Background/Aims: Participant non-adherence to study protocols can adversely affect clinical trials by reducing statistical power. Similarly non-adherence is an important issue in clinical practice. Ceasing to participate is one form of non-adherence. Subject characteristics appear to influence adherence. We evaluated relationships between subject characteristics and completion of educational sessions in IDEA, a clinical trial testing 2 modes of diabetes education in patients with sub-optimal diabetes control.
Methods: IDEA eligibility criteria included Type 2 diabetes, A1c =7%, and no recent diabetes education. Subjects (n=623) were randomly allocated to one of 3 diabetes education treatment arms; individual education, group education, or a comparison group with no active intervention. Individual education consisted of three 1-hour sessions, following an AADE program. Group education was delivered in four 2-hour group sessions that emphasized patient interaction. Subjects were considered adherent if they completed all assigned sessions. At the end of the study, tracking data indicated completion rates of 72.0% (175/243) in the group education intervention and 86.1% (211/245) in the individual treatment intervention. We sought to identify demographic, psychosocial and clinical characteristics that might explain non-completion. We hypothesized that, within each arm, baseline health status (SF-12 mental composite and physical composite scores), depression (PHQ9), personality type (TIPI Big 5; extraversion, agreeability, emotional stability, conscientiousness, and openness), and Hb A1c level were associated with completion.
Results: In the group education arm, subjects scoring higher on the emotional stability scale were more likely to complete (p<0.05). Generally, demographic factors were more strongly associated with completion of group education than were psychosocial or clinical factors: women and older subjects were more likely to complete than men and younger subjects (gender, p=0.008; age, p=<0.0005). Within the individual education arm, completion was predicted by higher physical health score (SF12-PCS; p=0.005), higher mental health scales (SF12-MCS, p=0.008), lower depression score (PHQ9, p=0.002), and lower Hb A1c (p<0.05). However, neither gender nor age was associated with completion.
Conclusions: Factors related to study completion appear to differ between the 2 diabetes educational interventions. These results may be relevant to diabetes educational programs considering strategies to improve “no show” and lack of completion rates.