Table 1—

Intervention studies

Author(s)Social disadvantageObjective(s)Study designSetting of interventionStudy populationMeasuresMain findings (intervention versus control groups)
Fanning et al. (2004) (20)HispanicTo compare treatment algorithms versus usual care for patients with type 2 diabetes.Comparative study. Three groups were assessed: 1) community clinic clinic following treatment, 2) university clinic following treatment algorithms, and 3) community clinic following usual care practices.Three community-based outpatient health care facilities in San Antonio, Texas, serving mainly low-income Spanish-speaking populations.A total of 106 participants were in group 1, 170 in group 2, and 82 in group 3; mean age 43–48 years, 41–60% Mexican American, 41–60% employed, and mean highest grade achieved 8–10.Primary end point was A1C. Secondary end points were fasting blood glucose, lipids, blood pressure, weight, and foot and eye examinations performed. Length of follow-up was 12 months.Treatment algorithm groups had greater reductions in A1C (group 1: 3.1% change; group 2: 3.3% change) versus control subjects (1.3% change) (P < 0.0001). Treatment algorithm groups had greater reduction in LDL cholesterol (P < 0.0001). No differences between the groups were noted for blood pressure or weight reduction. Treatment algorithm groups had more documented eye exams ( P < 0.0001) and foot exams (P < 0.0001) versus control subjects.
Rothman et al. (2004) (21)Low SESTo examine the role of literacy on the effectiveness of a comprehensive disease management program for patients with diabetes.Randomized controlled trial. Participants were randomly assigned into an intervention or usual care.University general internal medicine practice in North Carolina.A total of 105 participants were in the control group and 112 patients in the intervention group broken down into high- and low-literacy groups; mean age 51–56 years, ∼58% female, and more than one-third had low literacy (sixth grade level or lower)Literacy was assessed using the Rapid Estimate of Adult Literacy in Medicine. Outcome measures were A1C and systolic blood pressure. Length of follow-up was 12 months.Greater improvements in levels of A1C (−2.1%) versus control subjects (−1.2%) at 12 months (adjusted difference, −1.0% [95% CI −1.5 to −0.4]; P = 0.001). This difference was only significant among participants with low literacy (adjusted difference, −1.4% [−2.3 to −0.6]; P < 0.01). Greater improvement in systolic blood pressure (adjusted difference, −7.6 mmHg [−13.0 to −2.2]; P = 0.006). Differences were comparable for participants with low and higher literacy levels.
The California Medi-Cal Type 2 Diabetes Study Group (2004) (22)Low SESTo determine whether intensive diabetes case management using specific, population-directed, case management strategies could improve glycemic control in type 2 diabetesRandomized controlled trial. Participants were randomized to receive intensive case management or usual care.One community-based and two university-based sites in Santa Barbara, California; Medicaid, Los Angeles and San Diego counties.A total of 171 participants were in the intervention group and 146 in the control group; mean age 57 years, 71% female, ∼20% African American, 39% Hispanic, and 80% did not have education >12th grade.Primary end point: A1C. Secondary end points: weight, blood pressure, and lipids. The mean duration of follow-up was 25.3 months.Greater reduction in A1C (−1.88%) versus control subjects (−1.13% at 36 months) (P < 0.01 in every time period). Greater reduction in achieving specified A1C levels (≤6.5, ≤7.0, and ≤7.5%, P < 0.01). There were no statistically significant changes in secondary end points including weight, blood pressure, and lipid levels between the groups.
Davidson (2003) (23)African American and HispanicTo determine whether diabetes care directed by nurses following detailed protocols and algorithms and supervised by a diabetologist results in meeting evidence-based guidelines more often than under usual care.Comparative study. Participants were referred to the diabetes managed care program from two county clinics (clinics A and B). Clinic C patients were used as control subjects.Three county clinics in Los Angeles, California.A total of 252 participants were referred to clinic A, 252 to clinic B, and 209 to clinic C; 66–76% female, mean age 52 years, 2–19% African American, and 79–85% Hispanic.ADA process measures. The time period of the study was 2 years.Of 10 ADA process measures, 8 were significantly better performed among patients receiving nurse-directed care including A1C goal, lipid profile, eye exam, renal profile, foot exam, diabetes education, and nutritional counseling (all P values <0.01). Among patients followed for ≥6 months, A1C fell by 3.5% in the nurse-directed group and 1.5% in usual care (P < 0.001).
Echeverry et al. (2003) (24)Low SESTo determine whether a low-literacy diabetes educational reminder card would enhance process measures of diabetes care.Comparative study. Medical charts were reviewed for the intervention site and compared with diabetic patients seen on the same day at a nonintervention site.County clinics in Los Angeles, California, serving an underserved minority population.A total of 209 medical charts were reviewed in the intervention site and 218 in the nonintervention site; mean age 53 years, 67% female, and majority of the population was Hispanic and African American.ADA process measures including completed foot exam, urine protein, and lipid panel. Length of follow-up was 1 month.Intervention and nonintervention participants who met process measures were 48 and 95% for foot exam, 67 and 89% for urine protein, and 35 and 45% for lipid panel, respectively. No P values were reported.
Gary et al. (2003) (25)African AmericanTo determine whether a multifaceted, behavioral intervention implemented by an NCM and/or CHW could improve A1C and other indicators of diabetes control in type 2 diabetic subjects.Randomized controlled trial. Four arms: usual medical care (control), usual medical care + NCM, usual medical care + CHW, and usual medical care + NCM + CHM (combined team) interventions.East Baltimore, predominantly African-American inner-city community.A total of 34 participants completed the study in the usual care group, 38 in the NCM group, 41 in the CHW group, and 36 in the combined intervention group; mean age 59 years, mean level of education attained was grade 10, and 50% of the participants had a yearly income of ≤$7,500.A1C, LDL cholesterol, triglycerides, blood pressure, dietary risk score, leisure-time physical activity index, and BMI. Length of follow-up was 2 years.The combined team intervention group demonstrated a statistically significant improvement in triglycerides (−35.5 mg/dl, P = 0.041) and diastolic blood pressure (−5.6 mmHg, P = 0.042) from baseline compared with the control group. No other changes were significantly different between groups.
Miller et al. (2003) (26)African AmericanTo determine whether rapid-turnaround A1C availability would improve intensification of diabetes therapy and reduce A1C levels in patients with type 2 diabetes.Prospective controlled study. Patients were allocated by day of the week to a “rapid” group with A1C results at the time of the visit or a “routine” group.Neighborhood primary care clinic in Atlanta, Georgia.A total of 280 participants were enrolled in the routine group and 317 in the rapid group; mean age 61 years, 78% female, and ∼96% African American.Frequency of intensification of diabetes therapy; A1C levels. Average length of follow-up was 7 months.The average A1C level decreased from 8.4 to 8.1 in the rapid group (P = 0.04) and from 8.1 to 8.0 in the routine group (P = 0.31), but between-group results were not provided. Among participants with A1C ≥7%, providers intensified diabetes therapy in 51% of the rapid patients compared with 32% of the routine participants (P = 0.0003). There were no significant differences between the rapid and routine groups in the amount of change of medications.
Chapin, Williams, and Adair (2003) (27)Low SESTo develop and test an inexpensive visual tool, the THDR, to help patients with diabetes improve glycemic control.Prospective controlled study. Patients were assigned by the first letter of the last name of their primary care resident physician into either intervention or usual care groups.Primary care clinic in an inner-city neighborhood in Minneapolis, Minnesota.A total of 57 participants in the intervention group and 70 participants in the control group; mean age 52 years, 53% female, ∼60% had Medicaid and 20% had no insurance, 30% African American, 20% recent immigrant, and 5% Native American.A1C at least 3 months after the first use of the THDR. Length of follow-up was 15 months.Greater decrease in mean A1C versus control subjects (P = 0.047). A total of 51% achieved a decrease in A1C ≥0.9 vs. 18% of control subjects. The mean change in A1C was greater in patients who used the THDR four or more times.
Clancy et al. (2003) (28)Low SESTo evaluate group visits in the management of uninsured patients with uncontrolled type 2 diabetes.Randomized controlled trial. The intervention group received group visits and the control group received usual care.Adult Primary Care Centre at the Medical University of South Carolina in Charleston, South Carolina.A total of 120 participants: 77.5% African American; 78% female; average age 54 years (range 22–83); mean reported educational level 10.6 years; average literacy level grade 7.5; 44.3% had Medicaid, 19.1% had Medicare, and 30% had no insurance; and 23% reported working either full- or part-time.Trust in Physician Scale and Patient Care Assessment Tool and ADA standards of care indicators. A1C and lipids. Follow-up at 2 weeks, 3 months, and 6 months.No statistically significant differences were seen in diabetes or lipid control at 3 and 6 months. Statistically significant improvement was seen in compliance with the 10 ADA process-of-care indicators at the last group visit (P < 0.001). Significantly higher scores in the Trust Physician Scale (P = 0.02) versus control subjects.
Brown et al. (2002) (29)Mexican AmericanTo determine the effects of a culturally competent diabetes self-management intervention in Mexican Americans with type 2 diabetes.Randomized controlled trial. The experimental group was compared with a 1-year wait-listed control group who received usual care.Starr County, Texas, on the border of Mexico.A total of 126 participants were in the experimental group and 126 in the control group. Eligibility was aged 35–70 years, mean age 54 years. Most spoke Spanish at home and ∼50% read little or no English. Starr County is characterized by high unemployment (24.4%) and low per capita income ($8,225).Diabetes-related knowledge, health beliefs. A1C, fasting glucose, lipids and, BMI. Length of follow-up was 12 months.After controlling for baseline measures, A1C was significantly lower at the end of the study versus control subjects (P = 0.011). Lower fasting blood glucose and diabetes knowledge (P = 0.019 and <0.001, respectively versus control subjects). There was an increase in diabetes knowledge of 6.7 items correct (18%) on the diabetes knowledge scale (P < 0.001) versus control subjects. No statistically significant differences were observed between experimental and control groups in lipids, BMI, or health beliefs.
Keyserling et al. (2002) (30)African AmericanTo determine whether a culturally appropriate clinic- and community-based intervention for African-American women with type 2 diabetes increases moderate-intensity physical activity.Randomized controlled trial. Three groups: 1) clinic and community, 2) clinic only, and 3) minimal interventions.Five primary care practices in central North Carolina.A total of 200 participants: African-American women, mean age 59 years; ∼30% reported total household income of <$10,000/year; ∼30% were employed.The primary outcome was physical activity by Caltrac accelerometer for 7 days. Diabetes knowledge, dietary intake, diabetes health status instruments, A1C, total cholesterol, HDL, and weight. The length of follow-up was 12 months.The average 6- and 12-month difference in physical activity measured by kcal/day was 44.1 in group 1 compared with group 3 (P = 0.0055) and 33.1 in group 2 compared with group 3 (P = 0.029). At 12 months, physical activity was higher in group 1 compared with group 2 (P = 0.019) but not in group 2 compared with group 3 (P = 0.31). There was an increase of 1.6 points in diabetes knowledge in group 1, 1.2 points in group 2, and 0.7 points in group 3. The P value for the test of overall group effect at 6 and 12 months was 0.037. No statistically significant differences were found between the groups for dietary intake, total cholesterol, HDL, A1C and weight, and health status at 6 months.
Davidson et al. (2000) (31)Low SESTo evaluate care in a diabetes management program carried out by pharmacists in a general free medical clinic setting compared with usual care in the same clinic.Comparative study. The experimental group was referred to the pharmacist-run diabetes management program; the control group was randomly selected from diabetic patients not seen in the pharmacy clinic.Venice Family Clinic in Los Angeles County, California.A total of 89 participants in the experimental group and 92 participants in the control group. Patient representing the cohort selected: 92% no medical insurance, 15.6% homeless, 60.1% female, 64.3% Hispanic, 20.3% Caucasian, 9.5% African American, 2.5% Asian, and 0.6% Native American.A1C. The length of the study was 12 months.Greater prevalence of insulin use, diabetes complications, hypertension, and initial values of A1C versus control subjects (8.8 vs. 7.9). A1C improved significantly (0.8 reduction versus 0.05 reduction in the control group; P < 0.03).
Basch et al. (1999) (32)African AmericanTo evaluate a multicomponent educational intervention to increase ophthalmic examination rates.Randomized controlled trial. The intervention group received routine care plus the educational intervention; the control group received only routine care.General medicine clinics at five different sites in the New York City metropolitan area.A total of 280 African-American participants: 34% male, 70% unemployed, 50% high school graduates, 40% received Medicare and 20% received Medicaid, and 65% had a family income <$10,000/year.Documented retinal examination within 6 months of randomization.The examination rates across study sites were 54.7 vs. 27.3% in the control group. The odds ratio for examination status 6 months after randomization associated with the intervention was 4.3 (95% CI 2.4–7.8).
Agurs-Collins et al. (1997) (33)African AmericanTo evaluate a weight loss and exercise program designed to improve diabetes management in older African Americans.Randomized controlled trial. The intervention group consisted of group and individual sessions; the control group received usual care.Urban hospital in Washington, DC, other providers and local community.A total of 32 participants in the each of the intervention and usual care groups, 66–88% women, mean age 62 and 61 years (range 55–79), 56 and 59% high school graduates, and 59 and 56% not employed, respectively.Nutrition knowledge questionnaire. PASE, food frequency questionnaire, A1C, weight, and self-efficacy for diabetes management. The length of follow-up was 6 months.Mean weight at 6 months decreased versus an increase in control subjects, overall net difference from baseline of −2.8 kg for women (P = 0.007) and −2 kg for men (P = 0.26). Decreased A1C at 6 months, a net difference of −2.5% for women (P < 0.001) and −1.9% for men (P < 0.01) versus control subjects. Significant improvements in physical activity, nutrition knowledge, and dietary intake of cholesterol were observed at 3 months but decreased in magnitude at 6 months and were not statistically significant.
Elshaw et al. (1994) (34)Mexican AmericanTo assess the impact of a culturally specific, intensive diabetes education program on dietary patterns.Randomized controlled trial. The intervention group received education sessions; the control group did not.Clinics and community settings in rural boarder towns of Harlingen and Brownsville, Texas.A total of 31 Mexican-American men and 73 women age >50 years with type 2 diabetes; mean age 62 years for men and 60 years for women; low income area.Twenty-four-hour dietary recall; weight measured at 10 and 14 weeks.All groups experienced significant weight loss except for experimental female subjects. Weight loss between groups was similar. There was a trend toward decreased calorie intake in all groups over time. Cholesterol did not change.
Heath et al. (1987) (35)Zuni IndiansTo evaluate the effects of a community-based exercise program.Comparative study. Participants attended at least one exercise session; a matched comparison group was selected using a random-start method from the registry of patients with type 2 diabetes.Zuni reservation in western New MexicoA total of 30 participants and 56 nonparticipants: 79 and 80% female, mean age 42 and 44 years, and duration of diabetes 8 and 9 years, respectively.Weight, fasting blood glucose, and medication use. Study duration was 27 months.Mean weight loss of 4.1 vs. 0.9 kg among nonparticipants (P < 0.05). Decrease in BMI from 31 to 29.5 kg/m2 compared with no significant change among the nonparticipants. Between-group differences in weight were not significant for men. Participants attending sessions for the longest period (>52 weeks) showed the greatest amount of weight loss. Mean fasting blood glucose values dropped significantly from baseline (238 to 195 mg/dl) compared with an insignificant drop in the nonparticipant group (228 to 226 mg/dl). Participants were more than twice as likely to have decreased their medication as nonparticipants (RR 2.2 [95% CI 1.3–3.7]).
Mulrow et al. (1987) (36)Low SESTo determine whether an education program specifically designed for patients with type 2 diabetes and limited literacy could improve and sustain glucose and weight control.Randomized controlled trial. Three programs were tailored intervention, monthly groups, and single didactic lecture.St. Thomas’ Hospital Diabetes Clinic in central London, U.K.A total of 120 participants: mean age 53 years, 55% female, 50% West Indian, and mean education level 9 years.Weight, A1C, and lipids. Length of follow-up was 11 months.No statistically significant differences in changes in A1C, lipids, and weight were found among the three groups at the 11-month follow-up visit.
  • ADA, American Diabetes Association; CHW, community health worker; NCM, nurse case manager; PASE, Physical Activity Scale for the Elderly Questionnaire; RR, relative risk; THDR, Take Home Diabetes Record.