Validation using item response theory, and assessment of sensitivity to change of the Hypoglycemia Awareness Questionnaire in patients with type 2 diabetes treated with insulin

Abstract

Aims/Introduction

It is important to develop valid tools to evaluate hypoglycemia perception such as the Hypoglycemia Awareness Questionnaire (HypoA-Q) in patients with Type 2 Diabetes (T2D). The aim of the study is to validate the HypoA-Q in patients with T2D treated with insulin using item response theory.

Materials and Methods

Individuals with T2D treated with insulin were included by non-random convenience sampling. A partial credit model was used for validation using item response theory, infit, and outfit statistics were calculated, person-item map and item characteristic curves were plotted, and differential item functioning was assessed.

Results

The study included 502 participants, the mean age at diagnosis of diabetes was 47.8 ± 13.9 years, the median time with the diagnosis was 15 years (IQR: 9–22), and the mean HbA1c 71 ± 27.3 mmol/mL (8.6 ± 2.5%), 48.6% had Glomerular Filtration Rate >60 mL/min/min². Item fit was found with items covering the full range of the construct of the participant population, although response options could be simplified. The person-item map showed that the scale covers a wide range of the construct and that the scale has items to measure these different levels. Item bias was not evident when comparing subgroups by age, sex and treatment.

Conclusions

The HypoA-Q is a valid measure for assessing hypoglycemia awareness in insulin-treated T2D patients because it has items that fit the measurement process, measures a wide range of awareness, and is free of biases related to gender, age, treatment, and duration of diabetes.