Psychometric properties of the revised Diabetes Knowledge Test using Rasch analysis

Area of Science:

• Endocrinology
• Psychometrics
• Health Education

Background:

• The revised Diabetes Knowledge Test (DKT2) is widely used to assess diabetes knowledge.
• Its structural validity and psychometric properties have not been previously evaluated.
• Understanding the DKT2's psychometric properties is crucial for accurate diabetes knowledge assessment.

Purpose of the Study:

• To determine the psychometric properties of the revised Diabetes Knowledge Test (DKT2).
• To evaluate the structural validity of the DKT2 instrument.

Main Methods:

• Secondary data analysis of individuals with diabetes from university hospitals.
• Rasch analysis was employed to assess the DKT2's psychometric properties.

Main Results:

• One item from the 14-item DKT2 was removed due to poor fit, resulting in a 13-item instrument.
• The analysis indicated a unidimensional structure but revealed poor person reliability (0.50) and a low person separation index (1.01).
• Measurement invariance was not achieved for specific items concerning gender and age.

Conclusions:

• The 13-item DKT2 exhibits a unidimensional structure but suffers from poor person reliability and low person separation.
• Specific items (11 and 2) presented greater difficulty for females and older adults, respectively.
• The study highlights the need for careful consideration when using the DKT2 and suggests adding more challenging items in future iterations.