Table 3:

Summary of articles focused on Social and Community Context

Author/Year	Study Design	Objective	Number of Participants	Sample population	Setting	Impact on Outcome
Knight, 2008 (59)	Cross-sectional	Using screening by student pharmacists to identify depression, undiagnosed and under treated in patients with T2DM and to understand influence on DM control	45	Adults with primarily T2DM seen urban primary care clinic	5 Cincinnati Health Department Clinics	No significant differences were found on mean A1c or fasting blood glucose between depressed and non-depressed individuals.
Egede, 2010 (40)	Cross sectional	Assess differences in metabolic control and health-related quality of life among depressed and non-depressed adults in indigent population with type 2 diabetes	201	Low income adults with type 2 diabetes	Academic medical center in Southeastern United States	Depressed compared to non-depressed patients had a lower SF-12 PCS (36.1 vs. 39.0, p≤0.001) and MCS (41.6 vs. 46.8, p≤0.001); there was no significant difference in LDL and HbA1c between these two groups
Moayad, 2006 (68)	Cross-sectional	Evaluate effect of family cohesiveness, acculturation, socioeconomic position, and CVD risk factors on DM severity	275 (n=193 for patients deemed to have severe DM)	Mexican Americans with T2DM	Two medical clinics in north Fort Worth, TX for 2 years	HbA1c used to define disease severity (≥7=severe; <7=non-severe). Acculturation and family cohesiveness were not statistically associated with disease severity. Factors associated with disease severity: receiving food stamps (adjusted OR 3.8, 95% CI 1.5-9.6), having spent childhood in Mexico (adjusted OR 1.2, 95% CI 1.1-1.4), and smoking status (adjusted OR 3.2, 95% CI 1.1-10.0).
Venkatesh, 2013 (86)	Mixed methods	Examine impact of acculturation on glycemic control	30	1st generation Asian Indian (AI) adults with T2DM	Participants recruited from community (AI organizations, stores, restaurants, and listservs) in a mid-western state (MI)	Statistical difference in mean HbA1c (6.4±0.3 intervention, vs. 7.9±0.7 control; p<.01). Interactions among acculturation and annual household income, BMI, and DM duration significantly predicted higher HbA1c levels (R2 change=.368; F change=4.21; p=.02).
Osborne, 2010 (70)	Cross-sectional	To examine the relationship between health literacy, determinants of DM self-care and glycemic control	130	Adults with T2DM seen in local primary care clinic	University Internal Medicine Clinic of the Medical University of South Carolina	Glycemic control and DM self-care was indirectly effected by health literacy (r= −0.01), (r= −0.07)
Okuson, 2002 (69)	Secondary data analysis using the Third National Health and Nutrition Examination Survey (NHANES III)	To compare racial/ethnic differences in diabetes awareness, treatment and glycemic control between NHW, NHB, and HW Americans	6334	Individuals who were NHW, NHB, or HW, ≥40 years old and had data available for weight, height, waist circumference, had health insurance, and HbA1c values	N/A	NHB and HW men were 0.73 [CI 0.45-1.17] and 0.45 [95% 0.29-0.71] less likely than NHW to have poor glycemic control than NHW men; NHB and HW women were 0.39 [95% CI 0.26-0.61] and 0.43 [0.26-0.70] less likely than NHW to have poor glycemic control than NHW women
Toobert, 2007 (85)	Randomized controlled trial-intervention Mediterranean Lifestyle Program Duration: 24 months	To examine the long-term effects of interactions between lifestyle behaviors, psychosocial factors and social environment	270	Postmenopausal women with type 2 diabetes	N/A	Quality of life was not statistically significant between the intervention and usual care groups (F=3.45, p<0.1); the direct and indirect cost of the intervention group was $409,165 compared to $309,302 for the usual care group
Kirk, 2011 (58)	Chart review and cross-sectional survey	Examine the relationship between patient variables and glycemic control (HbA1c <7%) in a cohort of family medicine patients with type 2 diabetes	669	Individuals were selected if their chart had an International Classification of Diseases, Ninth Revision, Clinical Modification codes for diabetes	University-based family medicine teaching site	Of the study population, 43.2% had an HbA1c > 7%, 36.3% had an LDL < 100 mg/dl, 44.7% had a systolic blood pressure <130mmHg, and 69.1% had a diastolic blood pressure < 80mmHg
Fortmann, 2011 (42)	Cross-sectional path analysis	Investigate value of support resources in disease management	208	Latinos with type 2 diabetes	Low-income serving community clinics in San Diego	Path analyses revealed that greater support resources were linked to better self-management (p<0.001) and less depression (p<0.01). Once these factors were statistically controlled the pathway from support resources to HbA1c was markedly reduced (p=0.57)
Gary, 2000 (43)	Cross-sectional	Determine prevalence of depressive symptoms and relationship between depressive symptoms and metabolic control	183	African American adults aged 35-75	Two primary care clinics in Baltimore, MD	Depressive symptoms are marginally associated with suboptimal levels of HbA1c (p=0.104), blood pressure (p=0.073), and LDL (p=0.176), and significantly associated with suboptimal levels of total cholesterol and triglycerides (p<0.05).
Bird, 2007 (31)	Cross-sectional	Assess gender differences in quality of care	4,054	Enrollees in commercial and Medicare health plans	National sample	Gender differences found in variety of quality of care measures. Statistically significant differences in intermediate outcomes for diabetics includes lipid control for Medicare enrollees (p<0.01) and commercial enrollees (p<0.01), and blood pressure control for commercial enrollees (p<0.05). Relative risk for women relative to men for LDL was 0.84.
Chiu, 2010 (36)	Longitudinal cohort	Longitudinal association between depressive symptoms and glycemic control and extent to which it is explained by health behaviors	998	Adults with diabetes	National sample from Health and Retirement Study	Adults with higher levels of depressive symptoms at baseline showed higher HbA1c levels at 5 year follow-up. Health behaviors account for 13% of link.
Culica, 2007 (37)	Quasi-experimental	Evaluate cost and impact of educational model delivered to predominantly Spanish speaking Mexican American population with type 2 diabetes	162	Adults diagnosed with diabetes and no advanced complications	Clinic in Dallas, TX	Change in HbA1c at 6 mo was not significant, but change from baseline at 12 mo was significant (p<0.01). Blood pressure did not change at 6 or 12 mo.
Hayes, 2000 (50)	Prospective cohort within trial to assess provider adherence to diabetes management protocol	Examine applicability and relationship of locus of control to glycemic control in low-literacy, economically deprived population	312	African Americans with type 2 diabetes referred to diabetes unit of public hospital	Urban public hospital in Georgia	No significant association between locus of control and baseline HbA1c. Significant relationship between belief in chance and glycemic control at 6-mo follow up (p<0.05) and change in HbA1c (p<0.05). Patients with stronger believe in change had higher HbA1c levels and less change in HbA1c.
Hernandez-Tejada, 2012 (51)	Cross-sectional	Examine relationship between perceived control of diabetes and qualify of life	188	Adults with type 2 diabetes having been seen in clinic	Low-income clinic in Southeast US	Positive association between perceived control and quality of life (p<0.05) when controlling for sociodemographics and comorbidity.
Kacerovsky-Bielesz, 2009 (56)	Cross-sectional	Examine sex-specific differences and influences of social and psychological factors on glucometabolic control.	257	Adults ages 40-80 diagnosed with diabetes and seen at diabetes outpatient clinic at least twice in 6-mo	Metropolitan diabetes outpatient clinic	Mean HbA1c did not differ between genders. Variables predicting high HbA1c differed by gender. For women: pain (p=0.003), social isolation (p=0.008), internal control (p=0.039), external control (p=0.019), and depressive coping (p=0.031). For men: emotional support (p=0.007), satisfaction with support (p=0.054), and instrumental support p=0.066)
Wexler, 2006 (89)	Cross-sectional	Assess impact of medical comorbidities, depression, and treatment intensity on quality of life	909	Adults diagnosed with diabetes with continuous care	Two outpatient primary care clinics and hospital based primary care practice in Massachusetts	Emotion, pain and impaired vision were predominant drivers of overall health utility. Patients with depression had the lowest quality of life (p<0.0001) and differed by comorbidity. After adjustment the utility penalty diminished to some degree. Presence of depression was strongest correlate in multivariate model, decreasing utility by 0.37.