BACKGROUND: We do not yet know how to use blood glucose self-monitoring (BGSM) most effectively in the self-management of type 2 diabetes treated with oral medication. Training in monitoring may be most effective in improving glycaemic control and well being when results are linked to behavioural change. METHODS/DESIGN: DiGEM is a three arm randomised parallel group trial set in UK general practices. A total of 450 patients with type 2 diabetes managed with lifestyle or oral glucose lowering medication are included. The trial compares effectiveness of three strategies for monitoring glycaemic control over 12 months (1) a control group with three monthly HbA1c measurements; interpreted with nurse-practitioner; (2) A self-testing of blood glucose group; interpreted with nurse- practitioner to inform adjustment of medication in addition to 1; (3) A self-monitoring of blood glucose group with personal use of results to interpret results in relation to lifestyle changes in addition to 1 and 2. The trial has an 80% power at a 5% level of significance to detect a difference in change in the primary outcome, HbA1c of 0.5% between groups, allowing for an attrition rate of 10%. Secondary outcome measures include health service costs, well-being, and the intervention effect in sub-groups defined by duration of diabetes, current management, health status at baseline and co-morbidity. A mediation analysis will explore the extent to which changes in beliefs about self-management of diabetes between experimental groups leads to changes in outcomes in accordance with the Common Sense Model of illness. The study is open and has recruited more than half the target sample. The trial is expected to report in 2007. DISCUSSION: The DiGEM intervention and trial design address weaknesses of previous research by use of a sample size with power to detect a clinically significant change in HbA1c, recruitment from a well-characterised primary care population, definition of feasible monitoring and behaviour change strategies based on psychological theory and evidence, and measures along the hypothesised causal path from cognitions to behaviours and disease and well being related outcomes. The trial will provide evidence to support, focus or discourage use of specific BGSM strategies.