Interventions for Improving Nutrition and Physical Activity Behaviors in Adult African American Populations: A Systematic Review, January 2000 Through December 2011

Introduction The incidence of preventable chronic diseases is disproportionally high among African Americans and could be reduced through diet and physical activity interventions. Our objective was to systematically review the literature on clinical outcomes of diet and physical activity interventions conducted among adult African American populations in the United States. Methods We used the Preferred Reporting Items for Systematic Review and Meta Analysis construct in our review. We searched Medline (PubMed and Ovid), Cochrane, and DARE databases and restricted our search to articles published in English from January 2000 through December 2011. We included studies of educational interventions with clinically relevant outcomes and excluded studies that dealt with nonadult populations or populations with pre-existing catabolic or other complicated disorders, that did not focus on African Americans, that provided no quantitative baseline or follow-up data, or that included no diet or physical activity education or intervention. We report retention and attendance rates, study setting, program sustainability, behavior theory, and education components. Results Nineteen studies were eligible for closer analysis. These studies described interventions for improving diet or physical activity as indicators of health promotion and disease prevention and that reported significant improvement in clinical outcomes. Conclusion Our review suggests that nutrition and physical activity educational interventions can be successful in improving clinically relevant outcomes among African Americans in the United States. Further research is needed to study the cost and sustainability of lifestyle interventions. Further studies should also include serum biochemical parameters to substantiate more specifically the effect of interventions on preventing chronic disease and reducing its incidence and prevalence.


Introduction
Education and community-based programs in disease prevention and health promotion played an important role in achieving Healthy People 2010 (1) goals and are of continued importance in attaining Healthy People 2020 goals, especially among minority populations (2). Primary prevention interventions aimed at reducing the higher incidence and prevalence of chronic diseases among African Americans (3)(4)(5) include education about healthful lifestyle choices regarding diet and physical activity. Interest in and funding for research into the health effects of improved nutrition and physical activity is needed to develop interventions that would reduce disparities between African Americans and the US population overall.
Growing awareness of the role of a healthful diet and physical activity in reducing chronic disease has greatly increased the volume of recent literature on the subject, so that many reviews (6)(7)(8)(9) are now outdated. Researchers have recommended using more rigorous methodology (6,9) and have discussed validated and reliable instruments (6,9), theory-based interventions (9), culturally tailored interventions (6,7), and concepts for successful recruitment and retention of diverse study populations (8). Researchers have not fully addressed clinical outcomes. For example, a recent review (10) examined physical activity outcomes (ie, increasing frequency of exercise) among African Americans but did not evaluate various health-related outcomes (ie, weight loss or improved lipid profile) resulting from increased exercise. Few reviews focused on clinical outcomes from improved diet and nutrition (6)(7)(8)(9)(10). Our objective was to review available literature on the effect of educational interventions on clinical outcomes resulting from improved nutrition and increased physical activity among adult African Americans in the United States.

Data sources
We used the Preferred Reporting Items for Systematic Reviews and Meta Analysis (PRISMA) (11) for this review. We searched Medline (PubMed and Ovid) and Cochrane databases and DARE (Database of Abstracts of Reviews of Effects) to identify effective interventions in modifying dietary and physical activity behaviors among adult African Americans in the United States, January 2000 through December 2011.

Study selection
Nineteen studies (12-32) met our selection criteria, were confirmed by 2 reviewers, and were included in this review ( Figure); 1 study reported dietary and physical activity components of the intervention in 2 separate articles (20,21), and another reported intervention results in 1 publication (26) and recruitment and retention information in another (27). We restricted our search to articles published in English from January 1, 2000, through December 31, 2011. Articles were limited to the past 12 years to capture the most recent studies; earlier studies had been included in previous reviews (6)(7)(8)(9). Research and publishing on the effect of diet and physical activity interventions among minority and underserved populations have increased during the past 12 years in the United States, providing adequate literature for review.

Data extraction
Our initial search used a combination of the following key words: African American, nutrition, diet, physical activity, weight loss, and intervention. The inclusion criteria were 1) that the article addressed educational interventions in diet and physical activity with clinically relevant outcomes (eg, changes in weight, body mass index (BMI [kg/m ]), body fat percentage, cholesterol, triglycerides, blood pressure) and 2) that the educational interventions included at least 1 direct or indirect instructional strategy to educate participants on how to increase daily physical activity and fruit and vegetable consumption or how inactivity and poor dietary patterns can negatively affect health. We excluded studies that dealt with nonadult populations or populations with pre-existing catabolic or other complicated disorders (eg, cancer, cirrhosis, HIV/AIDs, heart failure); studies that did not focus on African Americans; studies that provided no quantitative baseline or follow-up data; and studies that included no diet or physical activity education or intervention. Data were extracted by 1 reviewer and verified by a second. Information on the following components were collected: study settings, theoretical principles incorporated into the intervention design, study outcomes, intervention details, recruitment strategies, attendance and retention rates of the participants, sustainability of the achieved changes in health behaviors, and provision of physical activity and nutrition education.
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). 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).

Study setting and recruitment strategies
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.

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.

Discussion
Because it has been confirmed that African Americans have an increased burden of chronic disease risk that is exacerbated by unhealthy diets and limited physical activity (34,35), interventions for diet and physical activity behavior modification may alleviate this burden. Several studies published since 2000 were available for review, and based on the current literature, diet and physical activity interventions targeting African Americans need to be refined. This review evaluated the most current literature on diet and physical activity interventions among African Americans, and we emphasize that clinical progress was a major consideration in our review. The evidence shows that nutrition and physical activity lifestyle modifications alone can significantly improve predictors of various clinical outcomes in African Americans (13)(14)(15)(16)(17)(18)(19)(20)(21)(23)(24)(25)(26)(27)(28), but more research is needed to substantiate the clinical significance of the outcomes.
In a more recent, similar review, Pezmeki and Jennings (10) examined physical activity outcomes without evaluating the effect of the behavior change on various health outcomes. Seven interventions (17,(19)(20)(21)23,26,27,30) included in our analysis overlapped with the previous review. Because the review by Pezmeki and Jennings (10) aimed to examine physical activity outcomes exclusively, we saw a need to fill a gap in the literature by presenting a comprehensive examination of studies that promoted diet and physical activity behavior change to improve health outcomes, including major clinical indicators.
Although all 19 of the studies we reviewed demonstrated significance in various clinical outcomes, the most frequently reported variables were weight, BMI, SBP, and DBP. It was apparent that the African American participants in populations targeted by selected interventions were obese and prehypertensive. Mean decreases in SBP and DBP were sufficient to decrease blood pressure but not the number of people diagnosed with prehypertension (16)(17)(18)(19)(23)(24)(25)(26)(27)(28)32). Although BMI remained in the obese range for participants in all 19 studies, mean weight loss reported amounted to approximately 3.0% of initial body-weight (13,(15)(16)(17)(18)(19)(22)(23)(24)(25)28,30,32). Unfortunately, we did not find enough evidence to comment on the effect of diet and physical activity interventions on other clinical outcomes, including lipid and blood glucose levels.
Our findings are consistent with published research results that support using community-based settings for implementing interventions to improve diet and physical activity among African Americans. Our findings also indicate that churches may be useful in reaching communities beyond their congregations for study participation; several studies successfully recruited and reported changes among participants who did not attend church at the study's location (18,23,29,31,32). In addition, we found that some of the highest participant retention rates were among programs conducted in churches (18,19).
Eleven studies (13,14,18,19,(24)(25)(26)28,29,31,32) did not report using behavior theory in their intervention design, and this omission may indicate a weakness in the analyzed results of those studies. Unfortunately, compared with a past review of physical activity interventions among African Americans little progress has been made in increasing the number of studies that use health behavior theory in intervention program design (6). Research designed with a theoretical foundation better predicts and validates whether lifestyle interventions will promote changes in health behaviors while also providing a means to evaluate change processes in health behavior. Only about one-third of health -behavior-change studies conducted from 2000 through 2005 and reviewed by Painter et al (36) reported using behavior theory in intervention design. Our results indicate a slightly higher proportion; 42.0% of studies used behavior theory models. Furthermore, of the 8 studies (12,(15)(16)(17)(20)(21)(22)(23)30) that used a behavior theory, 7 reported significant results (12,(15)(16)(17)22,23,30), whereas the remainder of studies (13,14,18,19,(24)(25)(26)29,31,32) all reported significant results. It is possible that studies examined may lack explicit reporting of the use of behavior theories. We suggest that common health behavior theories be incorporated into intervention design, applied more thoroughly in research beyond mere consideration of their definition, tested in African Americans, and modified as necessary to better suit the population (36).
This review found no differences in select clinical outcomes (eg, weight, SBP, and DBP) between programs that offered physical activity education and those that offered physical activity plus diet education, although few physical activityeducation only interventions were available for comparison. One study included in this review (14) was noteworthy in that it found no differences in body composition (weight, waist circumference, and BMI) at a 6-month interval between participants who received a diet education intervention and those who received a diet plus physical activity intervention. This lack of difference suggests that both diet and physical activity should be considered major components of a lifestyle behavior change intervention until further evidence is available and examined. Other comparison issues noted were the lack of homogeneity among the education tools used in interventions and our inability to determine which specific education tools were used. Without this information, it is hard to generalize, repeat, and retest successful education programs as the NIH DPP has done (13,18,25,32).
Compared with programs aimed at improving physical activity and diet in predominately white populations, mean attendance and retention rates at programs for African Americans were lower. The mean attendance rate for studies with predominantly white participants was 95.0% (37-39) compared with 58.0% for studies in our review with predominantly African American participants (12,15,17,18,22,24,31). In addition, studies with mostly white participants reported a mean retention rate of 87.0% (37,38) compared with a mean of 80.0% reported for studies with mostly African American participants (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32). Several studies (39-43) identified effective retention practices for African American research participants, but because culture plays a central role in behavior choices, generic health messages may conflict with the cultural beliefs of minority populations (44). The increasing prevalence of coronary vascular disease among African Americans illustrates the need for culturally relevant associations between food and health, including the use of culturally tailored intervention components, such as lifestyle-change education to lower coronary vascular disease risk by modifying, not eliminating, cultural foods to improve diet and by suggesting culturally acceptable and attractive modes of physical activity. It is still necessary to improve both African American attendance and retention rates for health interventions. However, more than half of the studies in the current investigation (n = 10) did not report an attendance rate (13,14,16,(19)(20)(21)(23)(24)(25)(26)32), thereby making attendance unavailable for evaluation. Not reporting attendance is notable because program attendance is imperative for participants to acquire the knowledge and skills needed to modify health behaviors (45).
Our study had limitations. Our review was only as good as the available studies analyzed. Not all included studies used randomization for participant allocation, and some of the results may show allocation bias. Another limitation is publication bias. Many relevant studies that were unpublished because of nonsignificant results would have been included in this review if published. Other analyzed studies did not use control groups for comparison, which may have introduced some experimenter bias for the overall conclusions.
This review provides researchers with current information on intervention program effectiveness and provides policy makers and health care practitioners with evidence about whether nutrition education and physical activity promotion interventions are valuable methods for improving the health of African Americans. The authors identified from the reviewed literature a common protocol for implementation of lifestyle interventions to prevent chronic disease in African American adults (Appendix). To our knowledge, there is no previously published analysis that encompasses all variables considered in this systematic review.
Nutrition and physical activity interventions promote positive changes in the health behaviors of adult African Americans by providing them with knowledge and resources about disease prevention. The reviewed studies have shown that these health interventions had a positive effect on the participants' dietary choices and physical activity habits, which translated to clinically relevant outcomes in communities, churches, and health clinics. Using health education and interventions designed to teach African American adults about healthy choices (eg, proper nutrition and adequate exercise) empowers them to make necessary lifestyle changes. Thus, potential exists for reducing preventable risks for diseases and comorbidities while positively affecting the health of the community. Future interventions should incorporate theoretical models appropriate for health-related issues and randomized controlled trial design into the basis of their programming, because this gives credence to evidence-based research.
The current literature suggests that more research is needed to determine the cost and sustainability of lifestyle intervention programs. We noticed in our review that the costs of designing and implementing interventions were rarely published, offering no guidance as to what the typical costs are of interventions. Additionally, a cost-benefit analysis would facilitate the awareness and spread of information about public policy and funding allocation. Sustainability of improved diet and physical activity behavior is also rarely noted. Programs are implemented and results are documented, but whether targeted populations are able to further or maintain progress after the cessation of the intervention is still in question.
Further research is required to substantiate the link between the intervention-induced diet and physical activity changes and related disease-risk biochemical markers. Such data could be documented on a clinical level, and the information on population disease-risk characteristics could become available. This type of data also adds to the potency of the argument for funding public education programs. Although many interventions have proven successful for African American populations, filling the information gaps will promote and substantiate further progress of intervention research and a subsequent improvement in overall health of African Americans and a reduction in population health disparities.   Intervention Duration: 12 months. Location: university. Design: group, individual, and telephone contacts (1st-6th month, 3 group and 1 individual contacts/month; 6th-12th month, 2 group sessions and 2 telephone contacts/month). Diet prescribed for both groups to achieve 8% to 10% weight loss over 12 months; meal replacements offered during first 6 months. PA goals for both groups to achieve 60 minutes, 5 days per week, moderate-intensity PA. Delayed PA group started PA at 6th month. Education: diet and physical activity. Follow-up: 6 months, and 12 months.

Attendance and retention
Attendance: not reported. Retention: initial group, 90% at 6 months, 73% at 12 months; delayed group, 90% at 6 months, 83% at 12 months. Study not African American-specific; however, no difference reported between white and African American participants.

Clinical outcomes and results
Weight: 12.1 kg decrease for initial group; 9.9 kg decrease for delayed group (P < .001). Loss in percentage body fat and waist circumference significant in both groups (P < .001); higher in initial group but not significant. SBP: approximately 15 mm Hg decrease at 12 months in initial and delayed groups (P < .001). DBP: approximately 6 mm Hg decrease in initial and delayed groups (P < .001 Yanek et al, 2001 (17) Sample Women (n = 529) ≥40 y (mean, 53.1 y; SD, 9.3 y). Three groups: group without spirituality (n = 188); group 2, group with spirituality (n = 267); group 3, self-help group, no spirituality (n = 74).

Clinical outcomes and results
Weight: 0.5 kg decrease in without-spirituality and spirituality groups (P ≤ .001); 0.8 kg increase in self-help group (P < .01). BMI, 0.2 kg/m decrease in without-spirituality and spirituality groups (P = .02); 0.1kg/m increase in self-help group (not significant). Waist circumference: 1.7 cm decrease in without-spirituality and spirituality groups (P < .001); 0.02 cm decrease in self-help group (not significant). SBP: 1.6 mm Hg decrease in without-spirituality and spirituality groups (P < .001); 0.9 mm Hg decrease in self-help group (no significance). DBP: 0.4 mm Hg decrease in without-spirituality and spirituality groups; 0.2 mm Hg increase in self-help group. Results based on 12th month follow up.

Clinical outcomes and results
Weight: mean loss 3.3 kg in both groups (P < .05); no significance between groups. BMI: 1.2 decrease in both groups; no significance compared with baseline or between groups. Percentage body fat: mean loss 0.5% in both groups (P < .05); no significance between groups. SBP: 1 mm Hg loss in both groups, no significance compared with baseline or between groups. DBP: 2 mm Hg loss in both

Study Components Description
groups; no significance compared with baseline or between groups. HDL, 3 mg/dL decrease. LDL: 4 mg/dL decrease (P < .05 at baseline); no significance between groups. Total cholesterol, triglycerides, FG: no significant differences from baseline or between groups.