Objectives. We tested the effectiveness of a community-based, literacy-sensitive, and culturally tailored lifestyle intervention on weight loss and diabetes risk reduction among low-income, Spanish-speaking Latinos at increased diabetes risk.

Methods. Three hundred twelve participants from Lawrence, Massachusetts, were randomly assigned to lifestyle intervention care (IC) or usual care (UC) between 2004 and 2007. The intervention was implemented by trained Spanish-speaking individuals from the community. Each participant was followed for 1 year.

Results. The participants’ mean age was 52 years; 59% had less than a high school education. The 1-year retention rate was 94%. Compared with the UC group, the IC group had a modest but significant weight reduction (−2.5 vs 0.63 lb; P = .04) and a clinically meaningful reduction in hemoglobin A1c (−0.10% vs −0.04%; P = .009). Likewise, insulin resistance improved significantly in the IC compared with the UC group. The IC group also had greater reductions in percentage of calories from total and saturated fat.

Conclusions. We developed an inexpensive, culturally sensitive diabetes prevention program that resulted in weight loss, improved HbA1c, and improved insulin resistance in a high-risk Latino population.

Type 2 diabetes is a serious disorder with many complications and is characterized by insulin resistance and relative insulin deficiency. The prevalence of diabetes in Latino Americans is 1.5 times that in non-Latino Whites.1 Between 1988 to 1994 and 2005 to 2006, the prevalence of diabetes increased from 9.6% to 12.6% in the adult Latino population.2,3 Prediabetes, as defined by impaired glucose tolerance or impaired fasting blood glucose, is often present 5 or more years before the development of type 2 diabetes.4 Several randomized clinical trials have shown that it is possible to prevent or delay the progression of the prediabetic state to clinical type 2 diabetes.5–7

The Diabetes Prevention Program (DPP) demonstrated that a lifestyle intervention incorporating dietary modification and increased physical activity produced an average weight loss of 5.6 kilograms at 1 year and by 4 years reduced the incidence of diabetes by 58% versus usual care.7 However, the intervention was intensive and costly, beginning with a 16-session curriculum that was delivered individually over 24 weeks and continuing with a number of follow-up individual and group sessions. The total intervention cost over the first year was $1399 per participant. The participants, although representing diverse American subpopulations, were generally well-educated, literate in English, and of average socioeconomic status.

The effectiveness of the DPP lifestyle intervention delivered in a lower cost, lower intensity format to high-risk populations is not known. We hypothesized that a community-based, culturally tailored, literacy-sensitive lifestyle intervention delivered in a primarily group-based format would facilitate weight loss and reduce the risk of type 2 diabetes among low-income Latinos at elevated risk of developing diabetes.

Between 2004 and 2007, the Lawrence Latino Diabetes Prevention Project (LLDPP) recruited 312 Latino participants (60% of Dominican origin and 40% Puerto Rican) from Lawrence, Massachusetts, who were at high risk for type 2 diabetes. The study methodology has been previously described.8

Study Setting

Lawrence, Massachusetts, is a primarily low-income, 60% Latino city. Latino fluency in English is approximately 49% and illiteracy in Spanish is approximately 30% (personal communication, International Institute of Greater Lawrence, 2004). The 2008 household income was $33 684 compared with $65 401 for Massachusetts overall. Diabetes prevalence is estimated at 11.8% among Lawrence Latino adults, compared with 6.4% among non-Hispanic Whites statewide.1

The implementation of the LLDPP involved a collaboration between the Greater Lawrence Family Health Center (GLFHC), the Lawrence Senior Center, the YWCA of Greater Lawrence (YWCA), and investigators from the Worcester and Lowell campuses of the University of Massachusetts. The GLFHC provides medical services to approximately 80% of the Lawrence Latino population. Most study activities (i.e., screening, recruitment, group intervention sessions, and follow-up assessments) were held at the Lawrence Senior Center, a centrally located social service facility. The Senior Center and the YWCA provided additional staff support and knowledge of and relationships with the community.


Of the participants, 78% were recruited from the GLFHC patient panel. Additional outreach methods included public service announcements on local radio and television stations, newspaper advertisements, and mailings to non-GLFHC physicians. Eligibility criteria included self-reported Latino/Hispanic ethnicity, age 25 years or older, body mass index (BMI; defined as weight in kilograms divided by the square of height in meters) greater than 24, and a 30% or greater likelihood of being diagnosed with diabetes over the succeeding 7.5 years (risk was calculated by using a validated predictive algorithm based on age, gender, ethnicity, fasting blood glucose, systolic blood pressure, high-density lipoprotein (HDL) cholesterol, BMI, and family history of diabetes).9 Exclusion criteria included the inability to walk 5 city blocks (one quarter mile), life-limiting medical conditions, and taking a medication or having a medical condition that interfered with the assessment of diabetes risk. Medical clearance from the participant's primary care provider was a requirement for study participation.

All time-varying screening measures included in the predictive formula were repeated at baseline to avoid regression to the mean. After completion of baseline data collection, individuals were randomly assigned to the lifestyle intervention care (IC) or usual care (UC) condition by use of a randomized block design.10 Participants from the same household were assigned to the same condition. Each participant received $25 cash incentives at the baseline and 6-month assessment visits and $50 at the 1-year assessment visit.


The LLDPP intervention included participation in 3 individual and 13 group sessions over a 12-month period. The duration of the first group session was 1.5 hours and the remaining group sessions were 1 hour. The first individual visit was 1 hour and the last 2 were 30 minutes each. Additional individual sessions were scheduled when the patients missed group sessions and were willing to schedule a makeup session. The number of sessions varied slightly depending on the start date of the group sessions in relation to the enrollment date. On 40 occasions (2.3% of the total), home visits replaced participation in a group session. The intervention goals included increasing intake of whole grains and nonstarchy vegetables and reducing sodium, total and saturated fat, portion sizes, and the intake of refined carbohydrates and starches. The physical activity goal was to increase walking by 4000 steps per day over baseline. Participants received a pedometer and instructions on its use and information on safe places for walking and exercise in the community.

We developed the intervention by using principles of social cognitive theory and patient-centered counseling,11,12 with emphasis on providing basic information on diabetes prevention, promoting positive attitudes toward dietary and physical activity change (i.e., confidence in ability to make changes or self-efficacy), and building skills for making dietary and physical activity changes (i.e., goal-setting, self-monitoring, problem-solving challenges, healthy cooking skills, and grocery shopping skills). The adaptation of the DPP intervention to this new population of Latinos at risk for diabetes involved the use of focus groups to identify knowledge gaps, attitudes toward diabetes prevention, and challenges to lifestyle change for weight loss. We then adapted the intervention to address the identified knowledge gaps (diabetes can be prevented or its onset delayed; weight loss requires a change in energy balance), attitudes toward diabetes prevention (“I can prevent developing diabetes”), and anticipated challenges to participation.

Individual sessions were conducted in the participants’ homes, and group-based sessions were conducted at a community site to which the participants had easy access (the Lawrence Senior Center). The intervention built on earlier research conducted by members of the research team. This work identified constructs to address to facilitate lifestyle change among Latinos with type 2 diabetes13 and served as the basis for intervention development.14,15 The intervention was tailored to the population by being culturally and literacy-sensitive. Cultural tailoring included dietary advice based on Latino foods, including the customization of Latino recipes; targeting cultural beliefs and attitudes toward diabetes prevention through a videotape novella (watching soap operas is a popular activity in this population); and delivery of the intervention in Spanish by bicultural and bilingual individuals from the community. Low literacy, which is prevalent in the population (even in Spanish), was addressed through visual adaptations of materials that simplified complex information and utilized hands-on experiences. A picture-based food guide that classified foods into 3 colors, green, yellow, and red, was used to assist participants in identifying the dietary quality of foods with regard to glycemic index, sodium, and saturated fat content. Participants used this food guide during a supermarket tour. Goal-setting and self-monitoring worksheets were designed to be simple and the information easy to record by individuals with little formal education. Other hands-on activities included demonstrations of healthy cooking methods, demonstration of portion sizes with real foods, and practice walking with pedometers during the sessions.

We modified the previously developed DPP intervention to be appropriate for Latino individuals at risk for type 2 diabetes. Focus groups16 were conducted to assess unique knowledge gaps regarding diabetes prevention, attitudes toward prevention, and challenges to weight loss in this population. Additional focus groups were used to pretest the acceptability of the intervention materials (e.g., soap opera, goal sheets). The intervention was modified to be less intensive (13 sessions instead of 20) and to include a flexible format to match the needs of this population (i.e., delivered in a group format with some of the sessions delivered in an individual format in the participant's home, as needed). Participants were encouraged to set realistic goals and to self-monitor their dietary intake and physical activity between sessions. At each session, participants weighed in, discussed their goal attainment, and discussed challenges and solutions to achieve their goals. A healthy meal was served at all sessions and preparation methods were discussed. Significant others were invited to attend each group session. Patients received session reminder calls and transportation.

A team of 3 Spanish-speaking community individuals with post–high school education delivered the intervention, and all had some previous undergraduate education in nutrition. None were registered dietitians. Intervention fidelity was facilitated through extensive training in the delivery of the intervention protocol, including the nutritional and exercise aspects of the intervention, the theoretical background, and training in motivational counseling and group management skills. The training included role-playing and mock intervention sessions and was led by a behavioral psychologist and a senior registered dietitian, who also provided ongoing supervision. Booster training sessions were scheduled semiannually.

Outcome Measures

Primary outcomes were weight loss and hemoglobin A1c (HbA1c), both of which were measured at baseline and at the end of year 1. To determine BMI, body weight and height were measured by using standardized procedures with the patients wearing only light clothing and no shoes. HbA1c was measured by using the Primus Diagnostics boronate affinity high-performance liquid chromatography method at Dr. Randie Little's Diabetes Diagnostic Laboratory, a national reference laboratory for HbA1c, at the University of Missouri, Columbia, Missouri.

Secondary outcomes at baseline and at 1 year included fasting lipids, glucose, and insulin concentrations; blood pressure; dietary assessment; physical activity measurements; and quality of life and depression scores. The fasting blood glucose and insulin measures were used to estimate insulin resistance by using the homeostasis model assessment (HOMA-IR) calculation.17 The study, with 312 particpants and 1 year of follow-up, was not powered for clinical diabetes outcome.

Serum glucose, total and HDL cholesterol, and triglyceride were analyzed by using the Cobas Mira Autoanalyzer (Roche Diagnostics, Indianapolis, IN). HDL cholesterol was measured in the supernatant after magnesium-phosphotungstate precipitation of apo B–containing lipoproteins. LDL cholesterol was calculated according to the Friedewald formula.18 Insulin concentration was measured by the Vanderbilt Diabetes Center Hormone Assay & Analytical Services Core (Vanderbilt University Medical Center, Nashville, TN) by use of a double-antibody radioimmunoassay. Study personnel blinded to study condition measured blood pressure by using the mean of 2 measures taken with a Dinamap XL automated blood pressure monitor (GE Medical Systems Information Technologies, Inc., Milwaukee, WI).

Diet was assessed by computer-assisted 24-hour dietary recall using the University of Minnesota's Nutrition Coordinating Center's Nutrition Data System for Research software (NDSR-2007; Nutrition Coordinating Center, Minneapolis, MN). At each time point, three 24-hour recalls were conducted on randomly selected days within a 3-week period (2 weekdays and 1 weekend day). These recalls were conducted by trained Spanish-speaking registered dietitians who were not involved in the intervention and who were blinded to the participant's condition. Participants were provided food portion images to facilitate portion size estimation. The data derived from the 24-hour dietary recall (e.g., total calories, carbohydrates, total fat, saturated fat, protein, and dietary fiber) were generated by using the same software.

Physical activity was assessed at baseline and at 1 year as part of the 24-hour dietary recall assessment. The physical activity recall has been validated against both accelerometers and standard questionnaires19 and is linked with the compendium of physical activities of Ainsworth et al. for energy expenditure calculation.20

Depressive symptoms were assessed by using the Center for Epidemiological Studies-Depression Scale (CES-D),21 which was validated for community samples22 and for use in Spanish-speaking populations.23 Quality of life was assessed by use of the Short Form-12 (SF-12).24

All questionnaire assessments were printed in both Spanish and English and were administered orally in Spanish. To estimate the cost of the IC, we recorded all direct expenses per participant. We used actual costs tracked throughout the study, except for the time spent by interventionists for individual sessions and phone calls, which were estimated by extrapolation of logs covering 2- to 4-week periods during the study.

Statistical Analyses

We compared the baseline characteristics of the participants in the 2 intervention arms by using t-tests for normally distributed measures, rank tests for measures with skewed distributions, and Fisher's exact statistic for categorical measures. We used the same statistics to compare those who dropped out with those who completed the study and to compare dropouts between the randomly assigned groups. We tested changes between 1 year and baseline in the randomly assigned groups by using regression analysis with standard errors adjusted for clustering on household. In outcomes with substantial kurtosis, we used ranks for the regression analyses. For the changes in the primary and secondary outcomes at 1 year from baseline, we used estimated medians and median differences.25 We estimated the association of changes in primary outcomes and changes in behavioral measures and with group and individual sessions by using linear regression analysis. The association of weight change and session attendance included a linear spline with a single knot. A Sobel-Goodman test of mediation was carried out to examine factors that might be mediators of the change in HbA1c.26 All analyses were carried out by using those who completed (94%). Analyses were repeated by using a return to baseline imputation for dropouts. Changes in weight and HbA1c were modeled over chronologic time of the study to examine whether the intervention effect differed during the study by testing the interaction of intervention and month of the study.

A total of 312 participants were recruited into the study and were randomly assigned to the control (UC; n = 150) or intervention (IC; n = 162) condition. Eight households had multiple participants: 2 UC and 6 IC. The flow of recruitment and participation is shown in Figure 1. The 1-year study completion rate was 94%; 11 participants (6.8%) dropped out of the IC group and 7 (4.7%) dropped out of the UC group (P = .25). We excluded 5 additional participants from the analysis after completion of the study: 2 underwent gastric bypass surgery and 3 had incomplete data. Those who dropped out tended to be male (35% vs 25% among those who did not drop out; P = .32), to have lower systolic blood pressure (mean 124 vs 129 mmHg, P = .08), to have higher reported caloric intake (1852 vs 1515 kcal/day; P = .01), and to have higher HOMA-IR (6.57 vs 5.13; P = .15).

The mean age of the participants was 52 years (range = 25–79 years): 74% were female, 59% did not complete high school, and 46% were employed. The baseline characteristics of the 2 randomly assigned groups are compared in Table 1. The groups had comparable probabilities of developing diabetes over the following 7.5 years (UC = 0.58; IC = 0.55) and similar body weight (UC = 191.2 lb; IC = 190.2 lb) and HbA1c levels (UC = 5.77%; IC = 5.76%). Although the study was not powered for clinical diabetes outcome, of the participants who did not have diabetes at baseline (HbA1c < 6.5), 5 developed diabetes (3.7%) by 1 year in the UC group versus 2 in the IC group (1.4%; P = .32). There were no significant differences in physical activity or fasting blood glucose between the 2 randomly assigned groups.


TABLE 1— Selected Baseline Characteristics of the Participants in the Control and Lifestyle Intervention Groups: Lawrence Latino Diabetes Prevention Project, 2004–2007

TABLE 1— Selected Baseline Characteristics of the Participants in the Control and Lifestyle Intervention Groups: Lawrence Latino Diabetes Prevention Project, 2004–2007

VariableControl (n = 150)Intervention (n = 162)Pa
Age, y, mean ±SD52.37 ±11.651.37 ±10.9.43
Female, no. (%)115 (76.67)117 (72.22).44
< high school education, no. (%)85 (57.05)97 (60.62).57
Married or living together, no. (%)73 (50.00)86 (53.4).37
Employed, no. (%)70 (46.67)73 (45.34).82
Smoked in past 3 mo, no. (%)14 (9.66)21 (13.04).37
Family history of diabetes, no. (%)95 (63.33)102 (62.96)> .99
Probability of diabetes in 7.5 y, %, mean ±SD 0.58 ±0.20.55 ±0.2.19
Body mass index, kg/m2, mean ±SD34.18 ±5.933.57 ±5.1.33
HbA1c, %, mean ±SD5.77 ±0.45.76 ±0.3.91
Weight, lb, mean ±SD191.16 ±36.3190.19 ±31.9.8
Waist circumference, cm, mean ±SD104.43 ±12.9104.31 ±10.4.93
Fasting blood glucose, mg/dl, mean ±SD105.61 ±12.3104.41 ±11.9.38
Insulin, μU/ml, mean ±SD19.90 ±13.820.10 ±13.5.9
HOMA-IR, mean ±SD5.21 ±3.85.24 ±3.8.95
Dietary intake, mean ±SD
 Energy, kcal/d1531.56 ±593.71546.78 ±604.9.82
 Energy from fat, %25.82 ±6.426.49 ±6.0.35
 Energy from saturated fat, %8.17 ±2.78.50 ±2.6.27
 Energy from carbohydrate, %55.92 ±8.555.36 ±7.8.55
 Energy from protein, %17.49 ±4.717.59 ±5.8.87
 Total fiber, g/d15.71 ±7.015.74 ±8.2.97
Leisure-time physical activity, min/wk, mean ±SD251.08 ±158.4247.50 ±164.1.85
CES-D score
 Mean ±SD15.20 ±10.316.37 ±12.3.37
 > 16, no. (%)60 (40.54)72 (44.72).49

Note. CES-D = Center for Epidemiological Studies-Depression Scale; HbA1c = hemoglobin A1c; HOMA-IR = homeostasis model assessment of insulin resistance.

aComparisons were by t-test for means and Fisher's exact test for proportions.

The median changes from baseline to 1 year for the primary and selected secondary outcomes are shown in Table 2. The IC participants lost significantly more weight than did the UC participants (−2.5 lb; P = .004). This was associated with a significant reduction in HbA1c (−0.10%, P = .009). A repeat HbA1c analysis including dropouts and imputing a return to baseline resulted in minimal change and remained significantly different between the 2 conditions (−0.09% in the IC vs −0.04% in the UC; P = .02). Similar analyses for the other endpoints of interest also produced no meaningful changes. A significant correlation was present between weight loss and HbA1C change (r = 0.41; P < .001). Of the IC participants 16% had an increase in HbA1c versus 32% of the UC participants (P = .008). Insulin resistance also significantly improved in the IC compared with the UC group (median HOMA-IR = −0.36 in the IC and −0.06 in the UC; P = .03), which correlated with weight change (r = 0.32; P < .001). The IC participants had significantly greater reductions in percentage of dietary calories from fat (−2.02% vs −0.24%; P = .04), and there was a trend for a greater reduction in percentage of calories from saturated fat (−0.65% vs −0.43%; P = .08) and an increase in dietary fiber intake (3.13 vs 1.98 g/day; P = .07).


TABLE 2— Changes in Primary and Secondary Outcomes at 1 Year From Baseline in the Control and Lifestyle Intervention Groups: Lawrence Latino Diabetes Prevention Project, 2004–2007

TABLE 2— Changes in Primary and Secondary Outcomes at 1 Year From Baseline in the Control and Lifestyle Intervention Groups: Lawrence Latino Diabetes Prevention Project, 2004–2007

Change From Baseline
Control, Median (95% CI) or %Intervention, Median (95% CI) or %Intervention Effect (95% CI)Pa
Weight, lb0.63 (−1.05, 2.00)−2.5 (−4.0, −1.5)−2.5 (−4.25, −0.75).004
Body mass index, kg/m20.11 (−0.22, 0.38)−0.40 (−0.76, −0.25)−0.46 (−0.76, −0.14).004
Mean change (Δ) in HbA1c−0.04 (−0.08, −0.002)−0.10 (−0.15, −0.06)−0.07 (−0.10, −0.04).009
 ΔHbA1c ≤ −0.15364.008
 ΔHbA1c = 01520
 ΔHbA1c ≥ 0.13216
HOMA-IR0.06 (−0.57, 0.38)−0.36 (−0.64, −0.09)−0.28 (−0.76, 0.20).03
Fasting blood glucose, mg/dl−1.5 (−3.0, 2.2)0.5 (−0.94, 2.88)1.0 (−2.0, 3.5).62
Insulin, μU/ml−0.66 (−2.50, 1.50)−1.68 (−2.50, −0.80)−1.25 (−3.01, 0.57).16
Dietary intake
 Energy, kcal/d3.8 (−57.3, 70.2)−21.8 (−103.6,55.3)−30.1 (−141.2, 76.9).57
 Energy from fat, %−0.42 (−1.38, 1.57)−2.02 (−3.77, −0.29)−1.77 (−3.48, −0.08).04
 Energy from saturated fat, %−0.43 (−0.75, 0.36)−0.65 (−1.03, −0.27)−0.59 (−1.28, 0.07).08
 Energy from carbohydrate, %0.41 (−0.94, 2.14)1.20 (−0.18, 3.54)1.73 (−0.23, 3.76).08
 Energy from protein, %−0.11 (−0.79, 0.88)0.61 (−0.62, 1.60)0.02 (−1.15, −1.22).97
 Total fiber, g/d0.48 (−2.10, 2.12)3.13 (0.88, 4.46)1.98 (−0.16, 4.01).07
Leisure-time physical activity, min/wk3.3 (−20.7, 18.3)5.8 (−12.8, 21.7)3.33 (−26.7, 33.3).82
CES-D score−1.0 (−2.0, 1.0)−1.0 (−3.0,1.0)0 (−2.0, 2.0).98

Note. CES-D = Center for Epidemiological Studies-Depression Scale; CI = confidence interval; HbA1c = hemoglobin A1c; HOMA-IR = homeostasis model assessment of insulin resistance.

aP values were based on rank test.

Both conditions showed significant reductions in depressive symptomatology (CES-D score) compared with baseline, with similar changes between the groups; there was no significant difference in the quality of life (SF-12) score. No serious adverse events were reported in either condition during the study.

The participants attended a median of 6 group sessions and 8 total sessions (group and individual). Attendance at group sessions was low, dropping from 60% at the first session to 20% at the last session. Weight change was significantly associated with attendance (r = −0.37, P < .001). A model of the association of weight change and total sessions attended found a threshold of greater than 7 sessions out of 16 for an intervention effect on weight loss (Figure 2). There was a median weight loss of 0.13 pounds with attendance at 7 or fewer sessions, but a median weight loss of 4.75 pounds with attendance at more than 7 sessions (P = .02). The cost of the intervention per participant was estimated to be $661.

We analyzed feedback from 77 participants in the IC. The major barriers for dietary change included stress (34%), lack of willpower (55%), home healthy food environment (30%), and knowledge (18%); major barriers for exercise change included time constraints (48%), bad weather (38%), physical illness or disability (29%), fatigue (27%), and lack of motivation (29%).

A Sobel-Goodman test of mediation examining factors that may have been mediators of the change in HbA1c showed only change in weight as a mediating factor, explaining 27.8% of the effect (P = .1). No dietary factors were associated with HbA1c change in this analysis. It is possible that dietary data were not sufficiently accurate in this population to contribute meaningfully to this type of analysis.

The lifestyle intervention tested in this study led to a modest but significant and clinically meaningful weight loss among the IC compared with the UC participants. Of particular importance was the improvement in HbA1c and insulin resistance observed among the IC participants at 1 year. The HbA1c change was equal to that seen at 1 year in the DPP, despite the much larger weight loss in the DPP at the 1year time point (LLDPP weight loss: −2.5 lb; DPP weight loss: −12.3 lb).7 These findings suggest that the effect of weight loss on improvement in HbA1c and HOMA-IR may be greater in Caribbean Latinos than in other populations. Population sensitivity to the development of an insulin-resistant state or diabetes with lesser amounts of weight gain is well described,27 but less is known about corresponding sensitivity to weight loss. If confirmed in other studies, sensitivity to modest weight loss in high-risk groups such as Caribbean Latinos has important clinical and public health implications. It will also be important to explore possible genetic underpinnings for such population sensitivity.

The DPP was a proof-of-concept efficacy study7 that used a screening methodology (an oral-glucose-tolerance test) not commonly used in clinical general practice and that used an intensive and relatively expensive intervention. The LLDPP used a simple screening methodology and a less intensive intervention tailored to the needs of our low-literacy, low–socioeconomic status Latino participants. The first-year costs per participant were $661 for the LLDPP versus $1399 for the DPP.28

There were several limitations to our study. First, we did not achieve any meaningful improvement in physical activity. This may have been related to a lack of appropriate emphasis on this facet of the intervention or may have been related to the characteristics of a less educated, low–socioeconomic status population living in neighborhoods not conducive to physical activity. Second, the intervention did not produce significant changes in fasting blood glucose. However, previous reports have noted that weight loss in a prediabetic population may have a greater effect on postprandial glucose, HbA1c, and HOMA-IR than on fasting blood glucose.29 Third, the relatively short-term follow-up (1 year) precluded clinical endpoint assessment of diabetes incidence as a primary outcome. We also could not evaluate long-term maintenance. Fourth, owing to measurement challenges (e.g., the lack of validated measures appropriate for this low-literate Spanish-speaking population and the time demands of oral administration), we were unable to measure knowledge, attitudes, and expectations. Fifth, attendance at the group sessions was low. Low session attendance to lifestyle interventions has also been found in previous studies of low socioeconomic status Latino populations.30 Finally, a greater reduction in diabetes incidence was observed in older individuals in the original DPP; however, our sample size of 312 was not sufficient to look for age effects.

A strength of this study was the participatory nature of the collaboration between the university team and the partnering community organizations. The community presence of the study contributed to the design of effective and efficient recruitment protocols and resulted in a very low dropout rate (94% of participants completed the 1-year follow-up). In addition, the input of the community ensured the acceptability of the study methods to potential participants.

In conclusion, this tailored intervention program developed and tested for a target population of low-income Caribbean Latinos at elevated risk for diabetes produced a modest but significant degree of weight loss associated with significant improvement in insulin resistance and HbA1c. Implementation of this intervention at similar community settings and populations has the potential to bring about important public health benefits.


This research was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) (grant R18 DK067549-01 to I. S. O.). M. C. R. was partly supported by the NIDDK (grant 1 R18 DK065985-04), and Y. M. was partly supported by the National Heart, Lung, and Blood Institute (NHLBI) (grant 1R01HL094575-01A1). The research was also supported in part by the NIDDK (center grant 5 P30 DK32520). I. S. Ockene, M. C. Rosal, J. Mordes, and Y. Ma are members of the UMass Diabetes and Endocrinology Research Center (DERC) (DK32520).

The work was presented in preliminary form at the American Heart Association meeting in 2009.

We thank the people of Lawrence, MA; the Lawrence Mayor's Health Task Force; and our community partners: the Lawrence Council on Aging, the YWCA of Greater Lawrence, and the Greater Lawrence Family Health Center.

Note. The contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIDDK or the NHLBI.

Human Participant Protection

The study protocol was approved by the institutional review boards of the University of Massachusetts Medical School and the Greater Lawrence Family Health Center (Trial Registration: NCT00810290).


1. CDC. Total prevalence of diabetes in the United States, all ages, 2002. Available at: http://www.cdc.gov/diabetes/pubs/estimates.htm#prev. Accessed October 13, 2011. Google Scholar
2. Cowie CC, Rust KF, Ford ES, et al. Full accounting of diabetes and pre-diabetes in the U.S. population in 1988-1994 and 2005-2006. Diabetes Care. 2009;32(2):287294. Crossref, MedlineGoogle Scholar
3. Banks J, Marmot M, Oldfield Z, Smith JP. Disease and disadvantage in the United States and in England. JAMA. 2006;295(17):20372045. Crossref, MedlineGoogle Scholar
4. Unwin N, Shaw J, Zimmet P, Alberti KG. Impaired glucose tolerance and impaired fasting glycaemia: the current status on definition and intervention. Diabet Med. 2002;19(9):708723. Crossref, MedlineGoogle Scholar
5. Tuomilehto J, Lindstrom J, Eriksson JG, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med. 2001;344(18):13431350. Crossref, MedlineGoogle Scholar
6. Pan XR, Li GW, Hu YH, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care. 1997;20(4):537544. Crossref, MedlineGoogle Scholar
7. Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346(6):393403. Crossref, MedlineGoogle Scholar
8. Merriam PA, Tellez TL, Rosal MC, et al. Methodology of a diabetes prevention translational research project utilizing a community-academic partnership for implementation in an underserved Latino community. BMC Med Res Methodol. 2009;9:20. Crossref, MedlineGoogle Scholar
9. Stern MP, Williams K, Haffner SM. Identification of persons at high risk for type 2 diabetes mellitus: do we need the oral glucose tolerance test?Ann Intern Med. 2002;136(8):575581. Crossref, MedlineGoogle Scholar
10. Ryan P. sxd1.1: Update to random allocation of treatment to blocks. Stata Tech Bull. 1999;50:3637. Google Scholar
11. Bandura A. Self-efficacy: The exercise of control. New York, NY: W.H. Freedman and Company; 1997. Google Scholar
12. Rosal MC, Ebbeling CB, Lofgren I, Ockene JK, Ockene IS, Hebert JR. Facilitating dietary change: the patient-centered counseling model. J Am Diet Assoc. 2001;101(3):332341. Crossref, MedlineGoogle Scholar
13. Rosal MC, White MJ, Restrepo A, et al. Design and methods for a randomized clinical trial of a diabetes self-management intervention for low-income Latinos: Latinos en Control. BMC Med Res Methodol. 2009;9:81. Crossref, MedlineGoogle Scholar
14. Rosal MC, Olendzki B, Reed GW, Gumieniak O, Scavron J, Ockene IS. Diabetes self-management among low-income Spanish speaking patients: a pilot study. Ann Behav Med. 2005;29(3):225235. Crossref, MedlineGoogle Scholar
15. Rosal MC, Ockene IS, Restrepo A, et al. Randomized trial of a literacy-sensitive, culturally-tailored diabetes self-management intervention for low-income Latinos: Latinos en Control. Diabetes Care. 2011;34(4):838844. Crossref, MedlineGoogle Scholar
16. Rosal MC, Borg A, Bodenlos J, Tellez T, Ockene I. Awareness of diabetes risk factors and diabetes prevention strategies among a sample of low-income Latinos with no known diagnosis of diabetes. Diabetes Educ. 2011;37(1):4755. Crossref, MedlineGoogle Scholar
17. Katsuki A, Sumida Y, Gabazza EC, et al. Homeostasis model assessment is a reliable indicator of insulin resistance during follow-up of patients with type 2 diabetes. Diabetes Care. 2001;24(2):362365. Crossref, MedlineGoogle Scholar
18. Friedewald WT, Levy RI, Frederickson DS. Estimation of plasma low-density lipoprotein cholesterol concentration without use of the preparative ultracentrifuge. Clin Chem. 1972;18(6):499502. MedlineGoogle Scholar
19. Matthews CE, Freedson P, Hebert J, Stanek E, Ockene I, Merriam P. Comparison of physical activity assessment methods in the Seasonal Variation of Blood Cholesterol Levels Study. Med Sci Sports Exerc. 2000;32(5):976984. Crossref, MedlineGoogle Scholar
20. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;32(suppl 9):S498S504. Crossref, MedlineGoogle Scholar
21. Devins G, Orme C. Center for Epidemiologic Studies Depression Scale. In: Sweetland R, ed. Test Critiques. Kansas City, MO: Test Corp of America, a subsidiary of Westport Publishers, Inc; 1985:144160. Google Scholar
22. Haringsma R, Engels GI, Beekman AT, Spinhoven P. The criterion validity of the Center for Epidemiological Studies Depression Scale (CES-D) in a sample of self-referred elders with depressive symptomatology. Int J Geriatr Psychiatry. 2004;19(6):558563. Crossref, MedlineGoogle Scholar
23. Salgado dSN, Maldonado M. Psychometric characteristics of the Center for Epidemiologic Studies Depression Scale in adult Mexican women from rural areas. Salud Publica Mex. 1993;36:200209. Google Scholar
24. Jenkinson C, Layte R. Development and testing of the UK SF-12 (short form health survey). J Health Serv Res Policy. 1997;2(1):1418. Crossref, MedlineGoogle Scholar
25. Newson R. Parameters behind “nonparametric” statistics: Kendall's tau, Somers’ D and median differences. Stata Journal. 2002;2:4564. Google Scholar
26. MacKinnon D. Introduction to Statistical Mediation Analysis. New York, NY: Psychology Press, Taylor and Francis Group; 2008. Google Scholar
27. Shai I, Jiang R, Manson JE, et al. Ethnicity, obesity, and risk of type 2 diabetes in women: a 20-year follow-up study. Diabetes Care. 2006;29(7):15851590. Crossref, MedlineGoogle Scholar
28. Hernan WH, Brandle M, Zhang P, et al. Costs associated with the primary prevention of type 2 diabetes mellitus in the Diabetes Prevention Program. Diabetes Care. 2003;26(1):3647. Crossref, MedlineGoogle Scholar
29. Perreault L, Ma Y, Dagogo-Jack S, et al. Sex differences in diabetes risk and the effect of intensive lifestyle modification in the Diabetes Prevention Program. Diabetes Care. 2008;31(7):14161421. Crossref, MedlineGoogle Scholar
30. Feathers JT, Kieffer EC, Palmisano G, et al. The development, implementation, and process evaluation of the REACH Detroit Partnership's Diabetes Lifestyle Intervention. Diabetes Educ. 2007;33(3):509520. Crossref, MedlineGoogle Scholar


No related items




Ira S. Ockene, MD, Trinidad L. Tellez, MD, Milagros C. Rosal, PhD, George W. Reed, PhD, John Mordes, MD, Philip A. Merriam, MSPH, Barbara C. Olendzki, MPH, RD, Garry Handelman, PhD, Robert Nicolosi, PhD, and Yunsheng Ma, MD, PhDIra S. Ockene is with the Division of Cardiovascular Medicine, Department of Medicine, University of Massachusetts Medical School, Worcester, MA. At the time of the study, Trinidad L. Tellez was with the Greater Lawrence Family Health Center, Lawrence, MA. Milagros C. Rosal, George W. Reed, Philip A. Merriam, Barbara C. Olendzki, and Yunsheng Ma are with the Division of Preventive and Behavioral Medicine, Department of Medicine, University of Massachusetts Medical School. John Mordes is with the Division of Diabetes, Department of Medicine, University of Massachusetts Medical School. Garry Handelman and Robert Nicolosi are with the Department of Health and Clinical Sciences, University of Massachusetts, Lowell. “Outcomes of a Latino Community-Based Intervention for the Prevention of Diabetes: The Lawrence Latino Diabetes Prevention Project”, American Journal of Public Health 102, no. 2 (February 1, 2012): pp. 336-342.


PMID: 22390448