Interventions for Improving Nutrition and Physical Activity Behaviors in Adult African American Populations

A Systematic Review, January 2000 Through December 2011

Jennifer Lemacks, PhD, RD; Brittny A. Wells, MEd, CHES; Jasminka Z. Ilich, PhD, RD; Penny A. Ralston, PhD


Prev Chronic Dis. 2013;10 

In This Article


Characteristics of Analyzed Studies

Among the 19 studies reviewed (5.3% of nonduplicated identified articles), 1 included participants of various races and ethnicities,[13] and 4 (12,14,15,32) included both white and African American participants; however, all 4 had sizable samples and distinguished the African American population (according to the US Census Bureau 2002, 13% of the total US population are African Americans [33]) (Table). Mean age of participants in all studies (N = 3,530) was 51.5 years; 2,997 (85.0%) were women and 533 (15.0%) were men. One study[18] did not report mean age of participants but enrolled only 10 participants. Another study[13] was not represented in the mean age because it reported age data in ranges. The majority of men (63.5%) and women (66.9%) were aged between 40 and 50 years. A study by Goodpaster et al[14] with 48 participants was not represented in the mean sample evaluation because it was not possible to distinguish African American men and women from the total sample number. At baseline, mean BMI was 34.7 (n = 2,613; range, 30.3–38.9), mean systolic blood pressure (SBP) was 135.1 mm Hg (n = 1,868; range, 120.8–151.4), and mean diastolic blood pressure (DBP) was 81.1 mm Hg (n = 1,868; range, 74.0–91.0) (17–21,23–29,31,32). We excluded 2 studies[12,14] from these calculations for not distinguishing African American population characteristics; 2 studies[13,16] for not including BMI information, and 3 studies (15,22,30) for not including blood pressure information.

Clinical Outcomes

Overall, the nutritional and physical activity interventions we reviewed reduced risk for chronic diseases by succeeding in improving clinically relevant outcome measures, including weight loss[13–19,22–24,30,32] and decreases in waist circumference,[14,17,29] SBP,[14,16,18,23–26,28] DBP,[14,16,18,23–25] fasting plasma glucose,[18,25] body fat percentage,[14,16,19,23] hemoglobin A1c,[16,31] and blood lipids (high-density lipoprotein [HDL], low-density lipoprotein [LDL], total cholesterol, and triglycerides).[14,24] Only 1 study, which was published in 2 separate articles as dietary and physical activity outcomes,[20,21] was unable to produce significant results in 2 clinical outcomes: reducing BMI and waist circumference. Among the most commonly reported outcome measures (weight, BMI, SBP, and DBP), mean weight loss was 6.4 kg (range, 0.5–32.6 kg),[13–19,22–25,28,30] BMI decrease was 1.0 kg/m2 (range, 0–1.7 kg/m2),[15,17–19,22–25,28–30,32] SBP decrease was 6.7 mm Hg (range, 1.0–12.5 mm Hg),[12,14,16–19,23–26,28,32] and DBP decrease was 4.5 mm Hg (range, 1.0–10.2 mm Hg).[12,14,16–19,23–26,28,31,32] Weight loss seemed to vary widely; however, 1 study reported a weight loss of 32.6 kg resulting from a prescribed 1,000 kilocalorie diet;[24] the next highest weight loss was 14.5 kg.[22]

The improvements in serum biochemical markers, including HDL, LDL, and some other health indicators (eg, triglycerides, blood glucose levels) were measured and observed in just a few of the studies.[17,19,32] Project Joy,[17] a, church-group–based healthy-lifestyle intervention, reported significant decreases in LDL cholesterol in the intervention group and no changes in the self-help control group. However, a pilot, church-based, weight-loss program for African American adults using church members as health educators observed no significant changes in HDL, LDL, triglyceride, or blood glucose levels.[19] The same was true for another study conducted in rural primary care health centers where the percentage of African American participants at each center ranged from 73.2% to 89.4%.[32]

Study Setting and Recruitment Strategies

The majority of the 19 studies we reviewed were conducted in churches (n = 5), clinical or community health centers (n = 5), or other community locations (ie, community centers or local gyms) (n = 4). Remaining intervention locations were home-based (n = 1), universities (n = 2), a residence (n = 1), and a hospital (n = 1).

Recruitment methods varied depending on the intervention setting. Researchers conducting faith-based interventions,[17–29] generally announced the program in church bulletins, in the upcoming events section of church newsletters, at church services on Sundays, or at other church group meetings. They also relied on word of mouth and community contacts or delegated church pastors to make announcements about the program. Researchers conducting community and clinic-based interventions recruited participants via mailing lists,[22] physician referrals,[13,16,20,21,27] staff presentations,[20,21] social networking and word of mouth,[20,21] and targeted mass media.[20,21] In an individual-based study,[27] newspaper advertisements, flyers, and referrals from friends and coworkers were reported to be more effective recruitment methods (accounting for 73.0% of responses) than recruiting via blood pressure screenings, physician referrals, and Internet advertisements. Similarly, other researchers[20,21] found it easier to use personalized methods (ie, word of mouth or social networking) to recruit African American women with high BMI than African American women with low BMI (P = .01). Generally, 2 weeks to a month were spent recruiting participants before the start of the study. However, Staffileno and Coke[27] reported a recruitment period of 19 months because of the greater specificity of their desired population.

Attendance and Retention Rates

Attendance rates at the 10 intervention programs that reported attendance varied from 33.0% to 95.0%.[15,17,18,20–22,24,29–31] Various methods were reported for calculating attendance rates. One study[28] reported 85.0% and another[22] reported 55% of participants attended 75.0% of the sessions. One study reported an attendance rate of 95.0% for all 8 sessions, but that figure was based on an attendance requirement of only "at least one session".[20,21] The mean attendance rate of intervention programs that defined attendance as "participants who attended all sessions"[15,17,18,24,29–31] was 58.0%. Retention rates tended to be much higher (mean, 80.0%; range, 43.0%–96.0%) than attendance rates (mean, 58.0%; range, 33.6%–79.0%), suggesting it is more difficult to promote consistent participation in a program.[16,18–22,25,26] Retention rates were also more commonly reported than attendance rates. Financial or other incentives were helpful for retention of participants; 1 study with a retention rate of 90.0% paid each participant $100.[19]

Sustainability of Outcomes

Although interventions show evidence of short-term (range, 2–6 months) improvements in diet and exercise habits, there was limited follow-up to prove that these changes were long-term and that they continued beyond the intervention. Of 6 studies that followed up with participants beyond the intervention, 4 studies[13,14,18,25] demonstrated sustainability of significant intervention-provoked outcomes at 12 months; others[20,21,23] did not. West et al[13] also observed that African American women had gained more weight than had whites, African American men, or Hispanics at a 30-month follow-up in a diabetes prevention program that used one-on-one counseling in a clinical center.

Incorporation of Behavior Theory and Randomization Into Study Design

Of the studies reviewed, 7[12,15,16,20–23,30] reported incorporating social cognitive theory as the theoretical framework for the intervention, and one[17] used social learning theory. Four studies combined social cognitive theory with a socioecological model,[20,21] health belief model,[22,23] transtheoretical model,[12,15] and a community-based participatory research model.[12] Eleven studies[13,14,18,24–26,29,31,32] did not report using a theory-based intervention design. Only 6 studies[15,18,22,23,25] did not report using a randomized clinical trial design.

Educational Interventions Description

All of the 19 educational interventions reviewed[12–32] reported implementing physical activity education components, including goals for, benefits of, and strategies for increasing daily physical activity. A majority (78.9%, n =15)[12–20,22,24,25,30–32] of the 19 studies also implemented a dietary education component, including topics such as calorie reduction for weight loss, dietary sodium and fat reduction strategies, portion sizes, food group topics, and strategies for healthy eating while dining out and during the holidays. Four studies[13,18,25,32] followed the National Institutes of Health Diabetes Prevention Program (NIH DPP) curriculum for both diet and physical activity topics, whereas 1 study[12] used DASH (Dietary Approaches to Stop Hypertension) diet principles to guide nutrition education. Interventions[23,26,28,29] that focused only on education about physical activity for African American women were not significantly better at reducing weight, SBP, or DBP than were interventions[12–20,22,24,25,30–32] that focused on education about both physical activity and diet.