TheLive Well, Be Well Study: A Community-Based,Translational Lifestyle Program to Lower Diabetes Risk Factors in EthnicMinority and Lower–Socioeconomic Status Adults
Alka M Kanaya,MD
Jasmine Santoyo-Olsson,MS
Steven Gregorich,PhD
Melanie Grossman,PhD
Tanya Moore,PhD
Anita L Stewart,PhD
Correspondence should be sent to Alka M. Kanaya, MD, 1545 Divisadero St,Suite 311, San Francisco, CA 94115 (e-mail:alka.kanaya@ucsf.edu). Reprints can be ordered at http://www.ajph.org by clickingthe “Reprints” link.
Peer Reviewed
Contributors
A. M. Kanaya and A. L. Stewart had full access to all of thedata in the study and take responsibility for the integrity of the data andthe accuracy of the data analysis. J. Santoyo-Olsson and S. Gregorichperformed data analysis, and S. Gregorich led interpretation of the studyresults. M. Grossman and T. Moore helped with participant recruitment, studyvisits, and critical revision of the article. All authors helped toconceptualize ideas, interpret findings, and write and review drafts of thearticle.
Corresponding author.
Accepted 2011 Sep 4; Issue date 2012 Aug.
Abstract
Objectives. We evaluated a community-based, translationallifestyle program to reduce diabetes risk in lower–socioeconomic status(SES) and ethnic minority adults.
Methods. Through an academic–public health departmentpartnership, community-dwelling adults at risk for diabetes were randomlyassigned to individualized lifestyle counseling delivered primarily viatelephone by health department counselors or a wait-list control group. Primaryoutcomes (6 and 12 months) were fasting glucose level, triglycerides, high- andlow-density lipoprotein cholesterol, weight, waist circumference, and systolicblood pressure. Secondary outcomes included diet, physical activity, andhealth-related quality of life.
Results. Of the 230 participants, study retention was 92%. The6-month group differences for weight and triglycerides were significant. Theintervention group lost 2 pounds more than did the control group(P = .03) and had decreasedtriglyceride levels (difference in change, 23 mg/dL;P = .02). At 6 months, the interventiongroup consumed 7.7 fewer grams per day of fat(P = .05) and more fruits and vegetables(P = .02) than did controlparticipants.
Conclusions. Despite challenges designing effectivetranslational interventions for lower-SES and minority communities, this programmodestly improved some diabetes risk factors. Thus, individualized,telephone-based models may be a promising alternative to group-basedinterventions.
The prevalence of type 2 diabetes continues to rise at an alarming rate in the UnitedStates. Approximately 25.6 million adults (11.3% of the US population aged 20 years orolder) have diabetes, and another estimated 79 million have prediabetes.1 Greater risk of diabetes is observedfor ethnic minority1–5 and lower–socioeconomic status (SES)groups6 compared with Whiteadults of similar ages.
Several clinical trials have tested intensive lifestyle interventions or pharmacologicalagents in preventing or delaying type 2 diabetes in adults at risk.7–9 These trials showed impressivediabetes risk reductions for lifestyle interventions associated with relatively modestamounts of weight loss and exercise.7–9 Translating this knowledge into lifestyleinterventions delivered in real-world settings is thus a major priority.10–12
To reduce observed disparities in risk of diabetes, translational studies need to becommunity-based and designed for lower-SES and ethnic minority populations. Althoughmany translational lifestyle interventions are available, most were designed forclinical settings;13–21 only a few are offered in community settings.22–26Of community-based translations, only 3 were designed specifically for lower-SES orminority populations,23–25 and only 1 of these—Project HEED, or HelpEducate to Eliminate Diabetes—was evaluated with a randomized controlled trialdesign.23 HEED was successful inobtaining significant group differences in weight loss at 12 months, but no othersignificant clinical or behavioral changes were observed.
We conducted a randomized controlled trial of a low-intensity lifestyle intervention forlower-SES, ethnic minority, Spanish- and English-speaking adults. This was acollaborative project between the University of California, San Francisco, and the Cityof Berkeley Division of Public Health. Public health departments are a good venue forcommunity-based translations to reduce disparities because they serve vulnerablepopulations most at risk for chronic disease and engage in chronic diseaseprevention.
METHODS
This 12-month randomized controlled trial compared a lifestyle intervention groupwith a wait-list control group. After completing the trial, control groupparticipants were offered the lifestyle program. We examined program effectivenessat 6 months (active intervention phase) and 12 months (after maintenance phase) toimprove clinical diabetes risk factors, behavioral risk factors, and health-relatedquality of life.
Study Participants
We focused on community-dwelling adults in 4 distinct low-income neighborhoods innorthern California cities: Berkeley, Oakland, and Richmond. Recruitment beganwith community-based, educational outreach to identify individuals at risk fordiabetes.27 Individualscompleted a brief self-administered diabetes risk appraisal assessing age,race/ethnicity, history of high blood pressure, abnormal cholesterol level,gestational diabetes, family history of diabetes, regular exercise, and bodymass index (BMI; defined as weight in kilograms divided by height in meterssquared). The diabetes risk appraisal, adapted from existing diabetes risk toolsfor use in community settings, used only self-reported variables and asimplified scoring system. Staff scored the diabetes risk appraisal andexplained the risk score (0–3 points = low,4–8 points = moderate, ≥ 9points = high). Individuals with a moderate or high score(> 4) were educated about their diabetes risk and invited tocomplete an 8-hour fasting finger-stick test (Accu-chek; Roche Diagnostics,Indianapolis, IN) to determine fasting capillary glucose level.
Individuals who had a capillary blood glucose value between 106 and 160milligrams per deciliter, who had a moderate to high diabetes risk appraisalscore, and who were aged 25 years or older were told about the lifestyle programand study as well as the research process and screened for exclusion criteria.We excluded individuals with diabetes (physician diagnosis, use of insulin orother diabetes medications); diagnosis in past 6 months of myocardialinfarction, congestive heart failure, or stroke; heart procedure or heartsurgery in past 6 months; implanted defibrillator; hip or knee replacement inpast 3 months; insufficient cognitive functioning; and pregnancy. We alsoexcluded individuals not conversant in English or Spanish, with plans to moveout of the area within 1 year, and whose spouse or partner had already enrolled.We required physician consent for those with a pacemaker, heart disease, heartrhythm abnormalities, or atrial fibrillation, as well as chest pain or faintnessor dizziness in the past 6 months.
Live Well, Be Well Intervention
The lifestyle program28 wasdesigned for lower-SES, minority, and low-literacy adults and adapted fromseveral interventions with established efficacy. It was delivered in Spanish andEnglish and consisted of a 6-month active intervention phase and a 6-monthmaintenance phase. Trained health department counselors provided education andskills training to modify diet and physical activity through primarilytelephone-based counseling (12 calls) with 2 in-person sessions and 5 optionalgroup workshops. In-person and group sessions were held in neighborhoodsettings. Self-selected and attainable goal-setting and action plans wereemphasized to enhance self-efficacy. Motivational interviewing techniques todevelop and enhance participants’ motivation were used during thetelephone calls. All program materials are available on theLive Well,Be Well Web site (http://iha.ucsf.edu/LiveWellBeWell).
Participation in each program component was tracked. The program consisted of 19possible “contacts” for a total of 15 possible hours: 1introductory session, which included a program binder; 1 in-person planningsession; 12 telephone counseling calls (10 in active phase, 2 in maintenancephase); and 5 group workshops. Participation was calculated as the total numberof contacts received. Minimum compliance was defined as completion of theintroductory session, the planning session, and at least 8 telephone calls overthe active phase.
We assessed all outcomes at baseline with 6- and 12-month follow-ups. Trained,bilingual research assistants administered questionnaires and performed clinicaland anthropometric measurements with standardized protocols. At baseline, weobtained demographic information and a brief medical history. All questionnaireswere translated into Spanish.
We selected 7 primary clinical and anthropometric outcomes related to diabetesrisk. Laboratory measures included fasting serum glucose, triglyceride, and low-and high-density lipoprotein (LDL and HDL) cholesterol levels measured byenzymatic calorimetric methods (Quest Diagnostics, San Jose, CA). Weight inpounds was measured on a digital scale (Detecto Balance Beam Scale; CardinalScale Manufacturing Co, Webb City, MO). We measured waist circumference with aGullick II (FitnessMart; Country Technology, Inc, Gays Mills, WI) tapespring-tension measure at the site of maximum circumference midway between thelower ribs and the anterior superior iliac spine (mean of 2 measurements).Systolic blood pressure was measured with an Omron (Omron Healthcare, Inc, LakeForest, IL) automated blood pressure monitor after sitting for 5 minutes (meanof 2 measurements). Participants provided additional consent for serum bankingat baseline and 12 months for metabolic biomarker assays. We measured serumfasting insulin level by radioimmunoassay (Millipore, St. Charles, MO) andcalculated homeostasis model assessment-insulin resistance (HOMA-IR).29
Secondary behavioral risk factor outcomes targeted diet and physical activity. Weassessed dietary intake with the Modified Block Food Frequency Questionnaireavailable in English and Spanish.30–32 Questionnaires were scored byNutritionQuest (Berkeley, CA). We used 4 measures targeted specifically in theLive Well, Be Well program: (1) total kilocalories per day;(2) total fat per day; (3) total fiber per day; and (4) daily frequency ofconsumption of fruits, fruit juices, and vegetables. Missing values wereassigned to daily energy intake values that did not fall within the range of 500to 5000 kilocalories per day (9 instances). Physical activity was measured withthe CHAMPS Physical Activity Questionnaire.33 BecauseLive Well, Be Wellemphasized increased participation in all physical activities, especiallywalking, we used 3 measures: (1) hours per week in any physical activity, (2)metabolic equivalent hours per week in any physical activity, and (3) hours perweek walking.
We measured 5 secondary health-related quality-of-life outcomes hypothesized toimprove with lifestyle changes. Participants rated their overall health from“poor” to “excellent” on a scale from 1 to 5. ThreeMedical Outcomes Study measures were (internal consistency reliability inparentheses) the Psychological Distress II and Psychological Well-Being IIindexes (0.90 and 0.78, respectively)34 and the Sleep Problems Index (0.77).35 We used the PerceivedStress Scale (0.86) to assess stress.36
Sample size estimates assumed equal allocation to experimental groups,correlation between repeated outcomes equaling 0.75, unit-standardized outcomes,90% retention at 12 months, 80% power, and 2-tailed α equal to .05.Modeled outcomes included difference scores computed by subtracting outcomevalues assessed at baseline from each corresponding follow-up value. Given thestudy sample size of 230, the minimum detectable effect corresponded to a groupdifference in the effect of time equal to 0.28 SDs. This is generally considereda small to small-to-medium effect, suggesting good power.
Eligible participants were stratified by self-reported race/ethnicity (AfricanAmerican, Latino, other) and age categories (25–39, 40–64,≥ 65 years). We generated stratum-specific sequentialidentification numbers to randomly allocate individuals to experimental groupsin blocks of 4. Study staff was blinded to the linkage between identificationnumbers and group assignment. At the end of the baseline visit, the participantopened a sealed, opaque envelope preprinted with the sequential identificationnumber to determine the experimental group assignment. Because of the behavioralnature of the intervention, counselors administering the intervention andparticipants were not blind to group assignment.
Statistical Methods
A randomization check compared experimental groups at baseline viaχ2 andt tests.Intention-to-treat linear regressions modeled repeated change scores: outcomesassessed at each follow-up (6 or 12 months) minus the corresponding baselinevalue. All models included effects of experimental groups, assessment time, andthe groups-by-time interaction. Initially, an unstructured residual covariancestructure was specified for the model of each outcome, which was subsequentlyrelaxed in 2 steps to consider separate covariance structures for (1) eachexperimental group and (2) each combination of experimental group andparticipant sex. The final residual covariance structure was chosen by referenceto deviance statistics. The likelihood-based approach to model estimationallowed models to be fit to all available data and invoked the assumption thatmissing values occurred randomly, conditional on observed values.37
RESULTS
We randomly assigned 238 individuals to an intervention or a control group betweenJuly 2006 and July 2008 (Figure 1) withfollow-up assessments from December 2006 through August 2009. Of the 119 individualsassigned to the intervention group, 6 were excluded because of use of diabetesmedications at enrollment or for interim pregnancy, leaving 113 for analysis. Fromthe control group, 2 were excluded because of use of diabetes medications, leaving117 for analysis. Twelve-month study retention was 105 (93%) in the interventiongroup and 107 (91%) in the control group.
FIGURE 1—
Flow of participants from screening to completion of the finalfollow-up assessment:Live Well, Be Well program,Berkeley, Oakland, and Richmond, California, July 2006–August2009.
Baseline characteristics indicated a primarily ethnic minority and lower-SES sample,and the average age was older than 55 years, and 73% were women (Table 1). Almost one third wereSpanish-speaking, and 77% were from an ethnic minority group. About 20% had nohealth insurance, about 40% were employed, and approximately 30% reported financialhardship within the prior year. Educational attainment was diverse; about 38% had ahigh-school diploma or less education. More than half (55%) the sample were obese,and an additional 32% were overweight.
TABLE 1—
Demographic and Socioeconomic Characteristics of Study Participants:Live Well, Be Well Program, Berkeley, Oakland, andRichmond, California, July 2006–August 2009
| Characteristics | Intervention Group (n = 113), %or Mean (SD) | Control Group (n = 117), % orMean (SD) | P |
| Female | 73 | 74 | .87 |
| Race/ethnicity | |||
| African American | 23 | 23 | .74 |
| Non-Hispanic White | 22 | 23 | |
| Latino | 35 | 39 | |
| Asian | 18 | 13 | |
| Native American/Pacific Islander | 1 | 1 | |
| Multiethnic/mixed | 2 | 1 | |
| Age, y | 58 (16) | 55 (17) | .29 |
| Language of interview | |||
| English | 72 | 66 | .34 |
| Spanish | 28 | 34 | |
| Immigrant to the United States | 47 | 48 | .93 |
| Of immigrants, speak English | |||
| Not at all/poorly/fairly well | 55 | 68 | .16 |
| Well/very well | 45 | 32 | |
| Educational achievement | |||
| < High school | 21 | 25 | .14 |
| High school/GED | 20 | 11 | |
| Some college/tech | 27 | 22 | |
| ≥ Bachelor's degree | 32 | 42 | |
| Health insurance type | |||
| Any private | 63 | 61 | .97 |
| Public | 15 | 17 | |
| None | 22 | 22 | |
| Employed full- or part-time | 35 | 43 | .22 |
| Financial hardship in past ya | 30 | 32 | .8 |
| Family history of diabetes | 53 | 45 | .24 |
| BMI, kg/m2 | 30.1 (5.3) | 29.9 (6.1) | .78 |
| BMI categoriesb | |||
| Normal | 10 | 17 | .2 |
| Overweight | 31 | 32 | |
| Obese | 59 | 50 | |
| Hypertensionc | 50 | 44 | .36 |
| Arthritisd | 35 | 34 | .85 |
Note. BMI = body mass index;GED = general equivalency diploma.
Reponded yes to “In the past 12 months, was there ever a time whenyou did not have enough money to meet your daily needs?”
BMI categories defined as follows: for non-Asian participants, normal(< 25.0 kg/m2); overweight (25.0 to ≤ 29.9 kg/m2); and obese(≥ 30.0 kg/m2); for Asian participants, normal(< 23.0 kg/m2); overweight(23.0 ≤ BMI to ≤ 24.9kg/m2); and obese (≥ 25.0kg/m2).
Hypertension classified if either systolic > 140 mm Hg ordiastolic > 90 mm Hg, or if participant reported using anyblood pressure medication.
Responded yes to “Has a health professional ever told you that youhad arthritis or other joint problems?”
Approximately 72% of the intervention group participants were minimally compliant. Of19 possible program components, a mean (SD) of 12.5 (4.9) were completed. Of 12total possible telephone calls, a mean of 8.9 (3.8) were completed. The averagenumber of workshops attended of 5 was 1.7 (1.5). The mean (SD) hours of contactreceived was 9.0 (3.3) of a possible 15.
Table 2 presents baseline mean scores and 6-and 12-month change scores for the intervention and control groups and thesignificance of between-group comparisons for the coprimary clinical outcomes. Nobaseline group differences were seen in these measures.
TABLE 2—
Effect of Intervention and Control Groups on Changes From Baseline inClinical Outcomes:Live Well, Be Well Program,Berkeley, Oakland, and Richmond, California, July 2006 to August2009
| Intervention Group(n = 113) | Control Group(n = 117) | Between-Group Comparison ofChange | ||||||
| Outcomes | Baseline, Mean (SE) | 6-Mo Within-Group Change, Mean (SE) | 12-Mo Within-Group Change, Mean (SE) | Baseline, Mean (SE) | 6-Mo Within-Group Change, Mean (SE) | 12-Mo Within-Group Change, Mean (SE) | 6 Mo,P | 12 Mo,P |
| Weight, lb | 177.85 (3.68) | −2.30***(0.66) | −1.34 (0.71) | 176.45 (3.68) | −0.44 (0.57) | −0.42 (0.84) | .03 | .4 |
| Waist, cm | 100.56 (1.34) | −0.56 (0.44) | −0.06 (0.44) | 99.30 (1.32) | 0.50 (0.53) | −0.15 (0.48) | .13 | .89 |
| Systolic blood pressure, mm Hg | 126.90 (1.74) | 0.73 (1.39) | 0.34 (1.38) | 127.58 (1.98) | −1.17 (1.32) | 0.27 (1.61) | .32 | .98 |
| Fasting glucose, mg/dL | 93.82 (1.05) | −0.70 (0.87) | −0.88 (1.02) | 93.50 (1.14) | 0.42 (1.04) | −1.39 (0.96) | .41 | .72 |
| Triglycerides, mg/dL | 148.26 (10.71) | −8.76 (7.66) | −1.57 (6.83) | 128.13 (8.56) | 14.39* (6.35) | 4.87 (4.99) | .02 | .45 |
| Low-density lipoprotein, mg/dL | 112.00 (2.95) | −6.62***(1.84) | −5.78* (2.34) | 114.76 (2.99) | −2.39 (2.07) | −3.61 (2.17) | .13 | .5 |
| High-density lipoprotein, mg/dL | 53.05 (1.56) | 1.76* (0.72) | 3.19*** (0.88) | 54.69 (1.56) | 0.61 (0.71) | 1.69* (0.80) | .26 | .21 |
Note. Each model estimated unstructured residualcovariances separately within each combination of experimental groupsand respondent sex.
*P < .05;***P < .001.
Group differences in 6-month change for weight and triglycerides were significant.The intervention group lost 1.9 pounds more than did the control group(P = .03), and a net difference in changein triglyceride levels was found between groups (a difference of about 23 mg/dLfavoring the intervention group;P = .02). Nosignificant group differences were observed for other clinical outcomes. However, 2significant within-group changes were observed in the intervention group: LDLdecreased at 6 months (P < .001) and 12 months(P < .05), and HDL increased at 6 months(P < .05) and 12 months(P < .001); HDL also increased in thecontrol group at 12 months (P < .05).
Approximately half of the study participants consented to optional blood banking forthe insulin assays. Among those with insulin data, the baseline HOMA-IR among theintervention group was 1.18 ±0.55 and among the control participants was 1.29±0.70. The 12-month HOMA-IR in the intervention group was 1.21 ±0.54and in the control group was 1.34 ±0.63 with no difference betweengroups.
Table 3 presents baseline mean scores and 6-and 12-month change scores for the intervention and control groups, as well asbetween-group comparisons, for the secondary behavioral and health-relatedquality-of-life outcomes. No baseline group differences were found.
TABLE 3—
Effect of Intervention and Control Group on Changes From Baseline inBehavioral and Health-Related Quality-of-Life Outcomes:LiveWell, Be Well Program, Berkeley, Oakland, and Richmond,California, July 2006 to August 2009
| Intervention Group(n = 113) | Control Group(n = 117) | Between-Group Comparison ofChange | ||||||
| Outcomes | Baseline, Mean (SE) | 6-Mo Within-Group Change, Mean (SE) | 12-Mo Within-Group Change, Mean (SE) | Baseline, Mean (SE) | 6-Mo Within-Group Change, Mean (SE) | 12-Mo Within-Group Change, Mean (SE) | 6 Mo,P | 12 Mo,P |
| Total calories, kcal/da | 1870.51 (78.12) | −264.29***(50.57) | −301.57***(64.69) | 1915.12 (81.01) | −216.59** (69.23) | −245.88***(52.68) | .58 | .51 |
| Total fat, g/da | 71.49 (3.58) | −12.95***(2.41) | −14.35***(2.89) | 67.93 (3.11) | −5.28 (3.13) | −7.81** (2.39) | .05 | .08 |
| Total dietary fiber, g/db | 17.84 (0.90) | −1.13 (0.70) | −1.97* (0.77) | 19.73 (1.07) | −1.30 (0.82) | −1.79** (0.66) | .88 | .86 |
| Daily frequency fruits and vegetablesc | 3.01 (0.15) | 0.25 (0.16) | 0.12 (0.14) | 3.12 (0.16) | −0.26 (0.15) | −0.30* (0.14) | .02 | .04 |
| Physical activity, h/wka | 7.99 (0.63) | 0.74 (0.62) | 0.68 (0.67) | 7.00 (0.54) | 0.44 (0.61) | 1.07 (0.58) | .73 | .66 |
| Physical activity, metabolic equivalent, h/wka | 25.59 (2.07) | 3.00 (2.20) | 2.24 (2.11) | 23.62 (2.16) | 0.36 (1.97) | 6.37* (2.77) | .37 | .24 |
| Walking, h/wkc | 4.40 (0.38) | 0.37 (0.41) | 0.57 (0.46) | 3.90 (0.36) | 0.31 (0.40) | 0.55 (0.46) | .92 | .97 |
| Self-rated healthb | 3.09 (0.08) | 0.31*** (0.07) | 0.13 (0.07) | 2.92 (0.08) | 0.07 (0.10) | 0.11 (0.09) | .05 | .84 |
| Perceived stressc | 2.26 (0.07) | −0.01 (0.06) | 0.01 (0.06) | 2.37 (0.07) | 0.03 (0.06) | 0.06 (0.06) | .6 | .56 |
| Sleep problemsb | 2.15 (0.07) | 0.06 (0.06) | 0.01 (0.06) | 2.13 (0.07) | 0.07 (0.06) | 0.17** (0.06) | .91 | .05 |
| Psychological well-beingc | 3.74 (0.07) | 0.01 (0.05) | 0.07 (0.06) | 3.75 (0.07) | −0.14* (0.05) | −0.09 (0.06) | .05 | .04 |
| Psychological distressc | 2.09 (0.07) | −0.06 (0.05) | −0.11* (0.05) | 2.21 (0.06) | −0.04 (0.05) | −0.00 (0.05) | .75 | .17 |
Residual covariance unstructured within each combination of experimentalgroups and respondent sex.
Residual covariance unstructured within each intervention group.
Simple unstructured residual covariance.
*P < .05;**P < .01;***P < .001.
The intervention group consumed 7.7 fewer grams per day of fat at 6 months(P = .05). Intervention group membersreported more frequent consumption of fruits and vegetables than did the controlgroup at 6 months (P = .02) and 12 months(P = .04). No significant groupdifferences were observed in total calories, dietary fiber, or physical activity.However, some within-group changes were significant: within both groups at 6 months,total consumed calories decreased by 264 kilocalories per day for the interventiongroup (P < .001) and 217 kilocalories per dayfor the control group (P < .01). Total fiberintake decreased in both groups at 12 months(P < .05 for the intervention group andP < .01 for the control group). Within thecontrol group, metabolic equivalent hours per week in physical activity increased at12 months (P < .05).
The intervention group had better psychological well-being than did the control groupat 6 months (P = .05) and 12 months(P = .04). The intervention group also hadgreater improvement in self-rated health at 6 months(P = .05) and fewer reports of sleepproblems at 12 months (P = .05). There were nogroup differences in perceived stress or psychological distress, but there was awithin-group change: psychological distress was reduced at 12 months in theintervention group (P = .05).
DISCUSSION
In this community-based translational trial of a low-intensity, individuallytailored, telephone-based lifestyle intervention in at-risk, lower-SES, ethnicminority adults, we had excellent study retention and program compliance andachieved small improvements in several clinical and behavioral risk factors. Theintervention group had significantly more weight loss than did the control group,and the net difference in triglycerides, dietary fat intake, and daily fruit andvegetable consumption favored the intervention group after 6 months ofintervention.
The weight loss difference of 1.9 pounds between groups at 6 months was small butstatistically significant. Nevertheless, it may be clinically relevant becauseweight loss was the main predictor of reduced diabetes incidence in the DiabetesPrevention Program: every kilogram of weight loss was associated with a 16%reduction in risk for diabetes.38 Additionally, the intervention group had decreasedtriglycerides, whereas the control group had increased triglycerides from baselineat both time points; hence we observed a large group difference in change intriglycerides. The magnitude of the group difference in triglyceride change(difference of 23.2 mg/dL at 6 months) was comparable to the difference of 17milligrams per deciliter in the Finnish Diabetes Prevention Study.8 All mean triglyceride levels,however, were lower than the standard 150 milligrams per deciliter threshold valueused to designate elevated triglyceride levels.
The reduction in total fat intake (7.7 g/day less for the intervention group at 6months) was notable because reduction in fat consumption was a strong predictor oflower diabetes risk in the Diabetes Prevention Program.38 We found changes in consumption of fruits andvegetables of 0.6 servings per day at 6 months. The observed effect sizes of these 2dietary changes were comparable to those found in a systematic review of otherphysical activity interventions.39 The improvement in self-rated health was notable becauseself-rated health consistently predicts future health.40–42
No significant group differences were found in fasting glucose or LDL- orHDL-cholesterol levels, waist circumference, and systolic blood pressure. We have 3possible explanations: (1) some clinical risk factors were in the normal range atbaseline; (2) for some factors, although risk reductions were observed in theintervention group, reductions also were seen in the control group; and (3) theintervention may not have been intensive enough to achieve change.
First, our baseline venous fasting glucose level was in the normal range(mean = 94 mg/dL), which may have precluded observing change inthis important risk factor. Other clinical risk factors in near-normal rangeincluded systolic blood pressure (mean = 127 mm Hg), LDL (112mg/dL), and HDL (53 mg/dL). These normal values occurred despite efforts to recruitparticipants at high risk for diabetes based on elevated fasting capillary bloodglucose level and self-reported risk factors. This suggests that community-basedstudies may have more difficulty recruiting people at greatest risk, in contrast tostudies in clinical settings where fasting venous blood test screening can be done.Nevertheless, our sample was at moderate to high risk on other risk factors: about85% were overweight (more than half were obese), 78% of women and 50% of men hadelevated waist circumference, 35% fulfilled metabolic syndrome criteria, and almost50% had a family history of diabetes.
Second, although the intervention group had significant improvements in LDL and HDLcholesterol and a significant reduction in total caloric intake, similar controlgroup improvements precluded observing between-group differences. This may haveoccurred because diabetes prevention educational materials were part of outreach andrecruitment, and control group participants may have changed behavior to some extentwithout the program. Indeed, at follow-up, 17% of the control group reported havingparticipated in another lifestyle program at some point during the year. Inaddition, completing the food frequency questionnaire may have raised awarenessabout diet and portion sizes. Control group improvements have been observed in otherdiabetes risk reduction interventions. The DEPLOY (Diabetes Education &Prevention with a Lifestyle Intervention Offered at the YMCA) study foundsignificant weight loss in the control group YMCA setting and similarly had provideddiabetes education during recruitment.22 Project HEED found significant control group weightloss; qualitative analyses indicated that control group participants believed theybenefited from learning that they were at risk and being given information aboutdiabetes.23 Generally,however, such minimal “interventions” are not effective.12
Third, theLive Well, Be Well program may not have been sufficientlyintensive to achieve broad changes. This underscores the substantial challenge todesign practical, sustainable programs for underserved populations that obtain riskreductions comparable to those of the Diabetes Prevention Program. One recentcommunity-based translation, Healthy-Living Partnerships to Prevent Diabetes (HELPPD), found significant reductions in fasting glucose level and several otherclinical risk factors, but it was a high-intensity program offered to relativelywell-educated participants with high levels of risk.26 Designing such programs to be more intensive toachieve greater risk reduction might jeopardize the likelihood of adoption andsustainability by community-based organizations.Live Well, Be Wellconformed to contemporary recommendations for translating diabetes preventionlifestyle programs—namely, to use individually tailored goals,self-monitoring, counselors, and other participants to provide support and aproblem-solving approach to overcome barriers.11,43Live Well, Be Well also conformed to criteria for an“individually-adapted health behavior change program,” stronglyrecommended by the Task Force on Community Preventive Services to increase physicalactivity,44 and behaviorchange strategies were similar to those in the diabetes prevention trials.12
By designing theLive Well, Be Well program specifically forlower-SES minority groups that are at higher risk for type 2 diabetes than theircounterparts,1–6 this study contributes to the small field ofcommunity-based translations aimed at reducing health disparities. Although minoritygroups are often underrepresented in intervention research,45,46 we obtained a high study retentionrate (92%). The other translational intervention designed for underserved groups andtested via a randomized controlled trial design, Project HEED, was a 10-weekpeer-led group-based program offered in community settings to lower-SES Latino andAfrican American overweight adults with prediabetes.23 Participants were thus at higher risk than werethose inLive Well, Be Well. Ninety participants were randomlyassigned with 73% study retention. Significant differences in weight loss werereported (at 12 months, the intervention group lost 7.2 pounds compared with 2.4pounds in the control group;P < .01);however, they found no other clinical or behavioral changes. Their results arenotable because the program was less intensive thanLive Well, BeWell and was delivered by trained peer educators.
Live Well, Be Well is the only community-based translation that usedan individually tailored, primarily telephone-based model rather than a group-basedmodel. The use of telephone counseling and neighborhood settings for in-personsessions made it convenient to participate, possibly allowing people to enroll whootherwise could not have participated. Indeed, completion of telephone calls wassubstantially higher than workshop attendance; thus, a group-based approach may nothave been feasible for this population. Also, in participant interviews afterprogram completion, telephone calls were rated as the most useful program feature.Use of public health department infrastructure and staff for intervention deliveryhas not been tested previously in diabetes risk reduction studies.
Our approach addressed 4 translation priority areas10 by (1) focusing on vulnerable, understudied groups;(2) having few exclusion criteria, thus being more generalizable; (3) being apartnership between researchers and a public health department, thus reflectingtheir shared perspectives; and (4) being designed to be sustainable by embedding theprogram in the public health department's chronic disease preventionprogram.
Limitations
The study had several limitations. Implementation in 1 city-level public healthdepartment setting limited generalizability. However, agencies with largerservice areas (e.g., county-level health departments) might find thetelephone-based counseling model more attractive than in-person health educationworkshops. Additionally, because we had difficulty recruiting men, our sampleincluded only 26% men. Use of fasting screening tests limited generalizabilityto adults available in the morning, similar to other studies.23 Finally, the significantbetween-group differences in weight and triglyceride levels were small and mayhave limited clinical benefit.
Conclusions
Our community-based translational study indicated that theLive Well, BeWell intervention was associated with small changes in a fewimportant diabetes risk factors in lower-SES ethnic minority adults, thusproviding a promising approach for future translational efforts to reducedisparities. Because so few community-based models for delivering lifestyleinterventions are available, our results suggest that individually tailoredprograms with telephone counseling should be considered along with the moretraditional group-based approaches. Testing lifestyle programs that areintegrated into a health department's chronic disease preventioninfrastructure and delivered by public health department counselors in localcommunity venues provides a novel and sustainable goal for translationalresearch.
Future research could adaptLive Well, Be Well and explore therelative effectiveness of variations in program delivery organization (healthdepartment, peer educators), delivery mode (group- vs telephone-based),intensity (number of contacts, duration), and risk-level eligibility(overweight, other diabetes risk factors) in terms of risk reduction and programcompliance.
Acknowledgments
Financial support for this project was through a translational research grant fromthe National Institute of Diabetes and Digestive and Kidney Diseases (R18DK067896-01A2) and by the Resource Centers for Minority Aging Research program ofthe National Institute on Aging (grant P30-AG15272).
We acknowledge the substantial contribution of the study staff at the City ofBerkeley Division of Public Health (Rainbow Schwartz, Elisa Gallegos, MariaGuerrero, LeConté Dill, Kristen Tehrani, and Kate Clayton) and University ofCalifornia, San Francisco, staff (Julissa Cabrera, Adriana Delgadillo, RachelFreyre, Natalie Alvarez, and Deepika Mathur).
Clinical Trials Registration:clinicaltrials.gov identifierNCT00770926.
Human Participant Protection
The research protocol was approved by the University of California, San Francisco,Committee on Human Research, and written consent was obtained from all studyparticipants.
References
- 1.Centers for Disease Control andPreventionNational Diabetes Fact Sheet: NationalEstimates and General Information on Diabetes and Prediabetes in the UnitedStates. Atlanta, GA:Centers for Disease Control and Prevention;2011 [Google Scholar]
- 2.Vega WA, Amaro H.Latino outlook: good health,uncertain prognosis. Annu Rev PublicHealth.1994;15:39–67 [DOI] [PubMed] [Google Scholar]
- 3.Liao Y, Tucker P, Okoro CA, Giles WH, Mokdad AH, Harris VB.REACH 2010 Surveillance for HealthStatus in Minority Communities—United States,2001–2002. MMWR Surveill Summ.2004;53(6):1–36 [PubMed] [Google Scholar]
- 4.Carter JS, Pugh JA, Monterrosa A.Non-insulin-dependent diabetesmellitus in minorities in the United States. AnnIntern Med.1996;125(3):221–232 [DOI] [PubMed] [Google Scholar]
- 5.Harris MI, Flegal KM, Cowie CCet al. Prevalence of diabetes, impaired fastingglucose, and impaired glucose tolerance in U.S. adults: the Third NationalHealth and Nutrition Examination Survey, 1988-1994.Diabetes Care.1998;21(4):518–524 [DOI] [PubMed] [Google Scholar]
- 6.Robbins JM, Vaccarino V, Zhang H, Kasl SV.Socioeconomic status and diagnoseddiabetes incidence. Diabetes Res ClinPract.2005;68(3):230–236 [DOI] [PubMed] [Google Scholar]
- 7.The Diabetes Prevention Program ResearchGroupReduction in the incidence of type 2 diabeteswith lifestyle intervention or metformin. N Engl JMed.2002;346(6):393–403 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Tuomilehto J, Lindstrom J, Eriksson JGet al. Prevention of type 2 diabetes mellitus bychanges in lifestyle among subjects with impaired glucosetolerance. N Engl J Med.2001;344(18):1343–1350 [DOI] [PubMed] [Google Scholar]
- 9.Pan XR, Li GW, Hu YHet al. Effects of diet and exercise in preventingNIDDM in people with impaired glucose tolerance: the Da Qing IGT andDiabetes Study. Diabetes Care.1997;20(4):537–544 [DOI] [PubMed] [Google Scholar]
- 10.Garfield SA, Malozowski S, Chin MHet al. Considerations for diabetes translationalresearch in real-world settings. DiabetesCare.2003;26(9):2670–2674 [DOI] [PubMed] [Google Scholar]
- 11.Fisher EB, Walker EA, Bostrom A, Fischhoff B, Haire-Joshu D, Johnson SB.Behavioral science research in theprevention of diabetes: status and opportunities.Diabetes Care.2002;25(3):599–606 [DOI] [PubMed] [Google Scholar]
- 12.Baker MK, Simpson K, Lloyd B, Bauman AE, Singh MA.Behavioral strategies in diabetesprevention programs: a systematic review of randomized controlledtrials. Diabetes Res Clin Pract.2011;91(1):1–12 [DOI] [PubMed] [Google Scholar]
- 13.Almeida FA, Shetterly S, Smith-Ray RL, Estabrooks PA.Reach and effectiveness of a weightloss intervention in patients with prediabetes in Colorado.Prev Chronic Dis.2010;7(5):A103. [PMC free article] [PubMed] [Google Scholar]
- 14.Pagoto SL, Kantor L, Bodenlos JS, Gitkind M, Ma Y.Translating the diabetes preventionprogram into a hospital-based weight loss program.Health Psychol.2008;27(1,suppl):S91–S98 [DOI] [PubMed] [Google Scholar]
- 15.Whittemore R, Melkus G, Wagner J, Dziura J, Northrup V, Grey M.Translating the diabetes preventionprogram to primary care: a pilot study. NursRes.2009;58(1):2–12 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.McTigue KM, Conroy MB, Bigi L, Murphy C, McNeil M.Weight loss through living well:translating an effective lifestyle intervention into clinicalpractice. Diabetes Educ.2009;35(2):199–204,208 [DOI] [PubMed] [Google Scholar]
- 17.Absetz P, Valve R, Oldenburg Bet al. Type 2 diabetes prevention in the“real world”: one-year results of the GOAL ImplementationTrial. Diabetes Care.2007;30(10):2465–2470 [DOI] [PubMed] [Google Scholar]
- 18.McBride PE, Einerson JA, Grant Het al. Putting the Diabetes Prevention Programinto practice: a program for weight loss and cardiovascular risk reductionfor patients with metabolic syndrome or type 2 diabetesmellitus. J Nutr Health Aging.2008;12(10):745S–749S [DOI] [PubMed] [Google Scholar]
- 19.Vadheim LM, Brewer KA, Kassner DRet al. Effectiveness of a lifestyle interventionprogram among persons at high risk for cardiovascular disease and diabetesin a rural community. J Rural Health.2010;26(3):266–272 [DOI] [PubMed] [Google Scholar]
- 20.Kramer MK, Kriska AM, Venditti EMet al. Translating the Diabetes PreventionProgram: a comprehensive model for prevention training and programdelivery. Am J Prev Med.2009;37(6):505–511 [DOI] [PubMed] [Google Scholar]
- 21.Smith-Ray RL, Almeida FA, Bajaj Jet al. Translating efficacious behavioralprinciples for diabetes prevention into practice.Health Promot Pract.2009;10(1):58–66 [DOI] [PubMed] [Google Scholar]
- 22.Ackermann RT, Finch EA, Brizendine E, Zhou H, Marrero DG.Translating the Diabetes PreventionProgram into the community: the DEPLOY pilot study.Am J Prev Med.2008;35(4):357–363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Parikh P, Simon EP, Fei K, Looker H, Goytia C, Horowitz CR.Results of a pilot diabetesprevention intervention in East Harlem, New York City: ProjectHEED. Am J Public Health.2010;100(suppl1):S232–S239 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Boltri JM, Davis-Smith M, Okosun IS, Seale JP, Foster B.Translation of the NationalInstitutes of Health Diabetes Prevention Program in African Americanchurches. J Natl Med Assoc.2011;103(3):194–202 [DOI] [PubMed] [Google Scholar]
- 25.Merriam PA, Tellez TL, Rosal MCet al. Methodology of a diabetes preventiontranslational research project utilizing a community-academic partnershipfor implementation in an underserved Latino community.BMC Med Res Methodol.2009;9:20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Katula JA, Vitolins MZ, Rosenberger ELet al. One-year results of a community-basedtranslation of the Diabetes Prevention Program: Healthy-Living Partnershipsto Prevent Diabetes (HELP PD) Project. DiabetesCare.2011;34(7):1451–1457 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Santoyo-Olsson J, Cabrera J, Freyre Ret al. Aninnovative multiphased strategy to recruit underserved adults into arandomized trial of a community-based diabetes risk reductionprogram. Gerontologist.2011;51(suppl1):S82–S93 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Delgadillo AT, Grossman M, Santoyo-Olsson J, Gallegos-Jackson E, Kanaya AM, Stewart AL.Description of an academic communitypartnership lifestyle program for lower income minority adults at risk fordiabetes. Diabetes Educ.2010;36(4):640–650 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC.Homeostasis model assessment: insulinresistance and beta-cell function from fasting plasma glucose and insulinconcentrations in man. Diabetologia.1985;28(7):412–419 [DOI] [PubMed] [Google Scholar]
- 30.Huang MH, Harrison GG, Mohamed MMet al. Assessing the accuracy of a food frequencyquestionnaire for estimating usual intake of phytoestrogens.Nutr Cancer.2000;37(2):145–154 [DOI] [PubMed] [Google Scholar]
- 31.Huang MH, Schocken M, Block Get al. Variation in nutrient intakes by ethnicity:results from the Study of Women's Health Across the Nation(SWAN). Menopause.2002;9(5):309–319 [DOI] [PubMed] [Google Scholar]
- 32.Block G, Mandel R, Gold E.On food frequency questionnaires: thecontribution of open-ended questions and questions on ethnicfoods. Epidemiology.2004;15(2):216–221 [DOI] [PubMed] [Google Scholar]
- 33.Stewart AL, Mills KM, King AC, Haskell WL, Gillis D, Ritter PL.CHAMPS physical activityquestionnaire for older adults: outcome for interventions.Med Sci Sports Exerc.2001;33(7):1126–1141 [DOI] [PubMed] [Google Scholar]
- 34.Stewart AL, Ware JE, Jr, Sherbourne CD, Wells K.Psychological distress/well-being andcognitive functioning measures. :Stewart AL, Ware JE, Jr, MeasuringFunctioning and Well-Being: The Medical Outcomes Study Approach.Durham, NC: Duke UniversityPress;1992:102–142 [Google Scholar]
- 35.Hays RD, Stewart AL.Sleep measures.: Stewart AL, Ware JE, Jr, MeasuringFunctioning and Well-Being: The Medical Outcomes Study Approach.Durham, NC: Duke UniversityPress;1992:235–259 [Google Scholar]
- 36.Cohen S, Kamarck T, Mermelstein R.A global measure of perceivedstress. J Health Soc Behav.1983;24(4):385–396 [PubMed] [Google Scholar]
- 37.Little RJA, Rubin DB.Statistical Analysis With MissingData. 2nd ed.Hoboken, NJ: John Wiley andSons; 2002 [Google Scholar]
- 38.Hamman RF, Wing RR, Edelstein SLet al. Effect of weight loss with lifestyleintervention on risk of diabetes. DiabetesCare.2006;29(9):2102–2107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Lin JS, O'Connor E, Whitlock EP,Beil TL. Behavioral counseling to promote physical activity and a healthfuldiet to prevent cardiovascular disease in adults: a systematic review forthe U.S. Preventive Services Task Force. Ann Intern Med.−≠+≠;153(11):736–750. [DOI] [PubMed]
- 40.Idler EL, Benyamini Y.Self-rated health and mortality: areview of twenty-seven community studies. J HealthSoc Behav.1997;38(1):21–37 [PubMed] [Google Scholar]
- 41.Miilunpalo S, Vuori I, Oja P, Pasanen M, Urponen H.Self-rated health status as a healthmeasure: the predictive value of self-reported health status on the use ofphysician services and on mortality in the working-agepopulation. J Clin Epidemiol.1997;50(5):517–528 [DOI] [PubMed] [Google Scholar]
- 42.Ferraro KF, Kelley-Moore JA.Self-rated health and mortality amongblack and white adults: examining the dynamic evaluationthesis. J Gerontol B Psychol Sci Soc Sci.2001;56(4):S195–205 [DOI] [PubMed] [Google Scholar]
- 43.Kriska AM, Delahanty LM, Pettee KK.Lifestyle intervention for theprevention of type 2 diabetes: translation and futurerecommendations. Curr Diab Rep.2004;4(2):113–118 [DOI] [PubMed] [Google Scholar]
- 44.Task Force on Community PreventiveServicesIncreasing physical activity: a report onrecommendations of the Task Force on Community PreventiveServices. MMWR Recomm Rep.2001;50(RR-18):1–14 [PubMed] [Google Scholar]
- 45.Yancey AK, Ortega AN, Kumanyika SK.Effective recruitment and retentionof minority research participants. Annu Rev PublicHealth.2006;27:1–28 [DOI] [PubMed] [Google Scholar]
- 46.Glasgow RE, Toobert DJ, Hampson SE.Participation in outpatient diabeteseducation programs: how many patients take part and how representative arethey?Diabetes Educ.1991;17(5):376–380 [DOI] [PubMed] [Google Scholar]