• Open Access

People United to Sustain Health (PUSH): A Community-Based Participatory Research Study



The prevention of weight gain to address the obesity epidemic rather than weight loss involves promoting small changes in food choices and physical activity. People United to Sustain Health (PUSH) was designed to increase fruit and vegetable consumption, physical activity, and food security to prevent weight gain in rural adults. Forty-nine participants were randomized into a treatment group which received access to a “Rolling Store,” nutrition education and physical activity, and a control group which received family coping classes. Forty-one (84%) of participants completed the study. At the end of 6 months, weight for all participants was maintained from baseline to completion with no significant differences between the groups. The mean fruit consumption over 6 months for the treatment group increased and was significantly greater than change in the control group (p = 0.01). This community-based participatory research study was considered successful because weight gain was prevented.


Residents in the Lower Mississippi Delta (LMD) suffer disproportionately from chronic illnesses such as obesity, heart disease, diabetes, and high blood pressure.[1-3] Collaboratively, six academic/research partners in Arkansas, Louisiana, and Mississippi, a local community and the cooperative extension service in each state examined the nutritional health problems in the LMD region and identified opportunities for conducting research interventions.[4] The LMD is predominantly rural, some areas with a higher concentration of African Americans, higher rates of poverty, and lower educational attainment.[1] Louisiana has the eighth highest prevalence of adult obesity in the nation, at 28.9% and the seventh highest of overweight youths (ages 10–17) at 35.9%.[5] In the LMD area of Franklin Parish and specifically the town of Winnsboro, Louisiana, the estimated median annual household income was approximately $19,918. These incomes fall within Louisiana's poverty rate of 19.2%, the second highest rate in the nation, and the highest in the South.[6]

Intervening with traditional research methods in the LMD may be problematic due to the rural setting, low literacy levels, lack of health services, difficulties in retaining minorities once recruited, and lack of local qualified researchers to monitor intervention activities.[7-9] Consequently, a community-based participatory research (CBPR) model that links community residents with academic/research partners to collaboratively identify and prioritize health issues, and work together to develop and implement intervention strategies, may prove beneficial in LMD communities. CBPR is defined as “a collaborative process that equitably involves all partners in the research process and recognizes the unique strengths that each offers. CBPR begins with a research topic of importance to the community with the aim of combining knowledge and action for social change to improve community health and eliminate health disparities.”[10-14]

Environmental factors such as the availability of healthy food choices and physical activity opportunities influence behaviors; and attempts to improve these behaviors may be implemented in various settings, such as the church, home, various community facilities, and worksites.[15, 16] The church is potentially an effective setting because it typically emphasizes health promotion, social support, and offers convenience in program participation and dissemination.[17-19] In this paper, results from 6 months of the People United to Sustain Health (PUSH) study are reported. The purpose of the PUSH study was to utilize a “Rolling Store” delivery medium plus nutrition education to increase fruit and vegetable consumption, physical activity, and food security to prevent weight gain in adults residing in Franklin Parish, Louisiana. It was hypothesized that the provision of fresh fruits and vegetables along with education relative to health benefits, recipes, and healthy cooking demonstrations, may lead to a direct improvement in diet quality and an adoption of healthy food choices that result in the prevention of weight gain.


Engaging the Franklin Parish community

Franklin Parish community residents were involved from inception utilizing the principles CBPR is built upon,[20] including assessing the health problems in their community, identifying and planning an intervention, hiring and training, recruitment and retention, assisting with data collection and submission of data for analyses. Community residents along with key community stakeholders such as pastors, mayors, councilmen, councilwomen, hospital staff, school administrators and teachers, council on aging, entrepreneurs, etc., and academic/research partners met biweekly at various community venues for approximately 1 year. During this time, community residents elected two chairpersons to chair and co-chair community and research meetings. The purpose of these meetings and discussions was to help identify, develop, and implement a specific intervention that would be most beneficial in improving the health of residents within Franklin Parish.

Several ideas for interventions were generated from community residents including establishing community gardens, walking trail interventions, heart disease and diabetes interventions, and working with seniors to prevent Alzheimer's disease, etc. During the discussion process, community residents expressed a desire to learn about research previously conducted by the academic/research partners. Therefore, research interventions previously conducted in other communities were presented to key community stakeholders and residents of Franklin Parish. For example, a 6-month church-based weight loss study was presented and the findings from the study revealed that a church setting is conducive for implementing a health and nutrition program, church members may be trained to conduct the intervention, and both were effective in inducing weight loss.[16] A second 6-month study was presented using trained community residents as peer educators to teach nutrition education and physical activity classes, plus a “Rolling Store” to provide fruits and vegetables to prevent weight gain in African American women. The “Rolling Store” model along with nutrition education and physical activity classes taught by peer educators produced weight loss, was feasible, accessible, and economical in producing satisfactory health and behavioral outcomes.[21] As a result, Franklin Parish community leaders and residents decided to combine both of the previous examples, tailoring them to meet their community needs.

In an effort to determine if PUSH would be achievable and produce the desired results, an 8-week feasibility phase of the study was implemented. After debriefing the peer educator, “Rolling Store” operator, and 10 of 11 participants completing the feasibility phase, it was determined by key community stakeholders, community residents, and the academic/research partners that a larger PUSH study was attainable.


Study participants were recruited according to detailed inclusion and exclusion criteria. Word of mouth was the primary method utilized to recruit participants; flyers and newspaper ads were secondary methods. Only one adult per household was eligible to participate in the study. In addition, to be eligible for PUSH, potential participants were adult men and women, residents of Franklin Parish, age 18 years or older, body mass index (BMI) 23–45 kg/m2, willing to participate and able to provide informed consent, and willing to make weekly or monthly visits to the study site for 24 months. However, due to the untimely death of one of the peer educators, and retention of participants as a result of the death, only the results for the first 6 months of PUSH are presented. Participants unwilling or unable to provide informed consent, unable or unwilling to regularly participate for the entire study period, previously participated in the feasibility phase, unable to communicate with the study staff, currently participating in a dietary and/or lifestyle modification program, likely to move away from the Parish or women definitely planning to become pregnant over the course of the study, were ineligible to enroll in PUSH.

During this time, the peer educators and “Rolling Store” operator were identified and trained by the research investigators. Two peer educators with master's degrees along with the community chairperson and community liaisons received 2.5 days of specific and intensive training in the study protocol, motivational interviewing techniques, behavioral modification techniques, and basic dietary assessment. The “Rolling Store” operator received 1 day of training in the study protocol. Upon agreeing to participate in the study, eligible participants were given a consent form describing details of the study and were scheduled to visit the study site for screening. Written informed consent was obtained at this visit. The study protocol, procedures, and consent form were reviewed and approved by the research partners’ Institutional Review Boards at the Pennington Biomedical Research Center (PBRC), and Southern University and A&M College. All participants received the provision of fresh fruits and vegetables for 6 months, recipes, and pedometers to encourage physical activity as incentives.


All study measurements were conducted by trained technicians. At baseline, participants were asked to provide demographic information such as age, gender, marital status, smoking history, alcohol consumption, and whether or not they had high blood pressure and diabetes, or a family history of both. Dietary assessment was conducted by telephone interviews through the Southern University Survey Research Center using a Food Frequency Questionnaire (FFQ), a modification of the Block questionnaire[22] at baseline and at the end of 6 months. Participants were given a copy of the FFQ at baseline and end of 6 months to take home and prepare to report all foods eaten during the previous 24 hours when called by the interviewers. Food models and measuring guides were used to estimate portion size, and appropriate probing questions were also used. Study participants were asked to respond to a physical activity readiness questionnaire (PAR-Q)[23] for safety reasons before beginning physical activity. The questionnaire contained seven questions and a response of “no” to each question was required. If participants responded “yes” to one or more questions, they were deemed ineligible and/or a physician statement was necessary before participating in physical activity.

Anthropometric assessments included height, waist circumference, and weight. Height was measured at baseline only without shoes to the nearest centimeter using a stadiometer. Waist circumference was determined in centimeters using a nonstretchable flexible measuring tape. Weight was measured in kilograms each month for 6 months for all participants. BMI was calculated as the weight in kilograms divided by height in meters squared (kg/m2). Blood pressure was measured in two replicates using the OMRON HEM-907XL (OMRON Healthcare, Inc., Vernon Hills, IL, USA). Each measurement was made after the participant rested for 10 minutes prior to the first blood pressure measurement, and an additional 5 minutes prior to the second blood pressure measurement. The average of the two measurements was used in the analysis. Fasting blood samples were obtained for total cholesterol (TC; mg/dL), high-density lipoprotein cholesterol (HDL; mg/dL), low-density lipoprotein cholesterol (LDL; mg/dL), triglycerides (TRG; mg/dL), glucose (GLU; mg/dL), and Hemoglobin A1c levels via the finger stick collection procedure using the Cholestech LDX (Cholestech Corporation, Hayward, CA, USA).

Participants were encouraged to use their pedometers at the onset of PUSH to begin walking 15 minutes daily and to gradually increase over the course of the study to at least 30 minutes per day. Walking and other physical activity were self-reported in a daily diary but was not recorded and analyzed. Household food security status was evaluated using the 18-question US Food Security Survey Module[24] to construct the 6-month food security scale that classifies households as food-secure or food-insecure with or without hunger.

Study design and intervention

The study was conducted in Franklin Parish specifically within an African American Baptist church located in Winnsboro, Louisiana. The Pastor and members of the congregation approved the use of the church for PUSH activities, including classroom space, kitchen and banquet areas, and locked storage space for study equipment. Forty-nine participants were randomized into one of two groups: treatment (n = 26), or control (n = 23).

The treatment group received 24 weekly visits to the “Rolling Store” plus nutrition education classes taught by the peer educators for 1½ hours each including cooking demonstrations and techniques, and encouragement to increase physical activity. Nutrition education and other intervention materials currently available at PBRC, Southern University, and LSU AgCenter Research and Extension were selected by the study investigators. Participants in the treatment group were also asked to keep a 7-day food and exercise diary during 1 week of each month for 6 months. Each of the assigned 7-day blocks consisted of 5 weekdays and 2 weekend days. Food and exercise diaries were used to provide feedback and guidance based on current recommendations to maintain a healthy weight by making healthier choices. At each monthly visit to the church site, the peer educators reviewed with participants the lesson plan, and provided feedback and guidance based on current recommendations to maintain or prevent weight gain. Examples of the treatment intervention lesson plans were: “Healthy Living Program, 5-A-Day Fruits and Vegetables, Eating Healthy When Eating Out, Physical Activity (What Works, Local Places, and Walking Techniques), and Food Security.”

The control group met each month for 6 months and was taught six lessons by the peer educators entitled: “Strong Families; Strong Futures, Family Communication, Family Values, Family History and Traditions, Family Coping, and Family Mealtime.”

The “Rolling Store”

The “Rolling Store” or “grocery truck,” was a method used in the 1950s to sell groceries, candy, tobacco, feed, and kerosene and is historically a part of the food culture around the world; particularly in rural American towns.[26] The owner of a truck with a detachable camper shell residing in Franklin Parish was employed to operate the “Rolling Store.” Easily removable magnetic banners labeled “Rolling Store” were displayed on each door of the truck to promote its appearance. The principal investigator and “Rolling Store” operator were responsible for stocking the store each week. A budget of $10 per participant equating to approximately $213 per week, was allotted for purchasing fruits and vegetables to stock the store. On the same day of each week, the “Rolling Store” parked under the overhang of the church site from 2:00 pm to 6:00 pm to allow enough time for participants in the treatment group to visit the store to obtain quality fresh fruits and vegetables.

Participants in the treatment group received 7–10 choices of fruits and vegetables, equivalent to approximately five to nine servings each week. Participants received an equally distributed amount of assorted fruits and vegetables that were bought in bulk quantities, and at cost from the local grocery store. Fruits and vegetables varied each week, and each week participants received one or more new choices. Each week a featured fruit or vegetable handout was prepared and distributed to participants that included recipes for the item. Some of the featured fruits and vegetables consisted of oranges, turnips, cabbage, mangos, zucchini, bananas, corn, peaches, eggplant, potatoes, green beans, grapes, tomatoes, cucumbers, sweet potatoes, turnip greens, okra, onions, and apples with nutritional benefits and facts, how to select and store, and preparation techniques for each item. In addition, the peer educators conducted cooking demonstrations to illustrate how to prepare healthier (low-fat) foods such as banana pudding, baked apples, eggplant stuffed peppers, Louisiana cucumber salad, stir fry (bell peppers, broccoli, cabbage, carrots, onions, yellow squash), strawberry smoothies, zucchini lasagna, and sweet potato salad.

Statistical analysis

Anthropometrics (height, waist circumference, and weight), blood pressure, laboratory measurements, dietary intake data for fruits and vegetables, and food security were assessed at baseline and at 6 months, and changes over 6 months were analyzed in all participants and separately within each intervention group. Results were summarized as least squares means and 95% confidence intervals. Significance of differences between least squares means for the treatment and control groups were assessed using the t-test. All analyses were conducted by using SAS version 9.3 (SAS Institute, Inc., Cary, NC, USA). Significance was defined as p ≤ 0.05.


Of the 49 adult men and women randomized, 41 (84%) completed the intervention. Eight participants (five treatment, and three control) dropped from the study; five moved out-of-state, two changed jobs, and one became ill. Selected baseline characteristics of participants are presented in Table 1. Participants in the control and treatment groups were comparable with nearly two-thirds of all participants within the 40–59-year age range. Overall, African Americans accounted for 93% of all participants, of which 15% were men. Fifty-one percent of all participants were married compared to 50% divorced in the control group. Two-thirds of all participants never smoked whereas only 15% were current smokers in the control group. Nearly 75% of all participants self-reported not having high blood pressure and 95% as not having diabetes. However, more than 75% and 66% of all participants reported having a family history of high blood pressure and diabetes, respectively.

Table 1. Baseline characteristics of PUSH participants
DemographicsControl n = 20Treatment n = 21Total n = 41
  1. *High blood pressure.

Age (%)
18–39 years301422
40–59 years656263
60+ years52415
Ethnicity (%)
African American909593
Gender (%)
Marital status (%)
Smoke status (%)
Alcohol status (%)
<2 per month252927
HBP* (%)
Family history of HBP (%)
Diabetes (%)
Family history of diabetes (%)

As shown in Table 2, within-treatment group changes in mean weight, BMI, and waist circumference over 6 months were not significantly different from corresponding changes in the control group. Mean systolic blood pressure decreased overall. The decrease was greater in the treatment group (the control group demonstrated trivial change), but the between-group difference in change was not significant. Mean TRG increased approximately 15 mg/dL across 6 months overall and within both treatment and control groups; however, the between-group difference in change was not significant. The mean LDL cholesterol decreased by 5.6 mg/dL in the treatment group (p = 0.07), but not significantly more than the increase of 11.7 mg/dL in the control group. Blood glucose and HbA1C means both increased but the change was not differentially significant between groups.

Table 2. PUSH participant anthropometrics and lab results from baseline to 6 months—least squares mean (95% confidence interval)
VariablesAll participants n = 41Control n = 20Treatment n = 21*p-value
 Baseline6 MonthsBaseline6 MonthsBaseline6 Months 
  1. BP = blood pressure; TC = total cholesterol; HDL = high-density lipoprotein cholesterol; LDL = low-density lipoprotein cholesterol; TRG = triglycerides; GLU = glucose; HbA1c= hemoglobin.

  2. *p-Value for the difference in change (from baseline to 6 months) between the control and treatment group.

  3. †Indicates significant change from baseline value (p < 0.05).

Weight (kg)93.593.695.395.291.792.00.72
 (84.8, 102.1)(85.0, 102.3)(81.4, 109.1)(81.4, 109.1)(81.4, 102.0)(81.7, 102.3) 
BMI (kg/m2)32.432.432.632.532.232.30.68
 (29.3, 35.5)(29.3, 35.5)(27.6, 37.5)(27.5, 37.5)(28.5, 36.0)(28.6, 36.1) 
Waist circ. (cm)98.996.799.697.398.396.20.91
 (93.2, 104.7)(91.0, 102.5)(90.4, 108.7)(88.1, 106.4)(91.4, 105.1)(89.3, 103.0) 
Systolic BP126.0121.9125.5125.4126.4118.30.12
(mm/Hg)(119.1, 132.8)(115.0, 128.7)(115.1, 136.0)(114.9, 136.0)(118.2, 134.5)(110.1, 126.4) 
Diastolic BP75.276.575.778.874.774.10.24
(mm/Hg)(71.0, 79.4)(72.2, 80.7)(69.3, 82.2)(72.3, 85.3)(69.7, 79.7)(69.1, 79.1) 
TC (mg/dL)212.3213.1212.9217.3211.6209.00.35
 (192.3, 232.2)(193.1, 233.2)(181.1, 244.6)(185.5, 249.1)(187.8, 235.4)(185.1, 232.8) 
HDL (mg/dL)46.743.648.144.345.243.00.80
 (40.5, 52.9)(37.4, 49.8)(38.9, 57.3)(35.1, 53.5)(38.0, 52.5)(35.6, 50.3) 
LDL (mg/dL)138.0141.1131.7143.4144.4138.80.07
 (118.9, 157.2)(121.9, 160.3)(101.5, 161.9)(113.2, 173.7)(121.4, 167.4)(115.8, 161.8) 
TRG (mg/dL)138.9153.7154.9169.9122.8137.50.99
 (95.1, 182.6)(110.1, 197.2)(88.9, 221.0)(104.2, 235.6)(70.5, 175.0)(85.6, 189.3) 
GLU (mg/dL)96.2104.497.9105.394.6103.60.73
 (88.8, 103.7)(97.0, 111.9)(86.4, 109.4)(93.7, 116.8)(85.7, 103.4)(94.7, 112.5) 
 (5.4, 7.8)(6.0, 8.4)(5.2, 8.8)(5.2, 8.8)(4.9, 7.7)(5.9, 8.8) 

Finally, the mean fruit consumption over 6 months for the treatment group increased and was significantly different from the change observed in the control group (p = 0.01) as shown in Table 3. Mean vegetable consumption remained about the same in both groups and differential change was not significant between the groups. Mean food security scores decreased in all participants but again the decrease was not significantly different for treatment and control groups.

Table 3. PUSH participant fruit and vegetable consumption and food security scores from baseline to 6 months—least squares mean (95% confidence interval)
VariablesAll participants n = 41Control n = 20Treatment n = 21 
 Baseline6 MonthsBaseline6 MonthsBaseline6 Months*p-value
  1. *p-Value for the difference in change (from baseline to 6 months) between the Control and Treatment intervention groups.

  2. †Indicates significant change from baseline value (p < 0.05).

Servings/day(0.6, 2.4)(0.4, 2.2)(0.3, 3.2)(−0.8, 2.1)(0.2, 2.2)(0.9, 3.0) 
Servings/day(1.2, 2.0)(1.2, 2.1)(1.4, 2.7)(1.5, 2.8)(0.6, 1.6)(0.7, 1.6) 
Food Security Score0.
 (0.003, 0.6)(−0.2, 0.4)(−0.1, 0.8)(−0.3, 0.5)(−0.1, 0.6)(−0.2, 0.4) 


The development and implementation of this CBPR study to increase fruit and vegetable consumption utilizing a “Rolling Store” delivery system plus nutrition education, physical activity, and food security in adults residing in Franklin Parish, Louisiana was considered successful because prevention of weight gain was accomplished. The results are presented only for the first 6 months of follow-up due to the enormous loss in retaining participants following the unexpected death of a key peer educator. Despite the loss, the community remained actively engaged with the academic/research partners by assisting with recruitment and retention of participants, serving as peer educator assistants, and as data collectors. As a result, community leadership emerged from those residents elected by their peers as chairpersons, and liaisons who “stepped forward” in and for their community.[4] This is indicative of a successful CBPR partnership with the community taking ownership and assuming responsibility for the fate of PUSH in the midst of adversity.

PUSH participants were demographically heterogeneous and consistent with the study's eligibility criteria. At the end of 6 months, BMI and weight changes were not significant within either the control or treatment group, and in this sense both groups were successful in preventing weight gain. Research has suggested that, rather than concentrating on producing weight loss or preventing obesity, the focus should be on prevention of the gradual excessive weight gain that allegedly occurs in people of all ages as done in the PUSH study.[26-28] Although not statistically significant, a noteworthy improvement in systolic blood pressure occurred in the treatment group at 6 months. TC increased across both groups and, although insignificant, borderline high cholesterol was evident at the end of 6 months. Overall, improvements were exhibited in waist circumference and HDL cholesterol at the end of 6 months. LDL cholesterol approached statistical significance at the end of 6 months in the treatment group by comparison to the control group. It has been suggested that a diet rich in fruits and vegetables is more effective in lowering LDL than a low-fat diet such as the one used in this study.[29] Elevated TRG, blood glucose, and HbA1c levels increased in both groups suggesting risks of metabolic syndrome also increased.[30] Perhaps future interventions should focus on reducing the risks of metabolic syndrome and heart disease especially in this LMD population.

Participants in the treatment group consumed significantly more servings of fruit whereas consumption of vegetable servings remained about the same. Although not significant, food security decreased and may be attributed to the provision of fresh produce from the “Rolling Store,” and knowledge gained from peer-led education topics on healthy eating and cooking demonstrations. Previous research has shown that a weight maintenance program provided by trained peer educators can be effective, as long as they have the necessary interests and skills.[21, 31, 32] Since an average of five to nine servings of fruits and vegetables were taken home each week, it is feasible that other family members, children, and adolescents of the household may have also benefitted.[33]

All participants were provided with pedometers and encouraged to walk daily; however, no formal physical activity measurements were obtained other than to provide feedback to participants. CBPR studies and others in general may benefit from collecting information in the future from diaries as official data for investigating protocol compliance, especially since pedometers are reasonably simple and inexpensive and may be useful for motivating and increasing physical activity to help maintain weight.[34]

The major limitation in this CBPR study was the death of one peer educator. Despite training two peer educators and others to continue directing PUSH especially in the event of an unforeseen circumstance, a large number of participants were too grief stricken to continue in the study. Death is an inevitable and uncontrollable limitation no one can cure. Perhaps PUSH would have produced better and more significant results if the same peer educators were held constant and remained unchanged throughout the entire study period. Although only the first 6  onths of PUSH is presented, a strong principle of CBPR was visualized when the community remained engaged, took ownership of PUSH, and worked together with the academic/research partners to complete the study at the end of 24 months as initially planned.

Results from this CBPR study suggest that the concept of a “Rolling Store” with the provision of fresh fruits and vegetables plus education relative to health benefits, recipes, and healthy cooking demonstrations may be an effective approach to the prevention of weight gain. Since this study was implemented by trained peers from the community, sustainability of the intervention is conceivable. The impact on family members, especially upon children, must be tested, and the sustainability of health gains must also be evaluated. Nevertheless, the “Rolling Store” overcame some of the barriers to accessibility of healthy food choices in this LMD population.


Special thanks to Dr. Jennifer Rood and staff at the PBRC for training phlebotomists (Ms. Shirley Doyle and Ms. Cassandra McCoy) from the Franklin Parish community to collect and submit blood work data including how to properly dispose of it; research associates at Southern University and A&M College (Ms. Crystal Johnson and Ms. Valerie Richardson) and the PBRC (Ms. Marlene Afton and Ms. Dawn Turner) for assisting with training of peer educators, telephone interviews, and for obtaining refreshments for the community and research committee meetings. In memory of Mrs. Josie Dugue’ Lain—deceased peer educator, and to peer educator Ms. Susie Smith, thank you for implementing the first 6 months of the PUSH study. The authors would like to also thank Ms. Connie Murla and Dr. H. Raymond Allen and their staff for PUSH data entry. Special thanks to Mrs. Meredith Shapiro for additional statistical analysis support. To the Franklin Parish community at large, community stakeholders and research committee chairpersons (Mrs. Bessie Baker and Mr. Glen Watkins), and Ford's Food Center for the fresh fruits and vegetables each week to stock the “Rolling Store,” we thank you. Thanks to the “Rolling Store” operator, Mr. Moses Baker. More important, thanks to all of the participants for completing the PUSH study.

This study was supported in part by the United States Department of Agriculture, Agricultural Research Service (USDA/ARS) Project No. 6251-530000-0020-00D), and supported in part by 1 U54 GM104940 from the National Institute of General Medical Sciences of the National Institutes of Health which funds the Louisiana Clinical and Translational Science Center.

Conflicts of Interest

The authors declare that there is no conflict of interest.