Social determinants of type 2 diabetes and health in the United States

Table 1 Summary of reviewed studies

Ref.	Purpose	Design/method	Sample	Findings	Strengths/limitations/implications
Chaufan et al[25] (2011)	To examine “upstream” social determinants of health in a Latino immigrant population focusing on T2DM risk and food environment	Mixed methods focus groups, Food store survey instrument California	Staff members (n = 6); clients (n = 15)	Poverty prevented Latino immigrants from access to adequate housing, high quality education, and culturally appropriate food choices	Community-based interviews elicited lived experience of Latino immigrant population Limitations: small sample size, nonrandomized, descriptive study Implications: Policy change to enable access to affordable culturally acceptable healthy foods
Carbone et al[26] (2007)	To describe factors influencing Latino (Puerto Rican) diabetes self-management	Qualitative-focus groups Health center in Massachusetts, United States	Healthcare providers (n = 15) Patients (n = 37)	Cultural influence of family through traditional gender roles; financial restraints Social network (family and friends) source of strength; spirituality and faith important; discrepancy between healthcare provider and patient focus in self-management	Community-based limitations: Small sample size, qualitative study, not generalizable to all Latino populations Implications: account for inclusion of family and social networks in diabetes self-management education, need for increased cultural awareness by Healthcare provider of influencing factors on patient self-management
Chaufan et al[27] (2011)	To gain understanding how food environments influence a low income immigrant population	Qualitative-focus groups	Staff (n = 6) Clients (n = 15)	Barriers to food access: transportation, language barriers, poverty, employment	Strengths: community based Limitations: small sample size, not generalizable Implications: Need for policy to increase food access for immigrant populations in lower socio-economic levels
Tjia et al[28] (2008)	To gain understanding of barriers to medication adherence among older adults with T2DM	Qualitative-semi structured interviews	Adults over age 65 (n = 22) Female n = 16	Older adults with T2DM had concerns about medication cost or medication burden but did not discuss with physician	Strengths: perspective of older adults Limitations: small sample size, not generalizable Implications: need for increased patient-physician communication; policy level for messages to encourage open patient-physician dialogue for medical treatment
Denham et al[29] (2010)	To explore perceived patient barriers to diabetes education among healthcare providers	Cross-sectional survey	Healthcare providers from three practice settings: federally qualified healthcare centers; health departments; clinics (n = 182)	Perceived barriers: transportation; fewer diabetes educators and physicians; lack of insurance; education materials not screened for literacy levels nor cultural appropriateness	Strengths: look at perceived barriers to diabetes education among healthcare providers Limitations: small sample size, multiple sites, sample bias Implications: policy need to increase diabetes education reimbursement to providers for all individuals nationally; need to screen materials for literacy levels
Heuer et al[30] (2006)	To describe Hispanic migrant farmworkers perceptions of diabetes	Qualitative phenomenological study	Migrant farmworkers (n = 12) Female n = 6	Cultural/folk beliefs that diabetes was caused by stress or emotions	Strengths: focused on hispanic explanatory model of diabetes Limitations: small sample size, not generalizable Implications: need to tailor diabetes education to address culture and health beliefs
Shigaki et al[31] (2010)	To examine how patients diagnosed with type 2 diabetes view role of nurses in disease management	Qualitative Semi-structured individual interviews	n = 13 Female n = 7 White n = 9 African American n = 3 Other n = 1	Nurse viewed as a positive partner in disease management, patients prefer team-based medical care, open communication between healthcare providers and patient	Strengths: Patient perspective on nurse as partner in healthcare Limitation: small sample size; sample bias Implications: Nurse partner may provide link between patient and improving health outcomes in individuals with multiple co-morbidities including diabetes
Fitzgerald et al[32] (2008)	To determine diabetes care perceptions in patient and provider	Quantitative cross sectional survey	Providers n = 71 Patients: n = 273 female 61% White 63% African American 33%	Patient and provider differences included patients with more positive attitudes family support, paying for diabetes care	Strengths: quantitative study using reliable and valid instrument Limitations: may not be representative for patients with no co-morbidities Implications: Importance of communication between patient and provider and not making assumptions about patients
Zulkowski et al[33] (2005)	To examine differences between patient self-report and provider documentation in rural health center	Quantitative cross sectional survey; medical record review	n = 149 Female n = 86	Statistical difference between patient and provider on medical management or diabetes knowledge; most patients had not been seen by diabetes educator or dietician	Strengths: good response rate (61%); limitations: race not documented Implications: Need for increased access to diabetes education in rural areas; need for open communication between patient and provider
Ford et al[34] (2002)	To determine relationship between socio-economic status and race on diabetes management	Quantitative-random sampling by socio-economic status and race	n = 50 African American n = 21 Caucasian n = 29 Female n = 22	Differences in perception of sense of loss by race but not socio-economic status	Strengths: Limitations: small sample size: not generalizable; underpowered (α = 0.34) Implications: Need for education to stress management to prevent disease related complications
Mani et al[35] (2011)	To determine the influence of social networks on diabetes self-management	Cross-sectional survey	n = 154 Female n = 88 White n = 78 African American n = 61	Diabetes concern increased when larger social network diagnosed with diabetes	Strengths: patient perspective of importance of social networks Limitations: not generalizable-sample from two inner city clinics, only English speaking, data skewed Implications: Important to consider the influence of social networks on diabetes self-management
Song et al[36] (2012)	To examine social networks among Korean Americans diagnosed with type 2 diabetes	Cross-sectional survey from a larger study	n = 83 Female n = 35 Married n = 73	Gender differences in source of support, men sought spouse, women had higher unmet needs for social support than men, self-efficacy negatively associated with unmet social support, education level strong predictor of self-care activities, unmet needs for social support negatively associated with diabetes self-care	Strengths: patient perception of social support Limitations: small homogenous sample; not generalizable Implications: need to recognize the role of gender in determining social support among Korean Americans
Kollannoor-Samuel et al[37] (2011)	To identify influence of social determinants of health on FPG and HbA1c among low income Latinos	Cross-sectional	n = 211 Female n = 155 Puerto Rican n = 171 Unemployed n = 178 Spanish speaking only n = 138	Lower socio-economic status had higher FPG and HbA1c levels; better long term glycemic control when insured; increased physical activity associated with lower FPG and HbA1c levels	Strengths: sample size; part of larger RCT Limitations: large number of female participants-not generalizable or transferable to other Latino populations; unknown length of time since diagnosis Implications: Importance of physical activity for improved glycemic control; need for health insurance, increase in education to improve glycemic outcomes over the long term
Chiu et al[38] (2011)	To determine how gender influences functional limitations with type 2 diabetes	Secondary data analysis	n = 1619 Female n = 861	Psychosocial factors mediator between biological and exercise factors	Strengths: consistent with previous research Limitations: participant self report; no causal pathway Implications: Need to incorporate gender and psychosocial factors in diabetes treatment plan
Iida et al[39] (2010)	Examine role of spousal support for individual diagnosed with type 2 diabetes	Mixed methods-longitudinal Computer diary survey; individual interviews	n = 129 married couples Caucasian 75%; African American 23.6%	Physical symptoms increased spousal support; women gave higher level of support when spouse worried about diabetes; less spousal support when negative emotional effect from previous day	Strengths: sample size Longitudinal study, diary Limitations: unable to determine causal effect Implications: Need to measure and include spousal support in diabetes self-management plan
Mayberry et al[40] (2005)	To assess determinants of glycemic monitoring by primary care provider among medicaid beneficiaries	Retrospective cohort study	n = 3321 African American (2025) 61%; White (1296) 39%; Female 74.6%; Urban 49.6%; rural 50.4%	African Americans diagnosed at younger age; frequency of physician visits and medication prescription strong predictor of meeting ADA recommendation guideline for HbA1c monitoring; low monitoring of HbA1c and FPG among medicaid beneficiaries with type 2 diabetes, no significance between Black and White Medicaid beneficiaries for diabetes monitoring	Strengths: provides insight into determinants of glucose monitoring in Medicaid beneficiaries Limitations: Secondary data analysis; inability to determine patient-physician relationship Implications: Need to establish ADA evidence guidance protocol in office practice settings
Ford et al[41] (2005)	To determine estimates of obesity and diabetes weight information from healthcare providers in 100 United States metropolitan service areas	Cross sectional survey: 2000 Behavioral Risk Factor Surveillance System	n =81917 individuals with BMI data Sample size varied due to missing data	Highest prevalence of obesity Appalachian region; OR of being diagnosed with diabetes 3.5 higher in Charleston WV than Santa Fe NM; weight loss or maintenance discussed with physician 11.7%-34.6%	Strength: study at community level Limitation: secondary data analysis, missing data; self report Implications: Policy level- need to examine determinants of obesity and diabetes at local metropolitan service areas. Healthcare providers need to include discussion of weight management for individuals diagnosed with obesity or diabetes
Adams et al[42] (2008)	To examine medication adherence as a factor of glycemic control based on race	Secondary data analysis newly diagnosed patients prescribed oral medications	n = 1806 Black n = 467 White n = 1339	Black patients higher A1c at diagnosis; insufficient evidence to determine medication adherence by race	Strengths: limitations: unable to determine causal effect; potential overestimation of medication adherence based on data; implications: Need at policy level to provide screening for earlier diagnosis of Black and female patients
Aikens et al[43] (2005)	To examine the relationship of patient-provider communication on diabetes self-management and outcomes	Cross-sectional Telephone survey	n = 736 White 51% Black 20% Hispanic 11.9% Female 31.6%	Primary care provider main manager of diabetes (70.9%) General patient-provider communication related to improved quality of life; diabetes specific patient-provider communication related to glycemic control	Strengths: sample across three health systems; limitations: cross-sectional study; self-report; implications: need for open patient-provider communication to enhance problem solving, need to tailor management plan to individual patient
Paris et al[44] (2001)	To determine if determinants of type 2 diabetes were present in personnel entering the military	Cross-sectional secondary data analysis	Diagnosed with diabetes n = 419 Not diagnosed with diabetes n = 627	Enlisted military rank as socioeconomic measure for diabetes diagnosis higher in lower rank, minorities higher level of diabetes; educational not significant variable	Strengths: adequate power Limitations: no causal effect; potential for misclassification of diabetes diagnosis in database; implications: need for focus on policy level for increased physical activity and body mass index monitoring

FPG: Fasting plasma glucose; HbA1c: Glycosylated hemoglobin A1c; BMI: Body mass index; T2DM: Type 2 diabetes mellitus; RCT: Randomized control trial; ADA: American diabetes Association.