Characterizing problematic hypoglycaemia: iterative design and preliminary psychometric validation of the Hypoglycaemia Awareness Questionnaire (HypoA‐Q). Issue 3 (14th July 2015)

• Record Type: Journal Article
• Authors: Speight, J.; Barendse, S. M.; Singh, H.; Little, S. A.; Inkster, B.; Frier, B. M.; Heller, S. R.; Rutter, M. K.; Shaw, J. A. M.
• Is Part Of: Diabetic medicine. Volume 33:Issue 3(2016:Mar.)
• Page Start: 376
• Page End: 385
• Publication Date: 2015-07-14
• Subjects: Diabetes -- Periodicals; 616.462
• DOI: 10.1111/dme.12824
• Languages: English
• ISSNs: 0742-3071

Abstract

Abstract: Aims: To design and conduct preliminary validation of a measure of hypoglycaemia awareness and problematic hypoglycaemia, the Hypoglycaemia Awareness Questionnaire. Methods: Exploratory and cognitive debriefing interviews were conducted with 17 adults (nine of whom were women) with Type 1 diabetes (mean ± sd age 48 ± 10 years). Questionnaire items were modified in consultation with diabetologists/psychologists. Psychometric validation was undertaken using data from 120 adults (53 women) with Type 1 diabetes (mean ± sd age 44 ± 16 years; 50% with clinically diagnosed impaired awareness of hypoglycaemia), who completed the following questionnaires: the Hypoglycaemia Awareness Questionnaire, the Gold score, the Clarke questionnaire and the Problem Areas in Diabetes questionnaire. Results: Iterative design resulted in 33 items eliciting responses about awareness of hypoglycaemia when awake/asleep and hypoglycaemia frequency, severity and impact (healthcare utilization). Psychometric analysis identified three subscales reflecting 'impaired awareness', 'symptom level' and 'symptom frequency'. Convergent validity was indicated by strong correlations between the 'impaired awareness' subscale and existing measures of awareness: (Gold: rs =0.75, P < 0.01; Clarke: rs =0.76, P < 0.01). Divergent validity was indicated by weaker correlations with diabetes‐related distress (Problem Areas in Diabetes: rs =0.25, P < 0.01) and HbA1c (rs =‐0.05, non‐significant). The 'impaired awareness' subscale and other items discriminated between those with impaired and intact awareness (Gold score). The 'impaired awareness' subscale and other items contributed significantly to models explaining the occurrence of severe hypoglycaemia and hypoglycaemia when asleep. Conclusions: This preliminary validation shows the Hypoglycaemia Awareness Questionnaire has robust face and content validity; satisfactory structure; internal reliability; convergent, divergent and known groups validity. The impaired awareness subscale and other items contribute significantly to models explaining recall of severe and nocturnal hypoglycaemia. Prospective validation, including determination of a threshold to identify impaired awareness, is now warranted. What's new?: Approximately 20–25% of adults with Type 1 diabetes have impaired awareness of hypoglycaemia (IAH), which is associated with a six fold higher risk of severe hypoglycaemia. Existing measures of IAH have acknowledged limitations. Comprehensive, collaborative and iterative qualitative work (including adults with Type 1 diabetes and clinicians) has generated a new questionnaire designed to characterize IAH and problematic hypoglycaemia when awake and asleep, as well as to assess changes over time. Preliminary psychometric validation has shown that the Hypoglycaemia Awareness Questionnaire has: good face and content validity; convergent, divergent and known‐groups validity; satisfactory scale structure and internal consistency reliability; and early indication of predictive validity.