We aimed to investigate predictors of insulin uptake, and change in insulin appraisals, among adults with type 2 diabetes mellitus (T2DM) who participated in the Stepping Up trial.The Stepping Up model of care, supporting timely insulin initiation in primary care, was evaluated in a two-armed cluster-randomised controlled trial. Participants were 266 adults (mean±SD age 62±10years; 39% women) with T2DM (median (IQR) duration 8.5 (5, 13)years) from 74 primary care practices (Stepping Up intervention: 57%, control 43%). At 12months, 47% (n=126) had commenced insulin. Controlling for randomisation, logistic regression was used to explore baseline predictors of insulin uptake, including: demographic and clinical characteristics, emotional wellbeing (depressive symptoms and diabetes-related distress), insulin treatment appraisals, and, 'willingness' to initiate insulin. Two-way analysis of variance examined effects of, and interaction between, randomisation and insulin uptake on 12-month change in insulin appraisals.Participants using insulin at 12months were more likely (all p<0.05) than those with non-insulin-treated T2DM to report: lower socioeconomic status, higher baseline HbA1c (median difference: 0.3%; 3mmol/mol), greater willingness to commence insulin (very willing: 27% vs 12%), and less negative and more positive insulin appraisals. All contributed significantly to the final model (χ2(8)=92.1, p<0.001) except insulin appraisals. Regardless of trial allocation, those initiating insulin reported significantly greater reductions in negative insulin appraisals.Controlling for randomisation, 12-month insulin use was predicted by higher baseline HbA1c and 'willingness' to use insulin if recommended. Negative insulin appraisals reduced following insulin initiation.