(PS2-D82) Diabetes-specific Typology of Family Functioning Moderates Intervention Effects and Is Associated with Intervention Engagement

Introduction: Family interventions to support diabetes self-management have inconsistent effects. Heterogeneous treatment effects may reflect differing needs, preferences, and social contexts of persons with type 2 diabetes (PWD). We previously identified a typology of diabetes-specific family functioning, consisting of four types: Critically Involved, Satisfied with Low Involvement, Want More Involvement, and Collaborative & Helpful. To explore how type may impact outcomes and engagement for a family-focused intervention, we examined whether type moderated intervention effects or was associated with differences in engagement in FAMS 2.0 (Family/friend Activation to Motivate Self-care). FAMS 2.0 consists of daily tailored text messages; optional support person co-enrollment; and monthly coaching calls to set self-care goals and build skills to involve family/friends in goals.

Methods: PWD (N=316) in a RCT of FAMS 2.0 were typed as C (15%), S (20%), W (30%) and H (35%). Outcomes were hemoglobin A1c, diabetes distress, global wellbeing, diabetes self-efficacy, dietary behaviors, medication adherence, physical activity, and family involvement. We used linear regression with 100 chained imputations to assess the interaction between condition and type for outcomes. Intervention engagement was assessed via coaching session completion, text message response rates, co-enrolling a support person, goal success (average number of days per week the PWD met their goal) and family/friend involvement (agreeing to involve family/friends in their goals each month). We used logistic and linear regression to examine associations between type and engagement. Our goal was hypothesis generation to inform future intervention design; we included differences at p < .10.

Results: PWD were 52% male, 37% minoritized race or ethnicity, mean age 56.9 (SD=10.8) years, 37% using insulin, and 11.5 (SD=8.2) years average duration of diabetes. Type moderated intervention effects on HbA1c, diabetes self-efficacy, medication adherence, and physical activity. Type was associated with enrolling a support person, involving family and friends, and goal success. C had smaller physical activity gains than H and S (ps < .08). S had larger improvements in medication adherence than H and W (ps < .10) and larger physical activity gains than all types (ps < .08) but was less likely to enroll a support person (84% vs 97%, p=.05) and involve family (M=74% vs M=84%, p=.03) than H. W had larger gains in diabetes self-efficacy than H and C (ps < .05) and larger A1c improvements than H and S (ps < .08) but met their goals less than H (3.7 vs 4.1, p=.02).

Discussion: These exploratory findings can inform future intervention tailoring to optimize outcomes toward a precision medicine approach. Type of diabetes-specific family functioning at baseline was associated with intervention engagement and moderated intervention effects. W had the most improvements in self-efficacy and HbA1c, suggesting FAMS 2.0 was a good fit but they may benefit from more support to meet daily self-care goals. S were less engaged in family-focused aspects of the intervention but experienced improvements in self-care, suggesting they may benefit more from individualized intervention experiences.