Resources for Health: A Primary-Care-Based Diet and Physical Activity Intervention Targeting Urban Latinos With Multiple Chronic Conditions

Health Psychology
Copyright 2007 by the American Psychological Association
2007, Vol. 26, No. 4, 392– 400
0278-6133/07/$12.00
DOI: 10.1037/0278-6133.26.4.392
Resources for Health: A Primary-Care-Based Diet and Physical Activity
Intervention Targeting Urban Latinos With Multiple Chronic Conditions
Elizabeth G. Eakin
Sheana S. Bull
Queensland Cancer Fund and Queensland University of
University of Colorado Health Sciences Center, Denver
Technology
Kimberley M. Riley
Marina M. Reeves
Center for Research Strategies
Queensland Cancer Fund
Patty McLaughlin and Silvia Gutierrez
Clinica Campesina Family Health Services
Objective: The Resources for Health trial evaluates a social-ecologically based lifestyle (physical activity
and diet) intervention targeting low-income, largely Spanish-speaking patients with multiple chronic
conditions. Design: A randomized controlled trial was conducted with 200 patients recruited from an
urban community health center and assigned to intervention and usual care conditions. Intervention
involved 2 face-to-face, self-management support and community linkage sessions with a health
educator, 3 follow-up phone calls, and 3 tailored newsletters. Main Outcome Measures: Primary
outcomes measured at 6-months were changes in dietary behavior and physical activity. Changes in
multilevel support for healthy living were evaluated as a secondary outcome. Results: After adjustment
for age, sex, language, and number of chronic conditions, significant intervention effects were observed
for dietary behavior and multilevel support for healthy lifestyles but not for physical activity. Conclu-
sion: The Resources for Health intervention provides an effective and practical model for improving
health behavior among low-income, Spanish-speaking patients with multiple chronic conditions.
Keywords: RCT, self-management, health disparities, comorbidity, RE-AIM
There is widespread agreement that patient involvement in
for prevention of disease complications (Von Korff, Glasgow, &
disease management (referred to as self-management or collabo-
Sharpe, 2002; Wagner & Groves, 2002). For the majority of
rative management) is required for control of chronic disease and
chronic conditions, this involves addressing multiple behavioral
risk factors (i.e., physical activity, diet, smoking and alcohol), as
well as monitoring and managing the signs of symptoms of dis-
ease, taking medications appropriately, maintaining regular con-
Elizabeth G. Eakin, Viertel Centre for Research in Cancer Control,
Queensland Cancer Fund, Brisbane, Queensland, Australia, and Centre for
tact with health care providers, and managing emotional and social
Health Research–Public Health, Queensland University of Technology,
sequelae (Clark, Becker, Lorig, Rakowski, & Anderson, 1991;
Brisbane, Queensland, Australia; Sheana S. Bull, Colorado Health Out-
Lorig et al., 1999; Wagner, 2001).
comes Program, University of Colorado Health Sciences Center, Denver;
The past two decades have witnessed a wealth of research
Kimberley M. Riley, Center for Research Strategies, Denver, Colorado;
demonstrating the efficacy of chronic disease self-management
Marina M. Reeves, Viertel Centre for Research in Cancer Control, Queens-
interventions (Barlow, Wright, Sheasby, Turner, & Hainsworth,
land Cancer Fund; Patty McLaughlin and Silvia Gutierrez, Clinica
2002; Goldstein, Whitlock, & DePue, 2004; Norris et al., 2002;
Campesina Family Health Services, Denver, Colorado.
Von Korff, Gruman, Schaefer, & Curry, 1997). These interven-
Marina M. Reeves is now at Cancer Prevention Research Centre, School
tions have most often focused on key lifestyle behaviors, and on
of Population Health, University of Queensland, Australia.
This work was funded by Robert Wood Johnson Foundation Grant
medication adherence, and have targeted patients with a range of
041862 for its national program on Improving Chronic Illness Care (www
chronic conditions, with an emphasis on asthma, arthritis, and
.improvingchroniccare.org). We are grateful to the staff and patients of the
diabetes (Barlow, Wright, Sheasby, Turner, & Hainsworth, 2002).
Clinica Campesina Family Health Services for their support of and partic-
Despite the growing evidence base, some key gaps remain. The
ipation in this study. We also acknowledge the work of Kate Troy in
better part of this literature has focused on White, middle class
assisting with data management and Russ Glasgow for his insightful
samples, thus telling us little about how best to deliver chronic
comments on a draft of this article.
disease self-management interventions to low-income, ethnic mi-
Correspondence concerning this article should be addressed to Elizabeth
nority and underserved patients, who bear the largest burden of
G. Eakin, who is now at Cancer Prevention Research Centre, School of
disease and often have more than one chronic condition and
Population Health, University of Queensland, Level 3, Public Health
Building, Herston Road, Herston, Queensland 4006, Australia. E-mail:
multiple self-management behaviors that need to be addressed
e.eakin@sph.uq.edu.au
(Agency for Healthcare Research and Quality, 2003; Eakin, Bull,
392

---

RESOURCES FOR HEALTH
393
Glasgow, & Mason, 2002; Institute of Medicine Committee on
speaking, and the population served by Clinica Campesina have
Quality of Health Care in America, 2001). Members of these
generally spent fewer than 5 years living in the United States.
special populations often do not or cannot access traditional health
While there are other measures of acculturation that are more
care settings where self-management supports are more likely to
precise (Cuellar, Arnold, & Maldonado, 1995), researchers have
be offered (Brown, Long, Gould, Weitz, & Milliken, 2000;
commonly used both English language and number of years of
Levkoff & Sanchez, 2003; Sorensen et al., 2003).
U.S. residency as a proxy for acculturation (Bethel & Schenker,
Self-management interventions have largely been guided by
2005; Grunbaum, Kann, & Kinchen, 2004; Minnis & Padian,
intervention models that emphasize the individual, often ignoring
2001).
the social environmental context in which patients live (Glasgow
Study participants were adults with greater than one or more
& Eakin, 2000; Orleans, 2000; Sorensen et al., 2003). In addition,
chronic conditions for which a lifestyle intervention focused on
the evaluation of self-management interventions has tended to
physical activity and diet would be appropriate (i.e., hypertension,
occur in the context of efficacy trials with an emphasis on issues
chronic pain, hypercholesterolemia, depression, type 2 diabetes,
of internal validity. The conduct of practical behavioral trials that
osteoarthritis, obesity, chronic lung disease, heart disease, osteo-
address the generalizability of findings, the reach and representa-
porosis, hepatitis, history of cancer, previous stroke, multiple
tiveness of study participants, the feasibility of implementation in
sclerosis). Inclusion criteria were diagnosis of one or more chronic
the real-world context of primary care, and the institutionalization
conditions as above, age 30 years and over, having a telephone,
of interventions at the systems level, lags far behind (Glasgow,
and not planning to move from the area during the study’s time
Bull, Gillette, Klesges, & Dzewaltowski, 2002; Glasgow, David-
frame.
son, Dobkin, Ockene, & Spring, 2006; Glasgow, Klesges, Dze-
waltowski, Bull, & Estabrooks, 2004).
Recruitment
The Resources for Health study evaluates a physical activity and
diet intervention targeting low-income, largely Spanish-speaking
Study recruitment began in February 2002 and concluded in
patients with multiple chronic conditions. It adds to the literature
August 2003. Names and contact information for all patients
on self-management interventions by: (a) targeting an underserved,
meeting age and chronic condition eligibility criteria were ob-
minority subgroup; (b) extending the largely individually focused
tained from the clinic medical records database (n
605). Letters
interventions in this area to include an emphasis on participants’
were sent from clinic providers to patients describing the study and
social environment consistent with the social ecological model;
recommending their participation. Included with the letter was a
(c) addressing multiple risk behaviors in patients with multiple
stamped, self-addressed postcard for patients to return to the clinic
comorbid conditions; and (d) using the RE-AIM framework
if they wanted to decline being contacted about the study. Patients
(reach, efficacy, adoption, implementation, maintenance) (Glas-
for whom postcards had not been returned were followed up with
gow, Vogt, & Boles, 1999) to guide an approach to evaluation of
a phone call.
outcomes that balances both internal and external validity in the
Recruitment calls were made by a bilingual research assistant,
context of a practical behavioral trial.
with study consent and baseline data collected over the phone.
Randomization occurred following collection of baseline data and
was determined on the basis of a computer-generated randomiza-
Method
tion scheme and the opening of sequentially numbered envelopes.
Study Design
Intervention
This was a randomized controlled trial of a lifestyle intervention
evaluating short-term (6 weeks) and medium-term (6 months)
The conceptual framework underpinning the intervention was
outcomes on the primary behavioral targets of physical activity,
based on a behavioral ecological approach to chronic disease
dietary behavior, and the secondary outcome of multilevel support
self-management that emphasizes assessment, feedback, goal set-
for healthy lifestyles. Participants were randomly assigned to in-
ting, and problem solving (Glasgow & Eakin, 2000), as well as on
tervention or usual care conditions, with stratification on gender
social ecological theory with a focus on identification of multi-
and primary language.
level/community supports for health behavior change (Green,
Institutional Review Board (IRB) approval was initially re-
Richard, & Potvin, 1996; Stokols, 1992).
ceived from the AMC Cancer Research Center IRB in January
The intervention was culturally adapted and translated into
2001 and renewed by the Colorado Multiple IRB (COMIRB #
Spanish for an urban, low-income, largely Latino patient popula-
02-789) in December 2002.
tion (Riley, Glasgow, & Eakin, 2001). The adaptation included the
translation and validation of measures and intervention materials.
Setting and Participants
The intervention was conducted by an experienced, bilingual,
health educator, and involved two face-to-face visits (60 90 min)
The study was conducted at Clinica Campesina Family Health
3 months apart, three follow-up phone calls, and three newsletters
Services, a community health center that provides primary health
tailored to the behavioral goals of each participant. The face-to-
care services to low-income and medically underserved individu-
face visits took place either at the clinic or in the participant’s
als in the Denver Metro area. The study took place at the urban
home, based on participant preference. Because of the very low
North Denver clinic, the one with the largest percentage of
literacy levels of study participants, and the fact that some were not
Spanish-speaking clientele. Although the acculturation of the sam-
able to read and write in either Spanish or English, the use of visual
ple was not formally assessed, the sample is primarily Spanish
aids was emphasized throughout the intervention.

---

394
EAKIN ET AL.
The intervention protocol followed the 5 As approach advocated
U.S. physical activity guidelines, a dichotomous variable indicat-
in multiple behavioral risk factor interventions (Ask, Assess, Ad-
ing whether participants met the guidelines (i.e., 30 min/day of
vise, Agree, Arrange; Glasgow et al., 2002; Glasgow, Toobert, &
moderate physical activity on at least 5 days/week or 20 min/day
Hampson, 1996; Goldstein et al., 2004). Participants received
of vigorous activity on at least 3 days/week), although statistical
education on national physical activity (U.S. Department of Health
power was not based on this dichotomous variable. A Spanish
and Human Services, 1996) and dietary recommendations (U.S.
language version of the Behavioral Risk Factor Surveillance Sur-
Department of Health and Human Services and U.S. Department
vey Physical Activity items was obtained from the Centers for
of Agriculture, 2005) along with feedback from their baseline
Disease Control and Prevention and was evaluated for face validity
assessment. Participants then chose a self-management goal re-
and clarity during formative work for this trial.
lated to physical activity or healthy eating, and— key to the em-
Dietary behavior was measured using the Kristal Fat and Fiber
phasis on external resources—identified one or two types of
Behavior Questionnaire (FFB). The FFB is a 20-item scale mea-
social environmental resources they could use to help them reach
suring behaviors related to low-fat and high-fiber eating patterns.
their goal (e.g., family and friends, health care team, neighborhood
Previous research has found this scale to be reliable and sensitive
resources). At the conclusion of the session, participants received
to change and to correlate well with other “gold standard” mea-
a one-page goal sheet summarizing their personal action plan.
sures such as food records and food frequency questionnaires
At 2 and 6 weeks after the initial visit, the health educator made
(Glasgow, Perry, Toobert, & Hollis, 1996; Kristal, Shattuck, &
a brief follow-up phone call to reinforce progress toward goal
Henry, 1990; Shannon, Kristal, Curry, & Beresford, 1997). A
attainment and to problem-solve barriers. During the second face-
Spanish-language version of the FFB was adapted for use in the
to-face visit, participants were encouraged to consider setting a
current study (Kristal, Shattuck, & Patterson, 1999). A total FFB
goal for the second target behavior. A third follow-up phone call,
score was calculated, with scores ranging from 1.00 to 4.00, lower
the last point of contact, occurred 2 weeks after this visit to address
scores indicating better fat and fiber-related dietary behavior.
the goals and barriers again, and to discuss strategies for mainte-
Multilevel support for healthy lifestyles was measured using the
nance of behavior change, with an emphasis on use of multilevel
Chronic Illness Resource Survey (CIRS). The CIRS is a 22-item
support resources. To reinforce behavior change goals, three tai-
scale that assesses support for healthy lifestyle behaviors and
lored newsletters were mailed to participants over the course of the
chronic illness self-management from multiple sources including
6-month intervention. The low-literacy focused newsletters re-
family and friends, health care providers, neighborhood and com-
minded participants of their physical activity or diet goals, ad-
munity, media and health policies (Glasgow, Strycker, Toobert, &
dressed participant-reported barriers and suggested examples of
Eakin, 2000; Glasgow, Toobert, Barrera, & Strycker, 2005). CIRS
multilevel support resources that could be used. Patients in the
subscale and total scores were calculated and ranged from 1.00 to
usual care condition were mailed a local area community resources
5.00, with higher scores indicating more support. The English
guide and three newsletters on basic financial management (i.e.,
version of the CIRS had been previously validated (Glasgow et al.,
careers and employment, budgeting skills, and establishing credit).
2000), and a Spanish version was validated during formative work
for this trial (Eakin et al., in press).
Measures
Adoption, implementation, and maintenance.
Adoption was
measured by the number of clinics and providers approached who
The RE-AIM (Dzewaltowski, Glasgow, Klesges, & Estabrooks,
agreed to participate in the trial. Implementation was measured by
2004; Goldstein et al., 1999) framework was used to guide eval-
tracking the delivery of the intervention protocol, including the
uation of study outcomes.
number of intervention sessions delivered, and the percentage of
Reach.
Reach was calculated as the percentage and represen-
patients setting goals on physical activity and dietary behavior
tativeness of eligible patients who took part in the study. Patient
change. As this was a 6-month trial, longer-term maintenance of
characteristics were assessed using a short set of self-reported
outcomes was not assessed.
demographic items, including gender, age, education, income,
race, primary language, marital status, smoking status, and number
Statistical Analyses
and type of chronic conditions.
Efficacy.
Efficacy was evaluated by improvement from base-
General descriptive and bivariate statistical analyses were car-
line on two primary outcome measures (physical activity and
ried out using SPSS for Windows (Version 12.0.1, 2003, SPSS,
dietary behavior) and one secondary outcome measure (multilevel
Chicago, IL) statistical software package. Continuous variables
support for healthy lifestyles) at 6 weeks and 6 months. Data were
were tested for normal distribution. Analyses were carried out on
collected by a bilingual research assistant over the telephone.
an intention-to-treat basis. Significance was set at p
.05 (two-
Physical activity was measured using the Behavioral Risk Fac-
tailed). Based on the sample size of 200, this study had 81% power
tor Surveillance Survey Physical Activity items (Centers for Dis-
( p
.05) to detect an effect size of 0.20 on the two primary
ease Control and Prevention, 2000). This set of 15 items assesses
behavioral targets (physical activity and dietary behavior) at 6
total minutes of vigorous and moderate activity and walking in the
months.
past week and has been shown to correlate highly with other
We used independent sample t tests and Pearson’s chi-square
measures of physical activity (Brownson et al., 2004; Brownson,
tests were used to assess differences between participants and
Jones, Pratt, Blanton, & Heath, 2000). Two physical activity
nonparticipants, the treatment group and usual care group, and
outcome variables were calculated: (a) total minutes of walking
dropouts and completers. To examine changes between the treat-
per week (as walking is the preferred activity of most middle-aged
ment and usual care group over time, and to maximize the amount
and older adults; Simpson et al., 2003), and (b) consistent with
of data that were used in the analysis, repeated measures analysis

---

RESOURCES FOR HEALTH
395
of variance regression models, using a generalized estimating
Results
equations approach, were carried out using the SUDAAN statisti-
Reach
cal package (Version 8.0.2, Cary, NC). Regression models were
adjusted for sex, age, language, and number of chronic conditions.
Figure 1 shows the flow of participants through the study.
Subgroup analyses were also conducted to assess outcome mea-
Names of 605 potentially eligible patients were identified from
sures for the intervention group, based on dose of intervention and
clinic medical records. Of these, 345 were reached by phone for
behavioral goals set, compared with the usual care group. The
determination of eligibility and recruitment, 258 were study eligi-
alpha level was not adjusted for these secondary analyses.
ble, and 200 agreed to participate (33% of the initial pool, 58% of
605 Names of Potentially-Eligible Patients
Not able to contact
Contacted
n = 260
n = 345
Refused by postcard
No phone contact
letters returned
n = 204
n = 56
Not Eligible
n = 87
Eligible
• Moved/not reachable, n = 34
n = 258
• No longer clinic patient, n =
25
• Mental incapacity, n = 9
• No chronic condition, n = 7
Declined
• Language skills lacking, n =
n = 58*
5
• Age < 30 yrs, n = 4
• Death, n = 3
* Reasons for declining:
• Not interested, n = 33
• Too busy, n = 17
Participated
• Too ill, n = 4
n = 200
• Never completed baseline, n
= 4
Randomized Assignment
Treatment
Usual Care
n = 101
n = 99
Received dose
of intervention:
Full, n = 48
Moderate, n = 30
Low, n = 9
None, n = 14
Follow-up
Follow-up
Measurements
Measurements
6 weeks, n = 72
6 weeks, n = 65
6 months, n = 84
6 months, n = 78
Figure 1.
Flow of participants through the study.

---

396
EAKIN ET AL.
those reached by phone, and 78% of those reached and eligible).
time points are shown in Table 2. Scores were adjusted for sex,
Compared with nonparticipants, participants were more likely to
age, language, and number of chronic conditions. A statistically
be female (data not shown). Complete data were available for 69%
significant difference between the groups over time was evident
of participants at 6 weeks and 81% at 6 months, and retention rates
for dietary behavior, Wald F(2)
4.64, p
.011. The treatment
did not differ significantly between intervention and usual care
group showed a significantly greater improvement in dietary be-
groups. Women and Spanish speakers were more likely to com-
havior compared to the usual care group at both 6 weeks (inter-
plete study assessments compared with men and English speakers
vention effect
0.16
0.08, p
.031) and 6 months (inter-
(data not shown).
vention effect
0.18
0.06, p
.003). There were no
Baseline characteristics of participants randomized to the treat-
significant differences between groups over time for physical
ment group and those in usual care are shown in Table 1. Groups
activity when assessed as change in minutes walked per week,
differed significantly with respect to primary language and ethnic-
Wald F(2)
2.04, p
.132, or as the percentage of those meeting
ity, and although not statistically significant, there appeared to be
national physical activity guidelines, Wald F(2)
1.37, p
.505.
differences with respect to number of chronic conditions ( p
A statistically significant difference between the treatment and
.08). The stratified, random assignment scheme did not result in
usual care groups over time was observed for multilevel support
equal proportions of Spanish speakers in each study condition as
for healthy lifestyles (CIRS total score), Wald F(2)
8.23, p
block randomization was not used and recruitment was stopped
.001. The treatment group reported significantly greater support at
earlier than anticipated because of limited resources.
both 6 weeks (intervention effect
0.32
0.08, p
.001) and 6
months (intervention effect
0.23
0.08, p
.003). Analysis of
Efficacy
changes for the individual CIRS subscales revealed significant
Changes in dietary behavior, physical activity, and multilevel
differences over time between the treatment and usual care groups
support for healthy lifestyles between the groups over the three
for the personal, Wald F(2)
9.67, p
.001, and family and
Table 1
Baseline Demographic and Health-Related Characteristics of Sample
Treatment
Usual care
Characteristic
(n
101)
(n
99)
Age (years; M
SD)
50
13
49
13
Sex
Female
80 (79.2)
77 (77.8)
Male
21 (20.8)
22 (22.2)
Language*
Spanish
77 (76.2)
56 (56.6)
English
24 (23.8)
43 (43.4)
Ethnicity**
Hispanic/Latino
81 (80.2)
69 (71.1)
Anglo
8 (7.9)
22 (22.7)
Other
12 (11.9)
6 (6.2)
Smoker
20 (20.2)
24 (24.2)
Married
67 (66.3)
53 (54.1)
Live alone
6 (5.9)
11 (11.2)
Yearly household income
Less than $10,000
32 (34.4)
36 (37.5)
$10,000–$30,000
51 (54.8)
44 (45.8)
Greater than $30,000
10 (10.8)
16 (16.7)
Education
Elementary/some high school
67 (66.3)
69 (70.4)
High school graduate
17 (16.8)
13 (13.3)
Some college/college graduate
17 (16.8)
16 (16.3)
No. of chronic conditions
One
8 (7.9)
17 (17.2)
Two
25 (24.8)
28 (28.3)
Three or greater
68 (67.3)
54 (54.5)
Physical activity level
Total minutes walking per week (median and range)
60 (0–840)
70 (0–840)
% meeting guidelines
19 (18.8)
17 (17.2)
Dietary behavior scorea
2.46
0.47
2.50
0.42
Multilevel support for healthy lifestylesb
2.67
0.52
2.60
0.49
Note.
Data are n (%) or mean
standard deviation.
a Fat and Fiber Behavior Questionnaire: Lower scores indicate better dietary behavior (range
1.00 –
4.00).
b Chronic Illness Resources Survey (total score): Higher scores indicate greater support for healthy
lifestyles (range
1.00 –5.00).
* p
.004.
** p
.009.

---

RESOURCES FOR HEALTH
397
Table 2
Adjusted Mean Outcome Measures (
Standard Errors) for Intervention and Usual Care
Groups at Baseline, 6 Weeks, and 6 Months Postintervention
Variable
Baseline
6 weeks
6 months
Dietary behaviora
Intervention
2.47
0.05
2.25
0.06*
2.24
0.05*
Usual care
2.48
0.04
2.42
0.05
2.43
0.05
Multilevel support for healthy lifestylesb
Intervention
2.67
0.05
2.98
0.06*
2.98
0.06*
Usual care
2.61
0.05
2.59
0.06
2.69
0.05
Change minutes of walking/week
Intervention
11
20
16
20
Usual care
47
23
11
23
Note.
Data were adjusted for age, sex, language, and number of chronic conditions.
a Lower scores indicate better dietary behavior (range
1.00 – 4.00).
b Higher scores indicate better support
(range
1.00 –5.00).
* p
.05.
friends, Wald F(2)
5.11, p
.007, subscales (data not shown).
sults to be interpreted not only in terms of their immediate impact,
When the dietary behavior data were reanalyzed including adjust-
but also in terms of their potential for generalization.
ment for changes in total CIRS score, the intervention effect
Reach.
Given the challenges of recruiting and retaining a
previously seen had disappeared, Wald F(2)
1.72, p
.182,
highly migratory, low-income, low literacy population, the ability
indicating that the CIRS mediated the observed change in dietary
to (a) enroll 78% of those contacted and eligible and (b) retain 81%
behavior.
at 6 months supports the feasibility of this approach. The overrep-
resentation of women and Spanish speakers suggests that recruit-
Adoption
ment efforts reached those most in need but also that it will take
more effort to engage Latino men in health behavior interventions.
All three of the community health center clinics of the Clinica
Efficacy.
Significant improvements occurred in one of two
Campesina Family Health Services were approached for partici-
primary outcome measures, namely dietary behavior, and in the
pation, and all agreed to participate, although because of limited
secondary outcome of multilevel support for healthy lifestyles. The
research resources, patients were recruited only from the largest of
lack of change in physical activity was disappointing, but more
the three clinics. In that clinic, all of the 12 providers agreed to
than likely reflects the difficulty in measuring physical activity via
participate.
a brief self-report instrument (Sallis & Saelens, 2000) and the
potentially limited sensitivity of the Behavioral Risk Factor Sur-
Implementation
veillance Survey Physical Activity population surveillance mea-
sure to detect change in the context of an intervention trial.
Forty-eight (47.5%) of the 101 intervention condition partici-
Adoption.
The trial was clearly and strongly supported by the
pants received both intervention visits, 39 (38.6%) received only
community health care clinic in which it was conducted, as indi-
one, and 14 (13.9%) were unable to be contacted for intervention
cated by the 100% provider participation.
visits or phone calls. Forty-six (45.5%) received the intended three
Implementation.
follow-up phone calls, 29 (28.7%) received two calls, 9 (8.9%)
This trial was challenging to implement, with
received one call, and a further 3 (3.0%) were never reached for
numerous missed and rescheduled appointments; however, overall
follow-up calls. During the intervention, 41 (40.6%) participants
77% received at least three of the five intended intervention
set goals for physical activity only, 13 (12.9%) set diet goals only,
contacts, again supporting the feasibility of the approach.
33 (32.7%) set both physical activity and diet goals, and 14
Maintenance.
The medium-term (6-month) outcomes pre-
(13.9%) did not set any goals (because they were never reached for
clude evaluation of longer term maintenance of behavior change.
intervention). Subgroup analyses to determine the effect of the
However, with regard to maintenance at the clinic/systems level,
number of intervention contacts and the selection of behavioral
the clinic intends to retain the study interventionist on staff as a
goals did not alter the results observed (data not shown).
chronic disease health educator.
A strength of the study was the use of the social ecological
model to guide the intervention and broaden the focus to take into
Discussion
account the multiple levels of influence on health behavior change
The Resources for Health trial, with its focus on improving
(Green, 2001; Stokols, 1996). This was not only implicit as an
physical activity and dietary behavior in low-income and predom-
underlying theory, but operationalized via the use of the CIRS to
inantly Spanish-speaking patients with multiple comorbid chronic
provide feedback to participants and to assist with goal setting
conditions, addresses a significant gap in the literature (Goldstein
around use of multilevel resources. The CIRS was also used as an
et al., 2004). Taking the broader public health perspective of the
outcome measure, and one on which significant between-groups
RE-AIM framework (Glasgow et al., 2004) allows the study re-
change was observed. In addition, the CIRS also mediated the

---

398
EAKIN ET AL.
effect of the intervention on dietary outcomes—an effect demon-
The Resources for Health intervention was adapted for the
strated in a previous trial (Glasgow et al., 2005)— highlighting the
largely Latino sample from previous chronic disease self-
importance of taking into account multiple levels of support in the
management interventions conducted by Glasgow and colleagues
delivery of health behavior change interventions.
that targeted primarily Caucasian samples (Riley, Glasgow, &
Study limitations include the use of uncorroborated, self-
Eakin, 2001; Glasgow & Eakin, 2000). The Resources for Health
reported, albeit validated, outcome measures; the use of a single
sample showed both poorer levels of health behaviors at baseline
interventionist; recruitment of Latino patients (primarily of Mex-
and similar to or greater intervention effects for both fat- and
ican descent) from a single clinic, which may limit the generaliz-
fiber-related dietary behavior (Glasgow et al., 1997; Glasgow &
ability of findings; and the lack of longer term follow-up. The
Toobert, 2000; Kristal, Curry, Shattuck, Feng, & Li, 2000) and
sample size was also too small to allow for analyses of patients
multilevel support for healthy lifestyles (Glasgow et al., 2005).
nested within providers. However, providers were not involved in
This suggests that there is merit in adapting and evaluating existing
the delivery of the intervention, and our previous work has shown
evidence-based primary-care-based health behavior interventions
the intraclass correlations for the measures used in this study to be
with Latino and other high-risk subgroups (Goldstein et al., 2004;
extremely small (Glasgow et al., 1997).
O’Malley, Gonzalez, Sheppard, Huerta, & Mandelblatt, 2003).
Although there is a growing literature on health behavior inter-
Results of the Resources for Health trial suggest that a practical
ventions targeting Latinos, studies of interventions directly com-
and behaviorally focused, social ecological model driven inter-
parable to the chronic disease self-management approach of Re-
vention to enhance chronic disease self-management can be suc-
sources for Health are few. Existing studies have often targeted
cessfully delivered in a community health care context to low-
Latinas as the keepers of family health (Amaro & de la Torre,
income, Spanish-speaking patients. Future research needs to
2002) and have used a variety of intervention modalities, including
address generalization to other disadvantaged groups, longer term
lay health advisors (or promotoras) to deliver intervention (Kim,
maintenance of outcomes, and evaluate cost-effectiveness and
Koniak-Griffin, Flaskerud, & Guarnero, 2004; Navarro, Rock,
changes across multiple behavioral risk factors.
McNicholas, Senn, & Moreno, 2000; Swider, 2002) using print
and mass media (Nestle & Cowell, 1990; Wechsler & Wernick,
References
1992), as well as multiple modality approaches (Elder et al., 1998).
The most comparable study involved multiple health centers and
Agency for Healthcare Research and Quality. (2003). National healthcare
targeted multiethnic adults (a small percentage of whom were
disparities report. Rockville, MD: U.S. Department of Health and Hu-
Latino) around diet and physical activity (Emmons et al., 2005).
man Services.
Amaro, H., & de la Torre, A. (2002). Public health needs and scientific
Like Resources for Health, it involved a session with a health
opportunities in research on Latinas. American Journal of Public Health,
educator, follow-up phone calls, and tailored mailings; unlike
92, 525–529.
Resources for Health, it also involved clinicians providing a tai-
Barlow, J., Wright, C., Sheasby, J., Turner, A., & Hainsworth, J. (2002).
lored health behavior change prescription. Similar to the present
Self-management approaches for people with chronic conditions: A
findings, Emmons and colleagues (2005) found significant
review. Patient Education and Counseling, 48, 177–187.
changes in dietary behaviors, but not in physical activity, perhaps
Bethel, J., & Schenker, M. (2005). Acculturation and smoking patterns
suggesting that physical activity may be a more difficult behavior
among Hispanics: A review. American Journal of Preventive Medicine,
to change in lower income, multiethnic subgroups.
29, 143–148.
Brown, B. A., Long, H. L., Gould, H., Weitz, T. A., & Milliken, N. (2000).
A recent study based in the primary health care setting evaluated
A conceptual model for the recruitment of diverse women into research
a physical activity and dietary intervention that targeted primarily
studies. Journal of Women’s Health, 9, 625– 632.
Latinas 50 years of age and over recruited from community health
Brownson, R. C., Chang, J. J., Eyler, A. A., Ainsworth, B. E., Kirtland,
clinics and that delivered across three conditions: provider coun-
K. A., Saelens, B. E., et al. (2004). Measuring the environment for
seling; provider counseling and health education; and provider
friendliness toward physical activity: A comparison of the reliability of
counseling, health education, and community health worker sup-
3 questionnaires. American Journal of Public Health, 94, 473– 483.
port (Staten et al., 2004). At the 12-month follow-up, although
Brownson, R. C., Jones, D. A., Pratt, M., Blanton, C., & Heath, G. W.
there were no significant between-groups differences, all three
(2000). Measuring physical activity with the Behavioral Risk Factor
Surveillance System. Medicine and Science in Sports and Exercise, 32,
groups showed an increase in physical activity, whereas women
1913–1918.
receiving the comprehensive intervention were more likely to
Centers for Disease Control and Prevention. (2000). Behavioral Risk
make dietary changes. In an intervention to improve fat and fiber
Factor Surveillance System Survey Questionnaire. Atlanta, GA: Author.
targeting Latinas recruited via random digit dialing, Elder et al.
Clark, N. M., Becker, M. H., Lorig, K., Rakowski, W., & Anderson, L.
(2005) evaluated three conditions: lay health advisors plus tailored
(1991). Self-management of chronic disease in older adults. Journal of
materials, tailored materials only, or off-the-shelf print materials.
Aging and Health, 3, 3–27.
At 12-weeks postintervention, although there were no significant
Cuellar, I., Arnold, B., & Maldonado, R. (1995). The Acculturation Rating
differences between conditions, the lay health advisor condition
Scale for Mexican Americans-II: A revision of the original ARSMA
achieved significant improvements across a number of dietary
scale. Hispanic Journal of Behavioral Sciences, 17, 275–304.
Dzewaltowski, D. A., Glasgow, R. E., Klesges, L. M., & Estabrooks, P.
outcomes. In general these studies provide modest support for the
(2004). RE-AIM: Evidence-based standards and a web resource to
efficacy of health behavior interventions among Latinos but indi-
improve translation of research into practice. Annals of Behavioral
cate that more needs to be done to elucidate the most effective
Medicine, 28, 75– 80.
intervention modalities and to facilitate change across both diet
Eakin, E. G., Bull, S. S., Glasgow, R. E., & Mason, M. (2002). Reaching
and physical activity behaviors.
those most in need: A review of diabetes self-management interventions

---

RESOURCES FOR HEALTH
399
in disadvantaged populations. Diabetes/Metabolism Research and Re-
Green, L. W., Richard, L., & Potvin, L. (1996). Ecological foundations of
views, 18, 26 –35.
health promotion. American Journal of Health Promotion, 10, 270 –281.
Eakin, E. G., Reeves, M. M., Bull, S. S., Riley, K. M., Floyd, S., &
Grunbaum, J., Kann, L., & Kinchen, S. (2004). Youth risk behavior
Glasgow, R. E. (in press). Validation of the Spanish-language version of
surveillance—United States, 2003. Morbidity and Mortality Weekly Re-
the Chronic Illness Resources Survey. International Journal of Behav-
port Surveillance Summaries, 53, 1–96.
ioral Medicine.
Institute of Medicine Committee on Quality of Health Care in America.
Elder, J. P., Ayala, G. X., Campbell, N. R., Slymen, D., Lopez-Madurga,
(2001). Improving the 21st-century health care system. In Crossing the
E. T., Engelberg, M., & Baquero, B. (2005). Interpersonal and print
quality chasm: A new health system for the 21st century (pp. 23–38).
nutrition communication for a Spanish-dominant Latino population:
Washington, DC: National Academy Press.
Secretos de la buena vida. Health Psychology, 24, 49 –57.
Kim, S., Koniak-Griffin, D., Flaskerud, J. H., & Guarnero, P. A. (2004).
Elder, J. P., Campbell, N. C., Candelaria, J. I., Talavara, G. A., Meyer,
The impact of lay health advisors on cardiovascular health promotion:
J. A., & Moreno, C. (1998). Project Salsa: Development and institution-
Using a community-based participatory approach. Journal of Cardio-
alization of a nutrition health promotion project in a Latino community.
vascular Nursing, 19, 192–199.
American Journal of Health Promotion, 12, 391– 401.
Kristal, A. R., Curry, S. J., Shattuck, A. L., Feng, Z., & Li, S. (2000). A
Emmons, K., Stoddard, A., Fletcher, R., Gutheil, C., Gonzalez Suarez, E.,
randomized trial of a tailored self-help dietary intervention: The Puget
Lobb, R., et al. (2005). Cancer prevention among working class, multi-
Sound Eating Patterns Study. Preventive Medicine, 31, 380 –389.
ethnic adults: Results of the Healthy Directions-Health Centers Study.
Kristal, A. R., Shattuck, A. L., & Henry, H. J. (1990). Patterns of dietary
American Journal of Public Health, 95, 1200 –1205.
behavior associated with selecting diets low in fat: Reliability and
Glasgow, R. E., Bull, S. S., Gillette, C., Klesges, L. M., & Dzewaltowski,
validity of a behavioral approach to dietary assessment. Journal of the
D. A. (2002). Behavior change intervention research in healthcare set-
American Dietetic Association, 90, 214 –220.
tings: A review of recent reports with emphasis on external validity.
Kristal, A. R., Shattuck, A. L., & Patterson, R. E. (1999). Differences in
American Journal of Preventive Medicine, 23, 62– 69.
fat-related dietary patterns between Black, Hispanic and White women:
Glasgow, R. E., Davidson, K. W., Dobkin, P. L., Ockene, J., & Spring, B.
Results from the Women’s Health Trial Feasibility Study in Minority
(2006). Practical behavioral trials to advance evidence-based behavioral
Populations. Public Health Nutrition, 2, 253–262.
medicine. Annals of Behavioral Medicine, 31, 5–13.
Levkoff, S., & Sanchez, H. (2003). Lessons learned about minority recruit-
Glasgow, R. E., & Eakin, E. G. (2000). Medical office-based interventions.
ment and retention from the Centers on Minority Aging and Health
In F. J. Snoek & T. C. Skinner (Eds.), Psychology in diabetes care (pp.
Promotion. Gerontologist, 43, 18 –26.
141–168). Chichester, England: Wiley.
Lorig, K. R., Sobel, D. S., Stewart, A. L., Brown, B. W., Bandura, A.,
Glasgow, R. E., Klesges, L., Dzewaltowski, D., Bull, S., & Estabrooks, P.
Ritter, P., et al. (1999). Evidence suggesting that a chronic disease
(2004). The future of health behavior change research: What is needed
self-management program can improve health status while reducing
to improve translation of research into health promotion practice? An-
hospitalization: A randomized trial. Medical Care, 37, 5–14.
nals of Behavioral Medicine, 27, 3–12.
Minnis, A., & Padian, N. (2001). Reproductive health differences among
Glasgow, R. E., La Chance, P., Toobert, D. J., Brown, J., Hampson, S. E.,
Latin American- and US-born young women. Journal of Urban Health,
& Riddle, M. C. (1997). Long term effects and costs of brief behavioural
78, 627– 637.
dietary intervention for patients with diabetes delivered from the medical
Navarro, A., Rock, C., McNicholas, L., Senn, K., & Moreno, C. (2000).
office. Patient Education and Counseling, 32, 175–184.
Community-based education in nutrition and cancer: The Por La Vida
Glasgow, R. E., Perry, J. D., Toobert, D. J., & Hollis, J. F. (1996). Brief
Cuidandome curriculum. Journal of Cancer Education, 15, 168 –172.
assessments of dietary behavior in field settings. Addictive Behaviors,
Nestle, M., & Cowell, C. (1990). Health promotion of low-income minor-
21, 239 –247.
ity groups: The challenge for nutrition education. Health Education
Glasgow, R. E., Strycker, M. A., Toobert, D. J., & Eakin, E. G. (2000). A
Research, 5, 527–533.
social-ecological approach to assessing support for disease self-
Norris, S. L., Nichols, P., Caspersen, C. J., Glasgow, R. E., Engelgau, M.,
management: The chronic illness resource survey. Journal of Behavioral
Jack, Jr., L., et al. (2002). Increasing diabetes self-management educa-
Medicine, 23, 943–950.
tion in community settings: A systematic review. American Journal of
Glasgow, R. E., & Toobert, D. J. (2000). Brief, computer-assisted diabetes
Preventive Medicine, 22(Suppl. 4), 39 – 66.
dietary self-management counseling. Medical Care, 38, 1062–1073.
O’Malley, A. S., Gonzalez, R. M., Sheppard, V. B., Huerta, E., & Man-
Glasgow, R. E., Toobert, D. J., Barrera, M., & Strycker, L. A. (2005). The
delblatt, J. (2003). Primary care cancer control interventions including
Chronic Illness Resources Survey: Cross-validation and sensitivity to
Latinos: A review. American Journal of Preventive Medicine, 25, 264 –
intervention. Health Education Research, 20, 402– 409.
271.
Glasgow, R. E., Toobert, D. J., & Hampson, S. E. (1996). Effects of a brief
Orleans, C. T. (2000). Promoting the maintenance of health behavior
office-based intervention to facilitate diabetes dietary self-management.
change: Recommendations for the next generation of research and
Diabetes Care, 19, 835– 842.
practice. Health Psychology, 19(Suppl. 1), 76 – 83.
Glasgow, R. E., Vogt, T. M., & Boles, S. M. (1999). Evaluating the public
Riley, K. M., Glasgow, R. E., & Eakin, E. G. (2001). Resources for health:
health impact of health promotion interventions: The RE-AIM frame-
A social-ecological intervention for supporting self-management of
work. American Journal of Public Health, 89, 1322–1327.
chronic conditions. Journal of Health Psychology, 6, 693–705.
Goldstein, M. G., Pinto, B. M., Marcus, B. H., Lynn, H., Jette, A. M.,
Sallis, J. F., & Saelens, B. E. (2000, June). Assessment of physical activity
McDermott, S., et al. (1999). Physician-based physical activity counsel-
by self-report: Status, limitations, and future directions. Research Quar-
ing for middle-aged and older adults: A randomized trial. Annals of
terly for Exercise and Sport, 71(Suppl. 2), S1–14.
Behavioral Medicine, 21, 40 – 47.
Shannon, J., Kristal, A. R., Curry, S. J., & Beresford, S. A. (1997).
Goldstein, M. G., Whitlock, E. P., & DePue, J. (2004). Multiple behavioral
Application of a behavioral approach to measuring dietary change: The
risk factor interventions in primary care: Summary of research evidence.
fat- and fiber-related diet behavior questionnaire. Cancer Epidemiology,
American Journal of Preventive Medicine, 27(Suppl. 2), 61–79.
Biomarkers & Prevention, 6(5), 355–361.
Green, L. W. (2001). From research to “best practices” in other settings and
Simpson, M. E., Serdula, M., Galuska, D. A., Gillespie, C., Donehoo, R. S.,
populations. American Journal of Health Behavior, 25, 165–178.
Macera, C., et al. (2003). Walking trends among US adults: The Behav-

---

400
EAKIN ET AL.
ioral Risk Factor Surveillance System, 1987–2000. American Journal of
workers: An integrative literature review. Public Health Nursing, 19,
Preventive Medicine, 25, 95–100.
11–20.
Sorensen, G., Emmons, K. M., Hunt, M. K., Barbeau, E., Goldman, R.,
U.S. Department of Health and Human Services. (1996). Physical activity
Peterson, K., et al. (2003). Model for incorporating social context in
and Health: A report of the Surgeon General. Atlanta, GA: Author.
health behavior interventions: Applications for cancer prevention for
U.S. Department of Health and Human Services and U.S. Department of
working-class, multiethnic populations. Preventive Medicine, 37, 188 –
Agriculture. (2005). Dietary guidelines for Americans. Washington, DC:
197.
U.S. Government Printing Office.
Staten, L., Gregory-Mercado, K., Ranger-Moore, J., Will, J., Giuliano, A.,
Von Korff, M., Glasgow, R. E., & Sharpe, M. (2002). Organising care for
Ford, E., et al. (2004). Provider counseling, health education, and com-
chronic illness. British Medical Journal, 325, 92–94.
munity health workers: The Arizona WISEWOMAN Project. Journal of
Von Korff, M., Gruman, J., Schaefer, J., & Curry, S. (1997). Collaborative
Women’s Health, 13, 547–556.
management of chronic illness. Annals of Internal Medicine, 127, 1097–1102.
Stokols, D. (1992). Establishing and maintaining healthy environments:
Wagner, E. H. (2001). Meeting the needs of chronically ill people. British
Toward a social ecology of health promotion. American Psychologist,
Medical Journal, 323, 945–946.
47, 6 –22.
Wagner, E. H., & Groves, T. (2002). Care for chronic diseases. British
Stokols, D. (1996). Translating social ecological theory into guidelines for
Medical Journal, 325, 913–914.
community health promotion. American Journal of Health Promotion,
Wechsler, H., & Wernick, S. M. (1992). A social marketing campaign to
10, 282–298.
promote low-fat mild consumption in an inner-city Latino community.
Swider, S. M. (2002). Outcome effectiveness of community health
Public Health Reports, 107, 202–207.

---