AIM: To describe experience from a behavioural medicine intervention among poorly controlled adult type 1 diabetes patients, in terms of feasibility, predictors and associations of improved glycaemic control.

METHODS: Data were collected on 94 poorly controlled adult type 1 diabetes patients who were randomised to a study evaluating the effects of a behavioural medicine intervention. Statistics covered descriptive and comparison analysis. Backward stepwise regression models were used for predictive and agreement analyses involving socio-demographic and medical factors, as well as measures of diabetes self-efficacy (DES), diabetes locus of control (DLOC), self-care activities (SDSCA), diabetes-related distress (Swe-PAID-20), fear of hypoglycaemia (HFS), well-being (WBQ), depression (HAD) and perceived stress (PSS).

RESULTS: The participation rate in the study was 41% and attrition was 24%. Of those patients actually participating in the behavioural medicine intervention, 13% withdrew. From the regression models no predictors or associations of improvement in HbA(1c) were found.

CONCLUSIONS: The programme proved to be feasible in terms of design and methods. However, no clear pattern was found regarding predictors or associations of improved metabolic control as the response to the intervention. Further research in this area is called for.