Multilevel Interventions Can Enhance Diabetes Outcomes: Mid-term Results from the Alliance to Reduce Disparities in Diabetes Megan A. Lewis1, Connie L. Hobbs1*, Shawn Karns1, Joe Burton1, Jeffrey Brenner2, Patria Johnson3, Kathy Langwell4, Monica Peek5, James Walton6, Noreen M. Clark7, and Douglas Kamerow1

1RTI International, 2Camden Coalition of Healthcare Providers, 3Memphis Healthy Churches, 4Sundance Research Institute, 5University of Chicago 6Baylor Healthcare System, 7University of Michigan

The Alliance to Reduce Disparities in Diabetes integrates innovative professional and patient education and quality of care improvements aimed at vulnerable patients. The Alliance programs focus on reducing disparities in diabetes care and enhancing diabetes outcomes through clinical and community interventions. The Alliance comprises five grantees, a National Program Office, and an external evaluator. The five sites enrolled a multiethnic and multiracial patient population with diabetes and implemented multilevel and multicomponent interventions to enhance patient skills, clinician cultural competencies, and health care systems changes to address disparities and enhance care. Each grantee site provided clinical outcomes (hemoglobin A1c [HbA1c], blood pressure [BP]) and patient-reported outcomes (diabetes competence, quality of life, resources and supports for self-management, and diabetes self-care behaviors) for program participants to the Alliance’s external evaluator, RTI International. At midterm of the 5-year program, participant cohort data show that 64% of participants are female; 7% are White, 33% are African American, 3% are Native American, 50% are Latino, 1% are Asian or Pacific Islander, and 6% are Other or unknown race. Among participants, 29% are aged 18−44, 33% are 45−54, and 38% are 55 or older. Baseline and 1-year follow-up clinical data are reported for more than 1,400 patients with type 2 diabetes, with a decrease of mean HbA1c values from 8.4% to 7.9% (p < .01). Multivariable regression analysis also showed that patients who participated in more than half of the program had greater changes in both HbA1c values and BP over time, controlling for race, age, and gender. Analyses of the patient-reported survey measures also showed significant improvements in perceived diabetes competence, mental functioning related to quality of life, resources and supports for self-management, and self-care behaviors.

The Alliance to Reduce Disparities in Diabetes integrates innovative professional and patient education and quality of care improvements aimed at vulnerable patients. It focuses on reducing disparities in diabetes care and enhancing diabetes outcomes through clinical and community interventions. The Alliance, sponsored by The Merck Foundation, is a consortium of five grantees, a National Program Office based at the University of Michigan, and an external independent evaluator, RTI International. To address disparities and enhance care, the grantee sites have enrolled a multiracial patient group and implemented multilevel and multicomponent interventions to enhance patient skills, clinician cultural competencies, and health care systems changes.

The five grantee sites include:

■ Improving Diabetes Care and Outcomes on the South Side of Chicago, University of Chicago, Illinois

■ Camden Citywide Diabetes Collaborative, Camden, New Jersey

■ Diabetes for Life Program, Memphis, Tennessee

■ Reducing Diabetes Disparities in American Indian Communities, Wind River Indian Reservation, Wyoming

■ The Diabetes Equity Project, Dallas, Texas

Interventions

Although interventions may vary across the sites, the overall focus is on three core components: patients, clinicians, and systems change.

■ Patient education included community, small group, and individual materials, classes, and discussions.

■ Clinician education included training in cultural competency, communication, and facilitation of behavior change.

■ Diabetes management via systems change included care coordination, use of diabetes registries, nurse or community health worker participation in care management, enhanced community partnerships, and policy changes.

As of April 2013, 7,618 encounters took place across the five sites with 1,488 patients, in which they discussed an average of three to four diabetes self-management topics (Table 1).

Table 1. Number of Times Diabetes Self-management Topics Were Addressed

Method

RTI was selected to conduct a cross-site evaluation of the Alliance. We collected clinical and patient-reported data from the five grantees six times over the past 3 years and conducted three site visits (one virtual) to document the interventions that were implemented. To understand the net effect of the Alliance programs on diabetes and health outcomes, all data reported are aggregated across the grantee sites.

Measures

Each grantee site provided clinical (HbA1c, BP) and patient-reported outcomes for program participants.

Patient-reported measures collected by two or more grantee sites included:

■ Perceived diabetes competence1: Average of 4 items, with higher scores indicating greater confidence in managing diabetes

■ Resources and supports for self-management2: Average of 6 items, with higher scores indicating more support from one’s health care team in learning how to manage diabetes

■ Quality of life (Veterans RAND 12 Item Health Survey [VR-12])3: 12 items split into two subscales indicating mental and physical functioning (scored via algorithm), with higher scores indicating better functioning

■ Diabetes self-care behaviors4: Average of 2 items for each of five behaviors, with higher scores indicating more frequent self-care behaviors

Program Participants

Table 2 presents the demographic characteristics for program participants who currently have baseline and follow-up measures, and are included in the cohort analysis. The average time between baseline and follow-up measures is 1 year.

Table 2. Demographic Characteristics of Program Participants (N = 1,488)

Results for Clinical and Patient-Reported Outcomes

For the participant cohort, we used t-tests and multivariable regression analyses to understand how health and diabetes outcomes changed over time spent in the program. Table 3 presents the descriptive statistics for the baseline and follow-up clinical measures, and indicates significant differences. For example, HbA1c improved from baseline (first measure) to follow-up (last measure). Table 4 presents similar information for patient-reported outcomes. All measures show improvement from baseline to follow-up (p < .05), except for quality of life measured by physical functioning.

Table 3. Changes in Clinical Measures Across All Grantees

Table 4. Changes in Patient-Reported Measures Across All Grantees

To further understand the potential impact of the grantee programs on changes over time, we regressed each of the changes in clinical measures on selected characteristics, including age (under 55 years vs. 55 or older), gender (male vs. female), and race (any underrepresented race vs. White), and program participation status (high intensity vs. low intensity). High-intensity program participants attended over half of the program sessions of the Alliance programs; low-intensity program participants attended less than half of the sessions. The results are shown in Table 5.

Table 5. Results of Clinical Measures Regressed on Patient Characteristics and Program Participation Status

1. Williams, G C., McGregor, H.A., Zeldman, A., Freedman, Z.R., & Deci, E. L. (2004). Testing a self determination theory process model for promoting glycemic control through diabetes self-management. Health Psychology, 23(1), 58–66.

2. McCormack, L.A., Williams-Piehota, P.A., Bann, C.M., Burton, J., Kamerow, D.B., Squire, C., … Glasgow R. E. (2008). Development and validation of an instrument to measure resources and support for chronic illness self-management: a model using diabetes. Diabetes Education, 34(4), 707–718.

3. Kazis, L.E., Miller, D.R., Skinner, K.M., Lee, A., Ren, X.S., Clark, J.A., … Fincke, B.G. (2006). Applications of methodologies of the Veterans Health Study in the VA Health Care System: Conclusions and summary. Journal of Ambulatory Care Management, 29(2), 182–188.

4. Toobert, D.J., Hampson, S.E., & Glasgow, R.E. (2000).The Summary of Diabetes Self-Care Measure: Results from 7 studies and a revised scale. Diabetes Care, 23(7), 943–950.

Abstract 1. Background 2.Cross-site Evaluation 3. Preliminary Results 3. Preliminary Results (continued)

References

Although this project is not a randomized controlled clinical trial, midway through the grantee programs, comparisons of clinical and patient-reported measures from baseline to follow-up show improvement across multiple measures, including both clinical and behavioral outcomes. This preliminary evaluation suggests that by using multilevel, multifocal interventions that include patient, provider, and system components, diabetes outcomes can be improved and disparities potentially decreased in groups that are most burdened by diabetes management.

The next step in the cross-site evaluation will be an analysis of clinical data collected from a comparison cohort. This will allow us to determine with more confidence whether disparities in diabetes were reduced by the program interventions. Each site will also be providing more clinical outcome and patient-reported outcome data as the programs conclude in late 2013.

4. Discussion

5. Next Steps

AcknowledgmentsWe would like to thank The Merck Foundation for funding the Alliance to Reduce Disparities in Diabetes; Leslie Hardy, Vice President of The Merck Foundation; and National Program Office members Julie Dodge and Belinda Nelson. We would also like to thank the additional members of the RTI project team, including Pam Williams, Rebecca Moultrie, Tania Fitzgerald, Sidney Holt and Olivia Taylor, as well as Denise Charron-Prochownik from the Alliance National Advisory Board.

More Information*Presenting author: Connie L. Hobbs Research Public Health Analyst, Health Communication919.990.8398 connieh@rti.org

RTI International 3040 Cornwallis Road Research Triangle Park, NC 27709

Presented at: The 73rd Scientific Sessions: American Diabetes Association, Chicago, IL, June 21-25, 2013

www.rti.org RTI International is a trade name of Research Triangle Institute.

Exposure TopicsHbA1c, BP, and cholesterol education 183Complications 1,021Coping 1,964Decision making/Goal setting 153Diabetes basics 1,448Doctor visits and self-exams 201Glucose monitoring/Glucose targets 6,902Healthy eating/Food planning 6,546Medication 6,143Physical activity 2,599Smoking and alcohol 1,532Total 28,692

Characteristic N Unweighted %Gender

Male 518 35.8Female 929 64.2

Race/EthnicityWhite 105 7.1African American 487 32.7Native American 45 3.0Hispanic 741 49.8Asian/Pacific 19 1.3Other 17 1.1Unknown 74 5.0

Age18−44 418 28.645−54 486 33.255+ 559 38.2

Regression Coefficients

Predictor Variable

HbA1c

(n=1,434)

Systolic BP

(n=1,366)

LDL

(n=439)Constant (intercept)a –0.28 4.1* 1.9Aged 55 or older 0.35** −0.4 −3.5Male −0.30** 0.1 2.1African American 0.06 −0.9 2.3Native American −0.27 −2.7 0.8Asian −0.32 −0.2 −4.4Hispanic/Latino 0.13 −0.9 −1.9Other race/ethnicities −0.72 −0.8 −15.3Unknown race/ethnicity 0.14 3.9 6.5High-intensity program participation −0.62** −6.3** −4.5** Significant at the 5% level.

*Significant at the 10% levela Represents mean changes for low-intensity participant, White females under age 55.

Weighted Mean HbA1c and Other Measures (First and Last Measurement)

Measure Site Measure Mean#

Participants Min Max Mediant-test (p-value)*

HbA1c All First 8.4 1,488 4.3 17.7 7.8 All Last 7.9 1,488 4.8 18.5 7.4 0.000BP All First 131/79 1,392 66/36 220/170 128/80 All Last 132/79 1,392 80/38 230/141 130/78 0.498LDL All First 99 480 32 233 95 All Last 98 480 34 248 93 0.543Note: Patient had to have at least one pre-intervention and one post-intervention HbA1c measurement

*BP t-test performed on mean arterial pressure

OutcomeNumber of Sites Time Mean N Min Max

t-test (p-value)

Diabetes Competence 3 Baseline 5.3 211 1 73 Follow-up 6.0 211 1.4 7 0.000

Resources and Support for Self-management

3 Baseline 2.5 103 1 43 Follow-up 2.8 103 1 4 0.000

Quality of Life (VR-12)

Physical 4 Baseline 40.6 296 17.5 59.94 Follow-up 40.5 296 17.2 60.1 0.892

Mental 4 Baseline 43.0 296 11.5 69.34 Follow-up 44.5 296 16.8 71.9 0.003

Diabetes Self-care

General diet 4 Baseline 3.9 288 0 74 Follow-up 4.6 288 0 7 0.000

Specific diet 4 Baseline 4.2 287 0 74 Follow-up 4.5 287 0 7 0.000

Exercise 4 Baseline 2.9 290 0 74 Follow-up 3.2 290 0 7 0.002

Blood-glucose testing 4 Baseline 4.2 292 0 74 Follow-up 5.0 292 0 7 0.000

Foot care 4 Baseline 4.3 290 0 74 Follow-up 5.3 290 0 7 0.000

Note: VR-12 = Veterans RAND 12 Item Health Survey