Diabetes Homeless Medication Support (D-HOMES)

This randomized pilot trial of the Diabetes Homeless Medication Support intervention vs. brief diabetes education will test the perception and feasibility of anticipated study procedures and refine randomization and blinding.

This study has an overall goal to develop and pilot test a collaborative care intervention using motivational interviewing and behavioral activation alongside education and psychosocial support to improve medication adherence tailored to the experiences of people experiencing homelessness and diabetes (DH). The investigators' central hypothesis is that medication adherence and diabetes self-care (and eventual glycemic control, health care use/cost) will improve with an intervention tailored to the unique context of DH. This study will involve addition of a randomization schema to compare D-Homes to brief diabetes education. The study team will enroll participants (n=54), and refine procedures for the randomized trial outcome measures. This includes an assessment of sustained impact of the program at 24-weeks post-treatment. These will complement 12-week post-treatment primary outcome measures.

The study team will use 1 to 1 allocation to intervention vs. comparison arms. They will stratify based on more and less stable housing in the 12 mo. prior to enrollment.

There will be 10 sessions offered within 12 weeks to participants. Sessions will last approximately 30 minutes. During sessions a diabetes wellness coach will use behavioral activation and motivational interviewing to get to know participants and set goals to improve diabetes care. The coach will encourage a focus on medication adherence behaviors to the extent that participants are willing. The coach will also help with resource and care coordination. The coach will also provide a tailored tool to the patient's needs/goals and tailored diabetes education as needed.

Trained diabetes wellness coaches will provide an approximately 15 minutes of instruction about the basic concepts of diabetes. They will use handouts aligned with American Diabetes Association guidelines. They will read these with participants and answer basic questions. Handouts will cover (1) general diabetes knowledge, (2) healthy eating with diabetes, (3) physical activity with diabetes. The coach will also provide a general tool to support medication adherence (e.g. pillbox).

The acceptability of the intervention to participants will be measured by the Client Satisfaction Questionnaire, 8-item version, with a score range from 8-32, higher score indicating higher satisfaction.

The study team will track enrolled participants who complete post-treatment assessment visits. They have planned one 12-16 week post-treatment assessment and a second 24-30 week assessment. Retention will be measured as the percentage of enrolled participants who complete each of these assessments.

The study team will measure glycemic control using hemoglobin A1c. This will be done on a consistent, validated point-of-care machine using fingerstick blood samples. The study team will compare glycemic control from baseline to 16 weeks and from baseline to 30 weeks.

Psychological wellness as measured by the Short Form (SF)-12. The SF-12 is a self-reported outcome measure assessing psychological wellness and the impact of health on an individual's everyday life. Scores range from 20 to 60. We will compare SF-12 scores from baseline to 16 weeks and from baseline to 30 weeks.

As measured by the Adherence to Refills and Medications Scales-Diabetes (ARMS-D), Total scores range from 12-48, with higher values indicating worse outcomes. The study team will compare ARMS-D scores from baseline to 16 weeks and baseline to 30 weeks.

Inclusion Criteria: Age 18 yrs. or older English-speaking Homelessness by federal definition (HEARTH ACT) in the past 24 mos. Self-reported diagnosis of type 2 diabetes with A1c >7.5%, later verified in medical record and study point-of-care lab test. Plan to stay in local area or be reachable by phone for the next 24 weeks Willingness to work on medication adherence and diabetes self-care Exclusion Criteria: Inability to provide informed consent (e.g., presence of a legal guardian, prisoners) Active psychosis or intoxication precluding ability to give informed consent Pregnant or lactating people

As this work is pilot data to inform treatment development for later efficacy testing, we will not share individual participant data. The study team will submit for publication an overall paper describing our results and how they informed future intervention development.