'Is Insulin Right for Me?': Web-based intervention to reduce psychological barriers to insulin therapy among adults with non-insulin-treated type 2 diabetes-A randomised controlled trial.

ABSTRACT

AIMS: To test 'Is Insulin Right for Me?', a theory-informed, self-directed, web-based intervention designed to reduce psychological barriers to insulin therapy among adults with type 2 diabetes. Further, to examine resource engagement and associations between minimum engagement and outcomes. METHODS: Double-blind, two-arm randomised controlled trial (11), comparing the intervention with freely available online information (control). Eligible participants were Australian adults with type 2 diabetes, taking oral diabetes medications, recruited primarily via national diabetes registry. EXCLUSION CRITERIA prior use of injectable medicines; being 'very willing' to commence insulin. Data collections were completed online at baseline, 2-week and 6-month follow-up. PRIMARY OUTCOME: negative insulin treatment appraisal scale (ITAS) scores; secondary outcomes: positive ITAS scores and hypothetical willingness to start insulin. ANALYSES intention-to-treat (ITT); per-protocol (PP) examination of outcomes by engagement. TRIAL REGISTRATION ACTRN12621000191897. RESULTS: No significant ITT between-arm (intervention n = 233; control n = 243) differences were observed in primary (2 weeks Mdiff [95% CI] -1.0 [-2.9 to 0.9]; 6 months -0.01 [-1.9 to 1.9]), or secondary outcomes at either follow-up. There was evidence of lower Negative ITAS scores at 2-week, but not 6-month, follow-up among those with minimum intervention engagement (achieved by 44%) compared to no engagement (-2.7 [-5.1 to -0.3]). CONCLUSIONS: Compared to existing information, 'Is insulin right for me?' did not improve outcomes at either timepoint. Small intervention engagement effects suggest it has potential. Further research is warranted to examine whether effectiveness would be greater in a clinical setting, following timely referral among those for whom insulin is clinically indicated.