease.

Additionally, by age 85, one out of every two persons will develop Alzheimer' s disease (Evan et

al., 1989). As the incidence of Alzheimer' s disease rapidly increases with the numbers expected

to more than double by the year 2050, upwards of sixteen million people will be affected

(Alzheimer's Association, 2004). If these predictions come true, Alzheimer's disease is likely to

become a devastating clinical and public health issue over the next 20 years.

The second change is the sudden increase of Latinos in the United States. The recent

growth of the Latino population in the United States is one of the most dramatic demographic









increases that has taken place in the history of the United States. No other racial or ethnic groups

have had as great an impact on the demography of the United States over the last century as

Latinos. The Pew Hispanic Center (2005) reports that "the 2000 Census marked the [Latino]

population at 35.3 million people, an increase of 58 percent over 1990 (p.2)." Moreover, they

predict that by the end of this decade, there will be over 47 million Latinos in the U. S. and over

60 million by 2020. Furthermore, it is estimated that by the year 2050, one out of every four

Americans will be of Latino heritage.

The combination of these two changes is significant because according to the 2000

census, Latinos are the fastest growing segment of the general population as well as the fastest

growing group of elders 65 years of age and older. Latinos make up almost ten percent of the

total baby boomer generation (Metropolitan Life Insurance Company, 2005). Projections

indicate that by the year 2050, there will be 12.5 million Latinos elders 65 and older in the U.S.

(Gallagher-Thompson, Solano, Coon, & Arean, 2003). These numbers are significant due to the

growing needs of elders and the higher risk of developing Alzheimer' s disease in the elder

population. Thus, the health care community will need to be able to meet the needs of this

growing segment of the population, especially the needs of caregivers, in order for them to

continue providing the vast maj ority of care for their elders suffering from Alzheimer' s disease.

Statement of the Problem

Due to the high incidence of Alzheimer' s disease in ethnic minority elders, the

prevalence of informal caregivers is on the rise. The Alzheimer' s Association (2004) reports that

Latinos are believed to be at a greater risk of developing progressive memory disorders due to a

higher incidence of cardiovascular disease, believed to be a significant risk factor for

cardiovascular dementia and Alzheimer's disease. Sociodemographic factors such as lower

education and limited access to health care, also are believed to be risk factors for Alzheimer' s









disease. As the numbers of Latino elders continue to climb, so does their need for care.

Although reliable estimates are difficult to find, current estimates by the Alzheimer's

Association (2004) indicate that close to 200,000 Latinos nationwide suffer from Alzheimer' s

disease and related disorders. These data illustrate the importance of continuing efforts to

advance knowledge about the sequelae of Alzheimer' s disease and underscore the need to

include culturally diverse groups in the investigation in order to develop effective methods to

meet their needs. However, a review of the literature shows that although minorities, specifically

Latinos, are the fastest growing segment of the population, they continue to be underrepresented

in research studies. Furthermore, the studies that have been successful at recruiting and retaining

Latino caregivers show mixed results.

Although there has been a boom in research studies on Alzheimer' s disease over the last

decades and efforts have been put forward to include culturally diverse populations, there

continues to be a gap in the literature on how this disease affects minority populations and their

families. One of the main problems associated with the paucity in research studies with minority

populations involves failures to recruit and retain older members of minorities into research

studies. Researchers maintain that this is particularly true for Latino caregivers due to the stigma

associated with mental illness and the time consuming and labor intensive burden placed on them

(Arean, Alvidrez, Nery, Estes, & Linkins, 2003; Gallagher-Thompson, Solano, Coon, & Arean,

2003). The steady increase of Latinos in the U. S. warrants the need to examine the Alzheimer' s

disease caregiving experience of this segment of the population as well as the barriers to

accessing services.

As the prevalence of Alzheimer' s disease and other progressive memory disorders grows

exponentially over the next decades, it will compromise the already strained healthcare system,









the long-term care system, and the economy. Due to the high incidence of Alzheimer' s disease

in the Latino community and the lower rates of health insurance coverage and healthcare-seeking

behaviors in this community, it is imperative for health care professionals to understand how

cultural values and beliefs affect the willingness of family caregivers to seek and maintain

connections for help in order to minimize the negative impact of Alzheimer' s disease on patients,

families, the healthcare system, and society at large. Though seeking help is an important

research issue, this study focused on the problem of barriers to retention, more specifically, the

barriers of retaining informal Latino Alzheimer' s disease caregivers in research studies.

Need for the Study

Research studies persistently show that while the numbers of Latino families affected by

Alzheimer' s disease continue to grow, researchers have not been able to recruit and retain

proportional numbers of Latino Alzheimer' s disease patients and caregivers in studies. By 2010,

it is expected that "ethnoculturally diverse persons will represent 25% of the dementia-affected

cohort and by 2050 they will number 33%, as compared to 16.7 % of the total dementia-affected

elderly population in 2000" (Valle & Lee, 2002, p. S64). If the predictions come true, this illness

is likely to become a devastating clinical and public health issue over the next decades.

Studies show that Latinos seek health services at a much lower rate compared to other

groups despite their high incidence of cardiovascular disease and related illnesses (Alzheimer' s

Association, 2004; Larkey, Hecht, Miller, & Alatorre, 2001; Liao, Tucker, & Giles, 2004).

Despite the rapid growth of the Latino population, public sector responses to the health service

needs of this population have developed slowly and do not commensurate with the growth,

needs, and grim circumstances of the population. In addition, the Kaiser Family Foundation

(2003) reported that Latinos have the lowest insurance rates when compared to other racial

groups in the U.S. The gap in health insurance coverage is due in part to the disadvantaged









economic position of ethnic groups with a greater toll on Latinos (Center for Studying Health

System Change, 2004). Research findings show that the Latino population in the U.S. is not

receiving appropriate health care necessary to reduce the risk or manage progressive memory

di orders. Professional groups, such as the Alzheimer' s Association (2004) warn that ignoring

the threat of dementia in the Latino community may lead to a health crisis for this segment of the

population in the United States. These discrepancies have maintained the vast racial and ethnic

disparities in Latino access to health care.

Therefore, due to the rapid increase in Latino elders and the alarming incidence of

progressive memory disorders (e.g., Alzheimer' s disease), it is imperative that health care

professionals identify barriers to recruitment and retention of Latinos into research studies and

subsequent treatment protocols necessary for improving the delivery of healthcare services to the

Latino population and thus, bridge the gap in health disparities. Unfortunately, research studies

tend to place an inordinate burden on already taxed research participants. Therefore, it is

necessary to advance the current knowledge about how the burden of measurement and barriers

to participation contribute to attrition from existing research studies. Also, it is vital to revise

current procedures into culturally sensitive and appropriate new research and applied practices.

Purpose of the Study

Due to the changing demographics of the U. S., especially the fast growing rate of Latinos

and the silver tsunami (Fryling, 2006) that is starting to take place, it is imperative to explore

these changes and the implications to the health care system and society at large in order to

advance our understanding about barriers to providing care in the Latino community. The need

for scientific examination about the obstacles in recruiting and retaining minority participants

into research studies has been well documented (Arean, Alvidrez, Nery, Estes, & Linkins, 2003;

Gallagher-Thompson et al., 2003; Warren-Findlow, Prohaska, & Freedman, 2003). Thus, it is










important to identify the factors that predict the attrition rate from research studies in order to

gain a better understanding about how these influence the attitudes of Latino Alzheimer' s disease

caregivers on accessing and utilizing services. This knowledge is needed in order to bridge the

gap in health and health care disparities among Latino Alzheimer' s disease caregivers.

This study examined the barriers of retention, more specifically retaining Latino

Alzheimer' s disease caregivers in research studies. Data collected for the Miami, FL, and Palo

Alto, CA, sites of the Resources for Enhancing Alzheimer' s Caregiver Health, also known as the

REACH I study, were analyzed to examine and contrast attrition rates among two different sets

of Latino Alzheimer' s disease caregivers. This researcher was interested in identifying the

factors that may contribute to Latino Alzheimer' s disease caregivers prematurely dropping out of

research studies that may yield subsequent treatment approaches to improve the well-being of

both the Alzheimer' s disease caregiver and the care recipient.

Theoretical Framework

This study used a combination of two stress process models to guide its inquiry. First, it

borrowed the stress process model used in the Resources for Enhancing Alzheimer's Caregiver

Health (REACH) study. The REACH study (Czaja, Schulz, Lee, & Belle, 2003) adapted a stress

process model of caregiving to include environmental stressors to health outcomes in dementia

caregiving. According to the investigators, this model establishes that the functional limitations

and behavioral problems of the care recipient in conjunction with social and environmental

stressors make up the primary stressors for dementia caregivers. As caregivers engage in their

caregiving responsibilities, they perceive and identify stressors associated with caregiving. Then

they evaluate these demands, as well as their ability and resources to meet these demands. If the

caregivers feel helpless and unable to cope with these demands, they experience stress. "The

appraisal of stress is assumed to contribute to negative emotional, psychological, and behavioral










responses that put the individual at risk for physical and psychiatric disease"(Czaj a et al., 2003,

p. 386).

The second relevant stress process model is the one advanced by Haley, Levine, Brown,

and Bartolucci (1987) that proposed a model adapted from the stress response model to explain

caregivers' help-seeking behaviors. According to the authors, caregivers' appraisal of stressors,

personal coping responses, and social support, mediate between stress and caregiver outcome.

"Use of coping mechanisms such as seeking information, using problem solving, and emotional

release may also help the caregiver to manage his or her relative more effectively, and lessen

emotional distress" (Haley et al., 1987, p. 323). If caregivers have access to and use available

resources, they may increase the probability of obtaining better outcomes in coping with the

chronic stress of caregiving. These stress and coping models are useful in the conceptualization

of the process of dementia caregivers' participation in research studies and their utilization of

services. Once stressors and the caregivers' assessment of these are identified, they may be

changed or modified (Schulz et al., 2003).

Overview of the Resources for Enhancing Alzheimer's Caregiver Health Study

The Resources for Enhancing Alzheimer' s Caregivers Health (REACH) was a 5-year,

multi-site research proj ect conducted with support from the National Institutes on Aging and the

National Institute of Nursing Research. Six sites, including Birmingham, Boston, Memphis,

Miami, Palo Alto, and Philadelphia, developed and tested the effectiveness of multiple

interventions for family caregivers of persons with Alzheimer' s disease and related disorders.

Each site investigated a different intervention or a combination of interventions. The

interventions included (1) individual information and support strategies, (2) group support and

family systems therapy, (3) psychoeducational and skills-based training approaches, (4) home-

based environmental interventions, and (5) enhanced technology support systems. "REACH was










designed to examine the feasibility and outcomes of multiple different intervention approaches, .

(Schulz et al., 2003). A total of 1,222 participants were recruited with a strong emphasis on

recruiting minority caregivers. For the purpose of this study, only data collected from the Miami

and Palo Alto sites were examined.

Hypotheses

This study was a secondary analysis of a multi-site prospective randomized controlled

trial in which data collected from the REACH project were used to examine predictors of

attrition of Latino Alzheimer' s disease (AD) caregivers from research studies. Data used for this

examination included (1) caregiver' s age, (2) caregiver' s level of education, (3) income, (4)

depression, (5) social support, (6), caregiver health, and (7) attrition rate.

The following six null hypotheses were tested in this study:

Hol: There is no relationship between attrition drop-out status (attrition) and Latino
caregivers' age.

Ho2: There is no relationship between drop-out status (attrition) and Latino caregivers'
achieved level of education.

Ho3: There is no relationship between drop-out status (attrition) and Latino caregivers'
income level.

Ho4: There is no relationship between drop-out status (attrition) and depression among
Latino caregivers as measured by Center for Epidemiologic Studies Depression Scale
(CES-D).

Ho5: There is no relationship between drop-out status (attrition) and social support
among Latino caregivers.

Ho6: There is no relationship between drop-out status (attrition) and Latino caregivers'
health status as measured by the Caregiver Health and Health Behaviors Questionnaire
from the REACH study.









Definition of Terms


Alzheimer's disease. A progressive brain disorder that impairs a person's intellectual

functioning, that is, memory and the person's ability to learn, reason, and communicate. It also

impairs the person's judgment, behaviors, and eventually, the ability to care for self (Alzheimer's

Association, 2006).

Dementia. An acquired set of symptoms that negatively affect a person's intellectual

functioning such as thinking, memory and judgment (Alzheimer' s Caregiver Support Online,

2006).

Progressive memory disorders. Refers to disorders of the brain, such as Alzheimer' s disease

(AD), that cause worsening of a person's intellectual abilities over time (Alzheimer' s Caregiver

Support Online, 2006).

Reach study. A 5 year study conducted in six cities across the U. S to measure the effectiveness

of different dementia caregiver interventions (Schulz et al., 2003).

Silver tsunami. A term used by many to describe the tidal wave of baby boomers that is about

to have a significant societal impact (Fryling, 2006).

Overview of the Remainder of Study

The remainder of this study consists of four chapters. Chapter 2 provides a review of the related

literature. Chapter 3 contains a description of the specific procedures for the study, including

methodology, subjects, and research design. In Chapter 4, the results of the study are presented.

Chapter 5 includes a discussion of the results, conclusions, implications, limitations, and

recommendations for future research.










CHAPTER 2
REVIEW OF RELEVANT LITERATURE

The purpose of this chapter is to review the published literature pertinent to the proposed

investigation. This chapter is organized into seven sections: (a) attrition rates and Latinos, (b)

Latino caregivers' age,(c) Latino caregivers' level of education, (d) Latino caregivers' income,

(e) Latino caregivers' depression, (f) Latino caregivers' social support, (g) social support of

Latino Alzheimer' s disease caregivers and (h) Latino caregivers' health status.

As the U.S population grows older and more diverse, Latinos are becoming one the

fastest growing segments of the population. According to the 2000 U.S. Census, (U.S. Census

Bureau, 2001) the Latino population has increased dramatically over the last decade, becoming

the largest minority group in the country. Census estimates indicate there are over 35 million

Latinos living in the U.S. Of those, the U. S. Administration on Aging (2004) reports 4% are

Latino elders 65 years of age and older and proj ects that by 2020 this percentage will increase to

over 14%. By 2025, one in six elders living in the U.S. are expected to be Latinos (Horizon

Project, 2000) and by the year 2050, Latinos 65 years of age and older will reach the 12.5 million

mark (Gallagher-Thompson et al., 2003). Current estimates indicate that 44% of Latino elders

seventy and older receive care at home by a family member or friend (Weiss, Gonzalez, Kabeto,

& Landa, 2005).

The graying and diversification of the U. S. demographics are noteworthy due to high

incidence of chronic conditions, such as Alzheimer' s disease, in this segment of the population

(Weiss, Gonzalez, Kabeto, & Langa, 2005). The Alzheimer's Association's (2004) projects an

increase from 200,000 Latinos with Alzheimer' s disease to over 13 million by 2050. There is a

common misconception that most elders who require long-term care receive assistance from

formal caregiving facilities, such as nursing homes; however, estimates indicate otherwise. Over









75 % of long-term care for disabled adults takes place at the home of a family member or friend

(Family Caregiving Alliance, 2001; Gallagher-Thompson et al., 2003). Currently, it is estimated

that 44 % of Latinos 70 years of age and older receive informal care at home (Weiss et al., 2005).

As the Latino population continues to expand and grow older, so will the number of informal

Latino caregivers.

Alzheimer' s disease is a devastating illness that adversely impacts the lives of patients

and family caregivers alike. Alzheimer's disease caregivers assume challenging responsibilities,

for example, assisting the patient with activities of daily living such as grooming, toileting, and

bathing. Often times, caregivers also engage in activities that involve role reversals such as

managing the finances for a parent, preparing a meal for a wife, or driving a husband to medical

appointments. In addition, caregiving responsibilities are time consuming. Thus, often

caregivers will neglect their own needs (Miller & Mukherjee, 1999; Cox & Monk, 1996), health

(Gallagher-Thompson, 2003; Pinquart & Sorensen, 2005), and social relationships (Cox &

Monk, 1996). Curry and Jackson (2003) report that compared to White caregivers, Latino

caregivers are more likely to care for a family member instead of caring for a friend. The

implications of providing care for someone with a chronic condition are significant. Research

findings suggest that compared to other racial groups, Latino caregivers provide more hours of

informal caregiving (Weiss et al., 2005) and care for longer periods of time (Pinquart &

Sorensen, 2005).

Consistent research findings suggest that Latino families have higher rates of caregiving

at home for persons with Alzheimer' s disease than any other groups (Covinsky, Newcomer, Fox,

Wood, Sands, Dane, & Yaffe, 2003; Markides, Rudkin, Angel, & Espino, 1997), despite higher

levels of physical and psychological distress associated with the physical and cognitive









limitations of the patient. These numbers are significant because providing care for a person

with Alzheimer' s disease has been associated with negative caregiver outcomes such as poor

health, higher levels of depression, and decreased support (Gallagher-Thompson, 2003), in

addition to increased mortality (Schulz, O' Brien, Bookwala, & Fleissner, 1995; Schulz,

Visintainer, & Williamson, 1990).

The existing literature on how Alzheimer' s disease impacts the lives of culturally diverse

persons and their participation in research studies is scarce compared to the available literature

on Whites. Inclusion of psychosocial and cultural variables such as age, socio-economic status,

educational attainment, health-status, and social support are crucial to understand the impact of

Alzheimer's disease on the Latino population. When looking at these psychosocial and cultural

variables, two important considerations need to be noted: first, differences among groups and

second, degree of acculturation.

First, it is important to be cognizant of the differences among Latino groups in order to

develop appropriate studies and interventions. Although these groups share many psychosocial

characteristics, differences among groups and their needs have been documented (Briones,

Ramirez, Guerrero, & Ledger, 2002; Gallagher-Thompson, Solano, Coon & Arean, 2003; Valle

& Lee, 2002). Acculturation to mainstream society also has significant implications for families

caring for relatives with progressive memory disorders (Gallagher-Thompson et al., 2003; Valle

& Lee, 2002). While these differences warrant careful examination, well-designed research

studies including Alzheimer's disease patients and their families across these cultures will add to

the knowledge base regarding the disease and its implications for Latino groups.

Attrition Rates and Latinos

It is well documented that members of diverse groups are underrepresented in research

studies (Arean, Alvidrez, Rowena, Estes, & Linkins, 2003; Arean & Gallagher-Thompson, 1996;










Curry & Jackson, 2003; Gallagher-Thompson, Solano, Coon, & Arean, 2003). The significance

of this issue was underscored in 1994 when the National Institutes of Health (NIH) issued

guidelines for the inclusion of minority participants in clinical research trials (Curry & Jackson,

2003). Although there has been a boom in research studies on the impact of Alzheimer' s disease

over the last decades, and efforts have expanded to include culturally diverse populations, there

continues to be a gap in the literature on how this disease affects minority populations and their

families. At the forefront of this issue is the failure to recruit and retain Latino caregivers into

research studies. The inability to recruit and retain Latino caregivers in research studies has a

detrimental effect on the ability of health care professionals to develop informed, culturally

relevant programs to assist and alleviate the burden associated with caregiving. Furthermore,

current Eindings that lack adequate numbers of ethnic participants cannot be generalized to the

population without questioning the validity of the researchers' conclusions (Levkoff & Sanchez,

2003). Few studies report successful implementation of recruitment and retention methods with

older populations (Arean, Alvidrez, Rowena, Estes, & Linkins, 2003; Gallagher-Thompson,

Solano, Coon, & Arean, 2003).

Arean and associates (Arean et al., 2003) examined recruitment and retention rates in two

separate studies on older minority populations in order to compare the traditional recruitment and

retention methods with a consumer-centered approach. They found that the mainstream older

research participants tend to respond positively to traditional recruitment and retention methods

that focus on eliminating practical barriers to research studies such as offering transportation to

and from the research site as well as Einancial incentives. Ethnic older adults responded more

favorably to the consumer-center recruitment model, which also resulted in greater retention

rates. This model proposes that in order to successfully attract the participation of minority










elders into research studies, the following conditions must be met: securing the support and buy-

in from local community leaders; enlisting research staff that either knows or belongs to the

same ethnic group as the target population; anticipating and alleviating participants' burden

associated with the research study; and providing feedback to the community once the study is

completed.

In a recent study by Gallagher-Thompson and colleagues (2004) aimed at comparing the

effectiveness of three different recruitment methods among White and Latino female caregivers,

results were consistent with the prior conclusions that favored consumer-centered methods over

traditional approaches. The authors believe that the Latinos' cultural values of trust (conflanza)

and close relationships (personalismo) demand the development and establishment of consumer-

centered recruitment and retention methods derived from partnerships with local Latino-specific

organizations.

Others such as Henderson, Gutierrez-Myka, Garcia, and Boyd (1993) have put forward

similar efforts and developed a program to recruit minority caregivers to participate in existing

support groups for Alzheimer' s caregivers who had no minority representation. This effort

consisted of the following: training group leaders about culturally diverse populations;

conducting purposeful advertisement and promotion using the media, presentations, and

endorsement from community professionals and organizations; and using appropriate meeting

locations. The authors indicated that 24 months later, attendance by African Americans and

Hispanics rose dramatically. In addition, a meta-analysis on the effectiveness of interventions

with caregivers has some significant revelations. According to Sorensen, Pinquart, and

Duberstein (2002) the average drop-out rate across the 78 studies they reviewed was less than










20%, and most caregivers had at least a high school education, and were ethnic spouses or adult

children.

Many explanations have been offered regarding the dismal participation of Latinos in

research studies including mistrust of ethnic groups toward the scientific community, language

barriers, and the burden placed on research participants (Arean et al., 2003; Gallagher-Thompson

et al., 2003; Levkoff & Sanchez, 2003). However, caregiver-specific factors such as age,

educational attainment, income level, depression, social support, and health status variables are

also believed to influence the limited participation, more specifically the drop-out rate of Latinos

in research studies. Currently, there is little information about how these psychosocial factors

(i.e., caregiver age, achieved educational level, health status, level of depression, and social

support) affect Latino caregivers' participation in long-term research studies (Covinsky,

Newcomer, Fox, Wood, Sands, Dane, & Yaffe, 2003 ). The results of studies with Latinos from

different countries of origin need careful consideration due to the heterogeneity among Latino

participants.

Latino Caregivers' Age

It is estimated that, by the year 2050, the number of Latino elderly could reach twelve

million persons, representing 15% of all Latinos living in the U.S. The age demographics of

dementia caregivers seem to be in flux. A study by Schulz and Beach (1999) found that most

caregivers are middle-aged adult children and spouses caring for a parent or spouse. In 2001, the

Family Caregiver Alliance reported that the average age for caregivers of persons 50 years of age

and older was 47 and the average age of caregivers of persons 65 and older was 63 years old.

More recently, an Alzheimer's Association (2004) publication about Alzheimer' s caregiving in

the U.S indicates that the average age for an Alzheimer's disease caregiver is 48 years old and

that most caregivers provide care for an older relative.









A comprehensive inquiry into caregiving in the U.S. conducted by the National

Caregiving Alliance and the American Association of Retired Persons (2005) estimates that over

44 million caregivers 18 years of age and older in the U.S. provide unpaid care to an adult family

member or friend who is also 18 or older. Of these, 8% of all caregivers reported caring for

someone with Alzheimer' s disease; 16% of caregivers are 50 years and older and an estimated

12% are Latinos.

A paucity of research studies look at the relationship between age and the Latino

caregivers' attrition rate from such studies. The limited available published studies show mixed

results concerning the association between age and attrition rates in Latino research participants.

Therefore, the present review of the literature examined the effects of age on attrition rates in

other areas and other segments of the population. This approach identified several studies across

racial and ethnic groups reporting an inverse relationship between age and drop-out rate (Honas,

Early, Frederickson, & O'Brien, 2003; Warren-Findlow, Prohaska, & Freedman, 2003). A study

on recruitment and retention of HIV participants found that younger Latino males were less

likely to follow through with the research study (Villarruel, Jemmott, Jemmott, & Eakin, 2006).

The investigators believed that conflict with j ob schedules and residency status mediated their

results. Research studies in other areas consistently reported low participation of elder

individuals in research studies, (Mirowsky & Reynolds, 2000; Murthy, Krumholz, & Gross,

2004; Cotter, Burke, Loeber, & Mutchka, 2005).

It appears that the ability of older caregivers to participate in research studies is

influenced by factors such as knowledge about progressive memory disorders such as

Alzheimer' s disease, health status, perceived burden associated with research involvement. Also

influential are practical issues such as transportation, taking time off from work, and the









availability of another person to tend to the care recipient. These inferences are drawn from the

existing literature on age and participation in research studies.

Latino Caregivers' Level of Education

The U.S. Administration on Aging maintains that when compared to other groups, Latino

adults have less formal education (2005). The study on Caregiving in the U.S. conducted by the

National Caregiving Alliance and the American Association of Retired Persons (2005) shows

that 29% of all caregivers have a high school education, 30% have some college education, and

35% are college graduates. A report by Schulz (2000) suggests that the median education of

dementia caregivers is some college, the same as for the non-dementia caregivers. In 2004, the

Alzheimer' s Association published a report indicating that almost 40% of dementia caregivers

were college graduate; 25% had some college education, and 37% had a high school diploma or

less schooling.

Due to the lack of studies aimed at measuring the relationship between educational level

and Latino caregivers' participation in research studies, some researchers infer that Latino

caregivers have lower educational attainment when compared to other groups. The basis for this

inference appears to be the poor understanding of Latinos about Alzheimer' s disease and other

progressive memory disorders (Coon, Gallagher-Thompson, & Thompson, 2003; Schulz, 2000).

However, making this assumption is problematic because the poor understanding may reflect

their language differences and cultural beliefs about health and illness.

An examination of the literature on other ethnic caregivers found mixed results regarding

recruitment and retention of African American into research studies. A study on recruitment and

retention of underrepresented populations, specifically African Americans, into health promotion

research, found that persons under 60 were less likely to enroll in their study (Warren-Findlow,

Prohaska, & Freedman, 2003). The authors believed that the higher drop-out rate of younger










participants was related to lower educational levels, whereas other publications indicate that

African Americans are highly suspicious of research due to a long history unethical medical and

research protocols (Blakley & Harrington, 1997; Doty, 2007; Jonas, 1993; Ludmerer, 1999).

Although this study did not include Latino participants, research Eindings like this may impact

future researchers investigating the relationship between attrition rates and Latino caregivers.

Latino Caregivers' Income Level

Lower educational level is typically associated with lower income. The Horizon Proj ect

(2000) reports that Latino elders are less well-educated than and more likely to live in poverty.

The proj ect estimates that 21% of Latino elders in the U. S. live at or below the federal poverty

level compared to Whites. This study also reports that when compared to Whites, Latino elders

are less likely to receive social security income. These Eindings are alarming considering the

Financial burden on caregivers who already struggle with a taxing financial situation.

The cost to government agencies for the care of patients who have Alzheimer's disease is

substantial at present and it is rising rapidly (Prigerson, 2003). However, most of the cost is

absorbed by the unpaid family caregiver. The unpaid portion accounts for a large proportion of

the costs of treating the patient. Family caregivers spare the health care system billions of dollars

annually, but depression, stress, and weakened physical health in these caregivers, if left

untreated, come not only at a great personal cost to the caregivers and their families but

potentially at a substantial Einancial cost to society. The negative impact on the health of

dementia caregivers and the restricted ability to be away from the care recipient negatively

impacts the caregiver's limited resources as well as their earning potential. Specific to Latino

elders, their lower educational accomplishments, English proficiency, and work experience

negatively influence their financial attainment. Limited financial resources typically translate









into inability to pay for companionship services, professional assistance, and even transportation.

These restrictions may in turn hinder the caregivers' ability to participate in research studies.

Depression in Latino Caregivers

Depression in later life has been identified as one of the most common mental health

problems affecting elders (Abu-Rayya, 2006; Blazer, 2003). Bourgeois, Schulz, and Burgios

(1996) highlighted the risks associated with the negative consequences of depression in caregiver

health when they reported that caregiver depression is correlated to decreased function and

quality of life, and even death among caregivers. Maj erovitz (2001) reports that caregivers who

perceive the availability of social support report lower levels of depression associated with

caregiving stress. Although research findings report mixed results about depression in Latino

caregivers (Janevic & Connell, 2001), most research findings suggest that the prevalence of

depression in Latino caregivers is equal or greater than found in other caregiver groups. (Aranda,

2001; Schulz, 2000).

The combination of poor health and limited financial resources also has been linked to

higher rates of depression (Covinsky, Newcomer, Fox, Wood, Sands, Dane, & Yaffe, 2003).

Since Latino caregivers tend to have lower financial resources and compromised health, it is not

surprising that they are also believed to have higher rates of depression. Weiss, Gonzalez,

Kabeto, and Langa (2005) found that older Latinos have a significant higher risk of depression

than White or African American caregivers. Coon, Gallagher-Thompson, and Thompson (2003)

maintain that the relationship between high levels of depression and adherence to traditional

cultural norms is linked to the caregivers' feelings of helplessness and difficulty managing

problematic behaviors.

In an examination of patient and caregiver characteristics associated with caregiver

depression, Covinsky and colleagues (2003) found that Latino caregivers reported significant










higher levels of depression when compared to Whites or Blacks. Despite higher levels of

depression, this study found that Latinos were more reluctant to place their loved one in a facility

compared to caregivers from other groups. Placement of a loved one with Alzheimer' s disease

appears to be more consistent with White conventions rather than Latino values such as the

importance of family and mutual assistance known as familismo despite the toll on family

caregivers (Schulz, 2000).

A study measuring the relationship between racial or ethnic concordance between

caregiver and researcher on caregivers' attrition, depression, and burden found no significant

difference in caregiver drop-out at the 12-month follow-up (McGinnis, Schilz, Stone, Klinger, &

Mercurio, 2006) regardless of concordance of race or ethnicity. No additional studies examining

the relationship between Latino caregivers, depression, and participation in research studies were

available at the time of this review. However, due to the demands placed on research

participants, it may be inferred that depressed caregivers are less likely to follow through with

research protocols if they feel the involvement increases the amount of burden they already

experience (Coon, Rubert, Solano, Mausbach, Kraemer, Arguelles, Haley, Thompson, &

Gallagher-Thompson, 2003; Eisdorfer, Czaja, Loewenstein, Rubert, Arquelles, Mitrani, &

Szapocznick, 2003; Miranda, Azocar, Organista, Mufioz, & Lieberman, 1996).

Social Support and Latino Caregivers

Responsibilities associated with caring for someone with a chronic illness may disrupt

caregivers' social networks. Caregivers usually obtain social support from two main sources, 1)

informal networks such as family and friends and 2) formal services from community

organizations. Both sources of social support may alleviate the strain associated with caregiving.

While informal support is believed to lessen the negative consequences of caregiving (Miller,

Townsend, Carpenter, Montgomery, Stull, & Young, 2001) there are mixed results on the









relationship between social support and levels of distress among caregivers. Cohen, Boyle,

Coakley, and Lawlor (2002) found that lack of social support in conjunction with limited

financial resources have a negative impact on caregivers. Caregivers from ethnic groups tend to

have lower financial resources and health insurance coverage to meet their needs and thus seek

formal supports such as day care centers and other programs. Keller, Gonzalez, and Fleuriet

(2005) studied issues of recruitment and retention of ethnic women in an exercise program and

found that participants who lacked social support did not follow through with the program.

The literature on Latino caregivers of persons with Alzheimer' s disease show that Latino

caregivers may be reluctant to seek support for several reasons including limited knowledge

about Alzheimer' s disease, the stigma associated with mental illness, feelings of shame, and the

Latino cultural value of family pride and mutual assistance or familismo (Coon, Gallagher-

Thompson, & Thompson, 2003; Schulz, 2000). Moreover, due to the strong sense of familial

obligation in the Latino community, the responsibility of caring for a loved one with progressive

memory problems rests primarily in the family, particularly on a female member (Coon et al.,

2003). Reliance on help outside the extended family is typically viewed in a negative light. Other

factors such as low educational attainment and employment may also contribute to Latino

caregivers' feelings of isolation and limited social support (Dilworth-Anderon & Gibson, 2002).

Informal social support is oftentimes possible because strong family connections are

maintained by the close proximity in which Latinos tend to live in regards to their relatives. It is

also possible due to their reliance on extended family connections which include members of the

extended family, friends, neighbors, and members of their religious groups (Sue & Sue, 1990).

Latinos' strong emphasis on the extended family and their tendency to live in close proximity to

relatives provide caregivers with the opportunity for to exchange aid. However, some studies









have found that the close connections typical in the Latino culture may also have negative

effects. Lefley (1996) points out that the large social networks of Latinos may become a source

of stress in the event of unmet expectations, poor advice, criticism, as well as the strong

emphasis on protecting the family's honor. Semple (1992) reports that interpersonal stress with

social networks has negative effects on the well being of caregivers. Furthermore, in a study on

the caregiving experience of Mexican-Americans, Phillips, Torres de ArdC~n, Komnenich,

Killeen and Rusinak (2000) found that despite the availability of an extended family system,

Latinos perceived that they had limited access to informal support and, therefore, used less

support than Whites. The relationship between the availability of social support and Latino

caregivers' participation in research studies has not been documented. However, due to the

cultural values of close relationships and privacy, it may be inferred that Latino caregivers with

strong social support would not have high rates of participation in research studies.

Latino Caregivers' Health Status

Not only has the drastic growth of the elderly population over the last few decades placed

exceptional demands in caregivers and other family members but also on the whole healthcare

system. Many researchers have discussed negative physical, mental, and emotional

consequences of caregiving (Connell, Janevic, & Gallant, 2001). Others have addressed the

impact of caregivers' inability to cope and the feelings of distress on their overall health

(Gallagher-Thompson et al., 2000). Pillemer and Suitor (2002) maintain that the negative

outcomes associated with caregiving are heightened for caregivers of persons with Alzheimer' s

disease and that poor social support exacerbates these problems. The health of caregivers is

influenced by a myriad of factors including age, chronic conditions, physical problems, as well

as psychological factors such as emotional well-being and stress. In a report by the U.S.

Administration on Aging (2005), the leading disease-related causes of death in the Latino









community included heart disease, cardiovascular disease, and diabetes, among others, all

believed to be risk factors for developing Alzheimer' s disease or other related dementias. A

study by Corbeil, Quayhagen, and Quayhagen (1999) found that caregivers tend to have negative

appraisals of their stress levels compared to non-caregivers. Other findings suggest that many

caregivers report having poor health (Navaie-Waliser, Feldman, Gould, Levine, Kuerbis, &

Donelan, 2002) such as increased rates of high blood pressure and high insulin levels

(Cannuscio, Jones, Kawachi, Colditz, Berkman, & Rimm, 2002) and high rates of cardiovascular

disease (Lee, Colditz, Berkman, & Kawachi, 2003). The risk associated with chronic disease

increases sharply with age (Manton, 1996).

Researchers such as Fox, Newcomer, Yordi, and Arnsberger (2000) maintain that the

caregiving responsibilities for a person with Alzheimer' s disease can contribute to the decline of

the caregivers' health. In addition, Schulz and Beach (1999) believe that the negative impact of

the combined chronic stress, sense of loss, physical demands, and the biological vulnerability of

older caregivers put them at a higher risk of death.

Hinton' s (2002) emphasis on the cultural nature of the caregiving experience stressed that

cultural factors impact the caregivers' perception of the illness, burden of care, ability to cope

with caregiving stress, and ability to care properly for themselves. Caregiver burden includes

physical, mental, emotional, and financial costs of providing care and close supervision to

someone and is accompanied by a negative appraisal of one' s health.

Several factors contribute to the poor health status of Latino caregivers such as the higher

prevalence of diabetes (U. S. Department of Health & Human Services, 2005) and cardiovascular

disease (Urizar & Sears, 2006). Moreover, the lower socioeconomic status of Latinos has been

correlated with poor health. Dilworth-Anderson and Gibson (1999) maintain that the Latinos'










disadvantaged socioeconomic status places them at higher risks for health conditions such as

cardiovascular disease and diabetes, which in turn are risk factors for developing cardiovascular

dementia, perhaps in combination with Alzheimer' s disease. Weiss, Gonzalez, Kabeto, and

Langa (2005) found that older Latinos have a high burden of some diseases and poor self-

perceived health and that they are at a higher risk for poor health outcomes due to higher rates of

poverty and lower rates of health insurance coverage. When compared to the white population,

Latinos are believed to suffer higher rates of morbidity and mortality. Compounding these

health disparities are the cultural and linguistic barriers that racial/ethnic minorities encounter

when seeking health care. Lack of communication and ease can occur between patient and

provider when cultural differences in perceptions of illness, disease, and medical roles are not

recognized and addressed, thereby adversely affecting health outcomes. Other barriers include

immigration status, discrimination, lack of access to high quality educational opportunities,

unavailability of transportation and childcare, and inconvenient and insufficient hours of

operation at health facilities. In addition to the existing unfavorable health status in the Latino

community, caregiving responsibilities often introduce adverse consequences to caregivers'

health.

Summary and Conclusions

Little information exists in the literature about how the person-based variables of age,

educational level, income, level of depression, amount of social support, and health status

influence the participation and drop-out rates from research studies of Latino caregivers of

persons with Alzheimer' s disease. Available literature on drop-out rates for members of diverse

groups, more specifically Latino research participants, is scare and suggests mixed results. It

appears that due to the changing demographics, many caregivers are adult children caring for a

parent. The existing demands from work and nuclear families on adult children caregivers may










prevent them from participating in research studies. Moreover, the relationship between health

status and Latino caregivers' participation in research studies does not appear to have been

examined. However, it appears that depressed caregivers might not participate or complete their

participation in research studies if they feel that doing so exacerbates their burden. Finally, since

Latinos are believed to rely on support from within their networks, it is reasonable to assume that

those caregivers with adequate social support will be less likely to participate in studies and,

thus, keep their personal affairs within the family unit.

Due to the numerous unanswered questions regarding barriers to research participation in

the Latino community, studies examining these barriers are warranted. The existing body of

knowledge on predictors of attrition for Latinos occurs mostly in the areas of substance abuse

and health promotion research, leaving a sea of opportunities for researchers.










CHAPTER 3
METHOD S

This study was an archival analysis of the Resources for Enhancing Alzheimer' s

Caregiver Health (REACH) project; a multi-site prospective randomized controlled trial. The

purpose of this investigation was to examine variables identified in the published literature as

relevant to the prediction of drop-out status (i.e., attrition) in a sample of Latino Alzheimer' s

disease caregivers.

Resources for Enhancing Alzheimer's Caregiver Health

REACH participants were recruited from memory disorder clinics, primary care clinics,

community organizations and physicians' offices. In order to recruit caregivers from diverse

populations, REACH investigators used local community media to target these groups. Criteria

for inclusion required potential participants to be over the age of 21 and living with and

providing care for a relative with Alzheimer' s disease or other memory progressive disorder for

at least four hours a day and for a minimum of six months. A total of 1,222 participants were

recruited and randomly assigned to an intervention condition. This study only examined the

Latino caregivers in the Miami and Palo Alto sites. Assessment instruments, as described by

Wisniewski, Belle, Marcus, Burgio, Coon, Ory, Burns, & Schulz (2003) were translated into

Spanish for these two samples.

The Miami site (as described by Wisniewski et al., 2003) examined the effectiveness of

family-based structural multisystem in-home intervention (FSMII) and a family-based structural

multisystem in-home intervention plus a computer telephone integration system (FSMII +

CTIS). The FSMII was based on a family systems approach, aimed at improving communication

between caregivers and other members of the family by identifying problems in communication









and facilitating interactions. The FSMII + CTIS was an enhanced version of the FSMII with the

additional use of screen phones as a tool to increase interactions among family members.

The Palo Alto site (as described by Wisniewski et al., 2003) examined the effectiveness

of the coping with a caregiving class (CWC) and an enhanced support group (ESG). The CWC

was a psychoeducational class based on cognitive behavioral theories that demonstrated the

consequences of negative thoughts and behaviors in caregivers' mood. The ESG was a

professionally led support group in which caregivers were encouraged to attend regular meetings

and commit to participate for one year. Participants in this site were randomized into the

intervention group based on language preference and availability.

In addition to two active treatments, each site also included a control condition. Both

control conditions consisted of a minimal support condition in which the participants regularly

received written educational material and check-in phone calls (Wisniewski et al., 2003). The

remainder of this chapter addresses the following topics: 1) sample description and

characteristics, 2) variables, 3) null hypotheses, 4) research procedures, and 5) data analysis, and

6) methodological limitations.

Sample Characteristics

The sample for this investigation consisted ofN= 223 Latino Alzheimer's disease

caregivers in the United States. The sample was obtained from the Latino caregivers who

participated in two of the six sites in the REACH study: Miami, FL and Palo Alto, CA. The

Miami site consisted of n =1 13 Latino caregivers whereas the Palo Alto site was comprised of

n = 110 Latino caregivers. Participants were randomly assigned to either the treatment group or

control measure.

Palo Alto Sample. All 110 participants in the Palo Alto sample were females with a mean age

of 51.14 years (SD=-12.91). Annual income was measured categorically and ranged from a low









of less than $5,000 to a high of $70,000 or more; the modal income was $15,000 to 19,999. The

average level of education achieved was 10.49 years (SD = 3.99). Care givers in this sample had

a mean of 35.62 years (SD = 17.27) living in the U. S. and a mean of 17.44 years (SD = 18.28)

living with the care recipient. The mean depression CES-D score was 17.60 points (SD=-12.60).

Miami Sample. In the Miami sample (n = 113), 86 (76%) participants were females and 27

(24%) were males. The mean age for this sample was 66.58 years (SD = 10.54). Caregiver

annual income measured categorically, ranged from a low of less than $5,000 to a high of

$70,000 or more; the modal income was $5,000 to $9,999. The average level of achieved

education was 1 1.67 years (SD = 3.79). The mean number of years that caregivers had lived in

the U. S. was 3 1.36 years (SD = 8.83) and the mean number of years caregivers had lived with a

care recipient was 35.0 years (SD = 19.76). Care givers in the Miami site had a mean CES-D

score of 20. 12 points (SD = 1 1.09) which falls above the "at risk for depression" cutoff score of

16.

Variables of Interest and Operational Definitions

The predictor variables in the study were: age, educational achievement, income, degree

of depressive symptomatology, social support, and health status. The criterion variable was drop

out status (i.e., attrition). In the sections that follow, each construct is described along with its

operational definition (i.e., measure) according to the REACH study.

Age. Refers to the time elapsed from date of birth to the time of the REACH study. This

measure was acquired from the Caregiver Sociodemographic Form from the REACH proj ect that

was used to obtain descriptive information about the caregivers. Age was measured in years.

This study set out to elucidate the relationship between the age of caregivers and attrition.

Depressive symptoms. This variable was operationally defined as the affective state of

the caregiver at the time of measurement with the Center for Epidemiological Studies Depression









Scale (CES-D). The CES-D has 20 items to measure depression and is used widely to identify

individuals who are prone to depression (Radloff, 1977). The CES-D scores range from 0- 60

points. Higher scores indicate greater depressive symptoms and a threshold of 16 or higher often

indicates a risk for clinical depression (Wisniewski et al., 2003). Cronbach's alpha was .90 (a =

.90). This study set out to measure the relationship between caregiver depressive symptoms and

attrition from the REACH project.

Drop-out status (attrition). This variable was operationally defined as the loss of research

participants (i.e., caregivers) through withdrawal or mortality during the course of the REACH

study. For this research, participants' drop-out status (i.e., drop-out or treatment complete) was

assigned according to their status at the completion of the REACH trial.

Educational achievement. This variable refers to the number of years of formal

education completed by caregivers. This measure was acquired from the Caregiver

Sociodemographic form used to obtain descriptive information about the caregivers. Educational

level was measured by a range, from a low of 1 indicating an educational level of less than high

school to 17 indicating a graduate degree. This study set out to measure the relationship between

the educational level achieved by caregivers and attrition from the REACH proj ect.

Health status. This variable was operationally defined as the physical health status of the

caregiver as measured by the Caregiver Health and Health Behaviors questionnaire from the

REACH study. This measured was modified by the original REACH investigators from its

initial conception by combining items from existing measures including the Short Form Health

Survey (SF-36), the Asset and Health Dynamics Among the Oldest Old, and the Nutrition

Screening Initiative (NSI) (Switzer, Wisniewski, Belle, Dew, & Schulz, 1999). This research

examined the relationship between health status and attrition from the REACH proj ect.









Income. Refers to the total amount (from all sources) of money earned on an annual

basis by an individual. This measure was acquired from the Caregiver Sociodemographic form

used to obtain descriptive information about the caregivers. The form used ten categories to

measure income level, from a low of less than $5,000 per year to a high of $70,0000r more per

year. This study set out to measure the relationship between caregivers' annual total income and

attrition from the REACH proj ect.

Social support. This variable refers to the amount of support that the caregivers received

from friends, neighbors, and other members of the family as measured by the Social Support

form. This measured was also modified from its initial conception by the original REACH

investigators. They changed the response scale and/or eliminated or added items to the scale

(Wisniewski et al., 2003). This study attempted to measure the relationship between social

support and attrition from the REACH proj ect.

Null Hypotheses

The following six null hypotheses were tested in this study:

Hol: There is no relationship between attrition drop-out status (attrition) and Latino
caregivers' age.

Ho2: There is no relationship between drop-out status (attrition) and Latino caregivers'
achieved level of education.

Ho3: There is no relationship between drop-out status (attrition) and Latino caregivers'
income.

Ho4: There is no relationship between drop-out status (attrition) and depression among
Latino caregivers as measured by Center for Epidemiologic Studies Depression Scale
(CES-D).

Ho5: There is no relationship between drop-out status (attrition) and social support
among Latino caregiver' s.
Ho6: There is no relationship between drop-out status (attrition) and Latino caregivers'
health status as measured by the Caregiver Health and Health Behaviors Questionnaire
from the REACH study.











Research Procedures

The REACH investigators provided permission for use of the data from the Miami and

Palo Alto sites for purposes of this study. The data were obtained from the Inter-university

Consortium for Political and Social Research (ICPSR), at the University of Michigan. ICPSR

maintains and provides access to a vast archive of social science data for research. Finally, an

application to the Institutional Review Board (IRB) for an Exempt Status was filed and granted

since this study involved a collection of existing data that is available for public use and has de-

identified information, recorded so that participants cannot be identified.

Measures

The original REACH study used several scales and modified others to measure the

variables and meet the needs of the original study. For the present study, original data were

used, however, two measures, social support and health status, were modified. Demographic

information for this study was obtained and used as reported by the REACH investigators.

These data were collected through the "Caregiver Sociodemographic Questionnaire" (Appendix

A). The data for the demographic variables of interest (i.e., age, educational level, and income)

were included in this study.

Caregivers' depressive symptoms were measured with the Center for Epidemiological

Studies Depression Scale (CES-D). The CES-D scores were obtained from the original data

(Appendix B). A score of 16points indicated the cut-off point for the presence of depression.

The higher the score, the higher the number of depressive symptoms reported by the caregivers.

In order to measure the contributions of the independent variable social support, a social

support index was obtained from one of the three original questionnaires used by the REACH

investigators; the "perceived social support questionnaire" (Appendix C). This subscale included









nine questions regarding the caregivers' perception about support available to them from friends

and family, and was rated using with a Likert-type scale. The lower the number, the less

perceived support the caregivers reported. The score for this subscale ranged from 9 to 32 points.

Similarly, in order to measure the contributions of the independent variable health, two

independent indices were created from the original data acquired from the REACH proj ect

(Appendix D). Questions from the two subscales, 1) perceived health, and 2) self-reported

health problem, were included. The first subscale, the perceived health questionnaire included a

total of three questions. This scale was rated on a Likert-type scale and measured the caregivers'

subjective assessment of their health. The scores ranged between 1 and 4.67 points. The higher

the score, the healthier the caregivers' perceived their health. The second subscale, the self-

reported health problems questionnaire, included a total of seven questions regarding the

presence or absence of medical conditions such as high blood pressure, heart condition, and

cancer. The scores ranged between a low of 0 and a high of 5. Higher scores reflect a higher

number of reported perceived and/or established health problems.

Data Analysis

To test the hypotheses for this study, a hierarchical logistic regression analysis procedure

was performed by the statistical program Statistical Package for the Social Sciences (SPSS).

Based on existing literature, the following sets of predictors were identified and used in the

hierarchical logistic regression model: demographic variables (age, level of educational

achievement, level of income) and person-based factors (depression, health status index, and

social support index) as defined in previously published REACH studies. Alpha (o) will be set

at 0.05.









In order to examine the unique contribution of the independent variables to the prediction

of attrition, the demographic variables age and educational level were entered first into the initial

block. For the Miami sample, the variable gender was also included in the first model. Income

was entered into the logistic regression next, followed by depression score (CES-D), and social

support. In the last model, the indices used to measure health status, perceived health status and

reported health problems were entered in the last block. The order of entry for the variables was

determined by two factors. First, demographic variables were entered into the model. The order

of entry for the remaining three variables (i.e., depression, social support, and health status) was

determined based on the review of the literature.

Methodological Limitations

There are limitations associated with archival analyses. First, as a secondary analysis it

does not lend itself for purposeful collection of data. Therefore, the data in this study are limited

to the original REACH investigators' research questions and data collection. Second, there is an

inability to control for levels of acculturation in this sample of Latinos, which may influence


caregivers' belief, attitudes, and willingness to participate and complete research studies.

Moreover, there are inherent limitations associated with self-report measures including

differences in comprehension levels or interpretation, memory deficiencies such as forgetfulness,

and using the forced-choice response categories which may simplify answers or distort the

information obtained along particular response (or choice) sets (Hilton, Harry, & Rice, 2003). In

addition, there is a danger associated with self-report assessment of physical and emotional

ailments, which is influenced by the culture of ethnic groups. Furthermore, there is an absence

of a control or comparison group. There are also limitations associated with the data analysis

procedures of hierarchical logistic regressions such as the less than optimal control of covariate









variables and their correlational nature; thus, caution must be employed when inferring a causal

link (Huck, 2004).










CHAPTER 4
RESULTS



The present study entailed a secondary analysis of the multi-site Resources for Enhancing

Alzheimer' s Caregiver Health (REACH) study. The purpose of this investigation was to

examine predictors of attrition from the REACH trial. Specifically, the importance of age,

educational achievement, income, self-reported levels of depressive symptoms, perceived social

support, and health status on the drop-out rates for Latino caregivers of persons with Alzheimer' s

disease from the REACH trial.

This chapter presents the results of the analyses completed for this study. The information

in this chapter is organized as follows. First, data on the subj ects who participated in this study

are presented with a comparison of the subj ects by research site and drop-out status (i.e.,

completion vs. non-completion). Then the hypotheses and the analyses are presented to address

the hypotheses. A summary concludes this chapter.

Preliminary Analyses

A total of 219 participants from the two research sites, Palo Alto and Miami, were included

in this study. These participants were from Palo Alto (n=-107, 48.8%) and Miami (n =112,

52. 1%). Subj ects with missing data and those with extreme scores on the variables of interest for

this study were deleted from the dataset. There were a total of 59 dropouts (26.9%) and 160

completers (82.1%). The mean age of the total group was 58.90 years, ranging from 25 to 90

years (SD = 14.08) and the mean years of completed education for the total group was 1 1.12

years, ranging from 2 to 17 years (SD = 3.86). The median and mode annual total income for the

participants was $15,000 to $19,999. Table 4-1 illustrates the data for all of the variables for the









total group as well as by location and Table 4-2 shows the data for all of the variables by

completion status.

It was also of interest to this study to determine whether there were significant

differences between sample sites (Palo Alto and Miami) and between drop-out status completer

and drop-out). In order to compare the mean values for the variables of interest between groups,

analysis of variance was used to compare respondents between location site and drop-out status

for the continuous variables (age, years of education, CES-D score, perceived social support,

perceived health status and reported health status). Chi Square was used to test for differences in

categorical data (income). The mode income for Palo Alto was between $15,000 to $19,999 and

the mode for Miami was $5,000 to $9,999. Participants from Miami had a lower income level

than did Palo Alto participants. Median income in Palo Alto was between two ranges, $15,000

to $19,999 and $20,000 to $29,999 while the median income for Miami was between the two

ranges of $10,000 to $14,000 and $15,000 to $19,999. The Chi Square for location was

significant for income (X2 (9) = 22.249, p = .008); however, there were six cells with counts less

than five and the test may not be valid.

The mode income for completers was $15,000 to $19,999 and the mode income for

dropouts was $5,000 to $9,999. It is interesting to note that income was noticeably less for

dropouts. The Chi Square test for dropouts versus completers by income found no significant

differences (X2 (9) = 16.514, p = .057), however this may not have been an appropriate test as

there were four cells with counts less than five bringing into question the validity of this test.

Table 4-3 shows the income breakdown by location and drop-out status.

To further test the differences between sites and drop-out status, ANOVA tests were

conducted. Table 4-4 shows the results of the ANOVA test between research sites. The










ANOVA test indicates there are significant differences between participants in the Palo Alto site

and the Miami site for age, years of education, CES-D score, social support, perceived health,

and reported health status. Inspection of the ANOVA Eindings for dropouts and completers

found no statistically significant differences between dropouts and completers for age, years of

education, CES-D score, perceived social support, perceived health, and reported health status.

Table 4-5 shows the results of the ANOVA test between drop-outs and treatment completers.

Dropouts and completers were remarkably similar; however, there were noticeable differences

between locations. A Spearman Rho Correlation was used to test for correlations between the

predictor variables and the criterion (i.e., drop-out status) in this study. These correlations are

discussed next.

Palo Alto Correlations

A Spearman Rho Correlation was calculated between the predictor variables and criterion

variable (i.e., drop-out status) in this study. The variable of primary interest was how the

independent variables (age, years of education, income, CES-D score, received support,

perceived health status, and health problems) were correlated with the dependent variable drop-

out status. Age was significantly correlated with drop-out status (r = .212, p = .028) although it

was a weak correlation. Years of education was inversely correlated with age (r = -.232, p =

.016). Income was strongly correlated with years of education (r = .252, p = .009). The most

powerful correlation in this study occurred between health problems and age (r = .393, p =

<.001). The CES-D (depression) score had a low correlation with health problems (r = .288, p =

.005) and with perceived health problems (r = .3 12, p = .003). All of the other correlations were

not strong correlations nor were they statistically significant. Table 4-6 shows the correlation

matrix for the Palo Alto sample.










Miami Correlations

A Spearman Rho Correlation was used to test for correlations between the variables used

in this study. The variable of primary interest was how the independent variables (age, years of

education, gender, income, CES-D score, received support, perceived health status, and health

problems) were correlated with the dependent variable of dropping out or staying in the REACH

program. There was a significant though low correlation between drop-out status and years of

education (r = .198, p = .037) for the Miami participants. There was a strong correlation

between years of education and income (r = .537, p = <.001). The correlation between years of

education and gender was significant but weak (r = .211, p = .025). The correlation between age

and health problem score was weak but statistically significant (r = .232, p = .015.). Age was

inversely correlated with income (r = -.424, p = <.001), moderately correlated with gender (r =

-.347, p <.001). There was a low moderate positive correlation between CES-D score and health

status (r = .314, p = .002) and between CES-D and health problems (r = .215, p = .041). There

was a low but significant correlation between gender and income (r = .188, p = .050) and

between gender and CES-D score (r = .273, p = .008). The correlation between received support

and health status was inverse and low but statistically significant (r = -.209, p = .032). Table 4-7

shows the correlation matrix for the Miami sample.

Hypothesis Testing

This study sought to address the following hypotheses:

Hol: There is no relationship between attrition drop-out status (attrition) and Latino
caregivers' age.

Ho2: There is no relationship between drop-out status (attrition) and Latino caregivers'
achieved level of education.

Ho3: There is no relationship between drop-out status (attrition) and Latino caregivers'
income.









Ho4: There is no relationship between drop-out status (attrition) and depression among
Latino caregivers as measured by Center for Epidemiologic Studies Depression Scale
(CES-D).

Ho5: There is no relationship between drop-out status (attrition) and social support
among Latino caregiver' s.

Ho6: There is no relationship between drop-out status (attrition) and Latino caregivers'
health status as measured by the Caregiver Health and Health Behaviors Questionnaire
from the REACH study.


Data Analysis

Logistic regression was conducted on research questions concerning the relative

importance of each block (set) of independent variables: demographics (age, gender, years of

education, income), CES-D score (CES-D), social support perceived (SSPerc), health status

perceived (HSPered), and actual health problem (HProb) served as statistically significant

predictors of dropping out of the REACH study. Tabachnick and Fidell (1996) maintain that

logistic regression does not require the adherence to any assumptions about the distribution of

the variables. Study data were checked for collinearity and found to be within an acceptable

range for tolerance and the variance inflation factor (Mertler & Vannatta, 2001). The data for the

two sites (Palo Alto and Miami) were also checked for outliers prior to commencing data

analysis. The Mahalamobis X2 a test for outliers, identified four outliers with scores in excess

of the test statistic. Three cases were deleted from the Palo Alto dataset and one case was

deleted from the Miami dataset. This resulted in Miami having an n = 112 and Palo Alto having

an n = 107. Missing data were treated as missing and no substitutions were made for missing

data. Since the participants from Palo Alto and Miami were statistically different on a number of

the independent variables as presented previously, each location was treated as a separate entity

for purposes of the logistic regression analysis.









Logistic regression using the enter method (simultaneous entry of predictors) was

conducted to determine which of the set of independent predictor variables would predict

whether the participant completed or dropped out of the REACH treatment program for each

location. Independent variables were entered in blocks to measure the effect of adding a

particular variable or set of variables into the model.

Palo Alto Sample

The first block used age and years of education as predictors of completing or dropping out

of the REACH program in the model. The [-2 Log likelihood (72.644) and X2 (2) = 4. 123, p =

.127] for this model indicated the model was not statistically significant. However, the model

using age and years of education correctly identified 83.9% of the participants. Wald statistics, as

shown in Table 4-8 indicated none of the variables served as a significant predictor of drop-out

status for the Palo Alto group of participants. The odds ratios for these variables suggest there is

little change in the likelihood of dropping out or completing the program when the predictors

variables (i.e., age and years of education) increased by 1. Table 4-8 shows the results of the

first model for the Palo Alto logistic regression analysis.

The second model added income to the model to determine the effects of income on the

model. The model was not statistically significant [-2 Log likelihood (71.679), X2 (3) = 5.088, p

=.165] and the model fit indices were fairly large indicating model fit was questionable. The

age, years of education, and income model did correctly identify 83.9% of the participants. The

Wald statistics presented in Table 4-9 indicated that none of the variables (i.e., age, years of

education, and income) served as a significant predictor of dropping out or completion for the

Palo Alto group of participants. The odds ratios for these variables found little change in the

likelihood of dropping out or completing the program when the predictors increased by 1. Table

4-9 shows the results of the second model of the Palo Alto logistic regression analysis.









The third model added the score on the Center for Epidemiologic Studies Depression Scale

(CES-D) as a predictor along with age, years of education and income. The model was not

statistically significant [-2 Log likelihood (71.674), X2 (4) = 5.094, p = .278] and the model fit

indices were fairly large indicating model fit was questionable. The age, years of education,

income, and CES-D model did correctly identify 83.9% of the participants. The Wald statistics

presented in Table 4-10 indicated none of the variables (i.e., age, years of education, income,

CES-D score) served as a significant predictor of dropping out or completion for the Palo Alto

group of participants. The odds ratios for these variables found little change in the likelihood of

dropping out or completing the program when the predictors increased by 1. Table 4-10 shows

the results of the third model of the Palo Alto logistic regression analysis.

The fourth model added the variable perceived social support (SSPerc) into the model with

age, years of education, income, and CES-D score. The model was not statistically significant

[-2 Log likelihood (71.666), X2 (5) = 5.101, p = .404] and the model fit indices were fairly large

indicating model fit was questionable. The model with age, years of education, income, CES-D

score, and received social support correctly identified 83.9% of the participants. The Wald

statistics presented in Table 4-11 indicated none of the variables (i.e., age, years of education,

income, CES-D score, and perceived social support) served as a significant predictor of dropping

out or completion for the Palo Alto group of participants. The odds ratios for these variables

found little change in the likelihood of dropping out or completing the program when the

predictors increased by 1. Table 4-11 shows the results of the fourth model of the Palo Alto

logistic regression analysis.

The fifth model added the variable perceived health status (HSPered) and reported health

status (HSProb) into the model predicting dropping out or completing the REACH program. The









model was not statistically significant [-2 Log Likelihood (69.3056) X2 (6) = 7.462, p =.3 82].

The model fit indices were fairly large indicating model fit was questionable and the Chi Square

statistic was not significant. The model correctly classified 82.8% of the cases. Regression

coefficients are presented in Table 4-12. Wald statistics indicated none of the variables (i.e.,

age, years of education income, CES-D score, received social support, perceived health status

and reported health status) served as a significant predictor of dropping out or completion for the

Palo Alto sample. The odds ratios for these variables indicate there is little change in the

likelihood of dropping our or completing the program when the predictors increased by 1. Table

4-12 shows the results of the fifth model of the Palo Alto logistic regression analysis.

Miami Sample

The first block used age, years of education, and gender as predictors of staying or

dropping out of the REACH program in the model. This model was not statistically significant

as indicated by [-2 Log likelihood (107.284) and X2 (3) = 2.407, p =.492]. However, the age and

years of education model did correctly identify 62.7% of the participants. Wald statistics

indicated none of the variables (age and years of education) served as a significant predictor of

dropping out or completion for the Miami sample. The odds ratios for these variables indicate

there is little change in the likelihood of dropping out or completing the program when the

predictors increased by 1. Table 4-13 shows the results of the first model of the Miami logistic

regression analysis.

The second model added income to the model to determine the effects of income on the

model. The model was not statistically significant [-2 Log likelihood (107.260), X2 (4) = 2.431,

p =.657] and the model fit indices were fairly large indicating model fit was questionable. The

age, years of education, gender, and income model did correctly identify 62.7% of the

participants. The Wald statistics presented in Table 4-14 indicated none of the variables (age,









years of education, and income) served as a significant predictor of dropping out or completion

for the Miami group of participants. The odds ratios for these variables found there was little

change in the likelihood of dropping out or completing the program when the predictors

increased by 1. Table 4-14 shows the results of the second model of the Miami logistic

regression analysis.

The third model added the depression score, as reported by the Center for Epidemiologic

Studies Depression Scale (CES-D), as a predictor along with age, years of education and income.

The model was not statistically significant [-2 Log likelihood (106.773, X2 (5) = 2.918, p = .713]

and the model fit indices were fairly large indicating model fit was questionable. The age, years

of education, income, and CES-D score model did correctly identify 60.2% of the participants.

The Wald statistics presented in Table 4-15 indicated none of the variables (age, years of

education, gender, income, and CES-D) served as a significant predictor of dropping out or

completion for the Miami sample. The odds ratios for these variables found little change in the

likelihood of dropping out or completing the program when the predictors increased by 1. Table

4-15 shows the results of the third model of the Miami logistic regression analysis.

The fourth model added Social Support Received (SSPerc) into the model with age, years

of education, income, and CSED. The model was not statistically significant [-2 Log likelihood

(106.708, X2 (6) = 2.983, p = .811] and the model fit indices were fairly large indicating model fit

was questionable. The age, years of education, income, CES-D score, and perceived social

support model did correctly identify 62.7% of the participants. The Wald statistics presented in

Table 4-16 indicated none of the variables (age, years of education, income, CES-D score, and

perceived social support) served as a significant predictor of dropping out or completion for the

Miami group of participants. The odds ratios for these variables found little change in the









likelihood of dropping out or completing the program when the predictors increased by 1. Table

4-16 shows the results of the fourth model of the Miami logistic regression analysis.

The fifth model added the health variables, perceived health status (HSPered) and reported

health status (HSProb) into the model predicting dropping out or completing the REACH

program. The model was not statistically significant [-2 Log Likelihood (105.783; X2 (8) = 3.908,

p =.865]. The model fit indices were fairly large indicating model fit was questionable and the

Chi Square statistic was not significant. The model correctly classified 61.4% of the cases.

Regression coefficients are presented in Table 4-17. Wald statistics indicated none of the

variables (age, years of education, income, CES-D, perceived social support, and perceived

health status and reported health status served as a significant predictor of dropping out or

completion for the Miami research site. The odds ratios for these variables indicate there is little

change in the likelihood of dropping out or completing the program when the predictors

increased by 1. Table 4-17 shows the results of the fifth model of the Miami logistic regression

analy si s.

Summary of Results

The outcomes of the study were not significant. None of the predictor variables

contributed significantly to the understanding of caregiver drop-out status (the criterion) from the

REACH trial. There were no significant differences between drop-out status though income

between the two samples groups varied significantly. In addition, there were significant

differences between the Palo Alto and Miami samples for age, years of education, CES-D score,

perceived social support, perceived health, and reported health status. There were no statistically

significant differences between dropouts and completers for age, years of education, CES-D

score, social support, perceived health, and reported health status.





Table 4-1. Variable means for total group and by location (Palo Alto and Miami)
Total Total Total Miami Miami Miami Palo Palo Palo
N Mean SD Mean SD N Alto Alto Alto
Mean SD N
Age 219 58.90 14.08 66.51 10.55 112 50.94 12.89 107


YrEduc 219 11.11 3.86 11.63 3.78 112 10.57 3.89

CES-D 186 18.69 11.96 19.98 11.06 99 17.39 12.72

SSPerc 210 15.81 8.51 13.68 8.63 106 17.99 7.85


107

93

104


HSPered 218


2.69 .69 2.74


.70 111 2.65 .67 107


HSProb 214 1.13 1.11 1.37 1.13 110


.88 1.03 104


Note: YrEduc =
Score; SSPerc =
reported Health


Years of Education; CES-D = Center for Epidemiologic


Studies Depression


Perceived Social Support; HSPered = Perceived Health Status; HSProb = Self-






















2.74 .75 58 2.67


.66 160


HS Prob 214


1.13 1.11 1.05 1.02 59 1.16 1.14


I


Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression
Score; SSPerc = Perceived Social Support; HSPered = Perceived Health Status; HSProb = Self-
reported Health; Drop = Drop-outs; Compl = Completers


Table 4-2. Variable means for total
Total Total Total
N Mean SD
Age 219 58.90 14.08

YrEduc 219 11.11 3.86


group and dropouts and completers
Drop Drop Drop Compl Compl
Mean SD N Mean SD
60.20 15.20 59 58.43 13.67


Compl
N
160


10.88 3.73


59 11.20 3.92 160

50 16.19 12.68 136


CES-D 186 18.69 11.96 20.04 9.72


SSPerc 210 15.81 8.51

HSPered 218 2.69 .69


14.58 8.08


55 16.25 8.64









Table 4-3. Income level group for completion status and location
Income Drop N Comp N Miami N Palo Alto N Total N



<$5,000 5 (8.6) 7 (4.4) 6 (5.5) 6 (5.6) 12 (5.6)

$5-9,999 11 (19.0) 21 (13.3) 23 (21.1) 9 (8.4) 32 (14.8

$10-14,999 15 (25.9) 20 (12.7) 21 (19.3) 14 (13.1) 35 (16.2)

$15-19,999 6 (10.3) 31 (19.6) 17 (15.6) 20 (18.7) 37 (17.1)

$20-29,999 9 (15.5) 19 (12.0) 16(14.7) 12 (11.2) 28 (13.0

$3 0-3 9,999 3 (5.2) 24 (15.2) 10 (9.2) 17 (15.9) 27 (12.5)

$40-49,999 3 (5.2) 18 (11.4) 3 (2.8) 18 (16.8) 21 (9.7)

$50-59,999 1 (1.7) 8 (5.1) 4 (3.7) 5 (4.7) 9 (4.2)

$60-69,999 3 (5.2) 4 (2.5) 5 (4.6) 2 (1.9) 7 (3.2)

$70,000 > 2 (3.4) 6 (3.8) 4 (3.7) 4 (3.7 ) 8 (3.7)

Total 58 (100) 158(100) 109(100) 107 (100) 216 (100)

Income reported in dollars










Table 4-4. ANOVA test between sites (Palo Alto and Miami)
Df F p


Age


1,217

1,217

1,184

1,208

1,216

1,212


95.966

4.129

2.196

14.233

.908

10.786


<.001

.043

.140

<.001

.342

.001


YrEduc

CES-D

S SPerc

HSPered

HSProb


Note: YrEduc =
Score; SSPerc =
reported Health.
p <.05.


Years of Education; CES-D = Center for Epidemiologic Studies Depression
Perceived Social Support; HSPered = Perceived Health Status; HSProb = Self-










Table 4-5. ANOVA test between treatment dropouts and treatment completers
Df F p


Age


1,217

1,217

1,184

1,208

1,216

1,212


.686

.303

.866

1.576

.424

.472


.408

.582

.353

.211

.516

.493


YrEduc

CES-D

S SPerc

HSPered

HSProb


Note: YrEduc =
Score; SSPerc =
reported Health.
p <.05.


Years of Education; CES-D = Center for Epidemiologic Studies Depression
Perceived Social Support; HSPered = Perceived Health Status; HSProb = Self-











Table 4-6. Palo Alto correlations
Drop Age Yrs Ed Income CES-D SSPerc HSPered HSProb
r r r r r r r r
p p p p p p p p


Note: Age = Years of Life; YrEduc = Years of Education; CES-D = Center for Epidemiologic
Studies Depression Score; SSPerc = Perceived Social Support; HSPered = Perceived Health
Status; HSProb = Self-reported Health
p <.05.


Drop


1.00


Age .212 1.00

.028

YrEduc -.052 -.232 1.00

.595 .016

Inome .172 -.159 .252 1.00

.079 .102 .009


CES-D -.035 -.087 -. 107 -.020 1.00

.740 .406 .306 .852


S SPerc


.077 .030 -.020 .017

.439 .752 .843 .860


.156


1.00


HSPered .033 .018 -.065


-.106 .288

.276 .005


.090

.365

.044

.665


1.00


.733 .852 .503


HSProb .127 .393 -. 178 -. 100

.199 <.001 .070 .312


.312


1.00


.003


.539

















































-.209

.032

-.043

.667


1.00


.059

.543


1.00


Note: Age = years of Life; YrEduc
Center for Epidemiologic Studies -


= Years of Education; Gender = Male or Female; CES-D
Depression Score; SSPerc = Perceived Social Support;


HSPered = Perceived Health Status; HSProb = Self-reported Health
p <.05.


SSPerc HSPered HSProb
r r r
P P P


I


Table 4-7. Miami
Drop

r

Drop 1.00

Age -.063

.509

YrEduc .198

.037

Gender .005

.958

Income .112

.247

CES-D -.120

.252

SSPerc .002

.984

HSPered -.043

.657

HSProb .064

.508


correlations
Age YrEduc Gender Income CESD
r r rr r
P P PP P


1.00


-.169

.075

-.347

<.001

.424

<.001

-.124

.238

-.004

.966

-.060

.530

.232

.015


1.00


.211

.025

.537

<.001

-.118

.258

-.025

.802

.038

.968

-.025

.786


1.00


.188

.050

.273

.008

.168

.085

.034

.720

-.013

.892


1.00


-.126

.236

.083

.401

.034

.726

-.164

.092


1.00


-.049

.652

.315

.002

.215

.041


1.00










Table 4-8 Palo Alto logistic regression, model one
B Wald Df p Odds/Ratio CI

Age .045 .028 1 .085 1.046 994, 1.101

YrEduc -.028 .086 1 .748 .973 .822, 1.151

Constant -.170 1.759 1 .923 .844

Note: YrEduc = Years of Education
p <.05.





Table 4-9 Palo Alto logistic regression, model two
B Wald Df p

Age .042 2.624 1 .105

YrEduc .047 .279 1 .598

Income .139 .934 1 .334

Constant -.352 .039 1 .843

Note: YrEduc = Years of Education
p <.05.


Odds/Ratio

1.043

000954

1.150

.703


CI

991, 1.097

.800, 1.137

.866, 1.525


I










Table 4-10. Palo Alto logistic regression, model three
B Wald Df p Odds/Ratio CI

Age .042 2.490 1 .115 1.042 .990, 1.098

YrEduc -.048 .281 1 .596 .953 .799, 1.137

Income .139 .932 1 .334 1.149 .866, 1.525

CES-D -.002 .006 1 .940 .998 .950, 1.049

Constant -.299 .025 1 .875 .741

Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression
Score
p <.05.










Table 4-11. Palo Alto logistic regression, model four
B Wald Df p Odds/Ratio CI

Age .042 2.493 1 .114 1.043 990, 1.098

YrEduc -.047 .278 1 .598 .954 .799, 1.138

Income .139 .931 1 .335 1.149 .866, 1.525

CES-D -.002 .004 1 .950 .998 .950, 1.050

SSPerc -.003 .007 1 .933 .997 .922, 1.077

Constant -.250 .016 1 .900 .779

Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression
Score; SSPerc = Perceived Social Support
p <.05.










Table 4-12. Palo Alto logistic regression, model five
B Wald Df p Odds/Ratio CI

Age .034 1.493 1 .222 1.035 .979, 1.094

YrEduc -.038 .161 1 .688 .963 .799, 1.159

Income .184 1.401 1 .237 1.202 .886, 1.632

CES-D -.006 .040 1 .842 .994 .939, 1.053

SSPerc -.017 .169 1 .681 .983 9.05, 1.068

HSPered .664 1.816 1 .178 1.942 .739, 5.102

HSProb .272 .380 1 .538 1.313 .552, 3.121

Constant -1.738 .557 1 .458 .175

Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression
Score; SSPerc = Perceived Social Support; HSPered = Perceived Health Status; HSProb = Self-
reported Health
p <.05.












B Wald Df p Odds/Ratio CI


Age .000 .000 1 .992 1.000 .953, 1.050

YrEduc .092 2.174 1 .140 1.097 .970, 1.240

Gender -.018 .001 1 .975 .982 .308, 3.126

Constant -.505 .043 1 .835 .603


Note: YrEduc = Years of Education; Gender = Male or Female
p <.05.


I


Table 4-13. Miami logistic regression,


model one










Table 4-14. Miami logistic regression, model two
B Wald Df p Odds/Ratio CI

Age -.001 .002 1 .965 .999 .949, 1.052

YrEduc .097 1.884 1 .170 1.102 .959, 1.267

Gender -.023 .002 1 .969 .977 .306; 3.115

Income -.019 .024 1 .876 .981 .772, 1.248

Constant -.401 .025 1 .873 .670

Note: YrEduc = Years of Education; Gender = Male or Female
p <.05.










Table 4-15. Miami logistic regression, model three
B Wald Df p Odds/Ratio CI

Age -.003 .012 1 .914 .997 .946, 1.050

YrEduc .096 1.818 1 .178 1.100 .958, 1.264

Gender .099 .026 1 .872 1.104 .330; 3.694

Income -.038 .091 1 .763 .963 .752, 1.232

CES-D -.015 .487 1 .485 .985 .943, 1.028

Constant .108 .002 1 .967 .898

Note: YrEduc = Years of Education; Gender = Male or Female; CES-D = Center for
Epidemiologic Studies Depression Score
p <.05.










Table 4-16. Miami logistic regression, model four
B Wald Df p Odds/Ratio CI

Age -.001 .003 1 .957 .999 .947, 1.053

YrEduc .093 1.695 1 .193 1.098 .954, 1.263

Gender .139 .048 1 .827 1.149 .331; 3.990

Income -.029 .050 1 .823 .971 .752, 1.254

CES-D -.016 .506 1 .477 .984 .943, 1.028

SSPerc -.007 .066 1 .798 .993 .940, 1.048

Constant .166 .004 1 .949 .847

Note: YrEduc = Years of Education; Gender = Male or Female; CES-D = Center for
Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support
p <.05.










Table 4-17. Miami logistic regression, model five
B Wald Df p Odds/Ratio CI

Age -.004 .017 1 .897 .997 .945, 1.051

YrEduc .094 1.697 1 .193 1.099 .954, 1.267

Gender .117 .033 1 .855 1.124 .321; 3.939

Income -.012 .009 1 .926 .988 .763, 1.279

CES-D -.013 .286 1 .593 .987 .942, 1.035

SSPerc -.010 .124 1 .725 .990 .937, 1.046

HSPered -.259 .505 1 .477 .772 .378, 1.576

HSProb .160 .526 1 .468 1.173 .762, 1.806

Constant .425 .025 1 .875 1.530


Note: YrEduc = Years of Education; Gender


Male or Female; CES-D = Center for


Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPered
Perceived Health Status; HSProb = Self-reported Health
p <.05.











CHAPTER 5
CONCLUSION

This chapter includes a discussion of the Eindings for each hypothesis as well as additional

conclusions and interpretations resulting from the analyses of data. The implications and

limitations of the study are also presented. Recommendations for future research conclude the

chapter.

In order to address current concerns in the literature about the low participation and

retention of Latino caregivers of persons with Alzheimer' s disease in research studies, the goal

of this research was to examine six variables believed to be relevant to the attrition rate of Latino

caregivers of persons with Alzheimer's disease from research studies. Specifically, this study set

out to examine the extent to which the variables caregiver age, caregiver attained educational

level, caregiver income, caregiver depression score (CES-D), caregiver perceived social support,

and caregiver health status help explained caregivers' drop-out status from two sites in the

Resources for Enhancing Alzheimer' s Caregiver Health (REACH) trial; Miami and Palo Alto.

The Miami site (as described by Wisniewski et al., 2003) examined the effectiveness of family-

based structural multisystem in-home intervention (FSMII) and a family-based structural

multisystem in-home intervention plus a computer telephone integration system (FSMII +

CTIS). The Palo Alto site (as described by Wisniewski et al., 2003) examined the effectiveness

of the coping with a caregiving class (CWC) and an enhanced support group (ESG).

Conclusions


The outcomes of the study were not significant. None of the predictors help to

significantly explain caregiver drop-out status. While other studies have suggested that these

variables may help explain why dementia caregivers drop-out prematurely from research studies,









this set of variables did not help predict attrition in these two samples of Latino caregivers of

persons with Alzheimer' s disease. Although the outcomes of this study did not produce

significant results, these findings contribute to the existing literature on Latino caregivers of

persons with Alzheimer' s disease and serve to inform future research studies by underscoring the

importance of an adequate sample size, purposeful instrumentation, and cultural considerations.

Thus, this study should be considered a pilot study that tested the feasibility of this investigation.

Palo Alto


No significance was found between caregivers' age and the likelihood of dropping out of the

REACH clinical trial (hypothesis 1). No significance was found between caregivers' level of

education and the likelihood of caregivers dropping out of the REACH trial (hypothesis 2). No

significance was found between caregivers' income level and the likelihood of caregivers

dropping out of the REACH trial (hypothesis 3). No significance was found between caregivers'

CES-D score (i.e., depressive symptoms) and the likelihood of caregivers dropping out from the

REACH trial (hypothesis 4). Similarly, no significance was found in caregiver perceived social

support and the likelihood of caregivers dropping out of the REACH trial (hypothesis 5).

Finally, no significance was found in caregivers' health status and the likelihood of caregivers

dropping out of the REACH trial (hypothesis 6).

Miami


No significance was found in caregivers' age and the likelihood of caregivers dropping out

of the REACH trial (hypothesis 1). No significance was found in caregivers' level of education

and the likelihood of caregivers dropping out of the REACH trial (hypothesis 2). In addition, no

significance was found in caregivers' income level and the likelihood of caregivers dropping out

of the REACH trial (hypothesis 3). No significance was found in caregivers' CES-D score (i.e.,










depressive symptoms) and the likelihood of caregivers dropping out of the REACH trial

(hypothesis 4). No significance was found in caregivers' perceived social support and the

likelihood of caregivers dropping out of the REACH trial (hypothesis 5). Finally, no

significance was found in caregivers' health status and the likelihood of caregivers dropping out

of the REACH trial (hypothesis 6).

Implications


The predictor variables in this investigation (i.e., caregiver' s age, achieved educational

level, income level, CES-D score, health status, and perceived social support) did not achieve

statistical significance for the criterion of drop-out status. However, for the Palo Alto sample, it

appears that there may be significance for the variables, caregivers' age, income, and perceived

health, if the sample size were larger or if the instruments were sufficiently powered to test these

specific variables. Similarly, in the Miami sample, the predictor variable of caregivers' achieved

educational level also indicates it may achieve significance with a larger sample size or better

instrumentation.

These Eindings are consistent with the existing mixed results in the current literature

regarding the participation of Latinos in research studies. A few studies have been successful

recruiting and retaining Latino participants (Gallagher-Thompson et al., 2004; Schulz, 2003).

However, many others have documented the challenges in recruiting and retaining Latino

participants in research studies (Bank, Arguelles, Rubert, Eisdorfer, & Czaj a, 2006; Levkoff &

Sanchez, 2003; Morano, & Bravo, 2002; Aranda, 2001; Fitten, Ortiz, & Pont6n, 2001; Wiliams,

Tappen, Buscemi, Rivera, & Lezcano, 2001). These mixed results may continue to be related to

the under-representation of Latinos in research studies. Although the participants in these two

samples were Latinos, these two sample sizes are considered small sample sizes for this type of










analysis. The samples were also different regarding the country of origin; while the Miami

sample was mostly Cuban-American, the Palo Alto sample was mostly Mexican-American. In

addition, these findings are inconsistent with other research findings that suggest the variables

included in this study help explain why Latino caregivers drop-out prematurely from research

studies. For example, Gallagher-Thompson, Solano, Coon, and Arean (2003) underscore the

need to examine the influence of physical health and mental well-being (i.e., depression) in

Latino caregivers in future research studies. Schulz et al. (2003) and Haley, Levine, Brown, &

Bartolucci, (1987) emphasize the role of social support. Consequently, recruitment and retention

of culturally diverse individuals have been areas of maj or concern for researchers. Schulz et al.

(2003) report that "one of the disappointments in the caregiving intervention research literature

has been the relative lack of success in achieving clinically meaningful outcomes" (p. 519).

Thus, it is critical to continue the efforts to advance the knowledge base about the impact of

these and other variables in the retention of Latino caregivers in research studies.

The aging of the population in combination with the rapid growth of the Latino population

and the increasing incidence of Alzheimer' s disease warrant the continuation of efforts to

advance knowledge about the impact of Alzheimer' s disease. Over one million Latinos are

expected to suffer from this devastating illness (Alzheimer' s Association, 2004). Thus, these

numbers underscore the need to include culturally diverse groups in investigations in order to

develop effective methods to meet their needs. Although numerous studies have been conducted

to advance the knowledge about the impact of Alzheimer' s disease, only a handful have included

ethnic minority participants, particularly Latinos (Bank et al., 2006). The debilitating and

progressive nature of Alzheimer' s disease demands careful consideration of and developing









strategies and resources to address the needs of caregivers, especially ethnic minorities, since the

generalizability of most studies cannot be extended to these groups.

Continued inquiry regarding the impact of these person-based variables (i.e., age,

education level, depressive symptoms, physical health, and social support), the impact of burden

associated with the caregiver role, and reasons for participation in research studies are crucial.

Moreover, inclusion and examination of other variables such as culture-specific variables (i.e.,

acculturation level, language proficiency, perception about the etiology of Alzheimer' s disease,

and the perception about research studies), care recipient characteristics (i.e., physical limitations

and degree of cognitive impairment), as well as different research delivery methods (i.e., use of

technology) may prove beneficial. The need for including cultural variables when examining

ethnic minorities has been well-documented in the literature (Mahoney, Cloutterbuck, Neary, &

Zhan, 2005; Moreno-John et al., 2004; Arean, Alvidrez, Nery, Estes, & Linkins, 2003; Levkoff

& Sanchez, 2003; Curry & Jackson, 2003; Aranda, 2001). Another area of inquiry that is

attracting much attention is the development of alternative service and research delivery options.

For example, as the baby boomers age, an increasing number of potential participants will

probably have had significant experience with telecommunication technology such as the

telephone and the internet, which may be excellent vehicles for gathering data (Bank, Arguelles,

Rubert, Eisdorfer, & Czaja, 2006).

Limitations


Due to the archival nature of this inquiry, this study did not lend itself for purposeful

collection of data nor allow for adding more research participants. The hypotheses in this study

were not powered to test the relationships of the variables of inquiry. The data in this study are

limited to the original REACH investigators' research questions and data collection which may









have negatively impacted this study's predictive variables. The sample size for this study was

also near the minimum for logistic regression analysis (Hsieh, Bloch, & Larsen, 1998).

Furthermore, there was no control or comparison group in the REACH trial.

Perhaps the most relevant limitation associated with this study was that it did not consider

cultural variables. This study did not examine the specific influence of culture in the retention of

the study's participants which may contribute to their motivation and investment in the research

outcome. In addition, there is a danger associated with self-report assessment of physical and

emotional ailments, which are subj ective and are influenced by the culture of ethnic groups.

This study was limited in its inability to control for levels of acculturation in the two samples of

Latinos, which may influence caregivers' belief, attitudes, and willingness to participate and


complete research studies. The study was also limited in its scope, in that no measures

considered important environmentally based variables, such as community resources and

transportation issues, were examined.

Another limitation of this study was that participants were recruited from physicians'

offices, Memory Disorder Clinics, and other clinical settings; recruitment did not include other

types of community-based caregivers. Additionally, the small number of males and caregivers

older than 65 years of age limited the power and applications of the findings. Thus, results

should only be generalized to younger female Latinas.

In addition, it is important to consider that the variables examined in this study may not

have been effective predictors of drop-out status (the criterion) because the participants may

have been more motivated, younger, and in better health than older caregivers. Typically, older

caregivers are at higher risk for potentially significant health problems associated with older age,

length of time caring for someone with progressive dementia (i.e., Alzheimer' s disease), and less










social support, which may influence their ability and willingness to participate in research

studies.

Recommendations


It is recommended that future research studies strengthen efforts to include Latino

caregivers of persons with Alzheimer' s disease in their studies. It is important to note that the

understanding of the variables examined in this study are embedded in the cultural context.

However, as mentioned earlier, this study did not measure any cultural variables. Thus, it is

recommended that future inquiries into Latino caregivers of persons with Alzheimer' s disease

are designed in a manner that allows for the measurement of cultural considerations such as level

of acculturation, language proficiency, and participants' perceptions about research, as well as

their perception about the variables of interest. In addition, because of the close-knit nature of

Latino families, it is recommended that future research examines the caregiving experience from

a family or systemic context. It is imperative that clinicians understand the obstacles hindering

Latinos' access to services (i.e., transportation issues, immigration status, awareness of

community services, etc). Religion is also a significant factor in the Latino community.

Religious belief may influence Latinos' belief and understanding about an illness or may be a

valuable source of support and endurance. The role of religion in the Latino population and its

influence on the caregiving experience warrants closer examination.

Many researchers have brought attention to the issue of language and language proficiency

(Aranda, 2001; Gallagher-Thompson et al., 2003; Levkoff & Sanchez, 2003; Gallagher-

Thompson, 1996). Therefore, conducting the study in Spanish when appropriate and having

interviewers who can speak the language fluently and accurately reflect the sample population

are crucial and may help foster trust. Latinos may drop-out prematurely from research studies if










they feel inadequate due to language limitations or if they do not feel at ease with the researcher.

Thus, conducting consumer-center recruitment and retention practices that are informed by the

culture of the target population, as proposed by Arean et al., (2003), is necessary to reach out

effectively to this group. These culturally-appropriate practices may prove beneficial in

alleviating the potential for stigma associated with participation in the research study and the

stress associated with speaking in another language.

Gallagher-Thompson, Solano, Coon, and Arean (2003) report that one aspect of their

success in recruiting and retaining Latino caregivers was taking into account the practical

barriers to participation in research studies. Research protocols should be designed to reduce the

burden of participation and to cover some of the cost such as travel and sitter services (Arean,

Alvidrez, Nery, Estes, & Linkins, 2003; Arean & Gallagher-Thompson, 1996). It is also

imperative that researchers are mindful of the time requirements. They must consider

conducting shorter interviews, and, if necessary, breaking down longer assessments into smaller

components. Researchers should also consider bringing the research to the Latino caregivers

instead of expecting them to come to the research site. This may be accomplished by partnering

with religious organizations or local businesses to conduct interviews in a neutral yet familiar

environment. This notion is supported by the work of Levkoff and Sanches (2003).

Consequently, attending to the barriers associated with the group of interest may prove beneficial

in minimizing any issues of mistrust regarding researchers.

In addition, research questions should also have practical value for the research

participants. It is important that Latino caregivers understand the value of participating in

research protocols, and, if possible, that they obtain services they normally would not access.

For example, many caregivers report poor health as a consequence of their role as caregiver.










Future research studies may benefit from addressing the impact of caregiving on the physical

health of caregivers. This notion is supported in the literature by the work of Gallagher-

Thompson et al., (2003) and Cox and Monk (1993), among others. It also important that the

research team provides feedback to the community once the study is completed (Arean et al.,

2003). Sharing the research Eindings and assisting professionals in the community to translate

the findings into practice serves to inform consumers, professionals, and to foment an

atmosphere of trust and collaboration.

Alternatively, future investigators should consider conducting an examination on the

influence of these variables through different data analysis methods. The relationships among

the variables age, income, health, and attrition warrant future investigation. The large scale

nature of research like the REACH study allows the best probability of examining the strength of

association between variables of interest. The REACH data are available for research from the

Inter-university Consortium for Political and Social Research (ICPSR) at the University of

Michigan. Future research is needed in order to develop better models using predictors of drop-

out status, including attitudes and beliefs, and to explore alternative ways to identify predicators

of attrition and recruitment and retention strategies for Latino caregivers of persons with

Alzheimer' s disease.

As the scientific community continues to strengthen efforts to increase the recruitment

and retention of Latinos in research studies, health professionals need to keep up with new

Endings. While more information is needed to inform future research and practice, current

health professionals should translate these recommendations into practice. Thus, ensuring that

Latino caregivers of persons with Alzheimer' s disease are receiving adequate and culturally-

relevant services. It is necessary that health professionals, such as counselors, working with









Latino families understand and incorporate into their practice the values of this cultural group.

These include: the close-knit and extended nature of the Latino family, degree of acculturation

into mainstream society, and language proficiency, role of religion, perceptions about

Alzheimer's disease and research studies, among others. However, it is also critical to consider

differences within Latino groups. As several investigators (Mohoney, Cloutterbuck, Neary, &

Zhan 2005; Gallagher-Thompson et al., 2003) have pointed out, members of the Latino

population have the most geographically diverse origins. Moreover, it is important to keep in

mind the generational differences within a single family unit which may manifest as different

language preferences and degree of acculturation to mainstream society. Given the paucity of

evidenced-based research findings, clinicians must take responsibility to ensure they are well-

prepared to provide Latino families with the best available services. In order to gain culturally-

appropriate skills, clinicians may participate in diversity training or incorporate Latino

consultants in the advisory boards. If these were not available, the clinician should find a more

appropriate referral source to work with this population.

The implications regarding the expected increase in cases of Alzheimer's disease in the

Latino community are overwhelming. Therefore, in order to keep up with the increasing

numbers of Latino families dealing with Alzheimer's disease, investigators must strengthen their

efforts to design appropriate research, to translate cultural sensitivity into culturally competent

recruitment and retention endeavors to enlist Latino participants into research studies. As the

cost of caring for persons suffering from Alzheimer's disease and related dementias continues to

soar, it is imperative to advance the knowledge base about how caregivers carry out the demands

associated with caregiving over the long-term. It is only through the acquisition and application

of this knowledge that society will effectively and sensitively manage this looming health crisis.









APPENDIX A
CAREGIVER SOCIODEMOGRAPHIC INFORMATION



Caregiver Sociodemographic Information Questionnaire.

1. During which visit is this interview taking place?
1 () Baseline
2 () 6 month follow-up visit
3 () 12 month follow-up visit
4 () 18 month follow-up visit
5 ()Other

1.1 Specify

2. Date of interview: / /
Month/day/year

Now I would like to obtain some general information about you.

4. What is your marital status?
Never married (0)
Married or living as married (1)
Widowed, not currently married (2)
Divorced, not currently married (3)
Separated (4)
Unknown (-3)
Refused (-4)

4. 1. What is the primary occupation your spouse has had most of his/her
working life? Since many people have more than one j ob at a given time,
we would like to know about the j ob that is/was your spouse's primary
source of income.


4.1.1 Job Code









5. How many years of formal education did you complete?
No formal education (0)
Grade 1 (1)
Grade 2 (2)
Grade 3 (3)
Grade 4 (4)
Grade 5 (5)
Grade 6 (6)
Grade 7 (7)
Grade 8 (8)
Grade 9 (9)
Grade 10 (10)
Grade 11 (11)
Grade 12/ High school diploma/ GED (General Education Diploma) (12)
Vocational/ training school after high school (13)
Some college/ associate degree (14)
College graduate (4 or 5 year program) (15)
Master's degree (or other post-graduate training) (16)
Doctoral degree (PhD, MD, Ed.D., D.V.M., DDS., JD, etc.) (17)
Unknown (-3)
Refused (-4)

6. What country did you reside in during the last year of formal education?
United States (1)
Canada (2)
Cuba (3)
Mexico (4)
Other (5

6.1. Specify
Unknown (-3)
Refused (-4)

7. How would you describe your primary racial or ethnic group?
White, Caucasian (1)
Black, African-American (2)
Native American, Eskimo, Aleut (3)
Asian or Pacific Islander (4)
Hispanic, Latino (5









7. 1. Would you describe yourself as
Mexican, Mexican American, Chicano (1)
Cuban or Cuban American (2)
Puerto Rican (3)
Dominican (4)
Other (5)
Unknown (-3)
Refused (-4)

7.1.1. Specify:
No primary group (6)

7.2. Specify:
Other (7)


7.3. Specify:
Unknown (-3)
Refused (-4)

8. In which country were you born?
United States (1)
Canada (2)
Cuba (3)
Mexico (4)
Other (5)

8.1 Specify:
Unknown (-3)
Refused (-4)

9. How many years have you lived in the United States? Years

10. What is the primary occupation you have had most of your working life? Since many people
have more than one j ob at a given time, we would like to know about the j ob that is/was your
primary source of income.


10.1 Job Code









11. What is your current employment status?
Employed at a job for pay, full-time (1)
Employed at a job for pay, part time (2)
Homemaker, not currently working for pay (3)
Not currently employed, retired (4)
Not currently employed, not retired (5)
Unknown (-3)
Refused (-4)

1 1.1. Are you employed outside of the home?
No (0) Yes (1) Unknown (-3) Refused (-4

1 1.2. How many hours per week do you work at your paid job?

Hours: minutes

1 1.3. Have you had to reduce the number of hours that you work in an average
week in order to provide care to (CR)?
No (0) Yes (1) Unknown (-3) Refused (-4)

11.3.1. How many hours have you had to reduce per week?

Hours: minutes

11.4. Did you stop working because of (CR)'s need for care?
No (0)
Yes (1)
Unknown (-3)
Refused (-4)

11.4. 1. Why?
You anticipated having to care for (CR) (1)
To provide additional care for (CR) (2)
Other (3)

11.4. 1.1. Specify
Next, I would like to ask you about your household income. Some people may not be
comfortable answering this question, but I want to assure you that your responses will be kept
strictly confidential. This information is very important to the proj ect because it helps us
understand how caregiving affects people with different incomes.









12. Which category on this card [give respondent card] best describes your yearly
household income before taxes? Do not give me the dollar amount, just give me the
category. Include all income received from employment, social security, support
from children or other family, welfare, Aid to Families with Dependent Children
(AFDC), bank interest, retirement accounts, rental property, investments, etc.
Less than $5000 (0)
$5000 $9,999 (1)
$10,000 $14,999 (2)
$15,000 $19,999 (3)
$20,000 $29,999 (4)
$30,000 $39,999 (5)
$40,000 $49,999 (6)
$50,000 $59,999 (7)
$60,000 $69,999 (8)
$70,000 or more (9)
Unknown (-3)
Refused (-4)

13. How hard is it for you to pay for the very basics like food, housing, medical care,
and heating? Would you say it is:
Not dimfcult at all (1)
Not very dimfcult (2)
Somewhat dimfcult (3)
Very dimfcult (4)
Unknown (-3)
Refused (-4)

14. How many people are living with you in your home excluding yourself!
persons

15. How long have you lived with (CR)? years

16. Did you and (CR) start living together so that you could take care of him/her?
No (0
) Yes (1)
Unknown (-3)
Refused (-4)










APPENDIX B
CENTER FOR EPIDEMIOLOGIC STUDIES DEPRESSION SCALE (CES-D)

CES-D Questionnaire.

1. During which visit is this interview taking place?
1 () Baseline
2 () 6 month follow-up visit
3 () 12 month follow-up visit
4 () 18 month follow-up visit
5 ()Other
1.1 Specify

2. Date of interview: / /
Month/day/year

This section deals with statements people might make about how they feel. Let me give you a
card with possible responses. [Give respondent card.] For each of the statements, please indicate
how often you felt that way during the past week.

4.1. I was bothered by things that usually don't bother me.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.2. I did not feel like eating; appetite was poor.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.3. I felt that I could not shake off the blues, even with help from my family and friends.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)









4.4. I felt that I was just as good as other people.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.5. I had trouble keeping my mind on what I was doing.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.6. I felt depressed.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.7. I felt that everything that I did was an effort.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.8. I felt hopeful about the future.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)









4.9. I thought my life had been a failure.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.10. I felt fearful.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4. 11. My sleep was restless.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.12. I was happy.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.13. I talked less than usual.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)









4.14. I felt lonely.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.15. People were unfriendly.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.16. I enjoyed life.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.17. I had crying spells.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.18. I felt sad.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)









4.19. I felt that people disliked me.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

4.20. I could not get going.
Rarely or none of the time [< 1 day] (0)
Some or a little of the time [1-2 days] (1)
Occasionally or a moderate amount of time [3-4 days] (2)
Most or almost all of the time [5-7 days] (3)
Unknown (-3)
Refused (-4)

5. CES-D Score

6. Is the CES-D score greater than or equal to 28?
No ()
Yes ()

6.1 Has the Principal Investigator or appropriate site
personnel been notified?
No ()
Yes ()

Please notify the Principal Investigator or appropriate site personnel.










APPENDIX C
SOCIAL SUPPORT



Social Support (SS) Questionnaire.

1. During which visit is this interview taking place?
Baseline (1)
6 month follow-up visit (2)
12 month follow-up visit (3)
18 month follow-up visit (4)
Other (5)

1.1 Specify

2. Date of interview: / /
Month/day/year

Social Networks
Now I would like to ask you some questions about your friends and family.

4.0. Overall, how satisfied have you been in the last month with the help you have
received from friends, neighbors, or family members?
Not at all (0)
A little (1)
Moderately (2)
Very (3)
Unknown (-3)
Refused (-4)

4.1. How many relatives other than (CR) do you see or hear from at least once a
month?
None (0)
One (1)
Two ()
Three or four (3)
Five to eight (4)
Nine or more (5)
Unknown (-3)
Refused (-4)











4.2. Think about the relative other than (CR) with whom you have the most contact.
How often do you see or hear from that person?
Less than monthly (0)
Monthly (1)
A few times a month (2)
Weekly (3)
A few times a week (4)
Daily (5)
Unknown (-3
Refused (-4)

4.3. How many relatives other than (CR) do you feel close to? That is, how many do
you feel at ease with, can talk to about private matters, or can call on for help?
None (0)
One (1)
Two (2)
Three or four (3)
Five to eight (4)
Nine or more (5)
Unknown (-3)
Refused (-4)

4.4. How many friends do you feel close to? That is, how many friends (not
including relatives) do you feel at ease with, can talk to about private matters,
or can call on for help?
None (0)
One (1)
Two (2)
Three or four (3)
Five to eight (4)
Nine or more (5)
Unknown (-3)
Refused (-4)

4.5. How many of these friends do you see or hear from at least once a month?
(not including relatives)
None (0)
One (1)
Two (2)
Three or four (3)
Five to eight (4)
Nine or more (5)
Unknown (-3)
Refused (-4)











4.6. Think about the friend (not including relatives) with whom you have the most
contact. How often do you see or hear from that person?
Less than monthly (0)
Monthly (1)
A few times a month (2)
Weekly (3)
A few times a week (4)
Daily (5)
Unknown (-3)
Refused (-4)

4.7. When you have an important decision to make, do you have someone other than
(CR) you can talk to about it?
Never (0)
Seldom (1)
Sometimes (2)
Often (3)
Very often (4)
Always (5)
Unknown (-3)
Refused (-4)

4.8. When other people you know have an important decision to make, do they talk
to you about it?
Never (0)
Seldom (1)
Sometimes (2)
Often (3)
Very often (4)
Always (5)
Unknown (-3)
Refused (-4)

Received Support and Satisfaction

5.1. In the past month, how often has someone, such as a friend, neighbor, or family
member other than (CR), provided transportation for you?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)










5.2. In the past month, how often has someone, such as a friend, neighbor, or family
member other than (CR), pitched in to help you do something that needed to get
done, like household chores or yardwork?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.3. In the past month, how often has someone helped you with shopping?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.4. Overall, how satisfied have you been in the last month with the help you have
received with transportation, housework and yardwork, and shopping?
Not at all (0)
A little (1)
Moderately (2)
Very (3)
Unknown (-3)
Refused (-4)

5.5. In the past month, how often was someone right there with you (physically) in a
stressful situation?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.6. In the past month, how often has someone provided comfort to you?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)










5.7. In the past month, how often has someone listened to you talk about your
private feelings?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.8. In the past month how often has someone expressed interest and concern in
your well-being?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.9. In the past month, how satisfied have you been with the support received during
difficult times, comforting from others, how others have listened, and interest
and concern from others?
Not at all (0)
A little (1)
Moderately (2)
Very (3)
Unknown (-3)
Refused (-4)

5.10. In the past month, how often has someone suggested some action you should
take in dealing with a problem you were having?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.11. In the past month, how often has someone made a difficult situation clearer and
easier to understand?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)










5.12. In the past month, how often has someone helped you understand why you did
not do something well?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.13. In the past month, how often has someone told you what they did in a situation
that was similar to one you were experiencing?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

5.14. Overall, how satisfied in the last month have you been with the suggestions,
clarifications, and sharing of similar experiences you have received from others?
Not at all (0)
A little (1)
Moderately (2)
Very (3)
Unknown (-3)
Refused (-4)

I'd like to ask you a few more questions about your relationship with others.
Remember, when the term "others" is used, it includes friends, neighbors, or family
members other than (CR).

Negative Interaction

6.1. In the past month, how often have others made too many demands on you?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)










6.2. In the past month, how often have others been critical of you?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

6.3 In the past month, how often have others pried into your affairs?
Never (0)
Once in awhile (1)
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)

6.4 In the past month, how often have others taken advantage of you?
Never ()
Once in awhile ()
Fairly often (2)
Very often (3)
Unknown (-3)
Refused (-4)




Full Text

PAGE 1

PREDICTORS OF ATTRITION IN LATINO ALZHEIMERS DISEASE CAREGIVERS IN THE REACH TRIAL: AN ARCHIVAL INVESTIGATION By IVETTE MAGALY FREYTES CACHO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007 1

PAGE 2

Copyright 2007 Ivette Magaly Freytes Cacho 2

PAGE 3

To my parents, mami & papi, with love 3

PAGE 4

ACKNOWLEDGMENTS I would like to thank Dr. Ed il Torres-Rivera and Dr. Sondra Smith, for chairing my committee and for their support and guidance throughout this study. I also want to thank Dr. Mary Fukuyama and Dr. Leilani Doty for their he lpful feedback and suggestions and Dr. Mirka Koro-Ljungberg for graciously agreeing, at the la st minute, to fill an unexpected vacancy on my committee. I am indebted to the original investigators of the Re sources for Enhancing Alzheimers Caregiver Health (R EACH) study for the contribution of the data and their expertise to this study. I would also lik e to thank Jocelyn Lee and Judy Callahan for their input regarding the data analysis. I am especially grateful to Betsy Pearman for her encouragement and her priceless contribution and assist ance with the daunting task of data analysis throughout the process. I would like to express my deepest apprecia tion to Dr. Timothy Ke tterson, who mentored me throughout the years and whose ongoing supp ort, guidance, and encouragement were essential in the undertaking and completion of this dissertation. He spent countless hours helping me through this journey and always believed in my ab ility to finish this task. I would also like to thank th e many friends who offered th eir support throughout all these years. Specifically, I want to thank Jim and Patty Probert for their ongoing support and love, and the countless and delicious meals they provided th roughout this process. I want to thank Dana Myers for giving me a social outlet when I was burned out and for cons istently showering me with her good vibes. I would like to thank Alice Martin and Debbie Gipple for always lending a listening ear and for helping me to find the humor in this process. 4

PAGE 5

Last but not least, I want to extend my deepes t appreciation to my pa rents for believing in me even when I doubted myself and encouraging me when I wanted to give up. Without their unconditional love and support, and their relentless faith in me, I ne ver would have come this far. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 ABSTRACT.....................................................................................................................................9 CHAPTER 1 INTRODUCTION................................................................................................................. .11 Scope of the Problem........................................................................................................... ...12 Statement of the Problem....................................................................................................... .14 Need for the Study..................................................................................................................16 Purpose of the Study........................................................................................................... ....17 Theoretical Framework.......................................................................................................... .18 Overview of the Resources for Enhanci ng Alzheimers Caregiver Health Study..................19 Hypotheses..............................................................................................................................20 Definition of Terms................................................................................................................21 Overview of the Remainder of Study.....................................................................................21 2 REVIEW OF RELEVANT LITERATURE...........................................................................22 Attrition Rates and Latinos.....................................................................................................24 Latino Caregivers Age...........................................................................................................27 Latino Caregivers Level of Education...................................................................................29 Latino Caregivers Income Level...........................................................................................30 Depression in Latino Caregivers............................................................................................31 Social Support and Latino Caregivers....................................................................................32 Latino Caregivers Health Status............................................................................................34 Summary and Conclusions.....................................................................................................36 3 METHODS...................................................................................................................... .......38 Resources for Enhancing Alzheimers Caregiver Health.......................................................38 Sample Characteristics............................................................................................................39 Variables of Interest an d Operational Definitions..................................................................40 Null Hypotheses................................................................................................................ ......42 Research Procedures............................................................................................................ ...43 Measures.................................................................................................................................43 Data Analysis..........................................................................................................................44 Methodological Limitations....................................................................................................4 5 6

PAGE 7

4 RESULTS...................................................................................................................... .........47 Preliminary Analyses..............................................................................................................47 Palo Alto Correlations.....................................................................................................49 Miami Correlations..........................................................................................................50 Hypothesis Testing............................................................................................................. ....50 Data Analysis..........................................................................................................................51 Palo Alto Sample.............................................................................................................52 Miami Sample.................................................................................................................54 Summary of Results............................................................................................................. ...56 5 CONCLUSION................................................................................................................... ....74 Conclusions.............................................................................................................................74 Palo Alto...................................................................................................................... ....75 Miami..............................................................................................................................75 Implications................................................................................................................... .........76 Limitations.................................................................................................................... ..........78 Recommendations................................................................................................................ ...80 APPENDIX A CAREGIVER SOCIODEMOGR APHIC INFORMATION..................................................84 B CENTER FOR EPIDEMIOLOGIC STUDI ES DEPRESSION SCALE (CES-D)................89 C SOCIAL SUPPORT............................................................................................................... 94 D CAREGIVER HEALTH AND HEALTH BEHAVIOR......................................................101 REFERENCE LIST.....................................................................................................................105 BIOGRAPHICAL SKETCH.......................................................................................................114 7

PAGE 8

LIST OF TABLES Table page 4-1 Variable means for total group and by location (Palo Alto and Miami)...........................57 4-2 Variable means for total group and dropouts and completers...........................................58 4-3 Income level group for co mpletion status and location.....................................................59 4-4 ANOVA test between sites (Palo Alto and Miami)...........................................................60 4-5 ANOVA test between treatment dropouts and treatment completers................................61 4-6 Palo Alto correlations........................................................................................................62 4-7 Miami correlations........................................................................................................ .....63 4-8 Palo Alto logistic regression, model one...........................................................................64 4-9 Palo Alto logistic regression, model two...........................................................................65 4-10 Palo Alto logistic regression, model three.........................................................................66 4-11 Palo Alto logistic regression, model four..........................................................................67 4-12 Palo Alto logistic regression, model five...........................................................................68 4-13 Miami logistic regression, model one................................................................................69 4-14 Miami logistic re gression, model two................................................................................70 4-15 Miami logistic re gression, model three..............................................................................71 4-16 Miami logistic re gression, model four...............................................................................72 4-17 Miami logistic re gression, model five...............................................................................73 8

PAGE 9

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREDICTORS OF ATTRITION IN LATINO ALZHEIMERS DISEASE CAREGIVERS IN THE REACH TRIAL: AN ARCHIVAL INVESTIGATION By Ivette Magaly Freytes Cacho May 2007 Chair: Edil Torres-Rivera Cochair: Sondra Smith Major: Mental Health Counseling Progressive memory disorders are becoming a national epidemic. It is estimated that 75% of persons with memory disorders are cared for informally, at home, by a friend or family member. Caring for someone with cognitive impairment is a demanding and challenging responsibility that places tremendous strain on families. Due to the changing demographics in the composition of the U.S. population, it is expected that a significant proportion of elders with dementia and their caregivers wi ll be of diverse backgrounds. Latinos, the largest and fastes t growing segment of the U.S. population, are expected to account for 25% of the dementia-affected cohort ; however, little is known about Latino dementia caregivers. Recruiting and retaini ng Latinos into research studies continues to be a challenge for investigators. This study was a secondary analysis of the Resources for Enhancing Alzheimers Caregiver Health (REACH) trial, a multi-site randomized trial that developed and tested the effectiveness of multiple interventions for family caregivers of persons with Alzheimers disease and related disorders. The focus of this study wa s to identify variables as sociated with attrition from the REACH trial in Latino participants at the Palo Alto and Miami sites. Six variables (i.e., age, educational level, income, depression score, social support, and health status) believed to predict attrition (i.e., drop-out status) of Latino dementia caregiv ers from the REACH trial were 9

PAGE 10

examined. None of these variables were predictive of attrition for these two samples. Although the outcomes of this study did not contribute new information concerning which socioeconomic, demographic, and client factors are associated with the attriti on of Latino dementia caregivers from the REACH trial, these findings indicate so me trends and thus contribute to the existing literature on Latino dementia caregivers and se rve to inform future research studies. The understanding of the variables examined in this study are embedded in the cultural context, thus, it is recommended that future inqui ries into Latino dementia caregivers allow for the measurement of cultural considerations. Due to the archival nature of this inquiry, the hypotheses in this study were not po wered sufficiently to test the re lationship of the variables of inquiry and data were not purposefully collected. 10

PAGE 11

CHAPTER 1 INTRODUCTION Progressive memory disorders, such as Alzheimers disease (AD), cardiovascular dementias, and many other related disorders, are becoming a national epidemic. Progressive memory disorders refer to a cluster of conditions that worsens over time, negatively impacts the brain, and impairs the functioning of individuals. Currently, it is estimated that there are more than 5 million people living with Alzheimers disease in the United States (Alzheimers Association, 2007; USA Today, 2007) and this number is expected to more than double in the coming years. In Florida, over 479,875 individu als are believed to suffer from Alzheimers disease (Department of Elder Affairs, 2006). As Alzheimers disease and related disorders continue to affect increasing num bers of people, it is estimated that a larger pr oportion of those will be culturally or ethnically diverse. Many of these individuals will be of Latino origin. The terms Hispanics and Latino are frequent ly used interchangeably and refer to a heterogeneous group that share Spanish as a comm on language. This study used the term Latino rather than the term Hispanic, as Latino is an inclusive term that has been used by people of Latin/Hispanic descent to empower themselves in political arenas, humanities, and literature (Oboler, 1995; Shorris, 1992; Torres, 2004; Torres-Rivera, Wilbur, Roberts-Wilbur, & Phan, 1999). Perhaps the biggest impact of Alzheimers disease (AD) is on the family and other caregivers. Due to the debilitating nature of progressive memory disorders (e.g., AD), individuals affected by these types of ailments require increasing assistance with daily living skills and close supervision. The majority of individuals suffe ring from Alzheimers disease receive care at home. The Family Caregiving Alliance (2001) reports th at 34 million adults, or 16 percent of the population, provide care to adults 50 years of age or older. This type of unpaid 11

PAGE 12

caregivers is referred to as informal or family caregivers. Statistical data provided by the Alzheimers Association (2005) in dicate that seven out of ten people with Alzheimers disease live at home in the care of family or friends. According to Schulz (2000), this translates into over 5 million households in the U.S. providing care for a loved one with dementia. A review of the caregiving lit erature shows that the terms caregivers and caregiving are broadly defined and refer to providing a wide ra nge of care and assistance. These range from providing care with activiti es of daily living such as personal hygiene, toil eting, eating, assisting with other routines such as ma naging money or assets, to providi ng total care to individuals who are not able to function independently (Ga llagher-Thompson, Coon, Solano, Amber, Rabinowitz, & Thompson, 2003; National Alliance for Caregiving & American Association of Retired Persons, 2005; Family Caregiver Alliance, 2001; Shultz, 2000). The demands associated with providing care to a person with chroni c medical conditions such as Alzheimers disease and related disord ers may place a significant burden on caregivers. Research findings show that due to the complex de ficits of a person with dementia (i.e., declines in intellectual functioning and self-care behaviors, personality changes, and problematic behaviors), caregivers of persons with Alzheimers disease and related disorders experience higher levels of stress and a number of additi onal negative outcomes such as poor health, depression, and emotional strain (Argimon, Lim on, Vila, & Cabezas, 2004; Family Caregiver Alliance, 2001; Haley, Levine, Brown, & Bart olucci, 1987; Gallagher-Thompson, Coon, Solano, Ambler, Rabinowitz, & Thompson, 2003; Gallagher-T hompson et al., 2003; Mitrani et al., 2006; Schulz & Beach, 1999). Scope of the Problem The Alzheimers Association (2004) reports that Latinos have a greater risk of developing Alzheimers disease and related disorders. In Florida, the Department of Elder 12

PAGE 13

Affairs estimates that there are currently over 34,000 Latinos diagnosed with Alzheimers disease (electronic communication, September 29, 2006). By 2050 the Alzheimers Association (2004) estimates that over one million La tinos nationwide will be afflicted with this devastating illness. Alzheimers disease, the most common progressive memory disorder, is an irreversible condition that adversely affects th e intellectual functioning of the brain and eventually results in major thinking and physical impairments as well as changes in personality. As the number of persons affected with this devastating illness increases, so does the number of informal caregivers helping a loved one with Alzheimers disease. Two demographic trends contribute to the gr owing numbers of people with Alzheimers disease in the Latino population. Fi rst, the natural chronol ogical process of aging appears to be a strong risk factor for Alzheimers disease. There has been a remarkable increase in life expectancy over the last century. In addition, the number of older individuals is steadily rising. By 2011, the first wave of adults in the ba by boomer generation (born between 1946 -1964) will turn 65 years of age (Center for Health Communication Harvard School of Public Health & MetLife Foundation, 2004). As the U.S. popula tion ages, the Alzheimers Association (2004) estimates that by age 65, one out of every ten persons will develop Alzheimers disease. Additionally, by age 85, one out of every two pers ons will develop Alzheimers disease (Evan et al., 1989). As the incidence of Alzheimers diseas e rapidly increases with the numbers expected to more than double by the year 2050, upwards of sixteen million people will be affected (Alzheimers Association, 2004). If these predictions come true, Al zheimers disease is likely to become a devastating clinical and public health issue over the next 20 years. The second change is the sudden increase of Latinos in the United States. The recent growth of the Latino population in the United States is one of the most dramatic demographic 13

PAGE 14

increases that has taken place in the history of th e United States. No other racial or ethnic groups have had as great an impact on the demography of the United Stat es over the last century as Latinos. The Pew Hispanic Center (2005) reports that the 2000 Census marked the [Latino] population at 35.3 million people, an increase of 58 percent over 1990 (p.2). Moreover, they predict that by the end of this decade, there will be over 47 million Latinos in the U.S. and over 60 million by 2020. Furthermore, it is estimated that by the year 2050, one out of every four Americans will be of Latino heritage. The combination of these two changes is significant because according to the 2000 census, Latinos are the fastest growing segment of the general population as well as the fastest growing group of elders 65 years of age and olde r. Latinos make up almost ten percent of the total baby boomer generation (Metropolitan Life Insurance Company, 2005). Projections indicate that by the year 2050, th ere will be 12.5 million Latinos elders 65 and older in the U.S. (Gallagher-Thompson, Solano, Coon, & Aren, 2003). These numbers are sign ificant due to the growing needs of elders and the higher risk of developing Alzheimers disease in the elder population. Thus, the health care community will need to be able to meet the needs of this growing segment of the population, especially the needs of caregi vers, in order for them to continue providing the vast majority of care for their elders suffering from Alzheimers disease. Statement of the Problem Due to the high incidence of Alzheimers disease in ethnic minority elders, the prevalence of informal caregivers is on the rise. The Alzheimers Association (2004 ) reports that Latinos are believed to be at a greater risk of developing progre ssive memory disorders due to a higher incidence of cardiovascular disease, be lieved to be a significant risk factor for cardiovascular dementia and Alzheimers diseas e. Sociodemographic fa ctors such as lower education and limited access to health care, also are believed to be risk factors for Alzheimers 14

PAGE 15

disease. As the numbers of Latino elders continue to climb, so does their need for care. Although reliable estimates are difficult to find, current estimates by the Alzheimers Association (2004) indicate that close to 200,000 Latinos nationw ide suffer from Alzheimers disease and related disorders. These data illustrate the impor tance of continuing efforts to advance knowledge about the sequelae of Alzh eimers disease and underscore the need to include culturally diverse groups in the investigation in order to develop effective methods to meet their needs. However, a review of the lite rature shows that although minorities, specifically Latinos, are the fastest growing se gment of the population, they c ontinue to be underrepresented in research studies. Furthermore, the studies that have been succe ssful at recruiting and retaining Latino caregivers show mixed results. Although there has been a boom in research studies on Alzheimers disease over the last decades and efforts have been put forward to include culturally diverse populations, there continues to be a gap in the li terature on how this disease aff ects minority populations and their families. One of the main problems associated w ith the paucity in research studies with minority populations involves failures to re cruit and retain older members of minorities into research studies. Researchers maintain that this is particularly true for Latino caregi vers due to the stigma associated with mental illness and the time consuming and labor intensive burden placed on them (Aren, Alvidrez, Nery, Estes, & Linkins, 2003; Gallagher-Thompson, Solano, Coon, & Aren, 2003). The steady increase of Latinos in the U.S. warrants the need to examine the Alzheimers disease caregiving experience of this segment of the population as well as the barriers to accessing services. As the prevalence of Alzheimers disease a nd other progressive memory disorders grows exponentially over the next decades, it will compromise the already strained healthcare system, 15

PAGE 16

the long-term care system, and the economy. Due to the high incidence of Alzheimers disease in the Latino community and the lower rates of health insurance coverage and healthcare-seeking behaviors in this community, it is imperative for health care professionals to understand how cultural values and beliefs affect the willingness of family caregivers to seek and maintain connections for help in order to minimize the negative impact of Alzheimers disease on patients, families, the healthcare system, and society at large. Though seeking help is an important research issue, this study focused on the problem of barriers to retention, more specifically, the barriers of retaining informal Latino Alzheime rs disease caregivers in research studies. Need for the Study Research studies persistently show that while the numbers of Latino families affected by Alzheimers disease continue to grow, researchers have not been able to recruit and retain proportional numbers of Latino Alzh eimers disease patients and caregivers in studies. By 2010, it is expected that ethnoculturally diverse persons will represent 25% of the dementia-affected cohort and by 2050 they will number 33%, as compar ed to 16.7 % of the total dementia-affected elderly population in 2000 (Valle & Lee, 2002, p. S64) If the predictions come true, this illness is likely to become a devastating clinical and public health issue over the next decades. Studies show that Latinos seek health services at a much lower rate compared to other groups despite their high incidence of cardiovascu lar disease and related illnesses (Alzheimers Association, 2004; Larkey, Hech t, Miller, & Alatorre, 2001; Li ao, Tucker, & Giles, 2004). Despite the rapid growth of the Latino population, public sector responses to the health service needs of this population have developed slowly and do not co mmensurate with the growth, needs, and grim circumstances of the populati on. In addition, the Kais er Family Foundation (2003) reported that Latinos have the lowest in surance rates when compared to other racial groups in the U.S. The gap in health insurance coverage is due in part to the disadvantaged 16

PAGE 17

economic position of ethnic groups with a greater toll on Latinos (Center for Studying Health System Change, 2004). Research findings show that the Latino population in the U.S. is not receiving appropriate health care necessary to reduce the risk or manage progressive memory disorders. Professional groups, such as the Al zheimers Association (2 004) warn that ignoring the threat of dementia in the Latino community may lead to a health crisis for this segment of the population in the United States. These discrepancies have maintained the vast racial and ethnic disparities in Latino access to health care. Therefore, due to the rapid increase in La tino elders and the alarming incidence of progressive memory disorders (e.g., Alzheimers disease), it is imperative that health care professionals identify barriers to recruitment and retention of Latinos into research studies and subsequent treatment protocols necessary for impr oving the delivery of healthcare services to the Latino population and thus, bridge th e gap in health disparities. Unfortunately, research studies tend to place an inordinate burden on already ta xed research participants. Therefore, it is necessary to advance the current knowledge abou t how the burden of m easurement and barriers to participation contribute to attrition from existin g research studies. Also, it is vital to revise current procedures into culturally sensitive and appropriate new research and applied practices. Purpose of the Study Due to the changing demographics of the U.S., especially the fast gr owing rate of Latinos and the silver tsunami (Fryling, 2006) that is star ting to take place, it is imperative to explore these changes and the implications to the health care system and society at large in order to advance our understanding about barriers to prov iding care in the Latino community. The need for scientific examination about the obstacles in recruiting and retaini ng minority participants into research studies has been well documented (Aren, Alvidrez, Nery, Estes, & Linkins, 2003; Gallagher-Thompson et al., 2003; Warren-Findlow, Prohaska, & Freedman, 2003). Thus, it is 17

PAGE 18

important to identify the factors that predict the attrition rate from research studies in order to gain a better understanding about how these influence the attitude s of Latino Alzheimers disease caregivers on accessing and utilizing services. This knowledge is needed in order to bridge the gap in health and health care disparities am ong Latino Alzheimers di sease caregivers. This study examined the barriers of retent ion, more specifical ly retaining Latino Alzheimers disease caregivers in research studies Data collected for the Miami, FL, and Palo Alto, CA, sites of the Resources for Enhancing Alzheimers Caregiver H ealth, also known as the REACH I study, were analyzed to examine and cont rast attrition rates among two different sets of Latino Alzheimers disease caregivers. This researcher was interested in identifying the factors that may contribute to Latino Alzheimers disease caregivers prematurely dropping out of research studies that may yield subsequent trea tment approaches to improve the well-being of both the Alzheimers disease caregiver and the care recipient. Theoretical Framework This study used a combination of two stress process models to guide its inquiry. First, it borrowed the stress process model used in the Resources for Enhancing Alzheimers Caregiver Health (REACH) study. The REACH study (Czaja, Sc hulz, Lee, & Belle, 2003) adapted a stress process model of caregiving to in clude environmental stressors to health outcomes in dementia caregiving. According to the investigators, this model establishe s that the func tional limitations and behavioral problems of th e care recipient in conjunction w ith social and environmental stressors make up the primary stressors for demen tia caregivers. As caregivers engage in their caregiving responsibilities, they perceive and identify stressors a ssociated with caregiving. Then they evaluate these demands, as well as their ability and resources to meet these demands. If the caregivers feel helpless and unable to cope with these demands, they experience stress. The appraisal of stress is assumed to contribute to negative emotiona l, psychological, and behavioral 18

PAGE 19

responses that put the individua l at risk for physical and psyc hiatric disease(Czaja et al., 2003, p. 386). The second relevant stress process model is the one advanced by Haley, Levine, Brown, and Bartolucci (1987) that propos ed a model adapted from the st ress response model to explain caregivers help-seeking behaviors. According to the authors, caregivers appraisal of stressors, personal coping responses, and social support, medi ate between stress and caregiver outcome. Use of coping mechanisms such as seeking in formation, using problem solving, and emotional release may also help the caregiver to manage his or her relative more effectively, and lessen emotional distress (Haley et al., 1987, p. 323). If caregivers have access to and use available resources, they may increase the probability of obtaining better outcomes in coping with the chronic stress of caregiving. These stress and co ping models are useful in the conceptualization of the process of dementia caregi vers participation in research studies and their utilization of services. Once stressors and th e caregivers assessment of th ese are identified, they may be changed or modified (Schulz et al., 2003). Overview of the Resources for Enhancing Alzheimers Caregiver Health Study The Resources for Enhancing Alzheimers Caregivers Health (REACH) was a 5-year, multi-site research project conducted with suppor t from the National Institutes on Aging and the National Institute of Nursing Research. Six sites, including Birmingham, Boston, Memphis, Miami, Palo Alto, and Philadelphia, developed and tested the effectiveness of multiple interventions for family caregivers of persons w ith Alzheimers disease a nd related disorders. Each site investigated a different interven tion or a combination of interventions. The interventions included (1) individual informa tion and support strategies, (2) group support and family systems therapy, (3) psychoeducational and skills-based training approaches, (4) homebased environmental inte rventions, and (5) enhanced techno logy support systems. REACH was 19

PAGE 20

designed to examine the feasibility and outcomes of multiple different intervention approaches, . . (Schulz et al., 2003). A total of 1,222 part icipants were recruited with a strong emphasis on recruiting minority caregivers. Fo r the purpose of this study, only data collected from the Miami and Palo Alto sites were examined. Hypotheses This study was a secondary analysis of a mu lti-site prospective randomized controlled trial in which data collected from the REACH project were used to examine predictors of attrition of Latino Alzheimers disease (AD) caregiver s from research studies. Data used for this examination included (1) caregivers age, (2) caregivers level of education, (3) income, (4) depression, (5) social support, (6), care giver health, and (7) attrition rate. The following six null hypotheses were tested in this study: Ho1: There is no relationship between attrit ion drop-out status (attrition) and Latino caregivers age. Ho2: There is no relationship between drop-out status (attrition) a nd Latino caregivers achieved level of education. Ho3: There is no relationship between drop-out status (attrition) a nd Latino caregivers income level. Ho4: There is no relationship between drop-out status (attrition) and depression among Latino caregivers as measured by Center for Epidemiologic Studies Depression Scale (CES-D). Ho5: There is no relationship between dropout status (attrition) and social support among Latino caregivers. Ho6: There is no relationship between drop-out status (attrition) a nd Latino caregivers health status as measured by the Caregive r Health and Health Behaviors Questionnaire from the REACH study. 20

PAGE 21

Definition of Terms Alzheimers disease. A progressive brain disorder that impairs a persons intellectual functioning, that is, memory and the persons abi lity to learn, reason, and communicate. It also impairs the persons judgment, behaviors, and eventu ally, the ability to care for self (Alzheimers Association, 2006). Dementia. An acquired set of symptoms that negatively affect a persons intellectual functioning such as thinking, memory and judgm ent (Alzheimers Caregiver Support Online, 2006). Progressive memory disorders. Refers to disorders of the br ain, such as Alzheimers disease (AD), that cause worsening of a persons intellec tual abilities over time (Alzheimers Caregiver Support Online, 2006). Reach study. A 5 year study conducted in six cities acr oss the U.S to measure the effectiveness of different dementia caregiver in terventions (Schulz et al., 2003). Silver tsunami. A term used by many to describe the tidal wave of baby boomers that is about to have a significant soci etal impact (Fryling, 2006). Overview of the Remainder of Study The remainder of this study consists of four chap ters. Chapter 2 provides a review of the related literature. Chapter 3 contains a description of the specific procedures for the study, including methodology, subjects, and research design. In Chapter 4, the result s of the study are presented. Chapter 5 includes a discussion of the results, conclusions, implica tions, limitations, and recommendations for future research. 21

PAGE 22

CHAPTER 2 REVIEW OF RELEVANT LITERATURE The purpose of this chapter is to review the published literatu re pertinent to the proposed investigation. This chapter is organized into se ven sections: (a) attrit ion rates and Latinos, (b) Latino caregivers age,(c) Latino caregivers leve l of education, (d) Lati no caregivers income, (e) Latino caregivers depression, (f) Latino caregivers social s upport, (g) social support of Latino Alzheimers disease caregivers and (h) Latino caregivers health status. As the U.S population grows older and more diverse, Latinos are becoming one the fastest growing segments of the population. A ccording to the 2000 U.S. Census, (U.S. Census Bureau, 2001) the Latino population has increased dramatically over the last decade, becoming the largest minority group in the country. Census estimates indicate there are over 35 million Latinos living in the U.S. Of those, the U. S. Administration on Aging (2004) reports 4% are Latino elders 65 years of age and older and projects that by 2020 this percen tage will increase to over 14%. By 2025, one in six elde rs living in the U.S. are exp ected to be Latinos (Horizon Project, 2000) and by the year 2050, Latinos 65 years of age and older will reach the 12.5 million mark (Gallagher-Thompson et al., 2003). Current estimates indicate that 44% of Latino elders seventy and older receive care at home by a family member or friend (Weiss, Gonzalez, Kabeto, & Landa, 2005). The graying and diversification of the U.S. demographics are noteworthy due to high incidence of chronic conditions, such as Alzheime rs disease, in this se gment of the population (Weiss, Gonzalez, Kabeto, & Langa, 2005). The Alzheimers Associations (2004) projects an increase from 200,000 Latinos with Alzheimers disease to over 13 million by 2050. There is a common misconception that most elders who re quire long-term care receive assistance from formal caregiving facilities, such as nursing hom es; however, estimates indicate otherwise. Over 22

PAGE 23

75 % of long-term care for disabled adults takes place at the home of a family member or friend (Family Caregiving Alliance, 2001; Gallagher-Thomps on et al., 2003). Currently, it is estimated that 44 % of Latinos 70 years of ag e and older receive informal care at home (Weiss et al., 2005). As the Latino population continues to expand and gr ow older, so will the number of informal Latino caregivers. Alzheimers disease is a devastating illness that adversely impacts the lives of patients and family caregivers alike. Alzheimers diseas e caregivers assume challenging responsibilities, for example, assisting the patien t with activities of daily living such as grooming, toileting, and bathing. Often times, caregivers also engage in activities that involve role reversals such as managing the finances for a parent, preparing a m eal for a wife, or driving a husband to medical appointments. In addition, caregiving respons ibilities are time consuming. Thus, often caregivers will neglect their own needs (Mille r & Mukherjee, 1999; Cox & Monk, 1996), health (Gallagher-Thompson, 2003; Pinqu art & Sorensen, 2005), and social relationships (Cox & Monk, 1996). Curry and Jackson (2003) report that compared to White caregivers, Latino caregivers are more likely to care for a family member instead of caring for a friend. The implications of providing care for someone with a ch ronic condition are significant. Research findings suggest that compared to other racial groups, Latino caregivers provide more hours of informal caregiving (Weiss et al., 2005) and ca re for longer periods of time (Pinquart & Sorensen, 2005). Consistent research findings suggest that La tino families have higher rates of caregiving at home for persons with Alzheimers disease than any other groups (Covinsky, Newcomer, Fox, Wood, Sands, Dane, & Yaffe, 2003; Markides, R udkin, Angel, & Espino, 1997), despite higher levels of physical and psychological distress associated with the physical and cognitive 23

PAGE 24

limitations of the patient. These numbers are significant because providing care for a person with Alzheimers disease has been associated with negative caregiver outcomes such as poor health, higher levels of depression, and decreased suppor t (Gallagher-Thompson, 2003), in addition to increased mortality (Schulz, O Brien, Bookwala, & Fleissner, 1995; Schulz, Visintainer, & Williamson, 1990). The existing literature on how Alzheimers disease impacts the lives of culturally diverse persons and their participation in research studies is scarce comp ared to the available literature on Whites. Inclusion of psychosocial and cultural variables such as age, socio-economic status, educational attainment, health-status, and social support are crucial to un derstand the impact of Alzheimers disease on the Latino population. Wh en looking at these psychosocial and cultural variables, two important consider ations need to be noted: first, differences among groups and second, degree of acculturation. First, it is important to be cognizant of the differences am ong Latino groups in order to develop appropriate studies and interventions Although these groups share many psychosocial characteristics, differences among groups and their needs have been documented (Briones, Ramirez, Guerrero, & Ledger, 2002; Gallaghe r-Thompson, Solano, Coon & Aren, 2003; Valle & Lee, 2002). Acculturation to mainstream soci ety also has significant implications for families caring for relatives with progressive memory disorders (Galla gher-Thompson et al., 2003; Valle & Lee, 2002). While these differences warrant careful examination, well-designed research studies including Alzheimers disease patients and their families across these cultures will add to the knowledge base regarding the disease a nd its implications for Latino groups. Attrition Rates and Latinos It is well documented that members of dive rse groups are underrepr esented in research studies (Aren, Alvidrez, Rowena, Estes, & Linkins, 2003; Aren & Gallagher-Thompson, 1996; 24

PAGE 25

Curry & Jackson, 2003; Gallagher-Thompson, Sola no, Coon, & Aren, 2003). The significance of this issue was underscored in 1994 when th e National Institutes of Health (NIH) issued guidelines for the inclusion of minor ity participants in clinical re search trials (Curry & Jackson, 2003). Although there has been a boom in research studies on the impact of Alzheimers disease over the last decades, and efforts have expanded to include culturally diverse populations, there continues to be a gap in the literature on how this disease aff ects minority populations and their families. At the forefront of this issue is the failure to recruit and retain Latino caregivers into research studies. The inability to recruit and re tain Latino caregivers in research studies has a detrimental effect on the ability of health care professionals to develop informed, culturally relevant programs to assist and alleviate the burden associated with caregiving. Furthermore, current findings that lack adequa te numbers of ethnic participants cannot be generalized to the population without questioning the validity of the researchers conclusions (Levkoff & Sanchez, 2003). Few studies report successful implementa tion of recruitment and retention methods with older populations (Aren, Alvi drez, Rowena, Estes, & Link ins, 2003; Gallagher-Thompson, Solano, Coon, & Aren, 2003). Aren and associates (Aren et al., 2003) exam ined recruitment and retention rates in two separate studies on older minority populations in order to compare the trad itional recruitment and retention methods with a consumer-centered a pproach. They found that the mainstream older research participants tend to respond positively to traditional recruitment and retention methods that focus on eliminating practical barriers to re search studies such as offering transportation to and from the research site as well as financial incentives. Ethnic older adults responded more favorably to the consumer-center recruitment mode l, which also resulted in greater retention rates. This model proposes that in order to successfully attract the participation of minority 25

PAGE 26

elders into research studies, the following conditions must be met: securing the support and buyin from local community leaders; enlisting rese arch staff that either knows or belongs to the same ethnic group as the target population; anti cipating and alleviating participants burden associated with the research study; and providi ng feedback to the community once the study is completed. In a recent study by Gallagher-Thompson and co lleagues (2004) aimed at comparing the effectiveness of three differen t recruitment methods among White and Latino female caregivers, results were consistent with th e prior conclusions that favored consumer-centered methods over traditional approaches. The authors believe that th e Latinos cultural values of trust (confianza) and close relationships (personalismo) demand th e development and establishment of consumercentered recruitment and retention methods derived from partnerships with local Latino-specific organizations. Others such as Henderson, Gutierrez-Myka, Garcia, and Boyd (1993) have put forward similar efforts and developed a program to recrui t minority caregivers to participate in existing support groups for Alzheimers caregivers who had no minority representation. This effort consisted of the following: training group l eaders about culturally diverse populations; conducting purposeful advertisement and promo tion using the media, presentations, and endorsement from community professionals and organizations; and usi ng appropriate meeting locations. The authors indicated that 24 mont hs later, attendance by African Americans and Hispanics rose dramatically. In addition, a meta -analysis on the effectiveness of interventions with caregivers has some signi ficant revelations. Accordi ng to Sorensen, Pinquart, and Duberstein (2002) the average dr op-out rate across the 78 studie s they reviewed was less than 26

PAGE 27

20%, and most caregivers had at least a high school education, and were ethnic spouses or adult children. Many explanations have been offered regardi ng the dismal participation of Latinos in research studies including mistru st of ethnic groups toward th e scientific community, language barriers, and the burden placed on research participants (Aren et al., 2003; Gallagher-Thompson et al., 2003; Levkoff & Sanchez, 2003). However, caregiver-specific factors such as age, educational attainment, income level, depression, social support, and health status variables are also believed to influence the limited participatio n, more specifically the drop-out rate of Latinos in research studies. Currently, there is little information about how these psychosocial factors (i.e., caregiver age, achieved educational level, health status, level of depression, and social support) affect Latino caregivers participation in longterm research studies (Covinsky, Newcomer, Fox, Wood, Sands, Dane, & Yaffe, 2003 ). The results of studies with Latinos from different countries of origin need careful cons ideration due to the heterogeneity among Latino participants. Latino Caregivers Age It is estimated that, by the year 2050, the number of Latino elderly could reach twelve million persons, representing 15% of all Latinos living in the U.S. The age demographics of dementia caregivers seem to be in flux. A study by Schulz and Beach (1999) found that most caregivers are middle-aged adult ch ildren and spouses caring for a parent or spouse. In 2001, the Family Caregiver Alliance reported that the average age for caregi vers of persons 50 years of age and older was 47 and the average age of caregiv ers of persons 65 and older was 63 years old. More recently, an Alzheimers Association (200 4) publication about Alzheimers caregiving in the U.S indicates that the average age for an Al zheimers disease caregiver is 48 years old and that most caregivers provide care for an older relative. 27

PAGE 28

A comprehensive inquiry into caregivi ng in the U.S. conducted by the National Caregiving Alliance and the American Association of Retired Persons (2005) estimates that over 44 million caregivers 18 years of age and older in the U.S. provide unpaid care to an adult family member or friend who is also 18 or older. Of these, 8% of all careg ivers reported caring for someone with Alzheimers disease; 16% of careg ivers are 50 years and older and an estimated 12% are Latinos. A paucity of research studies look at th e relationship between age and the Latino caregivers attrition rate from such studies. The limited av ailable published studies show mixed results concerning the association between age and attrition rates in Latino research participants. Therefore, the present review of the literature examined the eff ects of age on attrition rates in other areas and other segments of the population. This approach id entified several studies across racial and ethnic groups reporting an inverse relationship between age and drop-out rate (Honas, Early, Frederickson, & OBrien, 2003; Warren-Findlow, Prohaska, & Freedman, 2003). A study on recruitment and retention of HIV participants found that younger Latino males were less likely to follow through with the research study (Villarruel, Jemmott, Jemmott, & Eakin, 2006). The investigators believed that conflict with j ob schedules and residency status mediated their results. Research studies in other areas c onsistently reported low participation of elder individuals in research st udies, (Mirowsky & Reynolds, 2000; Murthy, Krumholz, & Gross, 2004; Cotter, Burke, Loeber, & Mutchka, 2005). It appears that the ability of older caregivers to particip ate in research studies is influenced by factors such as knowledge a bout progressive memory disorders such as Alzheimers disease, health status, perceived burden associated with research involvement. Also influential are practical issues such as tran sportation, taking time off from work, and the 28

PAGE 29

availability of another person to tend to the care recipient. These inferences are drawn from the existing literature on age and part icipation in research studies. Latino Caregivers Level of Education The U.S. Administration on Aging maintains th at when compared to other groups, Latino adults have less formal education (2005). Th e study on Caregiving in the U.S. conducted by the National Caregiving Alliance and the American A ssociation of Retired Persons (2005) shows that 29% of all caregivers have a high school education, 30% have some college education, and 35% are college graduates. A report by Schulz (2000) suggests th at the median education of dementia caregivers is some college, the same as for the non-dementia caregivers. In 2004, the Alzheimers Association published a report indicating that almost 40% of dementia caregivers were college graduate; 25% had some college education, and 37% had a high school diploma or less schooling. Due to the lack of studies aimed at measuring the relationship between educational level and Latino caregivers participa tion in research studies, some researchers infer that Latino caregivers have lower educational attainment when compared to other groups. The basis for this inference appears to be the poor understanding of Latinos about Alzheimers disease and other progressive memory disorders (Coon, Gallaghe r-Thompson, & Thompson, 2003; Schulz, 2000). However, making this assumption is problema tic because the poor understanding may reflect their language differences and cultural beliefs about health and illness. An examination of the literature on other ethnic caregivers found mixed results regarding recruitment and retention of Afri can American into research studies. A study on recruitment and retention of underrepresented popula tions, specifically African Americ ans, into health promotion research, found that persons under 60 were less likely to enroll in thei r study (Warren-Findlow, Prohaska, & Freedman, 2003). The authors believed that the higher drop-out rate of younger 29

PAGE 30

participants was related to lower educational levels, whereas other pub lications indicate that African Americans are highly suspic ious of research due to a l ong history unethical medical and research protocols (Blakley & Harrington, 1997; Doty, 2007; Jonas, 1993; Ludmerer, 1999). Although this study did not include Latino participan ts, research findings like this may impact future researchers investigating the relationship between attrition rates and Latino caregivers. Latino Caregivers Income Level Lower educational level is typically associated with lower income. The Horizon Project (2000) reports that Latino elders are less well-educated than a nd more likely to live in poverty. The project estimates that 21% of Latino elders in the U.S. live at or below the federal poverty level compared to Whites. This study also repo rts that when compared to Whites, Latino elders are less likely to receive social security inco me. These findings are alarming considering the financial burden on caregivers who already st ruggle with a taxing financial situation. The cost to government agencies for the care of patients who have Alzheimer's disease is substantial at present and it is rising rapidly (Prigerson, 2003). Ho wever, most of the cost is absorbed by the unpaid family caregiv er. The unpaid portion accounts for a large proportion of the costs of treating the patient. Family car egivers spare the health care system billions of dollars annually, but depression, stress and weakened physical hea lth in these caregivers, if left untreated, come not only at a great personal cost to the caregivers and their families but potentially at a substantial financial cost to society. The negative impact on the health of dementia caregivers and the restricted ability to be away from the care recipient negatively impacts the caregivers limited resources as well as their earning potentia l. Specific to Latino elders, their lower educational accomplishments English proficiency, and work experience negatively influence their financial attainment. Limited financial resources typically translate 30

PAGE 31

into inability to pay for companionship services, professional assistance, a nd even transportation. These restrictions may in turn hinder the caregivers ability to participate in research studies. Depression in Latino Caregivers Depression in later life has been identified as one of th e most common mental health problems affecting elders (Abu-Rayya, 2006; Blazer, 2003). Bour geois, Schulz, and Burgios (1996) highlighted the risks associated with the negative consequences of depression in caregiver health when they reported that caregiver depr ession is correlated to decreased function and quality of life, and even death among caregivers. Majerovitz ( 2001) reports that caregivers who perceive the availability of soci al support report lower levels of depression associated with caregiving stress. Although research findings re port mixed results about depression in Latino caregivers (Janevic & Connell, 2001 ), most research findings sugge st that the prevalence of depression in Latino caregivers is equal or greater than found in other caregiver groups. (Aranda, 2001; Schulz, 2000). The combination of poor health and limited fi nancial resources also has been linked to higher rates of depression (Covinsky, Newc omer, Fox, Wood, Sands, Dane, & Yaffe, 2003). Since Latino caregivers tend to ha ve lower financial resources and compromised health, it is not surprising that they are also believed to have higher rates of depres sion. Weiss, Gonzalez, Kabeto, and Langa (2005) found that older Latinos have a signifi cant higher risk of depression than White or African American caregivers. Coon, Gallagher-Thompson, and Thompson (2003) maintain that the relationship between high levels of depression and adherence to traditional cultural norms is linked to the caregivers feelings of helplessness and difficulty managing problematic behaviors. In an examination of patient and caregiver characteristics associated with caregiver depression, Covinsky and collea gues (2003) found that Latino car egivers reported significant 31

PAGE 32

higher levels of depression when compared to Whites or Blacks. Despite higher levels of depression, this study found that Latinos were more reluctant to place their l oved one in a facility compared to caregivers from other groups. Place ment of a loved one with Alzheimers disease appears to be more consistent with White conv entions rather than Latino values such as the importance of family and mutual assistance known as familismo despite the toll on family caregivers (Schulz, 2000). A study measuring the relationship between racial or ethnic concordance between caregiver and researcher on car egivers attrition, depression, and burden found no significant difference in caregiver drop-out at the 12-month follow-up (McGinnis, Sch ilz, Stone, Klinger, & Mercurio, 2006) regardless of concordance of ra ce or ethnicity. No additional studies examining the relationship between Latino care givers, depression, and participa tion in research studies were available at the time of this review. However, due to the demands placed on research participants, it may be inferred that depressed caregivers are less likely to follow through with research protocols if they feel the involvement increases the amount of burden they already experience (Coon, Rubert, Solano, Mausbach, Kraemer, Arguelles, Haley, Thompson, & Gallagher-Thompson, 2003; Eisdorfer, Czaja, Lo ewenstein, Rubert, Arquelles, Mitrani, & Szapocznick, 2003; Miranda, Azocar, Organista, Muoz, & Lieberman, 1996). Social Support and Latino Caregivers Responsibilities associated w ith caring for someone with a chronic illness may disrupt caregivers social networks. Caregivers usually obtain social support from two main sources, 1) informal networks such as family and frie nds and 2) formal services from community organizations. Both sources of social support may alleviate the strain asso ciated with caregiving. While informal support is beli eved to lessen the negative consequences of caregiving (Miller, Townsend, Carpenter, Montgomery, Stull, & Young, 2001) there are mixed results on the 32

PAGE 33

relationship between social support and levels of distress am ong caregivers. Cohen, Boyle, Coakley, and Lawlor (2002) found that lack of social support in conjunction with limited financial resources have a negati ve impact on caregivers. Caregive rs from ethnic groups tend to have lower financial resources and health insuran ce coverage to meet their needs and thus seek formal supports such as day care centers and other programs. Keller, Gonzalez, and Fleuriet (2005) studied issues of recruitment and retentio n of ethnic women in an exercise program and found that participants who lacked social suppo rt did not follow through with the program. The literature on Latino caregivers of persons with Alzheimers disease show that Latino caregivers may be reluctant to seek support for several reas ons including limited knowledge about Alzheimers disease, the stigma associated with mental illness, feelings of shame, and the Latino cultural value of family pride and mutual assistance or familismo (Coon, GallagherThompson, & Thompson, 2003; Schulz, 2000). More over, due to the strong sense of familial obligation in the Latino communit y, the responsibility of caring fo r a loved one with progressive memory problems rests primarily in the family, particularly on a female member (Coon et al., 2003). Reliance on help outside the extended family is typically viewed in a negative light. Other factors such as low educationa l attainment and employment ma y also contribute to Latino caregivers feelings of isolation and limited social support (DilworthAnderon & Gibson, 2002). Informal social support is oftentimes possible because strong family connections are maintained by the close proximity in which Latinos te nd to live in regards to their relatives. It is also possible due to their reliance on extended family connections which include members of the extended family, friends, neighbors, and members of their religious groups (Sue & Sue, 1990). Latinos strong emphasis on the extended family and th eir tendency to live in close proximity to relatives provide caregivers with the opportunity for to exchange aid. However, some studies 33

PAGE 34

have found that the close connections typical in the Latino culture may also have negative effects. Lefley (1996) points out that the large social networks of Latinos may become a source of stress in the event of unme t expectations, poor advice, criticism, as well as the strong emphasis on protecting the familys honor. Semple (1992) reports that inte rpersonal stress with social networks has negative effects on the well be ing of caregivers. Furthermore, in a study on the caregiving experience of Mexican-Ameri cans, Phillips, Torres de Ardn, Komnenich, Killeen and Rusinak (2000) found that despite the availability of an extended family system, Latinos perceived that they had limited access to informal support and, therefore, used less support than Whites. The relatio nship between the availability of social support and Latino caregivers participation in re search studies has not been documented. However, due to the cultural values of close relations hips and privacy, it may be infe rred that Latino caregivers with strong social support would not have high rates of participation in research studies. Latino Caregivers Health Status Not only has the drastic growth of the elderly population over the last few decades placed exceptional demands in caregivers and other fam ily members but also on the whole healthcare system. Many researchers have discusse d negative physical, mental, and emotional consequences of caregiving (Connell, Janevic, & Gallant, 2001). Others have addressed the impact of caregivers inability to cope and the feelings of distress on their overall health (Gallagher-Thompson et al., 2000). Pillemer an d Suitor (2002) maintain that the negative outcomes associated with caregiving are heightened for caregivers of persons with Alzheimers disease and that poor social support exacerbates these problems. The health of caregivers is influenced by a myriad of factors including age, chronic conditions, phys ical problems, as well as psychological factors such as emotional we ll-being and stress. In a report by the U.S. Administration on Aging (2005), th e leading disease-related caus es of death in the Latino 34

PAGE 35

community included heart disease, cardiovascular disease, and diabetes, among others, all believed to be risk factors for developing Alzheimers disease or other related dementias. A study by Corbeil, Quayhagen, and Quayhagen (1999) found that caregivers tend to have negative appraisals of their stress levels compared to non-caregivers. Other fi ndings suggest that many caregivers report having poor h ealth (Navaie-Waliser, Feldman, Gould, Levine, Kuerbis, & Donelan, 2002) such as increased rates of high blood pressure and high insulin levels (Cannuscio, Jones, Kawachi, Colditz, Berkman, & Rimm, 2002) and high ra tes of cardiovascular disease (Lee, Colditz, Berkman, & Kawachi, 2003). The risk associated with chronic disease increases sharply with age (Manton, 1996). Researchers such as Fox, Newcomer, Yordi, and Arnsberger (2000) maintain that the caregiving responsibilities for a person with Alzhei mers disease can contribu te to the decline of the caregivers health. In addition, Schulz and B each (1999) believe that the negative impact of the combined chronic stress, sense of loss, physical demands, and the biological vulnerability of older caregivers put them at a higher risk of death. Hintons (2002) emphasis on the cultural nature of the caregiving experience stressed that cultural factors impact the caregivers perception of the illness, burden of care, ability to cope with caregiving stress, and ability to care prope rly for themselves. Caregiver burden includes physical, mental, emotional, and financial cost s of providing care and close supervision to someone and is accompanied by a negative appraisal of ones health. Several factors contribute to the poor health st atus of Latino caregivers such as the higher prevalence of diabetes (U. S. Department of H ealth & Human Services, 2005) and cardiovascular disease (Urizar & Sears, 2006). Moreover, the lower socioeconomi c status of Latinos has been correlated with poor health. D ilworth-Anderson and Gibson (1999) maintain that the Latinos 35

PAGE 36

disadvantaged socioeconomic status places them at higher risks for health conditions such as cardiovascular disease and diabetes, which in turn are risk factors for de veloping cardiovascular dementia, perhaps in combination with Alzhei mers disease. Weiss, Gonzalez, Kabeto, and Langa (2005) found that older Latinos have a hi gh burden of some diseases and poor selfperceived health and that they are at a higher risk for poor health outcomes due to higher rates of poverty and lower rates of health insurance cove rage. When compared to the white population, Latinos are believed to suffer higher rates of morbidity and mortality. Compounding these health disparities are the cultural and linguistic barriers that racial/e thnic minorities encounter when seeking health care. Lack of communi cation and ease can occur between patient and provider when cultural differences in perceptions of illness, disease, and medical roles are not recognized and addressed, thereby adversely affecting health outco mes. Other barriers include immigration status, discrimination, lack of access to high quality educational opportunities, unavailability of transportati on and childcare, and inconveni ent and insufficient hours of operation at health facilities. In addition to the existing unfavorab le health status in the Latino community, caregiving responsibilities often in troduce adverse consequences to caregivers health. Summary and Conclusions Little information exists in the literature about how the person-base d variables of age, educational level, income, level of depression, amount of social support, and health status influence the participation and drop-out rates from research st udies of Latino caregivers of persons with Alzheimers disease. Available literature on drop-out rates for members of diverse groups, more specifically Latino research particip ants, is scare and suggests mixed results. It appears that due to the changing demographics many caregivers are adult children caring for a parent. The existing demands from work and nuclear families on adult children caregivers may 36

PAGE 37

prevent them from participating in research stud ies. Moreover, the relationship between health status and Latino caregivers part icipation in research studies does not appear to have been examined. However, it appears that depressed care givers might not particip ate or complete their participation in research studies if they feel th at doing so exacerbates th eir burden. Finally, since Latinos are believed to rely on support from within their networks, it is reasonable to assume that those caregivers with adequate social support will be less likely to participate in studies and, thus, keep their personal affairs within the family unit. Due to the numerous unanswered questions regarding ba rriers to research participation in the Latino community, studies examining these barriers are warranted. The existing body of knowledge on predictors of attrition for Latinos o ccurs mostly in the areas of substance abuse and health promotion research, leaving a sea of opportunities for researchers. 37

PAGE 38

CHAPTER 3 METHODS This study was an archival analysis of the Resources for Enhancing Alzheimers Caregiver Health (REACH) project; a multi-site prospective randomized controlled trial. The purpose of this investigation was to examine variables identified in the published literature as relevant to the prediction of dr op-out status (i.e., attr ition) in a sample of Latino Alzheimers disease caregivers. Resources for Enhancing Alzheimers Caregiver Health REACH participants were recruited from memory disorder clinics, primary care clinics, community organizations and physicians offices. In order to recruit caregivers from diverse populations, REACH investigators us ed local community media to ta rget these groups. Criteria for inclusion required potential participants to be over the age of 21 and living with and providing care for a relative with Alzheimers di sease or other memory progressive disorder for at least four hours a day and for a minimum of si x months. A total of 1,222 participants were recruited and randomly assigned to an intervention condition. This study only examined the Latino caregivers in the Miami and Palo Alto si tes. Assessment instruments, as described by Wisniewski, Belle, Marcus, Burgio, Coon, Ory, Bu rns, & Schulz (2003) were translated into Spanish for these two samples. The Miami site (as described by Wisniewski et al., 2003) examined the effectiveness of family-based structural multisystem in-home intervention (FSMII) and a family-based structural multisystem in-home intervention plus a com puter telephone integration system (FSMII + CTIS). The FSMII was based on a family syst ems approach, aimed at improving communication between caregivers and other members of the family by identifying problems in communication 38

PAGE 39

and facilitating interactions The FSMII + CTIS was an enhanced version of the FSMII with the additional use of screen phones as a tool to increase intera ctions among family members. The Palo Alto site (as described by Wisniews ki et al., 2003) examined the effectiveness of the coping with a caregiving class (CWC) and an enhanced support group (ESG). The CWC was a psychoeducational class ba sed on cognitive behavioral theo ries that demonstrated the consequences of negative thoughts and be haviors in caregivers mood. The ESG was a professionally led support group in which caregivers were encourag ed to attend regular meetings and commit to participate for one year. Participants in this site were randomized into the intervention group based on language preference and availability. In addition to two active treatments, each si te also included a control condition. Both control conditions consisted of a minimal support condition in which the participants regularly received written educational material and check -in phone calls (Wisniewski et al., 2003). The remainder of this chapter addresses the fo llowing topics: 1) sample description and characteristics, 2) variables, 3) null hypotheses, 4) research pro cedures, and 5) data analysis, and 6) methodological limitations. Sample Characteristics The sample for this investigation consisted of N = 223 Latino Alzheimers disease caregivers in the United States. The sample was obtained from the Latino caregivers who participated in two of the six sites in the RE ACH study: Miami, FL and Palo Alto, CA. The Miami site consisted of n =113 Latino caregivers wher eas the Palo Alto site was comprised of n = 110 Latino caregivers. Participants were random ly assigned to either the treatment group or control measure. Palo Alto Sample. All 110 participants in the Palo Alto sample were females with a mean age of 51.14 years ( SD =12.91). Annual income was measured categorically and ranged from a low 39

PAGE 40

of less than $5,000 to a high of $70,000 or more; the modal income was $15,000 to 19,999. The average level of educatio n achieved was 10.49 years ( SD = 3.99). Caregivers in this sample had a mean of 35.62 years ( SD = 17.27) living in the U.S. and a mean of 17.44 years ( SD = 18.28) living with the care recipient. The mean depression CES-D score was 17.60 points ( SD =12.60). Miami Sample. In the Miami sample (n = 113), 86 (76%) participan ts were females and 27 (24%) were males. The mean age for this sample was 66.58 years ( SD = 10.54). Caregiver annual income measured categor ically, ranged from a low of less than $5,000 to a high of $70,000 or more; the modal income was $5,000 to $9,999. The average level of achieved education was 11.67 years ( SD = 3.79). The mean number of year s that caregivers had lived in the U.S. was 31.36 years ( SD = 8.83) and the mean number of y ears caregivers had lived with a care recipient was 35.0 years ( SD = 19.76). Caregivers in the Miami site had a mean CES-D score of 20.12 points ( SD = 11.09) which falls above the at risk for depression cutoff score of 16. Variables of Interest and Operational Definitions The predictor variables in th e study were: age, educational achievement, income, degree of depressive symptomatology, soci al support, and health status. The criterion variable was drop out status (i.e., attrition). In the sections that follow, each construct is described along with its operational definition (i.e., measur e) according to the REACH study. Age Refers to the time elapsed from date of bi rth to the time of the REACH study. This measure was acquired from the Caregiver Sociodem ographic Form from the REACH project that was used to obtain descriptive information about the caregivers. Age was measured in years. This study set out to elucidate the relationshi p between the age of caregivers and attrition. Depressive symptoms. This variable was operationally defined as the affective state of the caregiver at the time of measurement with the Center fo r Epidemiological Studies Depression 40

PAGE 41

Scale (CES-D). The CES-D has 20 items to measur e depression and is used widely to identify individuals who are prone to de pression (Radloff, 1977). The CES-D scores range from 060 points. Higher scores indicate greater depressive symptoms and a threshold of 16 or higher often indicates a risk for c linical depression (Wisniewski et al., 2003). Cronbachs alpha was .90 ( = .90). This study set out to measure the relations hip between caregiver depressive symptoms and attrition from the REACH project. Drop-out status (attrition). This variable was operationally defined as the loss of research participants (i.e., caregivers) through withdrawal or mortality during th e course of the REACH study. For this research, participants drop-out status (i.e., dropout or treatment completer) was assigned according to their status at the completion of the REACH trial. Educational achievement. This variable refers to the number of years of formal education completed by caregivers. This measure was acquired from the Caregiver Sociodemographic form used to obtain descriptive information about the caregivers. Educational level was measured by a range, from a low of 1 indicating an educational level of less than high school to 17 indicating a graduate degree. This study set out to measure the relationship between the educational level achieved by caregivers and attrition from the REACH project. Health status. This variable was operationally defined as the physical health status of the caregiver as measured by the Caregiver Health and Health Behaviors questionnaire from the REACH study. This measured was modified by the original REACH investigators from its initial conception by combining items from exis ting measures including the Short Form Health Survey (SF-36), the Asset and Health Dynamics Among the Oldest Old, and the Nutrition Screening Initiative (NSI ) (Switzer, Wisniewski, Belle, Dew, & Schulz, 1999). This research examined the relationship between health st atus and attrition fr om the REACH project. 41

PAGE 42

Income. Refers to the total amount (from a ll sources) of money earned on an annual basis by an individual. This measure was acqui red from the Caregiver Sociodemographic form used to obtain descriptive information about the caregivers. The form used ten categories to measure income level, from a low of less than $5,000 per year to a high of $70,000or more per year. This study set out to measure the relations hip between caregivers annual total income and attrition from the REACH project. Social support. This variable refers to the amount of support that the caregivers received from friends, neighbors, and other members of the family as measured by the Social Support form. This measured was also modified from its initial conception by the original REACH investigators. They changed the response scale and/or eliminated or a dded items to the scale (Wisniewski et al., 2003). This study attempte d to measure the relationship between social support and attrition fr om the REACH project. Null Hypotheses The following six null hypotheses were tested in this study: Ho1: There is no relationship between attrit ion drop-out status (attrition) and Latino caregivers age. Ho2: There is no relationship between drop-out status (attrition) a nd Latino caregivers achieved level of education. Ho3: There is no relationship between drop-out status (attrition) a nd Latino caregivers income. Ho4: There is no relationship between drop-out status (attrition) and depression among Latino caregivers as measured by Center for Epidemiologic Studies Depression Scale (CES-D). Ho5: There is no relationship between dropout status (attrition) and social support among Latino caregivers. Ho6: There is no relationship between drop-out status (attrition) a nd Latino caregivers health status as measured by the Caregive r Health and Health Behaviors Questionnaire from the REACH study. 42

PAGE 43

Research Procedures The REACH investigators provided permission for use of the data from the Miami and Palo Alto sites for purposes of this study. The data were obtained from the Inter-university Consortium for Political and Social Research (ICPSR ), at the University of Michigan. ICPSR maintains and provides access to a vast archive of social science data for research Finally, an application to the Institutional Review Board (I RB) for an Exempt Status was filed and granted since this study involved a collectio n of existing data that is avai lable for public use and has deidentified information, recorded so th at participants cannot be identified. Measures The original REACH study used several scal es and modified others to measure the variables and meet the n eeds of the original study. For th e present study, orig inal data were used, however, two measures, social support and health status, were modified. Demographic information for this study was obtained and us ed as reported by the REACH investigators. These data were collected through the Careg iver Sociodemographic Questionnaire (Appendix A). The data for the demographic variables of inte rest (i.e., age, educational level, and income) were included in this study. Caregivers depressive symptoms were meas ured with the Center for Epidemiological Studies Depression Scale (CES-D). The CES-D sc ores were obtained from the original data (Appendix B). A score of 16points indicated the cut-off point fo r the presence of depression. The higher the score, the higher the number of depressive symptoms reported by the caregivers. In order to measure the contributions of the independent variable so cial support, a social support index was obtained from one of the thr ee original questionnaires used by the REACH investigators; the perceived social support que stionnaire (Appendix C). This subscale included 43

PAGE 44

nine questions regarding the care givers perception about support available to them from friends and family, and was rated using with a Likert -type scale. The lower the number, the less perceived support the caregivers repo rted. The score for this subscale ranged from 9 to 32 points. Similarly, in order to measure the co ntributions of the independe nt variable health, two independent indices were created from the or iginal data acquired from the REACH project (Appendix D). Questions from the two subscales, 1) perceived health, and 2) self-reported health problem, were included. The first subscale, the perceived health questionnaire included a total of three questions. This scale was rated on a Likert-type scale and measured the caregivers subjective assessment of their health. The scor es ranged between 1 and 4.67 points. The higher the score, the healthier the careg ivers perceived their health. The second subscale, the selfreported health problems questionnaire, include d a total of seven que stions regarding the presence or absence of medical conditions such as high blood pressure, heart condition, and cancer. The scores ranged between a low of 0 and a high of 5. Higher scores reflect a higher number of reported perceived and/or established health problems. Data Analysis To test the hypotheses for this study, a hierarchical logistic regression analysis procedure was performed by the statistical program Statistical Package for the Social Sciences (SPSS). Based on existing literature, the following sets of predictors were identified and used in the hierarchical logistic regre ssion model: demographic variables (age, level of educational achievement, level of income) and person-based factors (depression, heal th status index, and social support index) as defined in prev iously published REACH studies. Alpha ( ) will be set at 0.05. 44

PAGE 45

In order to examine the unique contribution of the independent variab les to the prediction of attrition, the demographic variables age and edu cational level were entere d first into the initial block. For the Miami sample, the variable gender was also included in the first model. Income was entered into the logistic re gression next, followed by depre ssion score (CES-D), and social support. In the last model, the indices used to measure health status, perceived health status and reported health problems were ente red in the last block. The orde r of entry for the variables was determined by two factors. First, demographic va riables were entered into the model. The order of entry for the remaining three variables (i.e., depression, social support, and health status) was determined based on the re view of the literature. Methodological Limitations There are limitations associated with archival analyses. First, as a secondary analysis it does not lend itself for purposeful collection of data Therefore, the data in this study are limited to the original REACH investigators research qu estions and data collection. Second, there is an inability to control for levels of accu lturation in this sample of Latinos, which may influence caregivers belief, attitudes, and willingness to participate and complete research studies. Moreover, there are inherent limitations asso ciated with self-report measures including differences in comprehension levels or interpreta tion, memory deficiencies such as forgetfulness, and using the forced-choice response categorie s which may simplify answers or distort the information obtained along particular response (o r choice) sets (Hilton, Harry, & Rice, 2003). In addition, there is a danger associated with se lf-report assessment of physical and emotional ailments, which is influenced by the culture of ethnic groups. Furt hermore, there is an absence of a control or comparison group. There are also limitations associated with the data analysis procedures of hierarchical logist ic regressions such as the less than optimal control of covariate 45

PAGE 46

variables and their correlationa l nature; thus, caution must be employed when inferring a causal link (Huck, 2004). 46

PAGE 47

CHAPTER 4 RESULTS The present study entailed a secondary analysis of the multi-site Resources for Enhancing Alzheimers Caregiver Health (REACH) study. The purpose of this investigation was to examine predictors of attrition from the REAC H trial. Specifically, the importance of age, educational achievement, income, self-reported leve ls of depressive symptoms, perceived social support, and health status on the drop-out rates fo r Latino caregivers of pe rsons with Alzheimers disease from the REACH trial. This chapter presents the results of the anal yses completed for this study. The information in this chapter is organized as follows. First, data on the subjects who participated in this study are presented with a comparison of the subjects by research site and drop-out status (i.e., completion vs. non-completion). Then the hypotheses and the analyses are presented to address the hypotheses. A summary concludes this chapter. Preliminary Analyses A total of 219 participants from the two research sites, Palo Alto and Miami, were included in this study. These participants were from Palo Alto ( n =107, 48.8%) and Miami ( n =112, 52.1%). Subjects with missing data and those with ex treme scores on the variables of interest for this study were deleted from the dataset. There were a total of 59 dropouts (26.9%) and 160 completers (82.1%). The mean age of the total group was 58.90 years, ranging from 25 to 90 years ( SD = 14.08) and the mean years of completed education for the total group was 11.12 years, ranging from 2 to 17 years ( SD = 3.86). The median and mode annual total income for the participants was $15,000 to $19,999. Table 4-1 illustrate s the data for all of the variables for the 47

PAGE 48

total group as well as by location and Table 4-2 shows the data for all of the variables by completion status. It was also of interest to this study to determine whether there were significant differences between sample sites (Palo Alto and Miami) and betw een drop-out status (completer and drop-out). In order to compare the mean valu es for the variables of interest between groups, analysis of variance was used to compare respon dents between location site and drop-out status for the continuous variables (age, years of edu cation, CES-D score, perceived social support, perceived health status and reported health status). Chi Square was used to test for differences in categorical data (income). The mode inco me for Palo Alto was between $15,000 to $19,999 and the mode for Miami was $5,000 to $9,999. Participants from Miami had a lower income level than did Palo Alto participants. Median in come in Palo Alto was between two ranges, $15,000 to $19,999 and $20,000 to $29,999 while the median in come for Miami was between the two ranges of $10,000 to $14,000 and $15,000 to $19,999. The Chi Square for location was significant for income ( X2 (9) = 22.249, p = .008); however, there were six cells with counts less than five and the test may not be valid. The mode income for completers wa s $15,000 to $19,999 and the mode income for dropouts was $5,000 to $9,999. It is interesting to note that income was noticeably less for dropouts. The Chi Square test for dropouts versus completers by income found no significant differences ( X2 (9) = 16.514, p = .057), however this may not have been an appropriate test as there were four cells with counts less than five bringing into question th e validity of this test. Table 4-3 shows the income breakdown by location and drop-out status. To further test the differences between sites and drop-out stat us, ANOVA tests were conducted. Table 4-4 shows the results of the ANOVA test be tween research sites. The 48

PAGE 49

ANOVA test indicates there are significant differences between participants in the Palo Alto site and the Miami site for age, years of education, CES-D score, social s upport, perceived health, and reported health status. Inspection of the ANOVA findings for dropouts and completers found no statistically significant differences betw een dropouts and completers for age, years of education, CES-D score, perceived social support, perceived healt h, and reported health status. Table 4-5 shows the results of the ANOVA test between drop-outs and treatment completers. Dropouts and completers were remarkably similar; however, there were noticeable differences between locations. A Spearman Rho Correlation was used to test for correlations between the predictor variables and the criter ion (i.e., drop-out status) in this study. These correlations are discussed next. Palo Alto Correlations A Spearman Rho Correlation was calculated betw een the predictor variables and criterion variable (i.e., drop-out status) in this study. The variable of primary interest was how the independent variables (age, y ears of education, income, CESD score, received support, perceived health status, and health problems) were correlated with the dependent variable dropout status. Age was significantly correlated with drop-out status ( r = .212, p = .028) although it was a weak correlation. Years of education was inversely correlated with age ( r = -.232, p = .016). Income was strongly correlat ed with years of education ( r = .252, p = .009). The most powerful correlation in this study occurred between health problems and age (r = .393, p = <.001). The CES-D (depression) score had a low correlation with health problems ( r = .288, p = .005) and with perceived health problems ( r = .312, p = .003). All of the other correlations were not strong correlations nor were th ey statistically significant. Table 4-6 shows the correlation matrix for the Palo Alto sample. 49

PAGE 50

Miami Correlations A Spearman Rho Correlation was used to test for correlations between the variables used in this study. The variable of primary interest was how the inde pendent variables (age, years of education, gender, income, CES-D score, received support, percei ved health status, and health problems) were correlated with th e dependent variable of droppi ng out or staying in the REACH program. There was a significant though low corre lation between drop-out status and years of education ( r = .198, p = .037) for the Miami participants. There was a strong correlation between years of education and income ( r = .537, p = <.001). The correlation between years of education and gender was significant but weak ( r = .211, p = .025). The correlation between age and health problem score was weak but statis tically significant (r = .232, p = .015.). Age was inversely correlated with income (r = -.424, p = <.001), moderately correlated with gender ( r = -.347, p <.001). There was a low moderate positiv e correlation between CE S-D score and health status ( r = .314, p = .002) and between CES-D and health problems ( r = .215, p = .041). There was a low but significant correlati on between gender and income (r = .188, p = .050) and between gender and CES-D score ( r = .273, p = .008). The correlation between received support and health status was inverse and low but statistically significant ( r = -.209, p = .032). Table 4-7 shows the correlation matrix for the Miami sample. Hypothesis Testing This study sought to address the following hypotheses: Ho1: There is no relationship between attrit ion drop-out status (attrition) and Latino caregivers age. Ho2: There is no relationship between drop-out status (attrition) a nd Latino caregivers achieved level of education. Ho3: There is no relationship between drop-out status (attrition) a nd Latino caregivers income. 50

PAGE 51

Ho4: There is no relationship between drop-out status (attrition) and depression among Latino caregivers as measured by Center for Epidemiologic Studies Depression Scale (CES-D). Ho5: There is no relationship between dropout status (attrition) and social support among Latino caregivers. Ho6: There is no relationship between drop-out status (attrition) a nd Latino caregivers health status as measured by the Caregive r Health and Health Behaviors Questionnaire from the REACH study. Data Analysis Logistic regression was conducted on rese arch questions concerning the relative importance of each block (set) of independent va riables: demographics (age, gender, years of education, income), CES-D score (CES-D), social support perceived (SSPerc), health status perceived (HSPercd), and actual health problem (HProb) served as statistically significant predictors of dropping out of the REACH study. Tabachnick and Fidell (1996) maintain that logistic regression does not requ ire the adherence to any assump tions about the distribution of the variables. Study data were checked for coll inearity and found to be within an acceptable range for tolerance and the variance inflation fact or (Mertler & Vannatta, 2001). The data for the two sites (Palo Alto and Miami) were also checked for outliers prior to commencing data analysis. The Mahalamobis X2 a test for outliers, identified four outliers with scores in excess of the test statistic. Three cases were delete d from the Palo Alto dataset and one case was deleted from the Miami dataset. This resulted in Miami having an n = 112 and Palo Alto having an n = 107. Missing data were treated as missi ng and no substitutions were made for missing data. Since the participants from Palo Alto a nd Miami were statistically different on a number of the independent variables as presented previously, each location was treated as a separate entity for purposes of the logistic regression analysis. 51

PAGE 52

Logistic regression using the enter method (simultaneous entry of predictors) was conducted to determine which of the set of i ndependent predictor variables would predict whether the participant completed or dropped out of the REACH treatment program for each location. Independent variables were entered in blocks to measure the effect of adding a particular variable or set of variables into the model. Palo Alto Sample The first block used age and years of educati on as predictors of co mpleting or dropping out of the REACH program in the model. The [-2 Log likelihood (72.644) and X2 (2) = 4.123, p = .127] for this model indicated the model was not statistically significant. However, the model using age and years of education co rrectly identified 83.9% of the pa rticipants. Wald statistics, as shown in Table 4-8 indicated none of the variable s served as a significan t predictor of drop-out status for the Palo Alto group of participants. The odds ratios for these variables su ggest there is little change in the likelihood of dropping out or completing the program when the predictors variables (i.e., age and years of education) increased by 1. Table 4-8 shows the results of the first model for the Palo Alto logistic regression analysis. The second model added income to the model to determine the effects of income on the model. The model was not statistica lly significant [-2 Log likelihood (71.679), X2 (3) = 5.088, p =.165] and the model fit indices were fairly la rge indicating model fit was questionable. The age, years of education, and inco me model did correctly identify 83.9% of the participants. The Wald statistics presented in Table 4-9 indicated that none of the variable s (i.e., age, years of education, and income) served as a significant predictor of dropping out or completion for the Palo Alto group of participants. The odds ratios for these variables found little change in the likelihood of dropping out or completing the program when the predictors increased by 1. Table 4-9 shows the results of the second model of the Palo Alto logistic re gression analysis. 52

PAGE 53

The third model added the score on the Center for Epidemiologic Studies Depression Scale (CES-D) as a predictor along w ith age, years of education and income. The model was not statistically significant [-2 Log likelihood (71.674), X2 (4) = 5.094, p = .278] and the model fit indices were fairly large indica ting model fit was questionable. The age, years of education, income, and CES-D model did correctly identify 83.9% of the participants. The Wald statistics presented in Table 4-10 indicated none of the variables (i.e., age, years of education, income, CES-D score) served as a signifi cant predictor of dropping out or completion for the Palo Alto group of participants. The odds ratios for these variables found lit tle change in the likelihood of dropping out or completing the program when the predictors increased by 1. Table 4-10 shows the results of the third model of the Palo Alto logistic regression analysis. The fourth model added the variable perceived social support (SSPerc) into the model with age, years of education, income, and CES-D score. The model was not statistically significant [-2 Log likelihood (71.666), X2 (5) = 5.101, p = .404] and the model fit in dices were fairly large indicating model fit was questionable. The mode l with age, years of education, income, CES-D score, and received social support correctly identified 83.9% of the participants. The Wald statistics presented in Table 411 indicated none of the variables (i.e., age, years of education, income, CES-D score, and perceived social suppor t) served as a signific ant predictor of dropping out or completion for the Palo Alto group of participants. The odds ratios for these variables found little change in the likelihood of dropping out or completing the program when the predictors increased by 1. Tabl e 4-11 shows the results of the fourth model of the Palo Alto logistic regression analysis. The fifth model added the variable perceived health status (HSPercd) and reported health status (HSProb) into the mode l predicting dropping out or completing the REACH program. The 53

PAGE 54

model was not statistically signi ficant [-2 Log Likelihood (69.3056) X2 (6) = 7.462, p =.382]. The model fit indices were fairly large indicating model fit was questionable and the Chi Square statistic was not significant. Th e model correctly classified 82 .8% of the cases. Regression coefficients are presented in Table 4-12. Wald statistics indicated none of the variables (i.e., age, years of education income, CES-D score, received social support, pe rceived health status and reported health status) served as a significant predictor of dropping out or completion for the Palo Alto sample. The odds ratios for these variables indicate there is little change in the likelihood of dropping our or completing the program when the predictors increased by 1. Table 4-12 shows the results of the fifth model of th e Palo Alto logistic regression analysis. Miami Sample The first block used age, years of educati on, and gender as predic tors of staying or dropping out of the REACH program in the model. This model was not st atistically significant as indicated by [-2 Log likelihood (107.284) and X2 (3) = 2.407, p =.492]. However, the age and years of education model did correctly identi fy 62.7% of the participants. Wald statistics indicated none of the variables (a ge and years of education) serv ed as a significant predictor of dropping out or completion for the Miami sample. The odds ratios for these variables indicate there is little change in th e likelihood of dropping out or completing the program when the predictors increased by 1. Table 4-13 shows the results of the fi rst model of the Miami logistic regression analysis. The second model added income to the model to determine the effects of income on the model. The model was not statistica lly significant [-2 Log likelihood (107.260), X2 (4) = 2.431, p =.657] and the model fit indices were fairly la rge indicating model fit was questionable. The age, years of education, gender, and income model did correctly identify 62.7% of the participants. The Wald statistics presented in Table 4-14 indica ted none of the variables (age, 54

PAGE 55

years of education, and income) served as a si gnificant predictor of dropping out or completion for the Miami group of participants. The odds ratios for these variable s found there was little change in the likelihood of dr opping out or completing the program when the predictors increased by 1. Table 4-14 shows the results of the second model of the Miami logistic regression analysis. The third model added the depression score, as reported by the Center for Epidemiologic Studies Depression Scale (CES-D), as a predictor along with age, years of education and income. The model was not statistically significant [-2 Log likelihood (106.773, X2 (5) = 2.918, p = .713] and the model fit indices were fairly large indica ting model fit was questionable. The age, years of education, income, and CES-D score model di d correctly identify 60.2% of the participants. The Wald statistics pres ented in Table 4-15 indicated none of the variables (age, years of education, gender, income, and CES-D) served as a significant predic tor of dropping out or completion for the Miami sample. The odds ratios for these variables found little change in the likelihood of dropping out or completing the program when the predictors increased by 1. Table 4-15 shows the results of the third model of the Miami logistic regression analysis. The fourth model added Social Support Received (SSPerc) into the model with age, years of education, income, and CSED. The model wa s not statistically signi ficant [-2 Log likelihood (106.708, X2 (6) = 2.983, p = .811] and the model fit indices were fairly large indicating model fit was questionable. The age, y ears of education, income, CES-D score, and perceived social support model did correctly identif y 62.7% of the participants. The Wald statistics presented in Table 4-16 indicated none of th e variables (age, years of edu cation, income, CES-D score, and perceived social support) served as a significant predictor of dropping out or completion for the Miami group of participants. The odds ratios for these variables found little change in the 55

PAGE 56

likelihood of dropping out or completing the program when the predictors increased by 1. Table 4-16 shows the results of the fourth model of the Miami logistic regression analysis. The fifth model added the health variables, pe rceived health status (HSPercd) and reported health status (HSProb) into the model pr edicting dropping out or completing the REACH program. The model was not statistically significant [-2 Log Likelihood (105.783; X2 (8) = 3.908, p =.865]. The model fit indices were fairly la rge indicating model fit was questionable and the Chi Square statistic was not significant. The model correctly classified 61.4% of the cases. Regression coefficients are pr esented in Table 4-17. Wald statistics indicated none of the variables (age, years of educat ion, income, CES-D, perceived social support, and perceived health status and reported health status served as a significant predictor of dropping out or completion for the Miami research site. The odds rati os for these variables i ndicate there is little change in the likelihood of dr opping out or completing the program when the predictors increased by 1. Table 4-17 shows the results of the fifth model of the Miami logistic regression analysis. Summary of Results The outcomes of the study were not signifi cant. None of the predictor variables contributed significantly to the unde rstanding of caregiver drop-out stat us (the criterion) from the REACH trial. There were no significant differences between drop-out status though income between the two samples groups varied signifi cantly. In addition, th ere were significant differences between the Palo Alto and Miami samp les for age, years of education, CES-D score, perceived social support, perceive d health, and reported health status. There were no statistically significant differences between dropouts and comp leters for age, years of education, CES-D score, social support, perceived health, and reported health status. 56

PAGE 57

Table 4-1. Variable means for total group and by location (Palo Alto and Miami) Total N Total Mean Total SD Miami Mean Miami SD Miami N Palo Alto Mean Palo Alto SD Palo Alto N Age 219 58.90 14.08 66.51 10.55 112 50.94 12.89 107 YrEduc 219 11.11 3.86 11.63 3.78 112 10.57 3.89 107 CES-D 186 18.69 11.96 19.98 11.06 99 17.39 12.72 93 SSPerc 210 15.81 8.51 13.68 8.63 106 17.99 7.85 104 HSPercd 218 2.69 .69 2.74 .70 111 2.65 .67 107 HSProb 214 1.13 1.11 1.37 1.13 110 .88 1.03 104 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPerc d = Perceived Health Status; HSProb = Selfreported Health 57

PAGE 58

Table 4-2. Variable means for tota l group and dropouts and completers Total N Total Mean Total SD Drop Mean Drop SD Drop N Compl Mean Compl SD Compl N Age 219 58.90 14.08 60.20 15.20 59 58.43 13.67 160 YrEduc 219 11.11 3.86 10.88 3.73 59 11.20 3.92 160 CES-D 186 18.69 11.96 20.04 9.72 50 16.19 12.68 136 SSPerc 210 15.81 8.51 14.58 8.08 55 16.25 8.64 155 HSPercd 218 2.69 .69 2.74 .75 58 2.67 .66 160 HS Prob 214 1.13 1.11 1.05 1.02 59 1.16 1.14 155 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPerc d = Perceived Health Status; HSProb = Selfreported Health; Drop = Dropouts; Compl = Completers 58

PAGE 59

Table 4-3. Income level group for completion status and location Income Drop N (%) Comp N (%) Miami N (%) Palo Alto N (%) Total N (%) <$5,000 5 (8.6) 7 (4.4) 6 (5.5) 6 (5.6) 12 (5.6) $5,999 11 (19.0) 21 (13.3) 23 (21.1) 9 (8.4) 32 (14.8 $10,999 15 (25.9) 20 (12.7) 21 (19.3) 14 (13.1) 35 (16.2) $15,999 6 (10.3) 31 (19.6) 17 (15.6) 20 (18.7) 37 (17.1) $20,999 9 (15.5) 19 (12.0) 16(14.7) 12 (11.2) 28 (13.0 $30,999 3 (5.2) 24 (15.2) 10 (9.2) 17 (15.9) 27 (12.5) $40,999 3 (5.2) 18 (11.4) 3 (2.8) 18 (16.8) 21 (9.7) $50,999 1 (1.7) 8 (5.1) 4 (3.7) 5 (4.7) 9 (4.2) $60,999 3 (5.2) 4 (2.5) 5 (4.6) 2 (1.9) 7 (3.2) $70,000 > 2 (3.4) 6 (3.8) 4 (3.7) 4 (3.7 ) 8 (3.7) Total 58 (100) 158(100) 109(100) 107 (100) 216 (100) Income reported in dollars 59

PAGE 60

Table 4-4. ANOVA test between sites (Palo Alto and Miami) Df F p Age 1,217 95.966 <.001 YrEduc 1,217 4.129 .043 CES-D 1,184 2.196 .140 SSPerc 1,208 14.233 <.001 HSPercd 1,216 .908 .342 HSProb 1,212 10.786 .001 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPerc d = Perceived Health Status; HSProb = Selfreported Health. p < .05. 60

PAGE 61

Table 4-5. ANOVA test between treatm ent dropouts and treatment completers Df F p Age 1,217 .686 .408 YrEduc 1,217 .303 .582 CES-D 1,184 .866 .353 SSPerc 1,208 1.576 .211 HSPercd 1,216 .424 .516 HSProb 1,212 .472 .493 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPerc d = Perceived Health Status; HSProb = Selfreported Health. p < .05. 61

PAGE 62

Table 4-6. Palo Alto correlations Drop r p Age r p Yrs Ed r p Income r p CES-D r p SSPerc r p HSPercd r p HSProb r p Drop 1.00 Age .212 .028 1.00 YrEduc -.052 .595 -.232 .016 1.00 Inome .172 .079 -.159 .102 .252 .009 1.00 CES-D -.035 .740 -.087 .406 -.107 .306 -.020 .852 1.00 SSPerc .077 .439 .030 .752 -.020 .843 .017 .860 .156 .141 1.00 HSPercd .033 .733 .018 .852 -.065 .503 -.106 .276 .288 .005 .090 .365 1.00 HSProb .127 .199 .393 <.001 -.178 .070 -.100 .312 .312 .003 .044 .665 .061 .539 1.00 Note: Age = Years of Life; YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPe rc = Perceived Social Suppor t; HSPercd = Perceived Health Status; HSProb = Self-reported Health p < .05. 62

PAGE 63

Table 4-7. Miami correlations Drop r p Age r p YrEduc r p Gender r p Income r p CESD r p SSPerc r p HSPercd r p HSProb r p Drop 1.00 Age -.063 .509 1.00 YrEduc .198 .037 -.169 .075 1.00 Gender .005 .958 -.347 <.001 .211 .025 1.00 Income .112 .247 .424 <.001 .537 <.001 .188 .050 1.00 CES-D -.120 .252 -.124 .238 -.118 .258 .273 .008 -.126 .236 1.00 SSPerc .002 .984 -.004 .966 -.025 .802 .168 .085 .083 .401 -.049 .652 1.00 HSPercd -.043 .657 -.060 .530 .038 .968 .034 .720 .034 .726 .315 .002 -.209 .032 1.00 HSProb .064 .508 .232 .015 -.025 .786 -.013 .892 -.164 .092 .215 .041 -.043 .667 .059 .543 1.00 Note: Age = years of Life; YrEduc = Years of Education; Gender = Male or Female; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPercd = Perceived Health Status ; HSProb = Self-reported Health p < .05. 63

PAGE 64

Table 4-8 Palo Alto logis tic regression, model one B Wald Df p Odds/Ratio CI Age .045 .028 1 .085 1.046 994, 1.101 YrEduc -.028 .086 1 .748 .973 .822, 1.151 Constant -.170 1.759 1 .923 .844 Note: YrEduc = Years of Education p < .05. 64

PAGE 65

Table 4-9 Palo Alto logis tic regression, model two B Wald Df p Odds/Ratio CI Age .042 2.624 1 .105 1.043 991, 1.097 YrEduc .047 .279 1 .598 000954 .800, 1.137 Income .139 .934 1 .334 1.150 .866, 1.525 Constant -.352 .039 1 .843 .703 Note: YrEduc = Years of Education p < .05. 65

PAGE 66

Table 4-10. Palo Alto logistic regression, model three B Wald Df p Odds/Ratio CI Age .042 2.490 1 .115 1.042 .990, 1.098 YrEduc -.048 .281 1 .596 .953 .799, 1.137 Income .139 .932 1 .334 1.149 .866, 1.525 CES-D -.002 .006 1 .940 .998 .950, 1.049 Constant -.299 .025 1 .875 .741 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score p < .05. 66

PAGE 67

Table 4-11. Palo Alto logist ic regression, model four B Wald Df p Odds/Ratio CI Age .042 2.493 1 .114 1.043 990, 1.098 YrEduc -.047 .278 1 .598 .954 .799, 1.138 Income .139 .931 1 .335 1.149 .866, 1.525 CES-D -.002 .004 1 .950 .998 .950, 1.050 SSPerc -.003 .007 1 .933 .997 .922, 1.077 Constant -.250 .016 1 .900 .779 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support p < .05. 67

PAGE 68

Table 4-12. Palo Alto logist ic regression, model five B Wald Df p Odds/Ratio CI Age .034 1.493 1 .222 1.035 .979, 1.094 YrEduc -.038 .161 1 .688 .963 .799, 1.159 Income .184 1.401 1 .237 1.202 .886, 1.632 CES-D -.006 .040 1 .842 .994 .939, 1.053 SSPerc -.017 .169 1 .681 .983 9.05, 1.068 HSPercd .664 1.816 1 .178 1.942 .739, 5.102 HSProb .272 .380 1 .538 1.313 .552, 3.121 Constant -1.738 .557 1 .458 .175 Note: YrEduc = Years of Education; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPerc d = Perceived Health Status; HSProb = Selfreported Health p < .05. 68

PAGE 69

Table 4-13. Miami logistic regression, model one B Wald Df p Odds/Ratio CI Age .000 .000 1 .992 1.000 .953, 1.050 YrEduc .092 2.174 1 .140 1.097 .970, 1.240 Gender -.018 .001 1 .975 .982 .308, 3.126 Constant -.505 .043 1 .835 .603 Note: YrEduc = Years of Educa tion; Gender = Male or Female p < .05. 69

PAGE 70

Table 4-14. Miami logistic regression, model two B Wald Df p Odds/Ratio CI Age -.001 .002 1 .965 .999 .949, 1.052 YrEduc .097 1.884 1 .170 1.102 .959, 1.267 Gender -.023 .002 1 .969 .977 .306; 3.115 Income -.019 .024 1 .876 .981 .772, 1.248 Constant -.401 .025 1 .873 .670 Note: YrEduc = Years of Educa tion; Gender = Male or Female p < .05. 70

PAGE 71

Table 4-15. Miami logistic regression, model three B Wald Df p Odds/Ratio CI Age -.003 .012 1 .914 .997 .946, 1.050 YrEduc .096 1.818 1 .178 1.100 .958, 1.264 Gender .099 .026 1 .872 1.104 .330; 3.694 Income -.038 .091 1 .763 .963 .752, 1.232 CES-D -.015 .487 1 .485 .985 .943, 1.028 Constant .108 .002 1 .967 .898 Note: YrEduc = Years of Education; Gende r = Male or Female; CES-D = Center for Epidemiologic Studies Depression Score p < .05. 71

PAGE 72

Table 4-16. Miami logistic regression, model four B Wald Df p Odds/Ratio CI Age -.001 .003 1 .957 .999 .947, 1.053 YrEduc .093 1.695 1 .193 1.098 .954, 1.263 Gender .139 .048 1 .827 1.149 .331; 3.990 Income -.029 .050 1 .823 .971 .752, 1.254 CES-D -.016 .506 1 .477 .984 .943, 1.028 SSPerc -.007 .066 1 .798 .993 .940, 1.048 Constant .166 .004 1 .949 .847 Note: YrEduc = Years of Education; Gende r = Male or Female; CES-D = Center for Epidemiologic Studies Depression Scor e; SSPerc = Perceived Social Support p < .05. 72

PAGE 73

Table 4-17. Miami logistic regression, model five B Wald Df p Odds/Ratio CI Age -.004 .017 1 .897 .997 .945, 1.051 YrEduc .094 1.697 1 .193 1.099 .954, 1.267 Gender .117 .033 1 .855 1.124 .321; 3.939 Income -.012 .009 1 .926 .988 .763, 1.279 CES-D -.013 .286 1 .593 .987 .942, 1.035 SSPerc -.010 .124 1 .725 .990 .937, 1.046 HSPercd -.259 .505 1 .477 .772 .378, 1.576 HSProb .160 .526 1 .468 1.173 .762, 1.806 Constant .425 .025 1 .875 1.530 Note: YrEduc = Years of Education; Gende r = Male or Female; CES-D = Center for Epidemiologic Studies Depression Score; SSPerc = Perceived Social Support; HSPercd = Perceived Health Status; HSProb = Self-reported Health p < .05. 73

PAGE 74

CHAPTER 5 CONCLUSION This chapter includes a discussion of the findi ngs for each hypothesis as well as additional conclusions and interpretations resulting from the analyses of data. The implications and limitations of the study are also presented. Reco mmendations for future research conclude the chapter. In order to address current concerns in th e literature about the low participation and retention of Latino caregivers of persons with Alzheimers diseas e in research studies, the goal of this research was to examine six variables believ ed to be relevant to th e attrition rate of Latino caregivers of persons with Alzheimers disease fro m research studies. Spec ifically, this study set out to examine the extent to which the variable s caregiver age, caregiver attained educational level, caregiver income, caregiver depression score (CES-D), careg iver perceived social support, and caregiver health status help explained care givers drop-out status from two sites in the Resources for Enhancing Alzheimers Caregiver H ealth (REACH) trial; Miami and Palo Alto. The Miami site (as described by Wisniewski et al., 2003) examined the e ffectiveness of familybased structural multisystem in-home interven tion (FSMII) and a family-based structural multisystem in-home intervention plus a com puter telephone integration system (FSMII + CTIS). The Palo Alto site (as described by Wi sniewski et al., 2003) examined the effectiveness of the coping with a caregiving class (CWC ) and an enhanced support group (ESG). Conclusions The outcomes of the study were not significan t. None of the predictors help to significantly explain caregiver drop -out status. While other studi es have suggested that these variables may help explain why dementia caregiver s drop-out prematurely from research studies, 74

PAGE 75

this set of variables did not help predict attriti on in these two samples of Latino caregivers of persons with Alzheimers disease. Alt hough the outcomes of this study did not produce significant results, these findings contribute to the existing lite rature on Latino caregivers of persons with Alzheimers disease and serve to inform future research studies by underscoring the importance of an adequate sample size, purposeful instrumentation, and cultural considerations. Thus, this study should be considered a pilot study th at tested the feasibility of this investigation. Palo Alto No significance was found between caregivers age and the likelihood of dropping out of the REACH clinical trial (hypothesis 1). No significance was found between caregivers level of education and the likelihood of caregivers dropping out of the REAC H trial (hypothesis 2). No significance was found between ca regivers income level and the likelihood of caregivers dropping out of the REACH trial (h ypothesis 3). No significan ce was found between caregivers CES-D score (i.e., depressive symptoms) and th e likelihood of caregivers dropping out from the REACH trial (hypothesis 4). Similarly, no signif icance was found in caregiver perceived social support and the likelihood of car egivers dropping out of the REACH trial (hypothesis 5). Finally, no significance was found in caregivers health status and the likelihood of caregivers dropping out of the REACH tr ial (hypothesis 6). Miami No significance was found in caregivers age and the likelihood of caregivers dropping out of the REACH trial (hypothesis 1) No significance was found in caregivers level of education and the likelihood of caregivers dropping out of the REACH trial (hypothesis 2). In addition, no significance was found in caregivers income le vel and the likelihood of caregivers dropping out of the REACH trial (hypothesis 3) No significance was found in caregivers CES-D score (i.e., 75

PAGE 76

depressive symptoms) and the likelihood of caregivers dropping out of the REACH trial (hypothesis 4). No significance was found in caregivers perceived social support and the likelihood of caregivers dropping out of the REACH trial (h ypothesis 5). Finally, no significance was found in caregiver s health status and the li kelihood of caregivers dropping out of the REACH trial (hypothesis 6). Implications The predictor variables in this investigati on (i.e., caregivers age, achieved educational level, income level, CES-D score, health status and perceived social support) did not achieve statistical significance for the criter ion of drop-out status. However, for the Palo Alto sample, it appears that there may be significance for the va riables, caregivers age, income, and perceived health, if the sample size were larger or if the instruments were sufficiently powered to test these specific variables. Similarly, in the Miami sample the predictor variable of caregivers achieved educational level also indicates it may achieve significance with a larger sample size or better instrumentation. These findings are consistent with the existing mixed results in the current literature regarding the participati on of Latinos in research studies. A few studies have been successful recruiting and retaining Latino participants (Gallagher-Thompson et al., 2004; Schulz, 2003). However, many others have documented the ch allenges in recruiting and retaining Latino participants in research studi es (Bank, Arguelles, Rubert, Ei sdorfer, & Czaja, 2006; Levkoff & Sanchez, 2003; Morano, & Bravo, 2002; Aranda, 2001; Fitten, Ortiz, & Pontn, 2001; Wiliams, Tappen, Buscemi, Rivera, & Lezcano, 2001). These mixed results may continue to be related to the under-representation of Latinos in research studies. Although the par ticipants in these two samples were Latinos, these two sample sizes are c onsidered small sample sizes for this type of 76

PAGE 77

analysis. The samples were also different regarding the country of origin; while the Miami sample was mostly Cuban-American, the Palo A lto sample was mostly Mexican-American. In addition, these findings are inconsistent with other research findings that suggest the variables included in this study help explain why Latino ca regivers drop-out prematurely from research studies. For example, Gallagher-Thompson, Solano, Coon, and Aren (2003) underscore the need to examine the influence of physical heal th and mental well-being (i.e., depression) in Latino caregivers in future res earch studies. Schulz et al. (2003) and Haley, Levine, Brown, & Bartolucci, (1987) emphasize the ro le of social support. Consequently, recruitment and retention of culturally diverse individuals ha ve been areas of major concern for researchers. Schulz et al. (2003) report that one of the disappointments in the caregiving interventi on research literature has been the relative lack of success in achieving clinically meaningful outcomes (p. 519). Thus, it is critical to continue the efforts to advance the knowle dge base about the impact of these and other variables in the retention of Latino caregivers in research studies. The aging of the population in combination wi th the rapid growth of the Latino population and the increasing incidence of Alzheimers di sease warrant the continuation of efforts to advance knowledge about the impact of Alzhei mers disease. Over one million Latinos are expected to suffer from this devastating illne ss (Alzheimers Association, 2004). Thus, these numbers underscore the need to in clude culturally diverse groups in investigations in order to develop effective methods to meet their needs. Although numerous studies have been conducted to advance the knowledge about the impact of Alzh eimers disease, only a handful have included ethnic minority participants, particularly Lati nos (Bank et al., 2006). The debilitating and progressive nature of Alzheimers disease de mands careful considera tion of and developing 77

PAGE 78

strategies and resources to addre ss the needs of caregivers, especi ally ethnic minorities, since the generalizability of most studies cannot be extended to these groups. Continued inquiry regarding the impact of these person-based va riables (i.e., age, education level, depressive symptoms, physical h ealth, and social support) the impact of burden associated with the caregiver role, and reasons for participation in research studies are crucial. Moreover, inclusion and examination of other vari ables such as culture-specific variables (i.e., acculturation level, language proficiency, percep tion about the etiology of Alzheimers disease, and the perception about research studies), care recipient characteristics (i.e., physical limitations and degree of cognitive impairment), as well as di fferent research delivery methods (i.e., use of technology) may prove beneficial. The need fo r including cultural variables when examining ethnic minorities has been well-documented in the literature (Mahoney, Cloutterbuck, Neary, & Zhan, 2005; Moreno-John et al., 2004; Aren, Alvidrez, Nery, Estes, & Linkins, 2003; Levkoff & Sanchez, 2003; Curry & Jackson, 2003; Aranda 2001). Another area of inquiry that is attracting much attention is the development of a lternative service and res earch delivery options. For example, as the baby boomers age, an increasing number of potential participants will probably have had significant experience w ith telecommunication technology such as the telephone and the internet, which may be excellent vehicles for gathering data (Bank, Arguelles, Rubert, Eisdorfer, & Czaja, 2006). Limitations Due to the archival nature of this inquir y, this study did not lend itself for purposeful collection of data nor allow for adding more rese arch participants. The hypotheses in this study were not powered to test the rela tionships of the variab les of inquiry. The da ta in this study are limited to the original REACH investigators re search questions and data collection which may 78

PAGE 79

have negatively impacted this studys predictive variables. The sample size for this study was also near the minimum for l ogistic regression analysis (Hsieh, Bloch, & Larsen, 1998). Furthermore, there was no control or comparison group in the REACH trial. Perhaps the most relevant limitation associated with this study was that it did not consider cultural variables. This study did not examine the specific influence of culture in th e retention of the studys participants which ma y contribute to their motivation and investment in the research outcome. In addition, there is a danger associated with self-re port assessment of physical and emotional ailments, which are subjective and are influenced by the culture of ethnic groups. This study was limited in its inability to control fo r levels of acculturation in the two samples of Latinos, which may influence caregivers belief, a ttitudes, and willingness to participate and complete research studies. The study was al so limited in its scope, in that no measures considered important environmentally based va riables, such as community resources and transportation issues, were examined. Another limitation of this study was that pa rticipants were recr uited from physicians offices, Memory Disorder Clinics, and other clin ical settings; recruitmen t did not include other types of community-based caregivers. Additionally, the small number of males and caregivers older than 65 years of age limited the power an d applications of the findings. Thus, results should only be generalized to younger female Latinas. In addition, it is important to consider that the variables examined in this study may not have been effective predictors of drop-out stat us (the criterion) becau se the participants may have been more motivated, younger, and in better health than older caregi vers. Typically, older caregivers are at higher risk for potentially significant health problem s associated with older age, length of time caring for someone with progressive dementia (i.e., Alzheimers disease), and less 79

PAGE 80

social support, which may influence their ability and willingness to part icipate in research studies. Recommendations It is recommended that future research studies strengthen efforts to include Latino caregivers of persons with Alzheimers disease in th eir studies. It is im portant to note that the understanding of the variables ex amined in this study are embedde d in the cultural context. However, as mentioned earlier, this study did no t measure any cultural variables. Thus, it is recommended that future inquiries into Latino caregivers of pe rsons with Alzheimers disease are designed in a manner that allows for the measur ement of cultural considerations such as level of acculturation, language proficiency, and particip ants perceptions about research, as well as their perception about the variables of interest. In addition, becau se of the close-knit nature of Latino families, it is recommended that future research examines the caregiving experience from a family or systemic context. It is imperative that clinicians unders tand the obstacles hindering Latinos access to services (i .e., transportation issues, immi gration status, awareness of community services, etc). Religion is also a significant factor in the Latino community. Religious belief may influence Latinos belief and understanding about an illness or may be a valuable source of support and endurance. The role of religion in the Latino population and its influence on the caregiving experien ce warrants closer examination. Many researchers have brought attention to th e issue of language and language proficiency (Aranda, 2001; Gallagher-Thompson et al ., 2003; Levkoff & Sanchez, 2003; GallagherThompson, 1996). Therefore, conducting the stu dy in Spanish when appropriate and having interviewers who can speak the language fluent ly and accurately reflect the sample population are crucial and may help foster trust. Latinos may drop-out prematurely fr om research studies if 80

PAGE 81

they feel inadequate due to language limitations or if they do not feel at ease with the researcher. Thus, conducting consumer-center recruitment and retention practices that are informed by the culture of the target population, as proposed by Aren et al., (2003), is necessary to reach out effectively to this group. These culturally-appropriate practices may prove beneficial in alleviating the potential for stigma associated with participation in th e research study and the stress associated with sp eaking in another language. Gallagher-Thompson, Solano, Coon, and Aren (2003) report that one aspect of their success in recruiting and retain ing Latino caregivers was taking into account the practical barriers to participation in rese arch studies. Research protocol s should be designed to reduce the burden of participation and to cove r some of the cost such as travel and sitter services (Aren, Alvidrez, Nery, Estes, & Linkins, 2003; Ar en & Gallagher-Thompson, 1996). It is also imperative that researchers are mindful of th e time requirements. They must consider conducting shorter interviews, and, if necessary, breaking down longer assessments into smaller components. Researchers should also consider bringing the research to the Latino caregivers instead of expecting them to come to the research site. This may be accomplished by partnering with religious organizations or local businesses to conduct interviews in a neutral yet familiar environment. This notion is supported by the work of Levkoff and Sanches (2003). Consequently, attending to the barr iers associated with the group of interest may prove beneficial in minimizing any issues of mist rust regarding researchers. In addition, research questions should also have practical value for the research participants. It is important that Latino careg ivers understand the valu e of participating in research protocols, and, if possibl e, that they obtain services they normally would not access. For example, many caregivers report poor health as a consequence of their role as caregiver. 81

PAGE 82

Future research studies may benefit from addressing the impact of caregiving on the physical health of caregivers. This notion is supporte d in the literature by the work of GallagherThompson et al., (2003) and Cox and Monk (1993), among others. It also important that the research team provides feedback to the commun ity once the study is completed (Aren et al., 2003). Sharing the research findings and assisting prof essionals in the comm unity to translate the findings into practice serves to inform consumers, professionals, and to foment an atmosphere of trust and collaboration. Alternatively, future inves tigators should consider conducting an examination on the influence of these variables through different da ta analysis methods. The relationships among the variables age, income, health, and attrition warrant future investigation. The large scale nature of research like the REACH study allows the best probab ility of examining the strength of association between variables of interest. The REACH data are available for research from the Inter-university Consortium for Political and So cial Research (ICPSR) at the University of Michigan. Future research is ne eded in order to develop better models using predictors of dropout status, including attitudes and beliefs, and to explore alternative ways to identify predicators of attrition and recruitment and retention strategies for Lati no caregivers of persons with Alzheimers disease. As the scientific community continues to strengthen efforts to increase the recruitment and retention of Latinos in research studies, health profession als need to keep up with new findings. While more information is needed to inform future research and practice, current health professionals should translate these reco mmendations into practice. Thus, ensuring that Latino caregivers of persons with Alzheimers disease are receiving adequate and culturallyrelevant services. It is necessary that health professionals, such as counselors, working with 82

PAGE 83

Latino families understand and incorporate into thei r practice the values of this cultural group. These include: the close-knit and extended nature of the Latino family, degree of acculturation into mainstream society, and language profic iency, role of religion, perceptions about Alzheimers disease and research st udies, among others. However, it is also critical to consider differences within Latino groups. As seve ral investigators (Mohoney, Cloutterbuck, Neary, & Zhan 2005; Gallagher-Thompson et al., 2003) ha ve pointed out, members of the Latino population have the most geographically diverse origins. Moreover, it is important to keep in mind the generational differences within a single family unit which may manifest as different language preferences and degree of acculturation to mainstream society. Given the paucity of evidenced-based research findings, clinicians must take responsibility to ensure they are wellprepared to provide Latino families with the best available services. In order to gain culturallyappropriate skills, clinicians may participate in diversity trainings or incorporate Latino consultants in the advisory boards. If these were not available, the clinician should find a more appropriate referral source to work with this population. The implications regarding the expected incr ease in cases of Alzheimer's disease in the Latino community are overwhelming. Therefore, in order to keep up with the increasing numbers of Latino families dealing with Alzheimer' s disease, investigators must strengthen their efforts to design appropriate research, to transl ate cultural sensitivity into culturally competent recruitment and retention endeavors to enlist Latino participants into research studies. As the cost of caring for persons sufferi ng from Alzheimer's disease and related dementias continues to soar, it is imperative to adva nce the knowledge base about how caregivers carry out the demands associated with caregiving over the long-term. It is only through the acq uisition and application of this knowledge that society wi ll effectively and sensitively mana ge this looming health crisis. 83

PAGE 84

APPENDIX A CAREGIVER SOCIODEMOGRAPHIC INFORMATION Caregiver Sociodemographic Information Questionnaire. 1. During which visit is this interview taking place? 1 ( ) Baseline 2 ( ) 6 month follow-up visit 3 ( ) 12 month follow-up visit 4 ( ) 18 month follow-up visit 5 ( ) Other 1.1 Specify ______________________________________ 2. Date of interview: __ __/ __ __/ __ __ Month/day/year Now I would like to obtain some general information about you. 4. What is your marital status? Never married (0) Married or living as married (1) Widowed, not currently married (2) Divorced, not currently married (3) Separated (4) Unknown (-3) Refused (-4) 4.1. What is the primary occupation your spouse has had most of his/her working life? Since many people have mo re than one job at a given time, we would like to know about the jo b that is/was your spouse's primary source of income. ______________________________________________________________ 4.1.1 Job Code __ __ __ 84

PAGE 85

5. How many years of formal education did you complete? No formal education (0) Grade 1 (1) Grade 2 (2) Grade 3 (3) Grade 4 (4) Grade 5 (5) Grade 6 (6) Grade 7 (7) Grade 8 (8) Grade 9 (9) Grade 10 (10) Grade 11 (11) Grade 12/ High school diploma/ GED (G eneral Education Diploma) (12) Vocational/ training school after high school (13) Some college/ associate degree (14) College graduate (4 or 5 year program) (15) Master's degree (or other post-graduate training) (16) Doctoral degree (PhD, MD, Ed.D ., D.V.M., DDS., JD, etc.) (17) Unknown (-3) Refused (-4) 6. What country did you reside in during the last year of formal education? United States (1) Canada (2) Cuba (3) Mexico (4) Other (5 6.1. Specify _______________________________________ Unknown (-3) Refused (-4) 7. How would you describe your prim ary racial or ethnic group? White, Caucasian (1) Black, African-American (2) Native American, Eskimo, Aleut (3) Asian or Pacific Islander (4) Hispanic, Latino (5 85

PAGE 86

7.1. Would you describe yourself as Mexican, Mexican American, Chicano (1) Cuban or Cuban American (2) Puerto Rican (3) Dominican (4) Other (5) Unknown (-3) Refused (-4) 7.1.1. Specify:__________________________________________ No primary group (6) 7.2. Specify: _____________________ Other (7) 7.3. Specify:____________________________ Unknown (-3) Refused (-4) 8. In which country were you born? United States (1) Canada (2) Cuba (3) Mexico (4) Other (5) 8.1 Specify:_____________________________________ Unknown (-3) Refused (-4) 9. How many years have you lived in the United States? __ __ Years 10. What is the primary occupation you have ha d most of your working life? Since many people have more than one job at a given time, we wo uld like to know about th e job that is/was your primary source of income. __________________________________________________________________ 10.1 Job Code __ __ __ 86

PAGE 87

11. What is your curren t employment status? Employed at a job for pay, full-time (1) Employed at a job for pay, part time (2) Homemaker, not currently working for pay (3) Not currently employed, retired (4) Not currently employed, not retired (5) Unknown (-3) Refused (-4) 11.1. Are you employed outside of the home? No (0) Yes (1) Unknown (-3) Refused (-4 11.2. How many hours per week do you work at your paid job? __ __ : __ __ Hours:minutes 11.3. Have you had to reduce the number of hours that you work in an average week in order to provide care to (CR)? No (0) Yes (1) Unknown (-3) Refused (-4) 11.3.1. How many hours have you ha d to reduce per week? __ __ : __ __ Hours:minutes 11.4. Did you stop working because of (CR)s need for care? No (0) Yes (1) Unknown (-3) Refused (-4) 11.4.1. Why? You anticipated having to care for (CR) (1) To provide additional care for (CR) (2) Other (3) 11.4.1.1. Specify _________________________________________ Next, I would like to ask you about your household income. Some people may not be comfortable answering this questi on, but I want to assure you that your responses will be kept strictly confidential. This information is very important to the project because it helps us understand how caregiving affects pe ople with different incomes. 87

PAGE 88

12. Which category on this card [give responde nt card] best describes your yearly household income before taxes? Do not give me the dollar amount, just give me the category. Include all income received from employment, social security, support from children or other family, welfare, Aid to Families with Dependent Children (AFDC), bank interest, retirement accounts, rental property, investments, etc. Less than $5000 (0) $5000 $9,999 (1) $10,000 $14,999 (2) $15,000 $19,999 (3) $20,000 $29,999 (4) $30,000 $39,999 (5) $40,000 $49,999 (6) $50,000 $59,999 (7) $60,000 $69,999 (8) $70,000 or more (9) Unknown (-3) Refused (-4) 13. How hard is it for you to pay for the very basics like food, housing, medical care, and heating? Would you say it is: Not difficult at all (1) Not very difficult (2) Somewhat difficult (3) Very difficult (4) Unknown (-3) Refused (-4) 14. How many people are living with you in your home excluding yourself? ___ ___ persons 15. How long have you lived with (CR)? ___ ___ years 16. Did you and (CR) start living together so that you could take care of him/her? No (0 ) Yes (1) Unknown (-3) Refused (-4) 88

PAGE 89

APPENDIX B CENTER FOR EPIDEMIOLOGIC STUDIES DEPRESSION SCALE (CES-D) CES-D Questionnaire. 1. During which visit is this interview taking place? 1 ( ) Baseline 2 ( ) 6 month follow-up visit 3 ( ) 12 month follow-up visit 4 ( ) 18 month follow-up visit 5 ( ) Other 1.1 Specify ______________________________________ 2. Date of interview: __ __/ __ __/ __ __ Month/day/year This section deals with statements people might make about how they feel. Let me give you a card with possible responses. [Give respondent car d.] For each of the statements, please indicate how often you felt that way during the past week. 4.1. I was bothered by things that usually don't bother me. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.2. I did not feel like ea ting; appetite was poor. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.3. I felt that I could not shake off the blues, even with help from my family and friends. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 89

PAGE 90

4.4. I felt that I was just as good as other people. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.5. I had trouble keeping my mind on what I was doing. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.6. I felt depressed. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.7. I felt that everything th at I did was an effort. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.8. I felt hopeful about the future. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 90

PAGE 91

4.9. I thought my life had been a failure. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.10. I felt fearful. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.11. My sleep was restless. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.12. I was happy. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.13. I talked less than usual. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 91

PAGE 92

4.14. I felt lonely. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.15. People were unfriendly. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.16. I enjoyed life. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.17. I had crying spells. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.18. I felt sad. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 92

PAGE 93

4.19. I felt that people disliked me. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 4.20. I could not get going. Rarely or none of the time [< 1 day] (0) Some or a little of the time [1-2 days] (1) Occasionally or a moderate amount of time [3-4 days] (2) Most or almost all of the time [5-7 days] (3) Unknown (-3) Refused (-4) 5. CES-D Score ___ ___ 6. Is the CES-D score greater than or equal to 28? No ( ) Yes ( ) 6.1 Has the Principal Investig ator or appropriate site personnel been notified? No ( ) Yes ( ) Please notify the Principa l Investigator or appr opriate site personnel. 93

PAGE 94

APPENDIX C SOCIAL SUPPORT Social Support (SS) Questionnaire. 1. During which visit is this interview taking place? Baseline (1) 6 month follow-up visit (2) 12 month follow-up visit (3) 18 month follow-up visit (4) Other (5) 1.1 Specify ______________________________________ 2. Date of interview: __ __/ __ __/ __ __ Month/day/year Social Networks Now I would like to ask you some questions about your friends and family. 4.0. Overall, how satisfied have you been in the last month with the help you have received from friends, nei ghbors, or family members? Not at all (0) A little (1) Moderately (2) Very (3) Unknown (-3) Refused (-4) 4.1. How many relatives other than (CR) do you see or hear from at least once a month? None (0) One (1) Two ( ) Three or four (3) Five to eight (4) Nine or more (5) Unknown (-3) Refused (-4) 94

PAGE 95

4.2. Think about the relative ot her than (CR) with whom you have the most contact. How often do you see or hear from that person? Less than monthly (0) Monthly (1) A few times a month (2) Weekly (3) A few times a week (4) Daily (5) Unknown (-3 Refused (-4) 4.3. How many relatives other than (CR) do you feel close to? That is, how many do you feel at ease with, can talk to abou t private matters, or can call on for help? None (0) One (1) Two (2) Three or four (3) Five to eight (4) Nine or more (5) Unknown (-3) Refused (-4) 4.4. How many friends do you feel close to ? That is, how many friends (not including relatives) do you f eel at ease with, can talk to about private matters, or can call on for help? None (0) One (1) Two (2) Three or four (3) Five to eight (4) Nine or more (5) Unknown (-3) Refused (-4) 4.5. How many of these friends do you see or hear from at least once a month? (not including relatives) None (0) One (1) Two (2) Three or four (3) Five to eight (4) Nine or more (5) Unknown (-3) Refused (-4) 95

PAGE 96

4.6. Think about the friend (not including relatives) with whom you have the most contact. How often do you see or hear from that person? Less than monthly (0) Monthly (1) A few times a month (2) Weekly (3) A few times a week (4) Daily (5) Unknown (-3) Refused (-4) 4.7. When you have an important decision to make, do you have someone other than (CR) you can talk to about it? Never (0) Seldom (1) Sometimes (2) Often (3) Very often (4) Always (5) Unknown (-3) Refused (-4) 4.8. When other people you know have an important decision to make, do they talk to you about it? Never (0) Seldom (1) Sometimes (2) Often (3) Very often (4) Always (5) Unknown (-3) Refused (-4) Received Support and Satisfaction 5.1. In the past month, how often has someone, such as a friend, neighbor, or family member other than (CR), pr ovided transportation for you? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 96

PAGE 97

5.2. In the past month, how often has someone, such as a friend, neighbor, or family member other than (CR), pitched in to he lp you do something that needed to get done, like household chores or yardwork? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.3. In the past month, how often has someone helped you with shopping? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.4. Overall, how satisfied have you been in the last month with the help you have received with transportation, hous ework and yardwork, and shopping? Not at all (0) A little (1) Moderately (2) Very (3) Unknown (-3) Refused (-4) 5.5. In the past month, how often was someone right there with you (physically) in a stressful situation? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.6. In the past month, how often has someone provided comfort to you? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 97

PAGE 98

5.7. In the past month, how often has someone listened to you talk about your private feelings? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.8. In the past month how often has some one expressed interest and concern in your well-being? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.9. In the past month, how satisfied have you been with the support received during difficult times, comforting from others, how others have listened, and interest and concern from others? Not at all (0) A little (1) Moderately (2) Very (3) Unknown (-3) Refused (-4) 5.10. In the past month, how often has someone suggested some action you should take in dealing with a problem you were having? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.11. In the past month, how often has someone made a difficult s ituation clearer and easier to understand? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 98

PAGE 99

5.12. In the past month, how often has someone helped you understand why you did not do something well? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.13. In the past month, how often has someone told you what they did in a situation that was similar to one you were experiencing? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 5.14. Overall, how satisfied in the last m onth have you been with the suggestions, clarifications, and sharing of similar experiences you have received from others? Not at all (0) A little (1) Moderately (2) Very (3) Unknown (-3) Refused (-4) Id like to ask you a few more questions about your relationship with others. Remember, when the term others is use d, it includes friends, neighbors, or family members other than (CR). Negative Interaction 6.1. In the past month, how often have ot hers made too many demands on you? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 99

PAGE 100

6.2. In the past month, how often have others been critical of you? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 6.3 In the past month, how often have others pried into your affairs? Never (0) Once in awhile (1) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 6.4 In the past month, how often have others taken advantage of you? Never ( ) Once in awhile ( ) Fairly often (2) Very often (3) Unknown (-3) Refused (-4) 100

PAGE 101

APPENDIX D CAREGIVER HEALTH AND HEALTH BEHAVIOR Caregiver Health and Health Behaviors Questionnaire. 1. During which visit is this interview taking place? Baseline (1) 6 month follow-up visit (2) 12 month follow-up visit (3) 18 month follow-up visit (4) Other (5) 1.1 Specify ______________________________________ 2. Date of interview: __ __/ __ __/ __ __ Month/day/year Now I would like to ask a few questions about your health. In general, would you say your health is: Poor (1) Fair (2) Good (3) Very good (4) Excellent (5) Unknown (-3) Refused (-4) How true or false are each of the following statements for you? Please refer to the responses listed on this card. 5.1 I seem to get sick a little easier than other people. Definitely false (1) Mostly false (2) Neither false nor true (3) Mostly true (4) Definitely true (5) Unknown (-3) Refused (-4) 101

PAGE 102

5.2. I am as healthy as anybody I know. Definitely false (1) Mostly false (2) Neither false nor true (3) Mostly true (4) Definitely true (5) Unknown (-3) Refused (-4) 5.3. I expect my health to get worse. Definitely false (1) Mostly false (2) Neither false nor true (3) Mostly true (4) Definitely true (5) Unknown (-3) Refused (-4) Do you currently have, or has a doctor told you that you have, any of the following health problems? No (0) Yes (1) Unknown (-3) Refused (-4) 6.1. Arthritis No (0) Yes (1) Unknown (-3) Refused (-4) 6.2. High Blood Pressure No (0) Yes (1) Unknown (-3) Refused (-4) 6.3. Heart Condition (specifically heart disease, heart attack, chest pain due to your heart, congestive heart fail ure, angina, MI). No (0) Yes (1) Unknown (-3) Refused (-4) 102

PAGE 103

6.4. Chronic Lung Disease such as chronic bronchitis or emphysema (not asthma). No (0) Yes (1) Unknown (-3) Refused (-4) 6.5. Diabetes No (0) Yes (1) Unknown (-3) Refused (-4) 6.6. Cancer No (0) Yes (1) Unknown (-3) Refused (-4) 6.7. Have you ever had or been told by a doctor that you had a stroke? No (0) Yes (1) Unknown (-3) Refused (-4) During the past two weeks, to what extent have you experienced the following symptoms. 7.1 Dizziness Not at all (0) A little (1) A lot (2) Unknown (-3) Refused (-4) 7.2. Headaches Not at all (0) A little (1) A lot (2) Unknown (-3) Refused (-4) 103

PAGE 104

7.3. Stomach or bowel problems Not at all (0) A little (1) A lot (2) Unknown (-3) Refused (-4) The following questions ask about your daily routine. 8.1. Do you typically eat fewer than 2 meals per day? No (0) Yes (1) Unknown (-3) Refused (-4) 8.2. Do you have 3 or more drinks of b eer, liquor or wine almost every day? No (0) Yes (1) Unknown (-3) Refused (-4) 8.3. Without wanting to, have you gained or lost 10 lbs or more in the last 6 months? No (0) Yes (1) Unknown (-3) Refused (-4) 8.4. Do you do vigorous exercises fo r 15-30 minutes or more at least 3 times a week? (Examples include running, sports, swimming, brisk walk ing, job that involves physical labor.) No (0) Yes (1) Unknown (-3) Refused (-4) 8.5. Do you smoke cigarettes now? No (0) Yes (1) Unknown (-3) Refused (-4) 8.5.1 About how many cigarettes do you usually smoke in a day now? cigarettes/day: ___ ___ 104

PAGE 105

REFERENCE LIST Abu-Rayya, H. M. (2006). Depression and social involvement among elders. The Internet Journal of Health, 5 Retrieved January 14, 2006, from, http://www.ispub.com/ostia/index.php?xmlFile Path=journals/ijh/vol5n1/depression.xml Alzheimers Association. (2007). Alzheimers di sease facts and figures 2007. Retrieved April 2, 2007, from http://www.alz.org/national/ documents/Report_2007FactsAndFigures.pdf Alzheimers Association. (2004). Hispanic/Latinos and Alzheimers disease. Retrieved August 20, 2006, from http://www.alz.org/national/documents/HispanicReportEnglish.pdf Alzheimers Association and National Alli ance for Caregiving. (2004). Families care: Alzheimers caregiving in the United St ates. Retrieved September 16, 2006, from http://www.caregiving.org/d ata/alzcaregivers04.pdf Alzheimers Caregivers Support Online (AlzOnlin e). (2006). Basics of dementia. Retrieved Ocotober 12, 2006 from, http://alzonline.phhp.ufl.edu/en/read ing/BasicsDemSess3UpdateFeb06.pdf Aranda, M. P. (2001). Racial and ethnic factors in dementia care-giving res earch in the US. Aging & Mental Health, 2001, 5 (S1), S116S123 Aren, P.A., Alvidrez, J., Nery, R., Estes, C. & Linkins, K. (2003). Recru itment and retention of older minorities in mental h ealth services research. The Gerontologist, 43 36. Aren, P.A. & Gallagher-Thompson, D. (1996). Issu es and recommendations for the recruitment and retention of older ethnic minority adults into clinical research. Journal of Consulting and Clinical Psychology. 64, 875. Argimon, J. M., Limon, E., Vila, J., & Cabezas, C. (2004). Health-related quality of life in carers of patients with dementia. Family Practice Vol. 21 454. Bank, A. L., Arguelles, S., Rubert, M., Eisdorfer, C., & Czaja, S. J. (2006). The value of telephone support groups among ethni cally diverse caregiv ers of persons with dementia. The Gerontologist, 46, 134. Blakeley, R. L. (1997). A clandestine past. In Blakeley, R. L., & Harrington, J. M. (Eds). Bones in the basement: Postmortem racism in nineteenth-century medical training (pp. 3). Washington, DC: Smithsonian Institution Press. Blazer, D. G. (2003). Depression in la te life: Review and commentary. The Journal of Gerontology: Medical Sciences, 58 249. 105

PAGE 106

Briones, D. F., Ramirez, A. L., Guerrero, M., & Ledger, E. (2002). Determining cultural and psychosocial factors in Alzheimers disease among Hispanic populations. Alzheimers Disease and Associated Disorders, 16 (S2), S86S88. Bourgeois, M. S., Schulz, R., & Burgio, L. (1996) Interventions for caregivers of patients With Alzheimers disease: A review and analys is of content, process, and outcomes. International Journal of Aging and Human Development, 43, 35. Cannuscio, C. C., Jones, C., Kawachi, I., Coldit z, G. A., Berkman, L. & Rimm, E. (2002). Reverberations of family illness: a longitudinal assessment of informal caregiving and mental health status in the nurses' health study. American Journal of Public Health, 92 1305. Center for Health Communication Harvard Scho ol of Public Health & MetLife Foundation. (2004). Reinventing Aging. Baby boomers a nd civic engagements. Retrieved August 31, 2006, from http://www.hsph.harvard.edu/chc/reinventingaging/Report.pdf Center for Studying Health System Change. (2004). Trends in health insurance coverage and access among Black, Latino and White Americans, 2001-2003. Retrieved August 31, 2005, from http://hschange.org/CONTENT/713/713.pdf Cohen, R. F., Boyle, C. A. Coakley, D., & Lawlor B. A. (1999). Dementia carer education and patient behaviour disturbance. International Journal of Geriatric Psychiatry, 14, 302. Connell, C. M., Javenic, M. R., & Gallant, M. P. (2001). The cost of caring: Impact of dementia on family caregivers. Journal of Geriatric Psychiatry and Neurology, 14, 179. Coon, D. W., Gallagher-Thompson, D. & Thompson, L. (Eds.) (2003). Innovative interventions to reduce dementia caregiver distress: A clinical guide. New York: Springer. Coon, D. W., Rubert, M., Solano, N., Mausbach, B., Kraemer, H., Arguelles, T. Haley, W. E., Thompson, L. & Gallagher-Thompson, D. (2004). Well-being, appraisal, and coping in Latina and Caucasian female dementia care givers: findings from the REACH study. Aging & Mental Health, 8 330. Corbeil, R. R., Quayhagen, M. P., Quayhagen, M. (1999). Intervention effects on dementia caregiving interaction: A stress -adaptation modeling approach. Journal of Aging and Health, 11 (1), 79. Cotter, R. B., Burke, J. D., Loeber, R. & Mutchka, J. (2005). Predictors of contact difficulty and refusal in a longitudinal study. Criminal Behaviour and Mental health, 15 126. Covinsky, K.E., Newcomer, R., Fox, P., Wood, J., Sands, L., Dane, K. & Yaffe, K. (2003) Patient and caregiver characteristics associated w ith depression in caregivers of patients with dementia Journal of General Internal Medicine; 18 1006. 106

PAGE 107

Cox, C., & Monk, A. (1996). Strain among caregiv ers: Comparing the experiences of African Americans and Hispanic caregivers of Alzheimers relatives. International Journal of Aging and Human Development, 43 (2), 93. Curry, L & Jackson, J. (2003). The science of including older ethnic and racial group participants in health-related re search. The Gerontologist, 43, 15. Czaja, S., Schulz, R., Lee, C.C. & Belle, S. H. (2003). A methodology for describing and decomposing complex psychosocial and behavioral interventions. Psychology and Aging, 18, 385. Department of Elder Affairs. (2006). Florida county prof iles 2006. Retrieved March 02, 2007, from http://elderaffairs.state.fl .us/english/Stats/DOCS/2006cp.pdf Dilworth-Anderson, P., & Gibson, B. E. (2002). The cultural influence of values, norms, meanings, and perceptions in understanding dementia in ethnic minorities. Alzheimers Disease and Associated Disorders, 16 (S2), S56S63. Doty, L., (2007). Renewing trust in regular (allopathic) medicine and research. SELAM International News, 9 (1), 1 (in press). Eisdorfer, C., Czaja, S. J., Loewenstein, D. A., Rubert, M. P., Arguelle s, S., Mitrani, V.B & Szapocznick, J. (2003). The effect of a fa mily therapy and technology-based intervention on caregiver depression. The Gerontologist, 43 521. Evans, D. A., Funkenstein, H.H., Albert, M.S., Scherr, P.R., Cook, N.R., Chown, M.J., Hebert, L.E., Hennekens,C.H. & Taylor, J.O., (1989). Estimated prevalence of Alzheimers disease in a community population higher than previously reported. Journal of the American Medical Association, 262 2251. Family Caregiving Alliance. (2001). Fact sh eet: Selected caregiver statistics. Retrieved September 15, 2006, from http://www.caregiver.org/caregiver /jsp/content_node .jsp?nodeid=439 Fox, P., Newcomer, R., Yordi, C. & Arnsberger, P. (2000). Lessons Lear ned from the Medicare Alzheimer Disease Demons. Alzheimer disease and associated disorders, 14 87. Fryling, K. (2006, August 17). Caregivers face challenges with "silver tsunami." University at Buffalo Reporter, 37. Retrieved October 03, 2006, from http://www.buffalo.edu/reporter/vol37/vol37n46/articles/WaldropLecture.html Gallagher-Thompson, D., Coon, D, Solano, N. Am ber, C., Rabinowitz, Y., & Thompson, L. (2003). Change in indices of distress among Latino and Anglo female caregivers of elderly relatives with dementia: Site-specific result s from the REACH nationa l collaborative study. The Gerontologist, 4,3 580. 107

PAGE 108

Gallagher-Thompson, D., Lovett, S., Rose, J., McKibbin, C., Coon, D. W., Futterman, A. & Thompson, L. (2000). Impact of psychoeduca tional interventions on distressed family caregivers. Journal of Clinical Geropsychology 6, 91. Gallagher-Thompson, D., Singer, L. S., Depp, .C., Mausbach, B. T., Cardenas, V. & Coon, D. W (2004). Effective recruitment strategies for La tino and Caucasian dementia family caregivers in intervention research. American Journal of geriatric Psychiatry, 12 484. Gallagher-Thompson, D., Solano, N., Coon, D, & Are n, P. (2003). Recruitment and retention of Latino dementia family caregivers in interven tion research: Issues to face, lessons to learn. The Gerontologist, 43 (1 ), 45. Haley, W.E., Levine, E.G., Brown, S. L., & Bartolucci,A. A. (1987). Stress, appraisal, coping, and social support as predictors of adapta tional outcome among dementia caregivers. Psychology & Aging, 2, 323. Henderson, J. N., Gutierrez-Myka, M., Garcia, J. & Boyd, S. (1993). A model for Alzheimers disease support group development in Afri can American and Hispanic populations. The Gerontologist, 33 (3), 409. Kaiser Family Foundation. (2003, June). Key facts: Race, ethnicity & medical care. Retrieved October 03, 2006, from http://www.kff.org/minorityhealth/upload/Key-FactsRace-Ethnicity-Medical-Care-Chartbook.pdf Hilton, N. Z., Harris, G.T., & Rice, M. E. (2003) Correspondence between self-report measures of interpersonal aggression. Journal of Interpersonal Violence, 18 223. Hinton, L. (2002). Improving care for ethnic minor ity elderly and their fa mily caregiver across the spectrum of dementia severity. Alzheimers Disease and Associated Disorders, 16 (S2), S50S55. Honas, J. J., Early, J. L., Frederickson, D. D., & O Brien, M. S. (2003). Predic tors of attrition in a large clinic-based weight-loss program. Obesity Research 11, 888. Horizon Project. (2000, May 12). A profile of Hispanic elders. San Antonio, TX: Valdez R. B. & Arce, C. Retrieved February 11, 2006, from http://latino.si.edu/virtua lgallery/GrowingO ld/Nationwide%20Demographic.pdf Hsieh, F. Y., Bloch, D., A., & Larsen, M. D. (1998). A simple method of sample size calculation for linear and logi stic regression. Statistics in medicine, 17 1623. Huck, S. W. (2004). Reading statistics and research Boston: Allyn & Bacon. Janevic, M. R. & Connell, C. M. (2001). Racial, ethnic, and cultural differ ences in the dementia caregiving experience: recent finding s. The Gerontologist, 41, 334. 108

PAGE 109

Jones, J. (1993). Bad blood. New York: The Free Press. Keller, C. S., Gonzalez, A., & Fleuriet, K. J. (2005). Retention of mi nority participants in clinical research studies. Western Journal of Nursing Research, 27, 292. Larkey, L. K., Hecht, M. L., Miller, K. & Alat orre, C. (2001). Hisp anic cultural norms for healthseeking behaviors in the face of symptoms. Health Education & Behavior, 28 65 80. Lee, S., Colditz, G. A., Berkman, L. F. & Kawachi, I. (2003). Caregiving and risk of coronary heart disease in U.S. women: A prospective study. American Journal of Preventive Medicine, 24 113. Lefly, H. P. (1996). Cross-cultural i ssues in family caregiving. In Family caregiving in mental illness. (pp. 149). Thousand Oaks: Sage Publications. Levkoff, S. & Sanchez, H. (2003). Lessons learned minority recruitment and retention from the centers on minority aging and health promotion. The Gerontologist, 43, 18. Liao, Y., Tucker, P. & Giles, W. H. (2004). Health status among REACH 2010 communities, 20012002. Ethnicity & Disease, 14 S1-9S1. Loomis, J. S., Freytes, M, & Ketterson, T.U. (A pril, 2005). Internet-based caregiver education support: Program evaluation. Presented at the annual meeting of the American Telemedicine Association (Published abstract). Ludmerer, K. M. (1999). Time to heal New York: Oxford University Press. Luna, I., Torres de Ardn, E., Mi Lim, Y., Crom well, S. L., Phillips, L. R., & Russell, C. K. (1996). The relevance of familism in cro ss-cultural studies of family caregiving. Western Journal of Nursing research, 18 (3), 267. Majerovitz S. D. (2001). Formal versus info rmal support: Stress buffering among dementia caregivers. Journal of Mental Health and Aging, 7 413. Manton, K. G., Corder, L. & Stallard, E. ( 1997). Chronic disability tr ends in elderly United States populations: 1982-1994. Proceedings of the National Academy of Sciences of the United States of America, 94 2593. Markides,K.S., Rudkin, L., Angel, R.J. & Espino, D.V. (1997). Health status of Hispanic elderly. In Martin, L.J. & Soldo, B. (Eds.). Racial and ethnic differences in the health of older Americans. Washington, DC.: National Academy Press. McGinnis,K. A., Schulz, R., Stone, R.A., Klinger, J. & Mercurio, R. (2006). Concordance of race or ethnicity of interventionists and caregivers of dementia patients: Relationship to attrition and treatment outcomes in the REACH study. The Gerontologist, 46 449. 109

PAGE 110

Metropolitan Life Insurance Company. (2005). Me tLife Mature market Institute Demographic Profile of American Baby Boomers. Retrieved August 8, 2006, from http://www.metlife.com/App lications/Corporate/WPS/C DA/PageGenerator/0,4132,P7328^S 701,00.html?FILTERNAME=@URL\&FILTERVALUE=/WPS/ Miller, B. & Mukherjee, S. (1999). Service use, caregiving mastery, and attitudes toward community services. Journal of Applied Gerontology, 18, 162. Miller, B., Townsend, A., Carpenter, E., Montgomer y, R. V. J., Stull, D. & Young, R. F. (2001). Social support and caregiver distress: A repl ication analysis. The Journal of Gerontology Series B: Psychological Sciences and Social Sciences, 56, S249S256. Miranda, J., Azocar, F., Organista, K. C., Muoz R. F. & Lieberman A., (1996). Recruiting and retaining low-income Latinos in Psychothe rapy research. Journal of Consulting and Clinical Psychology, 64, 868. Mirowsky, J. & Reynolds, J. R. (2000). Age, depre ssion, and attrition in th e national survey of families and households. Sociological Methods & Research, 28, 476. Mitrani, V. B., Lewis, J. E., Feaster, D. J ., Czaja, S. J., Eisdorfer, C., Schulz, R. & Szapocznick,J. (2006). The role of family f unctioning in the stress process of dementia caregivers: A structural family framework, The Gerontologist, 46, 97. Moreno-John, G., Gachie, A., Fleming, C. M., Npoles -Springer, A., Mutran, E., Manson, S. M. & Prez-Stable, E. J. (2004). Ethnic minority older adults participating in clinical research: Developing trust. Journal of Aging and Health, 16 93SS. Murthy, V. H., Krumholz, H.M. & Gross, C. P. (2004) Participation in cancer clinical trials: Race, sex-, and age-based disparities. Journal of the American Medical Association, 291 2720. National Caregiving Alliance and American Asso ciation of Retired Persons. (2005). Family Caregiving in the U.S. Findings from a natio nal study. Retrieved September, 2, 2006, from http://www.caregiving.org/data/Family %20Caregiving%20in%20the%20US.pdf Navaie-Waliser, M., Feldman, P. H., Gould, D. A ., Levine, C., Kuerbis, A. N. & Donelan,K. (2002). When the caregiver needs care: the plight of vulnerable caregivers. American Journal of Public Health, 9 409. Oboler, S. (1995). Ethnic labels, Latino lives: Identity and the politics of (re)presentation in the United States Minneapolis: University of Minnesota Press. Pew Hispanic Center. (2005). Hi spanics: A people in motion. Retrieved July 31, 2006, from http://pewhispanic.org/files/reports/40.pdf 110

PAGE 111

Phillips, L. R., Torres de Ardn, E., Komnenich, P., Killeen. M., & Rusinak, R. (2000). The Mexican American caregiving experience. Hispanic Journal of Behavioral Sciences, 22, 296. Pillemer, K & Suitor, J. J. (2002). P eer support for Alzheimers caregivers. Research on Aging, 24, 171. Pinquart, M. & Sorensen, S. (2005). Ethnic differences in stressors, resources, and psychological outcomes of family caregiving: A meta-analysis. The Gerontologist, 45 90 106. Prigerson, H. G. (2003). Cost to society of family caregiving for patients with end-stage Alzheimers disease. New England Journal of Medicine, 349 1891. Radloff, L. S. (1977). The CE S-D Scale: A self-re port depression scale for research in the general population. Journal of Applied Psyc hological Measures, 1 (3), 385. Semple, S. J. (1992). Conflict in Alzheimer' s caregiving families: Its dimensions and consequences. The Gerontologist, 32,648. Schulz, R. (2003). Resources for enhancing Alzheimers caregiver health 1996-2001(ICPSR version) [Computer file]. Ann Arbor, MI: Consortium for Political and Social Research. Schulz, R. (Ed.) (2000). Handbook on dementia caregiving: Eviden ced-based interventions for family caregivers. New York: Springer. Schulz, R. & Beach, S. R. (1999). Caregi ving as a risk factor for mortality. The Journal of the American Medical Association, 282, 2215. Schulz, R., Burgio, L., Burns, R., Eisdorfe r, C., Gallagher-Thompson, D., Gitlin, L.N. & Mohoney, D. F. (2003). Resources for enhancin g Alzheimer's caregiver health (REACH): Overview, site-specific outcomes, and future directions. The Gerontologist, 43 514 Schulz, R., OBrien, A, T., Bookwala, J., & Fleissner, K. (1995). Psychiatric and physical morbidity effects of Alzheimers disease caregiving: Prevalence, correlates, and causes. The Gerontologist, 35, 771. Schulz, R., Visintainer, P., & Williamson, G. M. (1990). Psychiatric and physical morbidity effects of caregiving. Journal of Gerontology: Psychological Sciences, 45 181. Shorris, E. (1992). Latinos: A biography of the people New York: Avon Books. Sorensen, S., Pinquart, M. & Duberstein, P. (20 02). How effective are interventions with caregivers? An updated meta-analysis. The Gerontologist, 42 356. 111

PAGE 112

Sue, D. W., & Sue, D. (1990). C ounseling Hispanic Americans. In Counseling the culturally different. (pp. 227). New York: John Wiley & Sons. Switzer, G. E., Wisniewski, S. R., Belle, S. H. Dew, M. A. & Schultz, R. (1999). Selecting, developing, and evaluating research instruments. Social Psychiatry & Psychiatry Epidemiology, 34, 399. Torres-Rivera, E., Wilbur, M. P., Roberts-Wil bur, J., & Phan, L. (1999). Group work with Latino clients. Journal for Specialists in Group Work, 24, 383. U.S. Administration on Aging (2005). Serving our Hispanic American Elders. Washington, DC: Department of Health and Human Services. U. S. Census Bureau. (2001 ). Mapping Census 2000: Th e geography of U.S. diversity (Series CENSR/01-1). Washington, DC : U.S. Government Printing Office. U.S. Department of Health and Human Services. (2005). National diabetes statis tics fact sheet: General information and national estimates on diabe tes in the United States Bethesda, MD: National Institutes of Diabetes a nd Digestive and Kidney Diseases. USA Today. (2007). Alzheimers cases up 10% to 5 million in half decade. Retrieved April 2, 2007, from http://www.usatoday.com/news/h ealth/2007-03-20-alzheimersreport_N.htm?csp=34 Urizar G. G. & Sears, F. S. (2006). Psychos ocial and cultural influe nces on cardiovascular health and quality of life among Hispanic cardiac patients in South Florida. Journal of Behavioral Medicine, 29 255. Valle, R. & Lee, B. (2002). Research priorities in the evolving demographic landscape of Alzheimers disease and associated dementias. Alzheimers Disease and Associated Disorders, 16 (S2), S64S76. Villarruel, A. M., Jemmott, L. S., Jemmott, J. B. & Eakin, B. L (2006). Recruitment and retention of Latino adolescents to a research study: Lessons learned from a randomized clinical trial. Journal for Specialist in Pediatric Nursing, 11 244. Warren-Findlow,J., Prohaska, T.R. & Freedman, D. (2003). Challenges and opportunities in recruiting and retaining underrepresented popul ations into health promotion research. The Gerontologis, 43, 37. Weiss, C. O., Gonzalez, H. M., Kabeto, M. U. & Langa, K. M. (2005). Differences in amount of informal care received by non-Hispanic whites and Latinos in a nationally representative sample of older americans. Journal of the American Geriatrics Society, 53 146. 112

PAGE 113

Wiliams, C., Tappen, R., Buscemi, C., Rivera, R., & Lezcano, J. (2001). Obtaining family consent for participation in Alzheimers research in a Cuban-American population: Strategies to overcome the barriers. American Journal of Alzheimers Disease and Other Dementias, 16, 183. Wisniewski, S. R.; Belle, S. H.; Coon, D. W.; Ma rcus, S. M.; Ory, M. G.; Burgio, L. D.; Burns, R., & Schulz, R. (2003). The resources fo r enhancing Alzheimer's caregiver health (REACH): Project design and baseline characteristics. Psychology & Aging, 18 (3), 375 374. 113

PAGE 114

BIOGRAPHICAL SKETCH Magaly Freytes was born and raised in Puerto Rico. She moved to Gainesville, Florida, in 1992 to attend the University of Florida wher e she earned a B.A. in sociology in 1996. After working for a year, she decided to pursue a graduate educatio n. In 1999 she earned a masters and a specialist in education degrees (M.Ed/Ed.S) in mental health counseling from the Department of Counselor Education from the same institution. After gaining clinical experience working as a counselor in a commun ity mental health facility for several years, she returned to graduate school to earn a Ph.D. in Counselor Education in May 2007. Since 2002 Ms. Freytes has been working for the Center for Telehealth & Healthcare Communications in the College of Public Health & Health Professions at the University of Florida. She is the Coordinator for the Alzheimers Caregiver Support Online project, where she develops and delivers information, education, an d support for informal and formal dementia caregivers in English and Spanish. In addition, she has been a volunteer at the Alachua County Crisis Center for over 10 years and has collaborated in a handful of proj ects geared toward the investigation of diverse populations. Ms. Freytes has been a presenter numerous times at local, state, and national conferences for academic, professional, an d general public audiences. 114