The Impact of Social Support on Outcomes in Adult Patients with Type 2 Diabetes: A Systematic Review

Burden of Diabetes

Diabetes is one of the most prevalent and fastest growing chronic illnesses, globally affecting more than 346 million people worldwide [1]. It is emerging as a global epidemic due to the rapid increase in overweight, obesity, and physical inactivity [1]. The combined impact of poor awareness, insufficient access, limited services, and inadequate resources makes diabetes the leading cause of blindness, amputation, and kidney failure worldwide [1]. Cardiovascular disease (CVD) accounts for 50%–80% of deaths in people with diabetes, a number expected to rise by more than 50% in the next 10 years [1]. For these reasons, diabetes has become a major cause of premature illness and death in most countries and is predicted to become the seventh leading cause of death in the world by 2030 [1]. Additionally, diabetes accounted for 11.6% of total health-care expenditure in 2010 [2]. Estimated global health-care expenditures to treat and prevent diabetes and its complications are expected to exceed US$490 billion by 2030; this equals approximately $561 billion (international dollars) by 2030 [2].

Currently in the U.S., 25.8 million people (or 8.3% of the population) have diabetes, and 7.0 million are still undiagnosed [3]. Nearly 1.9 million people 20 years of age or older were newly diagnosed with diabetes in 2010 in the U.S.; an estimated 79 million adults 20 years of age or older have prediabetes [3]. Men and minorities—non-Hispanic Blacks, Hispanics (Mexican Americans, Puerto Ricans, and Cubans), and American Indians/Native Americans/Pacific Islanders—are more affected, as compared with women and non-Hispanic Whites, respectively [4]. As has been observed globally, diabetes is the leading cause of complications, including heart disease and stroke, kidney failure, and nontraumatic lower-limb amputations [3, 4]. It is the cause of new cases of blindness among adults in the U.S. and, according to 2007 estimates, is the seventh leading cause of death in the U.S. [3, 4]. Regrettably, people diagnosed with diabetes have twice the risk of death of people of similar age without diabetes [3]. Given the vast numbers of people affected by this chronic disease, total estimated costs for diabetes in the U.S. are $174 billion annually [3].

Although preventable, type 2 diabetes (T2DM) accounts for 90% of all cases of diabetes worldwide [1, 3]. Evidence shows that at least 30 min of moderate exercise, along with healthy eating habits, aids in preventing and managing diabetes and diabetes-related outcomes [1, 3, 4]. These behaviors, in addition to other self-care activities such as home blood glucose monitoring, taking medications as prescribed, obtaining preventive services, and limiting alcohol intake and tobacco use, make living and adjusting to a life with diabetes more manageable. Self-management of T2DM, however, requires the assistance of multiple sources of support. In this article, we present the results of a systematic review of studies examining the effect of social support onT2DM and diabetes-related clinical outcomes.

Search Strategy and Eligibility Criteria

A reproducible strategy was used to search Medline/PubMed for articles published between January 2000 and June 2012. Four searches were conducted, yielding 1,023 articles. For all search strategies, social support was the main search term and was entered as the MeSH Major Topic and as a keyword in the abstract and title fields. For the initial strategy, the term health was entered; for the second, the term diabetes. Duplicates were removed, leaving 963 citations for assessment. Titles were eliminated if the research involved children, type 1 diabetes, or gestational diabetes. In addition, articles not written in English, not published in the U.S. (international), and using an international sample population were excluded. This produced 702 abstracts to examine for full article review. From this initial review, 33 articles of direct relevance to the systematic review (see inclusion criteria below) were selected for inclusion in this article.

An additional search, focusing on the last 2 years, was conducted to ensure a thorough review of the literature and inclusion of recent articles of interest and importance. Medline/PubMed was searched for articles published between January 2010 and June 2012; a total of 103 articles were found. After limiting search to the aforementioned time frame, humans, clinical trials/randomized clinical trials, English language, and adults, 19 years of age and older, 28 articles emerged. A more advanced strategy was employed, liming the search to U.S.-published only and eliminating articles where a specific social support measure was not identified. This resulted in nine additional items, which was narrowed further to 4 articles after removing duplicates.

The following inclusion criteria were used to determine eligible study characteristics: (1) The full article must have been published in English between 2000 and 2012; (2) the sample population had to include adults only with T2DM; (3) the studies could be cross-sectional, cohort, or intervention studies; and (4) the studies had to measure a change in a clinical (i.e., HbA1c or blood glucose, blood pressure (BP), lipids, mortality, etc.), psychosocial (i.e., depression, stress/distress), or behavioral (i.e., physical activity/exercise, diet/nutrition, self-monitoring/self-care/self-management, etc.) outcome.

Full articles were read and reviewed using a standardized checklist by the lead author (J.L.S.). The senior and corresponding author (L.E.E.) made the final decision regarding eligibility in case of uncertainty and indecision. Thirty-seven eligible studies were identified on the basis of the predetermined inclusion criteria.

Data Collection

Data collected from the eligible articles are shown in Tables 1 and 2 and have been placed in one of two categories: observational studies or interventional studies. Observational studies, including articles with cross-sectional and longitudinal designs, are shown in Table 1. Table 2 comprises studies that incorporate an interventional component and may have been designed as a randomized control trial (RCT), community experiment, or pilot study. The articles in both tables are arranged in descending chronological order by year of publication.

Data were extracted on the study design, social support measure(s), sample population, major findings, and study limitations. These tables include data from papers meeting inclusion criteria: social support, T2DM, indications for clinical outcomes (glycemic control, BP, lipid control, mortality), psychosocial effects (depression, stress), behavior adaptations (increased physical activity, adherence to a healthier diet), and perceptions of social support (increased, decreased, no support).

Social Support and Clinical Outcomes

On the basis of inclusion criteria of the database searches, 17 articles were reviewed that examined the impact of social support on clinical outcomes. Nine studies were cross-sectional, and 8 studies were interventional [21•, 22–26, 28••, 29•, 30–34, 35••, 36••, 37, 38]. Sample sizes ranged from 15 to 3,535 participants. In both designs, more studies suggested that higher levels of social support were associated with improved diabetes-related clinical outcomes (HbA1c, BP, lipids) [17, 22, 24, 26, 28••, 29•, 30, 33, 34, 35••, 36••, 37, 38]. This association held true regardless of the avenue for social support exchange or delivery (i.e., peer support, couples/spouse, and nurse manager) [22, 23, 34, 35••, 36••, 37, 38]. Three articles focused on diet support, while one (of the three) focused on exercise effects as well. Both of these behavioral-targeted support mechanisms improved clinical outcomes in diabetes management [24, 30, 33]. Furthermore, Epple and colleagues demonstrated that diet-related clinical outcomes (low-density lipoprotein [LDL], total cholesterol, triglycerides [TG], and HbA1c) were best when “other family members” (not the patient with T2DM) cooked the meals [33]. It should be mentioned as well that these “best outcomes” were observed in female participants [33]. Additionally, when the relationship between social support and mortality was examined, two studies were found. Both studies showed that higher levels of social support decreased mortality rates in patients with T2DM [21•, 25].

Contrary to those reports, however, three articles showed the opposite or no effect [24, 31, 32]. Two articles did not show an association between social support and clinical outcomes [31, 32]. In research exploring the relationship between perceived social support and CVD risk, McNeil and colleagues found that, despite having high levels of social support, no significant relationship existed between the two variables [31]. In another study, Chlebowy et al. were not able to show significance in the relationship between social support and self-care management or social support and glycemic control [32]. In addition, Chlebowy found that Blacks had less social support satisfaction [32]. In a similar study on race/ethnicity, social support, diabetes self-care, and clinical outcomes, Rees et al. did not find differences in social support by race, although Blacks were observed to have better clinical outcomes (decreased diastolic BP) and behavior modification (increased exercise, improved self-management) [28••]. In that same study, Whites had better LDL levels, and no significant clinical effects were observed in Latinos. Tang et al. demonstrated that negative support (and lower social support) increased the risk of medication noncompliance [24].

In this review of social support and clinical outcomes, there was strong evidence that higher levels of social support were associated with better clinical outcomes and behavior adaptations. Similarly, as observed in the literature, all articles in this review did not find a positive association between the two variables. Data from the opposing studies indicate that study limitations may have minimized the effects seen. For two of the studies, which were cross-sectional studies, small, homogeneous samples were studied [31, 32]. A third study, which was a randomized control trial of 200 participants, did show that there was a significant overall reduction in the mean HbA1c levels from baseline to 6 months but that there were no significant differences between groups [35••]. This outcome may have been due to study attrition, study design, and inherent biases.

Social Support and Psychosocial Outcomes

Two articles were reviewed that examined the impact of social support and psychosocial effects [17, 26]. One study was cross-sectional, and the other was interventional. Sample sizes ranged from 62 to 86 participants. Major findings were consistent across both studies, with participants having higher levels of social support experiencing fewer depressive symptoms and diabetes-related symptoms. In an international study that was excluded from the review, researchers concluded that participants with higher social support and, therefore, less depression developed better diabetes self-management skills and improved certain clinical outcomes (body mass index and TG) [39]. Despite the smaller sample size, Bond and colleagues were able to determine that Web-based interventions were effective in sustaining psychosocial well-being in older patients with diabetes [17].

However, given that the sample size of both of these studies was small, it is imperative that these studies be reproduced in larger, heterogeneous populations to ensure valid and reliable results. In addition, one of the studies was a recent randomized control trial using a Web-based intervention [17]. Findings demonstrated the effectiveness of sustained psychological well-being with the intervention. Given that result, Web-based interventions should be considered in those with lower social support and more positive psychological needs.

Social Support and Behavioral Modification

Eight cross-sectional studies and five interventional studies were evaluated to assess the role of social support on life-style modification [16, 18••, 20, 27, 40–46, 47•, 48••]. For these 13 studies, sample sizes ranged from 21 to 1,788. As was observed with social support and clinical outcomes, 7 studies demonstrated a positive association between social support and behavioral outcomes [16, 20, 40–42, 47•, 48••]. From a sample of 1,788 older adults with T2DM, Nicklett and colleagues found that increased diabetes support increased the likelihood of adherence to healthier outcomes (i.e., diabetes regimen)—thus, a negative association with health decline [20]. They added that the duration of diabetes diagnosis and the burden of illness were associated with a positive decline in health [20]. Within different study populations, Rosland et al., Wen et al., and Shaw et al. all demonstrated that family support improved glucose monitoring, diet self-management and exercise, and diabetes self-care, respectively [40–43]. Of importance, Rosland also showed that support from nonphysician health care professionals increased the likelihood of meal planning and preventive services, such as checking feet [40]. In a sample of older Mexican-American patients with T2DM, Wen found that increased family support decreased perceived barriers to diet self-care management [41, 42]. This was particularly true in older patients who lived with family members [42]. Participants also determined that higher social support was an indicator of positive family functioning [41]. Similarly, in a study conducted by King and colleagues, support from family, friends, and the community motivated participants to adopt healthier eating habits and increase physical activity [47•]. Having the support of the social environment increased the likelihood of diet and exercise adherence [47•]. Using an integrative health (IH) coaching intervention, Wolever et al. showed that participants in the IH group had increased perceptions of social support and increased self-report of medical adherence, physical activity, and perceived health status [48••]. As compared with participants in the usual care (control) group, individuals in the IH group reported less stress and improved clinically (decreased HbA1c). Although no physiologic outcomes changed in their study, McEwen et al. reported that social support improved self-care behaviors, increased diabetes knowledge, and reduced diabetes-related distress [16]. Results by Gleeson-Kreig and colleagues opposed previous findings, stating that social support was not related to self-management behaviors [27]. In an online community support group designed to facilitate increased walking patterns in patients with T2DM, Richardson et al. failed to demonstrate changes in behavior [46]. They were able to report, however, that those with lower baseline social support used the online community more by posting requests and viewing supportive statements. Given the recent advances in technology and measurable barriers to health care (i.e., access, transportation, finances, language, etc.), online communities may be of considerable interest when working with populations exhibiting lower levels of support.

As was detected in social support and clinical outcomes, results demonstrating the impact of social support on behavioral outcomes was also consistent with the literature. Major findings supported the hypothesis that higher levels of social support were associated with increased diabetes self-management and self-care behaviors. One study did find that social support was not related to self-management [27]. This study was conducted in a homogeneous sample of urban participants who lived at home with other relatives; this situation likely diminished the effect and limited the generalizability of the study.

Perceptions of Social Support

Five articles—three cross-sectional and two interventional—discussing the perceptions of social support were included for discussion [19, 49–52]. As was stated previously, the manner in which social support is perceived often varies from actual social support received [51]. This interpretation or receipt of support can differ on the basis of numerous factors, including gender, race/ethnicity, culture, or social environment. One study included in this review assessed the primary source of support in couples [18••]. In that study, males reported their female spouses as primary sources of support; the same was not true for females [18••]. In contrast, females rated nonspouse sources as their primary foundations for support [18••]. The source (or availability) of the support, the amount of the support, and manner of delivery may also contribute to perceptions of support [19, 52].

In a study of over 700 English- and Spanish-speaking U.S. patients with T2DM from four urban public hospital systems, Sarkar and colleagues found racial/ethnic differences in preference for self-management social support modes of delivery [49]. Urban Hispanics in this sample preferred telephone and group medical visits, as compared with Whites residing in urban areas [49]. As compared with Whites, Blacks, on the other hand, did not have a specific preference in delivery mode [49]. They would accept self-management support via telephone, group medical visits, and the Internet. In a small sample of White men and women with T2DM, Gleeson-Kreig found a preference for support by the media, followed by health care professionals, personal, and work [50]. Support from family, friends, and community members were least acceptable modes of delivery for this sample population [50]. Contrary to the findings by Gleeson-Kreig and associates, Shaw et al. demonstrated a higher preference for community, neighborhoods, and community organizations in adults living with T2DM in upstate New York [43]. Overall, this section has been included, in particular, to draw attention to differences in sources of support as received from different study populations, in the hope that tailored studies and interventions may be created that directly target areas of greatest need.