Health disparities in the U.S., including inequalities in the delivery of care and access to care across various racial, ethnic and socioeconomic groups, are of widespread concern, particularly in people with diabetes who require continuous, regular health care to effectively manage their disease.

Such disparities can greatly impact patients’ overall well-being and may lead to serious complications. Three studies that assessed ways to potentially decrease health disparities among people with diabetes were presented today at the American Diabetes Association’s 77th Scientific Sessions at the San Diego Convention Center.

Mobile health technology and Community Health Workers (CHWs) are two emerging strategies increasingly being used throughout the U.S. by health care teams. In the study, “Community Health Workers, Mobile Health, or Both for Management of Medicaid Patients with Diabetes” (365-OR), these approaches were evaluated to determine potential methods to improve diabetes management outcomes among minority patients. CHWs and the use of a mobile health technology app (mHealth) were tested both separately and together among 166 Medicaid patients with type 2 diabetes who receive care in Internal Medicine practices or diabetes clinics at three medical centers in Washington, D.C. At baseline, the patients had an average HbA1c level of 10.5 percent, and they were not meeting three or more of 13 wellness goals established by the study.

Patients in the 12-month study were randomly assigned to three different groups. Group 1 consisted of 56 patients who used an app—the Voxiva Care4Life (C4L) mHealth system. The C4L app helped patients manage their health with features that kept track of frequent measurements of blood sugar and blood pressure levels; provided alerts to remind them to take medications and keep doctor appointments; and offered tips on nutrition and exercise. Group 2 included 56 patients who were assigned CHWs. The CHWs were either educators or lay people who were integrated with the medical teams at each center and helped patients by providing services such as connecting them with primary care doctors and visits to see them; making home visits to help coordinate care and access to food resources and medications; providing language interpretation; helping to identify and address barriers to care; and advocating to ensure the patients received appropriate and culturally tailored health care services. Group 3 had 54 patients who were assigned both a CHW and the use of the C4L mHealth system/app.

Study endpoints included wellness/clinical goals, HbA1c levels, self-care behavior and diabetes distress. Prior to completion of the study, just 6 percent (n=11) of patients withdrew from the program.

Results indicated that within the 12 months, patients in all three groups had achieved on average 1.3 additional wellness/clinical goals from when they enrolled in the study. Additionally, HbA1c levels improved across all of the groups, and data showed that patients decreased their HbA1c levels by an average of 1.3 percent (p<0.0001). Overall, 30 percent of the patients achieved HbA1c levels of less than 8 percent—17 percent of Group 1 patients met that goal; 29 percent of the Group 2 patients; and 43 percent of the Group 3 patients; (p=0.02 vs. C4L alone). Significant improvements were also observed in all three groups of patients for numbers of hospitalizations (p=0.02); and numbers of urgent care visits (p=0.03). Diabetes distress also decreased in all groups (p<0.0001; NS between groups).

“Diabetes self-care is complex and can be a burden for many patients,” said study author Michelle Magee, MD, associate professor of medicine at Georgetown University, and the Director of the MedStar Diabetes Institute.  “When we provided the support of a CHW or a mobile health application, patients with type 2 diabetes experiencing challenges with their self-care were able to achieve important improvement in health measures and a reduction in distress secondary to living with this chronic condition. Evidence to show both the potential impact of CHWs and the potential use of mobile health applications to improve health outcomes, as detailed in this study, are needed in order for health care systems to comfortably invest dollars to these new patient support approaches. Our study shows that these two strategies can significantly improve patient health. In fact, the reduction in A1C levels in our study was as positive a change as what we typically see with the addition of another antihyperglycemic medication to patients’ treatment regimens. Additionally, the resulting increase in meeting wellness goals is important for patients’ daily health and for preventing long-term diabetes complications. And, reducing hospital admissions and acute care visits are important outcomes from both the patient and health economics perspectives.”

While the approach of combining a community health worker and mobile health technology was successful in this population of Medicaid patients, the strategies developed were designed to be adaptable for use by health care teams and the patients they care for at multiple locations. The study team recommends additional research into which programs are most successful and how best to expand them for broad implementation.

Teaching clinicians how best to assist patients with diabetes and their caregivers is an important aspect of continuing medical education. While many research studies and courses explain how clinical factors influence glycemic control, translating that knowledge into a patient care setting is often challenging. This study, “A Social Media Learning Collaborative Approach to Competency-Based Training in Diabetes” (368-OR), emphasized personalizing therapeutic options to fit the individual needs of patients by developing an online, case-based, interactive training toolkit. The study aimed to facilitate the interpretation of research results and to determine how patient-centered factors such as age, gender, socio-economic status, education, race and ethnicity, body weight and current glycemic control can impact the effectiveness of various diabetes treatments.

The study investigators pooled data from 19 clinical trials with a total of 6,954 patients on 38 diabetes regimens from 1,002 clinics, in addition to using Electronic Health Records from 233,627 diabetes patients, to estimate the odds that a particular patient would achieve good glycemic control with different treatment regimens, based upon individual personal characteristics.

Subsequently, eight of the 19 randomized clinical trials contained full quality-of-life and patient satisfaction data from 2,927 patients from 413 clinics. Researchers modeled the probability of achieving HbA1c levels of less than 8 percent and less than 7 percent using 12 regimens of insulin and oral agents alone or in combination during a 24 to 52 week period. Of the 2,927 patients analyzed, 22.6 percent had type 1 diabetes and an average HbA1c level of 8.0; and 77.4 percent of the patients had type 2 diabetes and an average HbA1C level of 9.2 percent.

The primary endpoint at 52 weeks (one year) was HbA1c levels of 7.7 percent. Patients’ socio-demographic information was assessed, and treatment satisfaction questionnaires and quality of life assessments were completed throughout the study. Outcomes of HbA1c levels of less than 8 percent and less than 7 percent were modeled with logistic regression, and resulting estimators were used to develop benchmarking calculators using WebOS, Android, iOS and Windows compatible WordPress software. Calculators were then tested and optimized within case-based learning exercises. During the exercises, the clinician could simultaneously modify patient characteristics to explore and visualize how individual patient profiles might influence the probability of reaching target glycemic goals.

The study determined that the interactive learning collaboratives tested could be beneficial in translating diabetes research findings into clinical practice, while providing a novel approach to competency-based training that meets both the American Diabetes Association’s and the American Association of Clinical Endocrinologists’ clinical care guidelines.

“Relying on the published literature and more passive online courses to translate research findings into concepts that can be applied in practice is not sufficient, and often does not result in knowledge retention or a change in behavior,” said study author Donald C. Simonson, MD, MPH, ScD of the Division of Endocrinology, Diabetes and Hypertension at Brigham and Women’s Hospital and Harvard Medical School in Boston. “Additionally, data on the effectiveness of various diabetes treatments are typically based upon the average effect estimated for a specific group of individuals in randomized clinical trials. However, there is large variability in treatment response that is not well quantified. Some patients respond very well to particular therapies, while others patient do not; and much of this variability can be explained by the personal characteristics of the patients. Our research emphasizes personalizing therapeutic options to fit the individual needs of patients so that clinicians can be made aware of how patients differ in their response to the same treatment based on various patient-centered demographic, socio-economic, behavioral and quality-of-life characteristics.”

The study group plans to continue refining the predictive models and intends to help communicate, disseminate and implement their findings and toolkit into practice by extending the social media learning collaborative to additional practitioners.

Type 1 diabetes (T1D) is now recognized by scientists to be heterogeneous, meaning it can be caused by varying factors and different genes. Understanding the differences in its causes among individuals of different racial/ethnic groups can help researchers and clinicians design improved prevention strategies and treatments. The study, “Ethnic Differences in Progression to Type 1 Diabetes in Relatives at Risk,” (285-OR) examined if there are racial/ethnic differences in how T1D develops by comparing the progression of islet autoimmunity and T1D among races/ethnicities in at-risk individuals.

Researchers used data from TrialNet’s Pathway to Prevention Study screening program, which offers screening for relatives of patients with T1D in the hopes of identifying the risk for type 1 diabetes up to 10 years before symptoms actually appear.

The trial evaluated data of 4,227 TrialNet Pathway to Prevention participants between 1 and 49 years old who did not have diabetes and were autoantibody [Ab] positive relatives of patients with T1D, and followed them prospectively. The trial participants consisted of the following racial/ethnic groups: 12 percent were Hispanic/Latino; 3 percent were African American of non-Hispanic origin; 1.4 percent were Asian/Pacific Islanders of non-Hispanic origin; 79.3 percent were white of non-Hispanic origin; and 4.3 percent were “other,” non-Hispanic origin.

The analysis indicates that race and ethnicity play a role in how T1D develops, and the study specifically demonstrated that the detrimental effect of obesity on T1D risk may differ by race/ethnicity. T1D develops in stages, where individuals first progress from having a single autoantibody (i.e. marker of T1D) to having multiple autoantibodies, and later develop symptoms of T1D. The participants of Hispanic/Latino origin had a 40 percent lower risk of progressing from single to multiple diabetes autoantibodies, compared to the non-Hispanic white participants (HR=0.59, 95% CI=0.40-0.88, p=0.01). Among lean children younger than 12 years of age with multiple positive autoantibodies, the Hispanic/Latino group had half the risk of developing T1D compared to the non-Hispanic white group (HR=0.50, 95% CI=0.27-0.93, p=0.028). However, in this age group, Hispanic/Latino children were more susceptible to the effect of overweight and obesity, which increased the risk of developing T1D by 34 percent among non-Hispanic whites (HR=1.34, 95% CI=1.01-1.79, p=0.046), but quadrupled the risk in the Hispanic/Latino (HR=2.03, 95% CI: 1.25-3.31, p=0.004).

“The differences in type 1 diabetes development among races/ethnicities discovered in this study are striking,” said Mustafa Tosur, MD, a fellow in the pediatric diabetes and endocrinology division of Texas Children’s Hospital at Baylor College of Medicine. “Especially of interest is the dramatic differential effect of being overweight/obese for Hispanic/Latino children younger than 12 years of age, compared to non-Hispanic white children in the same age group. The research demonstrates that racial and ethnic differences should be taken into consideration when counseling family members who are at-risk of developing type 1 diabetes, and when designing preventive care and treatment options. Considering the obesity epidemic in children, which is more prevalent among minorities, and the frequency of type 1 diabetes is growing most in Hispanics in the U.S., these findings have important public health implications.”

Tosur noted that because the study participants were autoantibody-positive relatives of patients, the results of the study are not necessarily representative of the general population. The study team plans to conduct further research on possible reasons for the differences among the various racial/ethnic groups.