BACKGROUND: Nurse prescribers are in a key position to promote medicine-taking in diabetes. Although patients' beliefs about medicines are important predictors of medicine-taking, evidence suggests nurses do not routinely explore these. OBJECTIVES: To evaluate a theory-based intervention designed to increase nurse prescribers' exploration of medicines' beliefs with people with diabetes. DESIGN: Mixed methods concurrent triangulation design. SETTINGS: Nurse prescribers were recruited from 7 Trusts in England. PARTICIPANTS: A purposive sample of 14 nurse prescribers attended four 1 day workshops. METHODS: Audio-recordings of each nurse prescribers' consultations with diabetes patients were collected at baseline, 1 week, 3 months and 6 months after the intervention. Nurse prescribers were interviewed at 1 month and 6 months post-intervention. Changes in medicines' discussion and participation in consultations were analysed using MEDICODE. Interview data were analysed using Framework Analysis. RESULTS: MEDICODE themes of 'attitudes towards medication' showed a significant rise at 1 week (p<0.01) and 3 months (p<0.05). 'Asks patient opinion about medication' significantly increased at 1 week (p<0.01). Discussion on 'concerns about medication' rose significantly at 1 week (p<0.001) and 6 months (p<0.01). Discussion on 'expected effects of medication', 'action of medication' and 'reasons for medication' showed no change. There were no significant changes in Dialogue Ratio. However, the Preponderance of Initiative moved towards more patient initiative at 1 week (p<0.0001), 3 months (p<0.0001), and 6 months (p<0.0001). In interviews, nurses reported increased attention to patients' medication beliefs and adoption of patient-centred skills. Contextual factors that positively influenced ability to explore medicines beliefs in practice settings were: support of colleagues and practicing new skills. Inhibiting factors included: patients' perceived lack of receptivity, time constraints, and concerns about opening a 'can of worms'. Six months interviews revealed using skills in practice enhanced nurses' confidence and sustainability of skills requires a nurse-patient relationship. Method triangulation illuminated how the intervention was implemented in practice contexts. CONCLUSIONS: The intervention was effective at changing some key dimensions of prescribing consultations. The use of a self-efficacy framework in the intervention, to promote nurses' confidence in working in a different way, may have been instrumental in effecting the changes found. Contextual factors influencing beliefs exploration in medicine-taking consultations were identified.