Reviews/Commentaries/Position Statements

R E V I E W A R T I C L E

Community-Based Lifestyle Interventions

to Prevent Type 2 Diabetes
DAWN W. SATTERFIELD, RN, PHD, CDE1 ED W. GREGG, PHD1 quality of life of those affected and their
MICHELE VOLANSKY, MPH2 LINDA S. GEISS, MA1 families and communities.
CARL J. CASPERSEN, PHD, MPH1 GWEN M. HOSEY, RNP, MSN, CDE1 To stem the rising tide of diabetes,
MICHAEL M. ENGELGAU, MD1 JEANNETTE MAY, MPH1 public health policies need to move up-
BARBARA A. BOWMAN, PHD1 FRANK VINICOR, MD, MPH1 stream toward prevention or at least a de-
lay in the onset of type 2 diabetes. A number
of recent studies offer scientific evidence
and new hope for curtailing the epidemic
of type 2 diabetes with support for inten-
OBJECTIVE — To conduct a literature review of community-based interventions intended to
prevent or delay type 2 diabetes. sive lifestyle modification and modest
weight loss as effective interventions
RESEARCH DESIGN AND METHODS — Recently published findings about the po- among adults at high risk for developing
tential to prevent or delay type 2 diabetes with intensive lifestyle interventions prompted a type 2 diabetes (5–7). The transitional
literature search for community-based diabetes prevention interventions. The literature review state in the natural history of diabetes
design was a search of databases for publications in 1990 –2001 that identified reports on when impaired glucose tolerance (IGT),
community-based interventions designed to prevent or modify risk factors for type 2 diabetes. impaired fasting glucose, or both are
present has recently become known as
RESULTS — The search revealed 16 published interventions, 8 of which were conducted in “pre-diabetes,” which affects 12 million
the U.S. and involved populations disproportionately burdened by diabetes (e.g., American
overweight Americans aged 45–74 years
Indians, Native Hawaiians, Mexican Americans, and African Americans). Of the studies report-
ing results among youth, there were posttest improvements in intervention groups in knowl- (8), also raising their risk for cardiovascu-
edge, preventive behaviors, and self-esteem. Among studies reporting results among adults, most lar disease (9). Detection of pre-diabetes
reported improvements in intervention groups in knowledge or adoption of regular physical is not a goal of most diabetes screening
activity. Several investigators offered important reflections about the process of engaging com- programs (10), but the lengthy develop-
munities and sharing decision making in participatory research approaches, as well as insights mental period of diabetes, coupled with
about the expectations and limitations of community-based diabetes prevention research. Many the potential to prevent or delay the onset
of the studies reported limitations in their design, including the lack of control or comparison of type 2 diabetes, offers an opportunity
groups, low response rates or lack of information on nonresponders, or brief intervention for multifaceted prevention efforts.
periods. Distinct preventive medicine strate-
CONCLUSIONS — There is a critical need to conduct and publish reports on well-designed
gies have been described by Rose (11) as
community-based diabetes prevention research and share information on the process, results, follows: 1) the “high-risk approach,”
and lessons learned. Armed with recent positive findings about diabetes prevention and litera- which identifies and focuses exclusively
ture documenting community-based efforts, advocates at local, state, and national levels can on individuals at highest risk for develop-
collaborate to stem the rising tide of diabetes in communities. ing diseases; and 2) the “population or
public health approach,” which attempts
Diabetes Care 26:2643–2652, 2003 to reduce risk factors for or causes of dis-
eases within communities, which are gen-
erally defined in terms of localities but can
also represent groups who share a com-

I
n a decade’s time (1990 –2001), the sociated with obesity and weight gain (3),
prevalence of self-reported (diag- which also increased over the same time mon cause or interest (12). Rose identi-
nosed) diabetes increased 61% in the period. One projection is that the 11 mil- fied advantages and disadvantages for
U.S. (1), including a startling 76% in- lion U.S. residents with diagnosed diabe- both approaches. For example, the high-
crease from 1990 to 1998 among people tes will increase to 29 million in 2050 (4). risk approach is generally cost-effective
in their 30s (2). Accounting for roughly The escalating prevalence of type 2 diabe- with a high likelihood of benefit for mo-
95% of all diabetes, type 2 diabetes is as- tes portends serious consequences for the tivated individuals. The population ap-
proach, often called the community-
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
based approach, offers a smaller benefit to
From the 1Division of Diabetes Translation, Centers for Disease Control and Prevention, Atlanta, Georgia; individuals but more potential for bene-
and 2Westat, Atlanta, Georgia.
Address correspondence and reprint requests to Dawn Satterfield, RN, PhD, CDE, CDC Division of fiting the larger population. Whereas the
Diabetes Translation, 2858 Woodcock Blvd., Davidson Building, Rm. 1028, Atlanta, GA 30341-4002. high-risk approach is palliative, the com-
E-mail: [email protected]. munity-based approach aims to address
Received for publication 14 October 2002 and accepted in revised form 11 May 2003. the underlying causes of ill health (11);
Abbreviations: BIA, bioelectrical impedance analysis; IGT, impaired glucose tolerance.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion
the latter is typically predicated on respect
factors for many substances. for community strengths, including cul-
© 2003 by the American Diabetes Association. tural practices and wisdom, with mean-

DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003 2643

2644

Community-based lifestyle interventions

Table 1—Population-based interventions designed to prevent type 2 diabetes in youth

Study descriptions Results

Geographic location Community

Target audience Study design and Goal(s) for prevention involvement/cultural Knowledge, attitudes, Prevalence of
Reference Sample size duration of type 2 diabetes Intervention relevance behaviors Clinical markers diabetes

Cook and Hurley (19) Gila River Indian Quasi-experimental; Maintain healthy body Classroom instruction for Program endorsed by Not reported Not reported Not reported
Community, no control group and weight to pro- 20 classes (10 weeks) tribal council
Arizona Initial pilot: 2 years mote diabetes about diabetes preven- Use of traditional
Akimel O’odham awareness tion with take-home symbols such as
(Pima) and Pee information saguaro cactus,
Posh (Maricopa) One-mile walk each day maize, and road-
children (Kinder- with teacher runner
garten-grade 2) Structured school break-
Sample size not fast/lunch
discussed
Gittelsohn et al. (21) Ontario, Canada Quasi-experimental; Prevent diabetes and Health and nutrition cur- Ethnographic data Not reported Not reported Not reported
First Nations Ojibwe- no control group its risk factors, par- riculum focused on collection to aid in
Cree Reserve Duration not ticularly obesity healthier food choices problem definition
school (grades discussed and increased physical and development
4–6) activity of culturally appro-
Sample size not Improved access to low- priate intervention
discussed fat foods where stu- strategies
dents and families eat Intervention strategies
and shop for groceries generated by and
pretested with
community
Holcomb et al. (31) Webb County, Quasi-experimental; Encourage healthier Curriculum-based pro- Not discussed Significant gains in Not targeted Not targeted
Texas, Mexico comparison lifestyles; improve gram that integrates knowledge about dia-
border (94% group knowledge, self- program’s goals into betes, self-efficacy, and
Hispanic) ⬇3 months efficacy, and reading, writing, math- healthy dietary-and
5th grade students behaviors ematics, science, and exercise-related behav-
(aged 10–12 years) physical education and ior from pretest to
in 14 schools encourages students to posttest; also from
DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003

n ⫽ 835 (completed eat low-fat foods and posttest through 4

both pre- and exercise regularly weeks of follow-up, but
posttests) on fewer variables
Macaulay et al. (24) Kahnawake, Quebec, Quasi-experimental; Long-term goal: de- Health education pro- Community advisory Not reported Not reported Not reported
Canada nonequivalent crease occurrence of grams for each grade board to advise on
Native Mohawk chil- comparison type 2 diabetes on nutrition, fitness, objectives, activi-
dren (grades 1–6) community Short-term goal: re- diabetes, the human ties, and traditions
n ⫽ 1,200 (different 3 years duce prevalence of body, with a focus on Incorporation of tra-
sample sizes in 63 obesity, high-calorie creating positive atti- ditional foods and
distinct interven- and high-fat diets, tudes and increasing activities
tions) and low physical self-esteem Study performed by
inactivity Family activities in the Native researchers
community: to expose
children to healthy life-
styles of adults (e.g.,
pow-wow dances)
Supportive environments
(e.g., walk/bike path)
DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003

Marlow et al. (23) Winnebago Indian Quasi-experimental; Use aspects of tradi- Educational intervention Program designed by On nine sets of matched Not targeted Not targeted
reservation, Ne- no control group tional Indian culture designed to provide Winnebago adoles- pretests and posttests
braska Half-day workshop to develop culturally information about dia- cents; coyote name of knowledge, 89%
Adolescents (aged (pilot) sensitive program betes prevention used by instructors achieved higher scores
13–18 years) Provide opportunity through physical activ- to symbolize indi- (no statistical analysis)
n ⫽ 24 for adolescents to ity and good nutrition rect education (i.e.,
take leadership role within a cultural one can learn from
in diabetes context the coyote’s mis-
education takes and
Design health educa- successes)
tion program ad- Use of Native Ameri-
dressing specific can stories, games,
needs of adolescents and legends
Teufel and Ritenbaugh (30) Zuni Pueblo reserva- Quasi-experimental; Reduce prevalence of Supportive social net- Formative assessment Significant reduction in Significant in- Not yet
tion, New Mexico no control group diabetes risk factors works established of target audience consumption of sug- crease in glu- measured
Zuni adolescents 4 years Support increased within community, before intervention, ared beverages cose-insulin
(grades 9–12) at physical activity, faculty, and youth including explora- Decrease in pulse rates, ratios
two high schools increased fruit and Teen wellness facility es- tion of attitudes suggesting improved
Baseline: n ⫽ 119; vegetable intake, and tablished to emphasize and beliefs on cardiovascular fitness
mid-project (2 reduced consump- physical activity health, food, and
years after base- tion of soft drinks Curriculum-based diabe- exercise
line) n ⫽ 173 tes education
introduced
Food supply in schools
modified
Trevino et al. (32) San Antonio, Texas Quasi-experimental; Primary goal to de- Parent health education Lessons developed by Diabetes health knowl- No significant Not targeted
Mexican-American no control group crease risk factors School classroom inter- use of culturally edge significantly changes
4th graders in two 2 years; results re- for diabetes: specifi- vention designed to relevant educa- improved
schools in poorest ported at 9 cally body fat supplement existing tional materials Significant decreases in
school district of months Secondary goal to in- health curriculum dietary fat servings and
San Antonio crease fruit and veg- School cafeteria total calories; and total
n ⫽ 102 etable intake, health component calories; significant
knowledge, self- After-school component increases in fruit and
efficacy, self-esteem, promoting physical vegetable servings
and activity levels activity
obesity, and exercise. Articles were
type 2 diabetes, cardiovascular disease,
“metformin”)—all in association with
used to prevent or control diabetes (e.g.,
development,” and the names of drugs
vention,” “prospective studies,” “program
“control,” “intervention,” “primary pre-
Search terms included “prevention,”
vention interventions in any country.
and 2001 that reported on diabetes pre-
language articles published between 1990

betes prevention.
interventions specifically focused on dia-
tion Database were searched to find
ter, and the Combined Health Informa-
Educational Resources Information Cen-
munity-based approaches. Medline, the
adults, whether from high-risk or com-
or modify its risk factors among youth or
signed to prevent or delay type 2 diabetes
examine all interventions specifically de-
was part of a larger review conducted to
METHODS — This literature review
RESEARCH DESIGN AND

vention at local, state, and national levels.

translation efforts for diabetes primary pre-
purpose of this review was to inform
community-based interventions. The
conducted a literature review to examine
ters for Disease Control and Prevention
community-based approaches, the Cen-
primary prevention of diabetes using
constituents. To evaluate the promise of
prevent and treat diabetes among their
comprehensive, sustainable programs to
tance to states and territories to develop
including providing support and assis-
findings into practice at multiple levels,
den of diabetes by translating research
tions, and others aim to reduce the bur-

vance of interventions (16,17).

cause they can ensure the cultural rele-
to be valued by local communities be-
communities (13). Such approaches tend
prevention of diabetes in well-defined
ularly suited to strategies for primary
sition of co-learners (15), may be partic-
studied (14), with researchers in the po-
ration of those affected by the issue being
action research, which involves collabo-
laying type 2 diabetes (13). Participatory
the challenging goals of preventing or de-
approaches are likely to be required for

from the beginning.

ingful participation of the community
Inclusion criteria included English-

Federal agencies, voluntary organiza-

Both high-risk and community-based

Satterfield and Associates

2645
2646

Community-based lifestyle interventions

Table 2—Population-based interventions designed to prevent or delay type 2 diabetes in adults

Study descriptions Results

Geographic location Community involvement

Target audience Study design and Goal(s) for prevention and culturally relevant Knowledge, Prevalence of
Reference Sample size duration of type 2 diabetes Interventions components attitudes, behaviors Clinical markers diabetes

Bjaras et al. (33, 34) Stockholm, Sweden Quasi-experimental; Reduce incidence of Community interventions: Not discussed Results reported for Not reported Not targeted
Population in three three intervention type 2 diabetes by ● Create supportive policy walking campaign
municipalities groups, two com- 25% by influencing environment for inter- conducted in one
n ⫽ 3,200 men and parison munici- risk factors (e.g., ventions municipality. in-
3,200 women palities reduce dietary fat ● Obtain media coverage creased (by one-
10 years intake, increase ● Inform public about third) the
physical activity) planned activities proportion of
Reduce the prevalence Intervention within com- adults participat-
of IGT munities (e.g., work- ing in regular
correspondingly places, residential exercise
areas):
● Implement strategies
related to dietary
change, weight control,
and physical activity
(e.g., walking groups)
Daniel et al. (20) British Columbia, Quasi-experimental; Reduce incidence of Physical activity events Interviews with commu- Cohort: significant Intervention com- Intervention
Canada one intervention type 2 diabetes by (e.g., 100-mile club), nity to examine reduction in pro- munity: community:
Okanagan Indian group, two com- reducing risk fac- cooking demonstra- knowledge, attitudes, portion of inter- ● Six new cases of ● No new cases
Band parison groups tors (e.g., over- tions, supermarket and and behaviors related vention IGT of diabetes
Cohort survey: 16 months weight enhancing) restaurant tours, media to diabetes and to participants who ● Substantially Comparison
mean of 105 and environmental campaign, environmen- identify prevention engaged in sweat- lower systolic communities:
adults at “high supports (e.g., tal support strategies producing activity blood pressure ● Six new cases
risk” (individuals walking groups, Sanction of Band Coun- Cross-sectional sur- and BMI than in of diabetes
with or at famil- hiring community cil for distribution of vey: significant comparison com-
ial risk for type 2 members) fliers. Use of commu- increases in munities
diabetes) nity facilities. Hired knowledge of Comparison com-
X-sectional survey: recreation coordinator diabetes and munities:
baseline: n ⫽ to promote physical prevalence of ● Four new cases of
DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003

150 adults; 16 activity sweat-producing IGT

months; n ⫽ 145 activity at least
once per week in
intervention
group
Engelgau et al. (18) Raleigh, North Quasi-experimental; Reduce modifiable Health promotion in gen- Involvement of commu- Not reported Not reported Not reported
Carolina comparison com- risk factors for dia- eral population: nity members in
African Americans munity betes in the general ● Increase in regular phys- project development
ⱖ18 years old 5 years population ical activity (e.g., walk- Work group of commu-
Intervention: n ⫽ Raise awareness (e.g., ing programs) nity members and
14,950; compari- use of radio and ● Decrease in fat intake representatives of local
son: n ⫽ 16,099 newspapers) about (e.g., cooking demon- organizations
risk factors for dia- strations)
betes and increase Outreach intervention:
use of screening Use of media, awareness
services activities and distribu-
tion of educational ma-
terials at key
community sites
DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003

Work with churches,

community centers,
civic groups, and public
housing community
groups to help with
screening promotion
Gittelsohnn et al. (21) Ontario, Canada Quasi-experimental Modify risk factors for Community education Ethnographic data col- Not reported Not reported Not reported
First Nations Duration not dis- diabetes, particu- about healthier lifestyle lection to aid in prob-
Ojibwe-Cree cussed larly obesity choices through mass lem definition and
Reserve, Canada media (e.g., radio and development of cul-
Sample size not television programs) turally appropriate
discussed Presentations at major intervention
community events (e.g., Intervention strategies
Treaty Days, annual generated by and pre-
health fairs) tested with commu-
One-to-one education nity
through series of home
visits includes nutrition
education and physical
activity component
Herbert (22) British Columbia, Study design not Develop culturally Briefly mentions interven- Significant involvement Not reported Not reported Not reported
Canada discussed sensitive ap- tion components, in- of the community in
Two First Nations 12 months proaches to diabe- cluding walking groups design of intervention
villages tes prevention and groups to gather
Sample size not traditional foods
discussed
Mau et al. (25) Two rural commu- Quasi-experimental Determine if lifestyle Lifestyle intervention in Development of cultur- Intervention group Not targeted Not targeted
nities in Hawaii, 4 years intervention with both communities ally responsive life- in pre-action
U.S. family support per- Trained family support style intervention stage more likely
Hawaiians at risk son improves life- person to attend diabe- to advance to ac-
for diabetes style behaviors tes-related functions tion stage for di-
n ⫽ 147 compared with with participant etary fat and physi-
standard interven- cal activity vs.
tion comparison group
Narayan et al. (26) Gila River Indian Randomized Test adherence to Intervention group: activi- Akime O’odham (Pima) Increased levels of Intervention group: Not measured
Community, 1 year specific lifestyle ty-based intervention community created physical activity ● Significant in-
Arizona interventions and involving structured intervention, includ- in both groups creases in BMI,
Akimel O’odham compare them for physical activity and ing local speakers on but no statistically weight, systolic
(Pima) adults changes in diabetes nutrition interventions; culture and history, significant differ- blood pressure,
who are over- risk factors (e.g., behavioral techniques, group facilitation by ences between diastolic blood
weight but nor- obesity) such as modeling and community members, groups pressure, and
moglycemic roleplay group problem and newsletters with plasma concentra-
n ⫽ 95 solving, and food prep- culture focus tions of glucose
aration demonstrations and insulin after
Comparison group: un- 2-hr fast
structured lifestyle in- ● Significantly greater

Satterfield and Associates

tervention based on weight gain and
self-directed learning, increase in plasma
facilitated by incorpora- concentrations of
tion of history and cul- glucose than com-
ture parison group
Comparison group:
● Significant de-
creases in waist
circumference and
2647

starch intake
Continued on followin page
2648

Community-based lifestyle interventions

Table 2—Continued

Study descriptions Results

Geographic location Community involvement

Target audience Study design and Goal(s) for prevention and culturally relevant Knowledge, Prevalence of
Reference Sample size duration of type 2 diabetes Interventions components attitudes, behaviors Clinical markers diabetes

Rowley et al. (27) Central Australia Quasi-experimental Examine trends in Informal education by Extensive discussions Not targeted Significant increases Prevalence of
Aboriginal popula- 7 years glucose tolerance physicians on diet and with community in mean BMI; diabetes unal-
tion ⬎15 years and coronary risk exercise council and commu- greater increase tered
old Health educator for 1-2 nity members on dia- for study partici-
n ⫽ 267 years betes and role of diet pants residing
Subsequent ongoing and exercise in treat- adjacent to a store
health promotion by ment and prevention compared with
health service staff in- those far from a
cluding trained diabetes store
community health edu- Significant decreases
cators in prevalence of
IGT and hyper-
cholesterolemia
among women
only
Simmons et al. (28) South Auckland, Quasi-experimental; Evaluate acceptability One educator presenta- Article states that project Significantly more Weight control in Not targeted
New Zealand comparison and impact of pilot tion, one video presen- is developing “cultur- retention of dia- intervention
Polynesian work- group program for diabe- tation, and 4-month ally appropriate strate- betes knowledge group; weight
force 4 months (ended tes awareness and exercise program gies” in intervention gain in compari-
Intervention: n ⫽ early) exercise Same ethnicity for diabe- group than in son group
108; comparison: tes community health comparison group;
n ⫽ 99 educators and study significant in-
participants crease in amount
DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003

of exercise in in-
tervention group
but decline in
comparison group
Simmons et al. (29) South Auckland, Quasi-experimental; Evaluate impact of Diabetes awareness ses- Evaluation and valida- Significant differ- Significant differ- Not targeted
New Zealand comparison diabetes-related sions followed by exer- tion of intervention ences between ences between
Two church con- group lifestyle program on cise groups materials among Pa- intervention and intervention and
gregations of 2 years (pilot) diabetes knowl- Reduced membership fees cific Island people in comparison comparison
Western Samo- edge, exercise hab- at local gymnasium area churches for dia- churches for
ans its, dietary habits, Cooking demonstrations betes knowledge, weight gain, BMI,
Intervention: n ⫽ and body size and local health promo- exercise amount, and waist circum-
78; comparison: tion services involving and dietary fat ference
n ⫽ 144 diabetes community intake
health educators
 Satterfield and Associates

grouped by their focus on type 2 diabetes pre-/posttest methodology. There was a diabetes knowledge score, although sta-
or other categories (e.g., prevention of wide range of sample sizes: from 24 ado- tistical analyses were not conducted.
obesity). Articles were then grouped by lescents on a U.S. Indian Reservation (23) In a program designed for students at
the approach to intervention (high-risk or to all residents living in several munici- two high schools in Zuni, New Mexico,
community-based). Studies using the palities in Stockholm, Sweden (33,34). Teufel and Ritenbaugh (30) targeted the
high-risk approach were defined as those The length of programs ranged widely, reduction of diabetes risk factors (e.g.,
based in a clinical trial setting. Commu- from a half-day workshop (23) to a 10- obesity and insulin resistance) through
nity-based approaches included those in year multiple-county intervention (33,34). various strategies, including integration
which authors described the intervention All interventions but one (28) com- of diabetes education into the school cur-
as a “community-based prevention effort” bined diet and exercise strategies in their riculum and modification of the school
or a “population-based approach.” program. Many of the interventions of- food supply. Knowledge, attitudes, and
This review reports on the original re- fered nutrition education that included behaviors surrounding type 2 diabetes
search articles that used community- cooking and food preparation demonstra- were targeted as secondary risk factors.
based approaches to prevent type 2 tions, grocery store tours, and recipe ex- The article we reviewed reported re-
diabetes. The interventions may have had changes. Exercise components included sults from baseline to mid-project (2 years
more than a primary prevention compo- residential walking programs, creation of after baseline). Risk factors were assessed
nent (i.e., a secondary or tertiary preven- exercise facilities, gentle exercise classes, using various measures, such as BMI, bio-
tion component). For example, Project and running clubs. electrical impedance analysis (BIA), 30-
DIRECT targeted the general population, The various program components min oral glucose tolerance test, activity,
as well as individuals with diabetes in the were designed to engage the target popu- and 24-h dietary recalls. Knowledge, atti-
community (18). However, because the lation in the development, implementa- tudes, and beliefs were measured using a
intent of this report was to review com- tion, and promotion of the interventions. questionnaire. The 24-h dietary recalls
munity-based interventions designed to Many incorporated culturally relevant showed nonsignificant increases in fiber
modify risk factors for or prevent type 2 messages, symbols, and strategies, with consumption and significant decreases in
diabetes, only Project DIRECT interven- respect for and inclusion of traditional consumption of sugary beverages. From
tion components targeted at primary pre- foods, activities, and knowledge. Many baseline to mid-project, BMI in both fe-
vention are discussed. were also based on a holistic view of males and males decreased in individuals
health, embracing spiritual, mental, emo- with a BMI ⱖ24 kg/m2; however, these
tional, and physical dimensions. results were not significant. Significant
RESULTS improvements in sitting pulse rates and
glucose-insulin ratios were also seen, thus
Description of studies Intervention outcomes suggesting improved cardiovascular fit-
The search revealed 16 reports published Results were reported for 11 of the 16 ness and a decline in hyperinsulinemia.
in peer-reviewed journals that met the in- interventions. The articles that did not The third study reporting results (31)
clusion criteria, and most included tar- present results focused on the process of targeted a predominantly Hispanic popu-
geted populations known to be at higher developing the program, or results were lation of fifth-grade students in schools on
risk for the development of type 2 diabe- not available at publication time. the Texas-Mexico border. The program
tes than the U.S. population at large (18 – Interventions targeting youth. For the was designed to encourage healthy life-
34). In the U.S., these populations four articles reporting results for youth styles through a curriculum focused on
included the Akimel O’odham (Pima) (23,30 –32), none of the studies included improving self-efficacy, knowledge and
(19,26), Ho-Chunk (Winnebago) (23), a control group; thus, all results are re- behaviors regarding type 2 diabetes, diet,
and Zuni Pueblo (30) peoples; Native Ha- ported for the intervention group(s), from and exercise. Two groups of teachers and
waiians (25); Mexican Americans baseline to follow-up periods. Below, we their students were involved, with both
(31,32); and African Americans (18). provide a brief description of each study groups using the curriculum and one
Four studies were conducted in Canada targeting youth and their reported results. group receiving training on implement-
(20 –22,24), two in New Zealand (28,29), Table 1 includes further details on each ing the program. Pre- and posttests were
and one each in Australia (27) and Swe- study. administered to determine knowledge of
den (33,34). Table 1 presents an overview Marlow et al. (23) worked with ado- diabetes, foods, exercise, exercise self-
of the methods and results (if reported) lescents residing on an Indian reservation efficacy, and related behaviors. Use of the
for each of the studies targeted at youth; in Nebraska to develop a culturally appro- program showed significant improve-
Table 2 presents the same information for priate education program to improve ments in knowledge and self-efficacy and
adult-focused studies. healthy eating and physical activity diet- and exercise-related behavior
Of the 16 interventions, 6 targeted among adolescents. Four adolescents led change. The program was effective with
youth, 9 targeted adults, and 1 targeted a half-day workshop using Native Amer- or without the training.
both groups. Because Gittelsohn et al. ican stories and activities to support be- Trevino et al. (32) targeted at-risk
(21) targeted both youth and adults, this havioral adaptations. Pre- and posttest Mexican-American fourth graders attend-
intervention is included in both tables but comparisons were used to measure ing schools in San Antonio to reduce
is counted as only 1 of the 16 interventions. knowledge change, although only 9 sets overweight and dietary fat intake by im-
The majority of studies reported us- of the 24 questionnaires were completed. plementing educational and behavioral
ing a quasi-experimental design and a Eight of the nine achieved an increased change components within four influen-

DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003 2649

Community-based lifestyle interventions

tial social systems for youth (parents, the lations within each community; clinical ferences were not statistically significant,
classroom, the cafeteria, and after-school markers such as blood pressure, choles- members of the Action group gained
care). Body fat was measured by BIA and terol, and glucose were also assessed more weight on average than members of
BMI; dietary fat intake was measured among cohort groups. Among cohort the Pride group.
through three 24-h dietary recalls. Pre- populations, both BMI and systolic blood Rowley et al. (27) conducted a com-
liminary results were presented for their pressure significantly decreased for the munity-wide prevention program among
2-year intervention at a 9-month period, intervention community relative to com- indigenous people in Australia to examine
and significant decreases in dietary fat parison communities; however, no other trends in glucose tolerance and risk of
servings and percent calories, significant significant changes were seen. Among coronary heart disease. The intervention
increases in fruit and vegetable servings, cross-sectional populations, the interven- included discussions held in clinical set-
and significant increases in diabetes tion group showed a significant increase tings on the benefits of diet and physical
knowledge were found. in knowledge of diabetes and an increased activity in preventing diabetes. Results in-
Interventions targeting adults. Seven prevalence of sweat-producing activity. cluded decreases in IGT prevalence; how-
of the ten interventions that targeted Among Native Hawaiians with or at ever, BMI increased significantly during
adults reported results (20,25–29,33). risk for diabetes, a family support inter- the 7-year period. The increase in BMI
Many of the studies targeting adults did vention was compared with a standard in- was greater among participants who lived
include comparison groups, although tervention in Hawaii to examine any in close proximity to a store compared
these may not have been used in all com- association between the Stages of Change with those residing far from a store.
ponents of the intervention. Below are de- construct and diet and exercise behaviors Two groups of hospital workers, di-
scriptions of the seven interventions. (25). Both groups received a lifestyle in- vided between an intervention and com-
Bjaras et al. (33) targeted three major tervention for 6 months, and participants parison group, participated in a pilot
risk factors for diabetes among residents in the family support intervention re- diabetes awareness and exercise program
of several Stockholm municipalities: im- ceived a trained self-identified family sup- in New Zealand (28). At 6 months from
proving physical activity levels, improv- port person. Although mean changes in baseline, significant differences were seen
ing healthy food intake, and decreasing diet and exercise behaviors from baseline between groups in the report of regular
obesity levels. The article reviewed for to follow-up varied widely and were not exercise activity (increases in the inter-
this work focused on the results of walk- significant for either group, patterns of vention group, decreases in the control
ing campaigns targeted at residents in one change based on the Stages of Change group). No differences occurred in BMI or
municipality who were not regularly ex- model for individuals in pre-action stages weight change.
ercising. Information on knowledge and were encouraging for the idea of includ- An urban program conducted among
attitudes about health was collected via ing a family support person in a lifestyle Western Samoans in New Zealand in-
self-report questionnaires after the walk- intervention. Participants in the family volved an intervention and comparison
ing programs. Knowledge about the rela- support group who progressed from pre- group from two different churches (29).
tionship between physical activity and action to action/maintenance stage gener- The intervention involved diabetes
several chronic diseases was fairly high, ally made healthier changes than the awareness sessions, exercise groups, and
although only 50 –57% of all participants standard intervention group. cooking demonstrations. Baseline and re-
thought that exercise could prevent dia- In a randomized pilot trial of lifestyle peat assessments involving clinical mark-
betes. One-third of the survey respon- interventions in an Akimel O’odham ers (e.g., glucose or fructosamine, anthro-
dents had previously not been exercising community in Arizona, normoglycemic pometric measurements) and a diabetes
regularly. obese adults were randomized to an “Ac- knowledge questionnaire were used to
Daniel et al. (20) conducted an inter- tion” group and a “Pride” group for 12 assess outcomes. Results for the interven-
vention to reduce prevalence of risk fac- months of intervention to reduce risk fac- tion church included stability of weight
tors for and the development of diabetes tors for diabetes (26). Participants in the contrasted to a weight gain in the compar-
among an indigenous population living Action group were guided by a structured ison church. The intervention group
on a reservation in British Columbia, Can- activity and nutrition intervention, demonstrated a significant reduction in
ada. Both behavior change and environ- whereas participants in the Pride group waist circumference as well as an increase
mentally supportive interventions were engaged in self-directed learning experi- in diabetes knowledge and regular exercise.
used, including walking groups, cooking ences grounded in an appreciation for
demonstrations, a media campaign, and their culture and history.
hiring people in the community to pro- At 6 and 12 months, both groups re- CONCLUSIONS — R e s e a r c h o n
mote the intervention. The intervention ported increased levels of physical activity community-based prevention of diabetes
community was matched with two com- and the Pride group reported a decreased is in its beginning phases, reflected in the
parison communities, and cohort (rela- intake of starch; however, group differ- paucity of studies found by this review.
tives of people with diabetes) and cross- ences were not statistically significant. Af- Among the studies that have been pub-
sectional (adults in the general ter 12 months, weight, BMI, systolic and lished, most have been conducted among
community) populations were surveyed diastolic blood pressure, 2-h plasma glu- populations disproportionately affected
in all communities. Questionnaires on cose, and 2-h insulin had significantly in- by diabetes, with their communities ei-
physical activity and dietary behavior, di- creased in the Action group, whereas ther initiating or collaborating with re-
abetes knowledge, and health beliefs were waist circumference had decreased signif- searchers. This finding, in itself, is
used in cohort and cross-sectional popu- icantly in the Pride group. Although dif- important and likely reflects the concern

2650 DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003

 Satterfield and Associates

of leaders about diabetes from these (e.g., plasma glucose levels, HbA1c lev- the United States. JAMA 286:1195–1200,
communities. els), are needed. Studies that include ex- 2001
Many researchers and collaborative amination of community change 4. Boyle JP, Honeycutt AA, Narayan KMV,
communities are breaking ground by im- indicators such as store buying patterns Hoerger TJ, Geiss LS, Chen H, Thompson
TJ: Projection of diabetes burden through
plementing culturally relevant prevention or the use of walking paths might be re- 2010. Diabetes Care 24:1936 –1940, 2001
programs in settings where many socio- vealing, particularly in programs that use 5. Tuomilehto J, Lindstrom J, Eriksson JG,
economic and environmental challenges an ecological framework. Valle TT, Hamalainen H, Ilanne-Parikka
exist. The inclusion of community health Although recent clinical trials have P, Keinanen-Kiukaanniemi S, Laakso M,
workers, traditional practices (e.g., bush shown that intensive lifestyle modifica- Louheranta A, Rastas M, Salminen V,
food), and the use of metaphors and tion and moderate weight loss can pre- Uusitupa M, the Finnish Diabetes Preven-
stories in a number of these studies pro- vent or delay the development of type 2 tion Study Group: Prevention of type 2
vides examples of genuine community in- diabetes (5–7), many of these clinical tri- diabetes mellitus by changes in lifestyle
volvement and application of cultural als were conducted in resource intensive among subjects with impaired glucose
knowledge. settings; adopting preventive measures tolerance. N Engl J Med 344:1343–1350,
2001
Although the studies we have reviewed on a population-wide basis will be more 6. Pan XR, Li GW, Hu YH, Wang JX, Yang
are to be lauded for using participatory challenging. Community-based interven- WY, An ZX, Hu ZX, Lin J, Xiao JZ, Cao
approaches, said to be the new gold stan- tions can dovetail with high-risk ap- HB, Liu PA, Jiang XG, Jiang YY, Wang JP,
dard for federally funded research (35), proaches and are valuable for reasons that Zheng H, Zhang H, Bennett PH, Howard
most had a number of limitations. Only differ from the high-risk approach. In ad- BV: Effects of diet and exercise in prevent-
one of the interventions we found used an dition to promoting lifestyle adaptations, ing NIDDM in people with impaired glu-
experimental design (26), likely because population-based approaches, governed cose tolerance: the Da Qing IGT and
of the cultural unacceptability of this ap- by the community, can identify and sup- Diabetes Study. Diabetes Care 20:537–
proach (25). Some comparison groups port protective factors within the culture 544, 1997
find this design discomfiting enough to that can be supported in meaningful 7. Diabetes Prevention Program Research
Group: The Diabetes Prevention pro-
mount their own interventions, even ways. They may also help garner social gram: reduction in the incidence of type 2
when promised a delayed intervention support among family and community diabetes. N Engl J Med 346:393– 403,
(31). The successful Pride group (26) was members and have far-reaching influ- 2002
essentially meant to be a comparison ences that, along with environmental 8. Benjamin SM, Rolka DB, Valdez R,
group that chose to incorporate tradi- changes, can help support adaptive re- Narayan KMV, Geiss LS: Estimated num-
tional ways (36). Interventions that show sponses among people at various points ber of adults with prediabetes in the U.S.
the most promise were associated with along a continuum of risk. in 2000: opportunities for prevention. Di-
well-designed research combined with However, to document these benefits, abetes Care 26:645– 649, 2003
participatory approaches. community-based interventions should 9. American Diabetes Association: The pre-
Other common study limitations in- use strong research designs with partici- vention or delay of type 2 diabetes. Dia-
betes Care 25:742–749, 2002
cluded the shortness of intervention du- patory approaches. These two ap- 10. Rolka DB, Narayan KMV, Thompson TJ,
ration, large numbers of nonresponders, proaches, used together, can help confirm Goldman D, Lindenmayer J, Alich K, Ba-
and the inability to match pre- and post- the potential effectiveness of population- call D, Benjamin EM, Lamb B, Stuart DO,
test data or to link self-reported lifestyle based endeavors to foster conditions that Engelgau MM: Performance of recom-
changes to health outcomes/indicators allow populations to be healthy, make mended screening tests for undiagnosed
(e.g., BMI, prevalence of IGT). Few stud- healthy choices, and prevent diabetes. diabetes and dysglycemia. Diabetes Care
ies demonstrated positive outcomes in all 24:1899 –1903, 2001
the intermediate outcomes of interest 11. Rose G: Sick individuals and sick popula-
(e.g., healthy eating behaviors and physi- Acknowledgments — The authors would tions. Int J Epidemiol 14:32–38, 1985
cal activity/exercise). Further, few studies like to acknowledge the thoughtful reviews of 12. Green L, Daniel M, Novick R: Partner-
Dr. David Simmons of Hamilton, New Zea- ships and coalitions for community-based
assessed whether the interventions are ef-
land, and Dr. Susan Norris of Atlanta, Georgia. research. Public Health Reports 11:20 –31,
fective in reducing plasma glucose levels 2001
or other diabetes risk factors among target 13. Harris SB, Zinman B: Primary prevention
populations. of type 2 diabetes in high risk populations
The limitations of the studies also References (Editorial). Diabetes Care 23:879 – 881,
provide us with the gaps in the literature 1. Mokdad AH, Ford ES, Bowman BA, Dietz 2000
and directions for future research. Re- WH, Vinicor F, Bales VS, Marks JS: Prev- 14. Green LW, George MA, Frankish CJ,
searchers should be encouraged to use alence of obesity, diabetes, and obesity- Daniel M, Herbert CJ, Bowie WR, O’Neill
more rigorous designs to evaluate com- related health risk factors. JAMA 289:76 – M: Study of Participatory Research in Health
munity- and population-based interven- 79, 2003 Promotion: Review and Recommendations
2. Mokdad A, Ford ES, Bowman BA, Nelson for the Development of Participatory Action
tions, including pre- and posttest designs. DE, Engelgau MM, Vinicor F, Marks JS: Research in Health Promotion in Canada.
In addition, more community-based Diabetes trends in the U.S.: 1990 –1998. Ottowa, Royal Society of Canada. Insti-
studies that examine proximal outcomes Diabetes Care 23:1278 –1283, 2000 tute of Health Promotion, 1995
such as self-reports or measured reports 3. Mokdad AH, Bowman BA, Ford ES, Vini- 15. Minkler M: Using participatory action re-
of physical activity (e.g., pedometer) and cor F, Marks JS, Koplan JS: The continu- search to build healthy communities.
weight loss, as well as clinical outcomes ing epidemics of obesity and diabetes in Public Health Rep 115:191–197, 2000

DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003 2651

Community-based lifestyle interventions

16. Macaulay AC, Commanda LE, Freeman American adolescents in the prevention of Zealand. Diabet Med 15:136 –142, 1998
WL, Gibson N, McGabe ML, Robbins diabetes. Diabetes Educ 24:441– 450, 1998 30. Teufel N, Ritenbaugh C: Development of
CM, Twohig PL: Participatory research 24. Macaulay AC, Paradis G, Potvin L, Cross a primary prevention program: insight
maximizes community and lay involve- EJ, Saad-Haddad C, McComber A, Desro- gained in the Zuni Diabetes Prevention
ment. BMJ 319:774 –778, 1999 siers S, Kirby R, Montour LT, Lamping Program. Clin Pediatr 37:131–141, 1998
17. Simmons D, Voyle J, Swinburn B, O’Dea DL, Leduc N, Rivard M: The Kahnawake 31. Holcomb J, Lira J, Kingery P, Smith D,
K: Community-based approaches for the Schools Diabetes Prevention Project: In- Lane D, Goodway J: Evaluation of Jump
primary prevention of non-insulin-de- tervention, evaluation, and baseline re- Into Action: a program to reduce the risk
pendent diabetes mellitus. Diabet Med 14: sults of a diabetes primary prevention of non-insulin dependent diabetes melli-
519 – 626, 1997 program with Native community in Can- tus in school children on the Texas-Mex-
18. Engelgau MM, Narayan KM, Geiss LS, ada. Prev Med 26:779 –790, 1997 ico border. J Sch Health 68:282–289, 1998
Thompson TJ, Beckles GL, Lopez L, Hart- 25. Mau MK, Glanz K, Severino R, Grove JS, 32. Trevino RP, Pugh JA, Hernandez AE,
well T, Visscher W, Liburd L: A project Johnson B, Curb JD: Mediators of lifestyle Menchaca VD, Ramirez RR, Mendoza M:
to reduce the burden of diabetes in the behavior change in native Hawaiians: ini- Bienestar: a diabetes risk-factor preven-
African-American community: Project tial findings from the Native Hawaiian Di- tion program. J Sch Health 68:62– 67, 1998
DIRECT. J Natl Med Assoc 90:605– 613, abetes Intervention program. Diabetes 33. Bjaras G, Harberg LK, Sydhoff J, Ostenson
1998 Care 24:1770 –1775, 2001 CG: Walking campaign: a model for de-
19. Cook V, Hurley J: Prevention of type 2 26. Narayan KM, Hoskin M, Kozak D, Kriska veloping participation in physical activ-
diabetes in childhood. Clin Pediatr 37: AM, Hanson RL, Pettitt DJ, Nagi DK, Ben- ity? Experiences from three campaign
123–130, 1998 nett PH, Knowler WC: Randomized clin- periods of the Stockholm Diabetes Pre-
20. Daniel M, Green LW, Marion SA, Gamble ical trial of lifestyle interventions in Pima vention Program (SDPP). Patient Educ
D, Herbert C, Hertzman CP, Sheps S: Ef- Indians: A pilot study. Diabet Med 15:66 – Counsel 42:9 –14, 2001
fectiveness of community-directed diabe- 72, 1998 34. Bjaras G, Ahlbom A, Alvarsson M,
tes prevention and control in a rural 27. Rowley K, Gault A, McDermott R, Knight Burstrom B, Diderichsen F, Efendic S,
Aboriginal population in British Colum- S, McLeay T, O’Dea K: Reduced preva- Grill V, Haglund BJ, Norman A, Persson
bia, Canada. Soc Sci Med 48:815– 832, 1999 lence of impaired glucose tolerance and PG, Svanstrom L, Tillgren P, Ullen H, Os-
21. Gittelsohn J, Harris SB, Whitehead S, no change in prevalence of diabetes de- tenson CG: Strategies and methods for
Wolever TMS, Hanley AJG, Barnie A, spite increasing BMI among aboriginal implementing a community-based diabe-
Kakegamic L, Logan A, Zinman B: De- people from a group of remote homeland tes prevention program in Sweden. Health
veloping diabetes interventions in an communities. Diabetes Care 23:898 –904, Promot Internation 12:151–160, 1997
Ojibwa-Cree community in northern On- 2000 35. Lindholm L, Rosen M: What is the “gold-
tario: linking qualitative and quantitative 28. Simmons D, Fleming C, Cameron M, en standard” for assessing population-
data. Chronic Dis Can 16:157–164, 1995 Leakehe L: A pilot diabetes awareness and based interventions? Problems of dilution
22. Herbert C: Community-based research as exercise programme in a multiethnic bias. J Epidemiol Community Health 54:
a tool for empowerment: the Haida Gwaii workforce. N Z Med J 109:373–376, 1996 617– 622, 2000
Diabetes Project example. Can J Public 29. Simmons D, Fleming C, Voyle J, Fou S, 36. Gregg EW, Narayan KMV: Culturally ap-
Health 87:109 –112, 1996 Feo S, Gatland B: A pilot urban church- propriate lifestyle interventions in minor-
23. Marlow E, Melkus G, Bosma A: STOP Di- based programme to reduce risk factors ity populations: more than what meets the
abetes! An educational module for Native for diabetes among W Samoans in New eye? (Letter). Diabetes Care 21:875, 1998

2652 DIABETES CARE, VOLUME 26, NUMBER 9, SEPTEMBER 2003