OBJECTIVES: To study evidence for construct validity, the aim was to describe the outcome from the recently developed Diabetes Questionnaire, assess the associations of that outcome with clinical variables and generic health-related quality of life, and study the sensitivity to differences between clinically relevant groups of glycaemic control in adults with type 1 and type 2 diabetes in a nation-wide setting.

DESIGN: Cross-sectional survey.

SETTING: Swedish diabetes care clinics connected to the National Diabetes Register (NDR).

PARTICIPANTS: Among 2479 adults with type 1 diabetes and 2469 with type 2 diabetes selected at random from the NDR, 1373 (55.4%) with type 1 and 1353 (54.8%) with type 2 diabetes chose to participate.

OUTCOME MEASURES: The Diabetes Questionnaire, the generic 36-item Short Form version 2 (SF-36v2) health survey and clinical variables.

RESULTS: Related to the prespecified assumptions, supporting evidence for construct validity for the Diabetes Questionnaire was found. Supporting divergent validity, the statistically significant correlations with the clinical variables were few and weak. In relation to the SF-36v2 and in support of convergent validity, the strongest correlations were seen in the Diabetes Questionnaire scales General Well-being and Mood and Energy. In those scales, machine learning analyses showed that about 40%-45% of the variance was explained by the SF-36v2 results and clinical variables. In multiple regression analyses among three groups with differing levels of glycated haemoglobin adjusted for demographics, other risk factors, and diabetes complications, the high-risk group had, in support of sensitivity to clinically relevant groups, statistically significant lower scores than the well-controlled group in most Diabetes Questionnaire scales.

CONCLUSIONS: This nation-wide study shows that the Diabetes Questionnaire captures some generic health-related quality-of-life dimensions, in addition to adding diabetes-specific information not covered by the SF-36v2 and clinical variables. The Diabetes Questionnaire is also sensitive to differences between clinically relevant groups of glycaemic control.