Community Interventions to Improve Glycemic Control in African Americans With Type 2 Diabetes: A Systemic Review

Purpose: The purpose of this study was to conduct a systematic review of published community interventions to evaluate different components of community interventions and their ability to positively impact glycemic control in African Americans with T2DM. Methods: Medline, PsychInfo, and CINAHL were searched for potentially eligible studies published from January 2000 through January 2012. The following inclusion criteria were established for publications: (1) describe a community intervention, not prevention; (2) specifically indicate, in data analysis and results, the impact of the community intervention on African American adults, 18 years and older; (3) measure glycemic control (HbA1C) as an outcome measure; and (4) involve patients in a community setting, which excludes hospitals and hospital clinics. Results: Thirteen studies out of 9,233 articles identified in the search met the predetermined inclusion criteria. There were 5 randomized control trials and 3 reported improved glycemic control in the intervention group compared to the control group at the completion of the study. Of the 8 studies that were not randomized control trials, 6 showed a statistically significant change in HbA1C. Conclusion: In general, the community interventions assessed led to significant reductions in HbA1C in African Americans with type 2 diabetes. Community health workers did not have a greater impact on glycemic control in this sample. The findings of this study provides insight for designing community-based interventions in the future, such as including use of multiple delivery methods, consideration of mobile device software, nutritionist educator, and curriculum-based approaches.

cost is expected to reach $192 billion by 2020 (NIDDK, 2011). Major complications and comorbid illnesses result from T2DM, including blindness and vision problems, nervous system disorders, kidney disease, amputations, periodontal disease, heart disease, and stroke (NIDDK, 2011). In addition, T2DM is the seventh leading cause of death based on US death certificates in 2007 (NIDDK, 2011).

Burden of Diabetes on African Americans
Minority populations, particularly African Americans, are disproportionately affected by T2DM (Narayan et al., 2003). African Americans are 1.6-times more likely to develop T2DM and suffer from complications compared to non-Hispanic Whites (Narayan et al., 2003). The average years of life lost for African American males due to T2DM is 9.3 years in contrast to 8 years in non-Hispanic White males (Narayan et al., 2003). Similarly, the average number of years lost for females due to T2DM is 12 years in African Americans compared to 10.3 years in non-Hispanic Whites (Narayan et al., 2003). These statistics become even more pertinent based on projections that T2DM will increase 3.0-fold in African Americans, and 1.2-fold in non-Hispanic Whites by 2020 (Hogan, Dall, & Nikolov, 2003).
Evidence shows that minority populations have higher mortality rates due to complications associated with T2DM (Campbell, Walker, Smalls, & Egede, 2012). Poor self-management behaviors and poor T2DM clinical outcomes are suggested as the reasons for higher mortality and complication rates. Based on the American Diabetes Association (ADA) guidelines for glycemic control, blood pressure and lipids, on average, African Americans have suboptimal T2DM-related outcomes and have an increased risk for microvascular and macrovascular complications (Campbell et al., 2012). For the purposes of this review will focus on glycemic control, as the primary outcome, measured by hemoglobin A1C (HbA1C), which is a known indicator of T2DM severity (ADA, 2013).

Barriers to Optimal Health Outcomes in African Americans
Traditional approaches to T2DM management, based on studies in primarily non-Hispanic White populations, may not be as effective in African Americans due to differing social determinants. Disparities in optimal T2DM management have been attributed to barriers at the patient-, provider-, and health systems-levels, although many of these barriers have not been adequately studied (Syler & Oddo, 2002;Marmot, 2005;Brown et al., 2004). At the patient-level, three important barriers to optimal outcomes have been identified: lack of T2DM specific knowledge, poor self-management skills, and poor motivation to make lifestyle behavior changes (Syler & Oddo, 2002). These barriers are thought to explain the lower likelihood of African Americans adhering to a healthy diet, engaging in regular physical activity, and participating in weight loss programs (Tang et al., 2005). Provider-and health system-level barriers to proper T2DM care include lower rates of screening, perceived complexity and difficulty of treating patients and lack of adequate time and resources (Narayan et al., 2003;Tang et al., 2005). Although several novel interventions are available to manage T2DM, minority groups are less likely to attain target glycemic levels set by the American Diabetes Association (Syler & Oddo, 2002).

Community Interventions and Diabetes Self-Care
Efforts to implement interventions targeting African American with T2DM in a health care setting have shown to reduce HbA1C by 0.8% (Ricci-Cabello et al., 2013). Yet, in recent years, health interventions implemented in nonclinical settings has increased, allowing for novel patient focused study components as can be seen in community interventions. One way this multi-level problem has been addressed is through community interventions (Ricci-Cabello et al., 2013).
For the purpose of this review, community interventions are techniques that introduce and promote diabetes self-care management outside of the hospital or hospital clinics. These interventions may be more effective than traditional methods because they address sociocultural and environmental factors that positively influence T2DM health outcomes (Nine, Lakies, Jarrett, & Davis, 2003;Norris et al., 2002;Englegau et al., 1998;Liburd, Jack, Williams, & Tucker, 2005;Jack, Liburd, Vinicor, Brody, & Murry, 1999;Brody, Jack, Murry, Landers-Potts, & Liburd, 2001;Egan, Tannahill, Petticrew, & Thomas, 2008;Hausmann, Ren, & Sevick, 2010). It has been suggested that systematically addressing socioeconomic and environmental factors can attenuate observed minority disparities in T2DM health outcomes, and may address some of the barriers to optimal outcomes (Fry, approach has the ability to address specific challenges and ways to overcome poor T2DM self-management (Englegau et al., 1998). Additionally, the approach is more comprehensive and usually tailored to accommodate the intended participants by focusing on reducing risk factors, prevention of complications and disabilities, and improving quality of care (Englegau et al., 1998).
The use of community health workers in community interventions has been shown to improve health behaviors and outcomes (Spencer et al., 2011). Community health workers are characterized as members of the target community who have been trained to be liaisons between their communities and healthcare providers (Spencer et al., 2011). Their duties include educating patients, identifying resources, and becoming part of the patients' social support network (Quinn et al., 2008). Community health workers seem to have a positive influence in racial and ethnic minority groups who have historically had restricted access to health care services (Spencer et al., 2011).

Systematic Review Objective
Though, theoretically, community health workers may factor more importantly in the success of community interventions than other elements of community interventions, this comparison has not been investigated. This systematic review was conducted in an effort to evaluatevariouscommunity interventions and their ability to positively impact glycemic control in African Americans with T2DM.

Search Strategy and Databases
A search for articles published from January 2000 through January 2012 using 3 databases, Medline, PsychINFO, and CINAHL, was conducted for potentially eligible studies using a reproducible strategy. The search was limited to 12 years because the majority of interventions targeting African Americans with T2DM were published starting in 2000 (Ricci-Cabello et al., 2013). The following separate searches were conducted using medical subject heading (MeSH) terms to maximize the search results. MeSH terms, as defined by the Cochrane Library (Cochrane), are a set of terms naming descriptors in hierarchical structure that allows one to search at various levels of specificity. The first search used the search terms diabetes and ethnic groups. The second search used the terms diabetes and lifestyle. The third search used the terms diabetes and community. The three separate searches, combined, resulted in 9,233 citations. Duplicates were removed.

Selection Criteria
The following inclusion criteria were implemented: (1) describe an intervention whose aim is to improve health outcomes, not prevention of T 2 DM; (2) specifically indicate in data analysis and results the impact of the community intervention on African American adults, 18 years and older; (3) indicate HbA1C as an outcome measure to determine successful diabetes management; (4) involve patients in a community setting which excludes hospitals and hospital clinics; and (5) articles must be published, no grey material were included. No articles were excluded based on study design. However, articles were excluded if titles indicated that the study was focused on gestational diabetes or type 1 diabetes. Full articles were read and reviewed by 2 independent researchers (BS, RW) using a study specific checklist that outlined the inclusion and exclusion criteria to assist with consistency between reviewers (Ueffing, Tugwell, Welch, Petticrew, & Kristjannson, 2011). In the case of a disagreement regarding article inclusion, a senior investigator (LE) was asked to make the final decision. Thirteen eligible studies were identified based upon the predetermined inclusion criteria (Figure 1).

Data Collection
Data was obtained on the number of study participants; race/ethnicity of the sample population; duration of the intervention; setting of the intervention; and intervention description including the community intervention being implemented, theoretical basis for study, study design, type of control, and outcome (HbA1C) ( Table 1). Each article was reviewed for statistical significance of the outcome variable, HbA1C. Mean baseline of A1C, community intervention group mean change in HbA1C, control group mean change in A1C, and statistical significance (p<0.05) were documented ( Table 2). The target threshold for glycemic control was an HbA1C less than 7%, based on ADA guidelines (Hogan et al., 2003). Finally, each article was reviewed for the use of community health workers, in addition to other relevant characteristics of the intervention, including use of a culturally tailored approach, curriculum-based approach, involvement of nurse educators, diabetes educators, nutritionist educators, one-on-one counseling, group counseling, physician involvement, supervised exercised, telemedicine, nurse case managers, and/or mobile device software (Table 3). No quantitative data synthesis was performed because of the heterogeneous interventions across studies and diversity of the study designs and outcome measures. Figure 1 shows the search results. 9,233 citations resulted from the search after duplicates were removed.

Study Selection
Reviewing titles produced 1537 abstracts, after which 34 articles were determined eligible for full review. Of the 34 full articles reviewed, 13 met the predetermined inclusion criteria.

Characteristics of Community Interventions
The community interventions utilized various delivery methods, such as software, telemedicine, and in-person sessions. Additionally, various types of healthcare professionals delivered the intervention information, including nurse case managers, community health workers, and diabetes educators. The type of information provided, if specified, was curriculum-based or culturally tailored. These community interventions were implemented via community clinics, community centers, home visits (i.e., nurse case managers or community health workers visiting the homes of individuals with T2DM to provide diabetes self-management education), or independent in-person interactions (i.e., mobile phone software that assists in monitoring HbA1C). The racial composition of study participants varied and included non-Hispanic Whites, Hispanic Whites, Latinos, and African Americans.

Discussion
Thirteen articles met the inclusion criteria set for this review. The reviewed articles showed a variety of community intervention elements. In general, community interventions are shown to help manage T2DM in African Americans; however, no specific component is superior in achieving glycemic control. Overall, community interventions had a similar impact on glycemic control despite different intervention elements, ranging from 0.55% to 1.23% decrease in HbA1C, all being statistically significant (p<0.05) (Nine et al., 2003;Quinn et al., 2011;Quinn et al., 2008;Sekhobo et al., 2008;Two Feathers et al., 2005;Magee et al., 2011). A reduction in HbA1c greater than 0.5% is clinically significant and is associated with reduced risk of T 2 DM-related complications, specifically cardiovascular disease (Osende et al., 2001).

Implications of Community Interventions in African Americans
Our findings do not suggest that community health workers have a greater impact in attaining glycemic control compared to other community intervention elements. Rather, these findings suggest that several elements of community intervention studies that target African Americans with T2DM can have a positive effect on glycemic control. Our findings are in line with Norris and colleagues who indicated that diabetes self-management education, presented in a community setting, is an effective way to lower HbA1C and improve other T2DM related outcomes (Norris et al., 2002). Better glycemic control has been linked to improved T2DM outcomes, decreased complications, and improved quality of life (Tang et al., 2005). Therefore, the fact that community intervention studies are successful in reducing HbA1C is a significant finding.
Overall, this review provides evidence that community interventions are effective in African Americans with T2DM. However, the evidence shows that community interventions in the population of interest should not be heavily dependent on community health workers as the sole delivery mechanism to facilitate glycemic control.
There are 3 elements of community interventions that were used in interventions that found statistically significant decreases in HbA1C: mobile device software, nutritionist educator, and curriculum-based approach (Table 3). Though no one study combined these 3 elements, future studies in this population should consider the use of these elements as part of community interventions. Additionally, future research should be conducted solely on African Americans to investigate the effectiveness of specific characteristics of community interventions in this population. Future studies may also find it important to use additional outcomes as a measure of effectiveness. For the purposes of replicating successful results, studies should provide more detailed descriptions of intervention techniques and measure the long-term impact on T2DM outcomes. Lastly, an intervention that includes a health system and community-based component may exponentially reduce HbA1C in African Americans.

Limitations
There are few noteworthy limitations of this systematic review. First, the review was limited to studies that were published in the English language between 2000 and 2012. Second, the review was limited to articles that had glycemic control as an outcome measure. There are other important diabetes outcomes that could be evaluated, such as risk factors associated with micro-and macrovascular disease. Third, publication bias and selective outcome reporting was possible and could have biased our conclusions. Fourth, there were various community intervention elements in the articles reviewed, which did not allow us to aggregate estimates of intervention effects or conduct a meta-analysis. Lastly, there are also limitations that can be derived from the 8 studies that were not RCTs, including increased inability to identify confounders, social desirability, and disproportionate number of participants in study groups.
In this review, risk of bias was likely present but no articles were excluded due to potential bias. The risk of bias could be introduced as a result of studies that did not show a statistically significant decrease in HbA1C due to small sample population, small number of potential study participants, inappropriate matching of control and intervention participants, researcher bias, and brief duration of the community intervention (Nine et al., 2003;Ueffing et al., 2012;Gary et al., 2009). Furthermore, studies with positive and statistically significant findings may be more likely to be published compared to studies who have negative or null findings.

Conclusion
The majority of the articles reviewed suggest that community interventions for T2DM self-management are effective in improving HbA1C in African Americans. Further research should be conducted where the study population is exclusively African Americans with T2DM so that the impact of various characteristics of community interventions on T2DM management can be assessed in this population. In the interim, the findings of this study provides insight for designing community-based interventions in the future, such as including use of multiple delivery methods, consideration of mobile device software, nutritionist educator, and curriculum-based approaches.