Glycemic control in patients with type 2 diabetes was improved through a shared medical appointment program focusing on lifestyle education and behavior change.
Published Online: December 14, 2016
Melissa S. Hernandez, MD; Ruth Nutting, PhD, LCMFT; Andrew J. Vasey, MD; Susan K. Burbach, BSN-RN; Jason F. Shiffermiller, MD, MPH
Precis : Glycemic control in patients with type 2 diabetes was improved through a shared medical appointment program focusing on lifestyle education and behavior change.
Over 29 million Americans are currently living with diabetes, a disease that ranks among the top 10 causes of death.1,2 Although the quality of care provided to these patients has improved, it remains suboptimal.3 Patient education and engagement are significant shortcomings: nearly half of patients report receiving no diabetes education at the time of their diagnosis.3 In traditional models of care, such barriers as limited time, lack of access to an interprofessional team, and patient reluctance to discuss self-care behaviors with a physician account for failure to deliver diabetes education.4 A redesign of clinic visits to make them more collaborative and patient-centered is warranted.5
The shared medical appointment (SMA) model optimizes patient and provider time through an interprofessional, team-based approach.6 Over the last 20 years, SMAs have been shown to reduce glycated hemoglobin (A1C),7,8 blood pressure,7 and low-density lipoprotein cholesterol8; boost adherence to the American Diabetes Association (ADA) standards of care9,10; and decrease hospitalization.7 Although existing literature supports the role of SMAs in improving objective outcomes, there has been less attention on patient-centered outcomes, such as satisfaction with SMAs and their components.7,11,12 The patient perspective is important to understand, as patient activation is imperative to chronic disease management.13
Since no standard exists for the format, components, or the team of healthcare providers involved, there is significant variability in the design of diabetes SMA interventions. A recent systematic meta-analysis concluded that heterogeneity among interventions makes it difficult to evaluate which group-visit components lead to successful SMA interventions.7 We undertook a mixed-methods study exploring how and why an SMA program achieved success.
Research Design and Methods
Setting and Participants
This study, approved by the university’s Institutional Review Board, took place at a Joint Commission–certified Primary Care Medical Home (PCMH), in which integrated care is provided to an underserved population. This PCMH serves as the continuity clinic for Internal Medicine residents who receive training at the affiliated academic medical center.
Patients with poorly controlled type 2 diabetes, which we defined as having an A1C greater than 8.0%, were invited to participate in the project. Patients were excluded if they lived in long-term care facilities, had cognitive impairment/dementia, and or did not speak English.
We designed a hybrid SMA structure by combining elements from other models and creating a more fundamental role for the behavioral medicine provider.14 Four SMA group cohorts were offered. Participants were invited to join whichever meeting time was most convenient for them. In an effort to maintain cohesiveness, new patients were added to the groups only once per year. Patients signed a Health Insurance Portability and Accountability Act document prior to participation, which informed them that protected health information would be shared within their group. Four senior Internal Medicine resident physicians led each of the 4 groups. A nurse diabetes educator, social worker, pharmacist, nursing staff, and behavioral medicine provider were present at every visit. The SMA sessions lasted 120 minutes and were held quarterly. Participants were required to meet with their primary resident physician for a traditional clinic visit outside of the SMA program at least once per year.
The SMA meetings were divided into 4 sections: intake, education, break-out, and closing. The education sessions followed the AADE7 Self-Care Behaviors™ curriculum, created by the American Association of Diabetes Educators, and covered nutrition, exercise, mental health, foot care, and other topics.15 During the breakout session, patients met one-on-one with the pharmacist, social worker, medical assistant, Internal Medicine resident, diabetes educator, and behavioral health provider. They then reconvened in a group led by the resident physician and attended by all interprofessional team members to collaboratively review lab work, glycemic control, barriers to adherence, and any changes in diabetes management. During this part of the SMA, each patient had the opportunity to ask questions of the group, share experiences, and participate in the development of their own care plans. Immediately following SMA visits, the resident physician presented all assessments and plans to an attending physician.
We conducted a mixed methods study. We began by performing a quantitative pre-post study of clinical outcomes and survey responses at the conclusion of the first year of the program. Focus groups were convened at the conclusion of the second year to explore patient insights into how the SMA visits influenced their diabetes management.
Data Collection and Outcome Measures
At each visit, we administered the Patient Health Questionnaire-9 (PHQ-9) to screen for and monitor depression.16,17 Self-management behaviors and knowledge were assessed by The Diabetes Lifestyle Assessment Tool, a publicly available, but nonvalidated, measure of diet, activity, self-monitoring, and medication adherence, among other domains.18 This assessment was administered at the first group session of the first year and at the final session of the first year. In addition to collecting A1C measurements from SMA participants, we collected contemporaneous A1C values for all clinic patients receiving usual care. Process measures were extracted from the electronic health record.
Informed by the results of the quantitative analysis, we developed a semi-structured focus group agenda through an iterative process. It contained 6 open-ended questions with 3 subquestions (TABLE 1). The purpose was to reveal patients’ insights into which features, or parts, of our SMA provided the greatest value. In addition to asking how SMAs were most valuable, questions investigated the effect of SMAs on a variety of factors, including adherence and mood. Two researchers attended each focus group: 1 moderator and 1 transcriber. The focus group interviews were recorded and transcribed verbatim by the moderator. Our qualitative data comprises the 4 transcripts.
At the conclusion of the first year, SMA attendance, survey responses, A1C, and process measures were analyzed; pre- and postintervention values were compared; and demographics and retention were examined using descriptive statistics. Paired survey responses and A1C levels were analyzed with the Wilcoxon signedrank test, abstracted data from chart reviews were analyzed using the Chi-squared test, and students’ t-tests were used to analyze A1C values in clinic patients receiving usual care. We set alpha equal to 0.05 for all comparisons.
A phenomenological method was employed to analyze focus group interviews.19 We attempted to minimize bias by having the researcher begin with a reflection on her own experience with SMAs. Each focus group transcript was then reviewed 3 times in order to identify noteworthy patient statements. Statements that pertained to patient perception of, and satisfaction with, SMAs were excerpted and coded, and then clustered into common themes. From each theme, a textural description was written identifying “what” SMA subjects experienced and a structural description was created identifying “how” the subjects experienced SMAs. Finally, a composite description incorporating the textural and structural descriptions, as well as participant examples, was generated. A second researcher reviewed themes and descriptions for clarity.
Quantitative Primary Outcomes
Of 214 eligible patients, 21 accepted the invitation to participate and were enrolled in the study. Group sizes ultimately ranged from 4 to 6 patients, the average participant age was 57 years, female subjects outnumbered males 16 to 5, and the majority of subjects were unmarried. Four of 21 subjects (19%) were uninsured, and 13 of 21 (62%) were enrolled in Medicare or Medicaid. Details of the demographic analysis are presented in TABLE 2. The first SMA visits occurred in May of 2013. Eighteen of the 21 subjects (86%) who attended the first SMA were still participating at the end of year 1. The mean A1C in SMA participants decreased from 10.1 % to 7.9% over the course of the first year (P = .003). This represents an absolute reduction of 2.2 percentage points in A1C and a relative reduction of 21.8 %. A1C in non-SMA clinic patients with diabetes (usual care) decreased from 7.5% (n = 109) to 7.0% (n = 117) (P = .079) over a similar time period. Overall, PHQ-9 scores in SMA participants decreased during the first year, but did not reach statistical significance (P = .11). Using The Diabetes Lifestyle Assessment Tool, we were unable to detect any impact of SMAs on diet, exercise, glucose monitoring, medication adherence, or social support.
Quantitative Secondary Outcomes
PDF is available on the last page.