Development and Feasibility Testing of DM-BOOST Intervention. (DM-BOOST)

Development and Feasibility Testing of a Diabetes Mellitus Program Using Behavioral Economics to Optimize Outreach and Self-management Support With Technology.

DM-BOOST uses clinical informatics tools to identify types of patients with gaps in diabetes care and deploy tailored, proactive outreach methods rooted in behavioral economics to nudge them towards increased engagement with diabetes self-management training and leverage patient-facing technologies to enhance longitudinal patient self-management support.

In DM-BOOST, the Principal investigator will deploy a mixed-methods, patient-centered approach to intervention development and initiate a multiphase optimization strategy (MOST) to learn how to maximize patient engagement and support of self-management training. In this pilot, study team will complete the first phase (Preparation), and initiate feasibility piloting of the second phase (Optimization). Completion of optimization and MOST's final phase (Evaluation), will occur in a subsequent project. In the preparation phase, Principal investigator will first analyze EHR and claims data in the UMCCTS data lake to identify sociodemographic characteristics associated with gaps in diabetes care to develop patient persona archetypes (Aim 1). Next, Principal investigator will selectively recruit patients of identified persona types as consultants, elicit stakeholder feedback during community engagement studios and conduct usability testing to iteratively design the intervention (Aim 2). Study team will then conduct a feasibility pilot (Aim 3) to assess user experience of the intervention implementation and collect exploratory outcome data to be used to inform a subsequent, complete optimization trial.

The purpose of this study is to develop and usability test a patient-centric intervention designed to improve implementation of diabetes self-management training. To accomplish this, 3 specific aims will be completed. Aim 1 - Retrospective data from the UMass Medical School EHR data repository will be analyzed to identify different clusters of patients with diabetes. Aim 2 - To facilitate a patient-centric design of the DM-BOOST intervention, Patient Research Expert Panel (PREP) members (n</=10) will be recruited from various patient types identified in Aim 1 (Aim 2a), participate in Community Engagement Studios to inform intervention conceptualization (Aim 2b) and usability test the intervention (Aim 2c). Aim 3 - The intervention will be pilot tested in n</=70 patients with type 2 diabetes (T2D). Participants will be randomized to either intervention or comparison groups.

After completing the informed consent, study staff will enter the participant's information into pre-populated REDCap identification numbers. This will assign allocation based on the randomization table. Using this technique, participants will be blinded to allocation. However, research staff will not be blinded to provide personalized training for intervention and control. The investigator will be blinded to randomization for all participants during the study.

Intervention group participants will complete a baseline survey, receive a referral to DSMT from the research team, a mailed welcome letter and self-care education sent via a series of personalized patient portal secure messages, text messages, and video call. They will be sent text messages with information about one of the American Association of Diabetes Educators 7 self-care behaviors and will receive encouragement to author their own self-management behavioral goals. Participants will also complete a telehealth training video call with research staff and review the goals that the participant replied with. The participant will then be encouraged to send a patient portal message to their DSMT CDCES that includes their personalized goals prior to their scheduled DSMT session. They will then complete a 3-month follow-up survey and qualitative interview.

Comparison Group participants will complete a baseline survey, receive a DSMT referral request from research team to their primary care provider and a mailed welcome letter. The mailed letter will welcome the participant to the study and contain general information about diabetes self-care behaviors and goal setting. They will complete a DSMT session. They will then complete a 3-month follow-up survey and qualitative interview.

Patient-reported assessment of intervention acceptability via usability testing. Qualitative data collection informed by the Technology Acceptance Model with assessment of perceived usefulness, ease of use, behavioral intention to use and external factors. No quantitative data measured.

Rate of clinical utilization as measured by number of visits per participant to primary, specialty care, and emergency/hospital care visits measured 6-months after follow-up visit.

Diabetes self efficacy will be measured at baseline and 3 months after enrolling in the study using the Diabetes Management Self-Efficacy Scale. Participants will provide feedback on set of questions, using a 5-point Likert scale( with 1=Strong Disagree, 2=Somewhat Disagree, 3= Neutral, 4=Somewhat Agree, 5= Strongly Agree)

Diabetes Treatment Satisfaction will be measured at 3 months after enrolling in the study using the Diabetes Treatment Satisfaction Questionnaire Change tool. Participants will be asked to share how their experience of current treatment has changed from their experience of treatment before the study began. They will answer each question by choosing 3 for Much More Satisfied Now up to -3 for Much Less Satisfied Now. (3,2,1,0,-1,-2,-3)

Self-management skills will be measured at 3 months after enrolling in the study. Participant will be asked questions about their diabetes self-care activities during the past seven days using the Summary of Diabetes Self-Care Activities Measure

Engagement data will be collected by research staff. It will be measured by the numbers of patients who request contact, are reached, enrolled in the study and scheduled DSMT appointment.

Measurement of HbA1c values to determine impact of intervention. HbA1c values at baseline visit will be compared with values at 3-6 months after participant's enrollment. These data will be obtained through EHR chart review.

Retrospective analysis of EHR data to identify clusters of sociodemographic predictors of guideline-concordant of diabetes care will be identified. Retrospective data will be requested from UMMS Data Lake through the Data Science Core. Data requested for adult patients with T2D since Epic EHR roll-out in October 2017 will include: • Sociodemographic characteristics (gender, date of birth, race/ethnicity, zip code, language, marital status, insurance type)

Retrospective analysis of EHR data to identify clusters of clinical predictors of guideline-concordant of diabetes care will be identified. Retrospective data will be requested from UMMS Data Lake through the Data Science Core. Data requested for adult patients with T2D since Epic EHR roll-out in October 2017 will include: • Clinical characteristics as measured by the level of HbA1c

Retrospective analysis of EHR data to identify clusters of clinical predictors of guideline-concordant of diabetes care will be identified. Retrospective data will be requested from UMMS Data Lake through the Data Science Core. Data requested for adult patients with T2D since Epic EHR roll-out in October 2017 will include: • Clinical characteristics as measured by the level of BMI. Weight and height will be combined to report BMI in kg/m^2

Retrospective analysis of EHR data to identify clusters of clinical predictors of guideline-concordant of diabetes care will be identified. Retrospective data will be requested from UMMS Data Lake through the Data Science Core. Data requested for adult patients with T2D since Epic EHR roll-out in October 2017 will include: • Clinical characteristics as measured by the smoking status

Retrospective analysis of EHR data to identify clusters of clinical predictors of guideline-concordant of diabetes care will be identified. Retrospective data will be requested from UMMS Data Lake through the Data Science Core. Data requested for adult patients with T2D since Epic EHR roll-out in October 2017 will include: • Clinical characteristics as measured by the the level of cholesterol

Retrospective analysis of EHR data to identify clusters of clinical predictors of guideline-concordant of diabetes care will be identified. Retrospective data will be requested from UMMS Data Lake through the Data Science Core. Data requested for adult patients with T2D will include: • Clinical utilization as measured by number of visits per participant to primary care, specialty visits, emergency room, hospitalizations, education/training, patient portal use, care management engagement since Epic EHR roll-out in October 2017

Inclusion Criteria: Adults (age 18+) Cognitively able to consent (Aims 2 and 3) Diagnosed with type 2 diabetes (Aims 1-3) Receive primary care at UMMHC in past 12 months at time of initial analysis (Aims 1-3) English speaking (Aims 2 and 3) Have access to patient portal or a smart phone (Aim 3) Exclusion Criteria: Adults unable to consent (lacking cognitive capacity) (Aims 2 and 3) Individuals who are not yet adults (infants, children, teenagers) (Aims 1-3) Pregnant women (Aims 1-3) Prisoners (Aims 1-3) Non-English speaking (Aims 2 and 3)