Revisiting the Geriatric Depression Scale: An IRT-Based 10-Item Screen Outperforms the GDS-15 in Diagnostic Accuracy...

Area of Science:

• Geriatric Psychiatry
• Psychometric Measurement
• Item Response Theory (IRT)

Background:

• Existing abbreviated Geriatric Depression Scales (GDSs) often compromise diagnostic accuracy for brevity.
• Classical Test Theory (CTT) based scale reductions may exclude highly discriminating items.
• A need exists for a brief, accurate screening tool for geriatric depression.

Purpose of the Study:

• To develop a minimum-item Geriatric Depression Scale (GDS) using Item Response Theory (IRT).
• To achieve diagnostic performance equivalent to the full 30-item GDS (GDS30).
• To enhance screening efficiency for late-life depression.

Main Methods:

• A 5:5 split-sample cross-validation was conducted on 6525 older adults (≥60 years).
• Item Response Theory (IRT) models (2-parameter logistic) estimated item parameters.
• Sequential item reduction and DeLong tests identified the minimum items for GDS30-equivalent Area Under the Curve (AUC).

Main Results:

• The 10-item IRT-optimized scale (GDS10-IRT) achieved an AUC of 0.856, not significantly different from GDS30 (AUC=0.883, p=0.396).
• The GDS10-IRT demonstrated superior efficiency (AUC/items) compared to GDS30.
• The widely used GDS15 showed significantly lower diagnostic accuracy than GDS30 (p < 0.001).

Conclusions:

• Item Response Theory (IRT) enables the development of a 10-item GDS (GDS10-IRT) with diagnostic accuracy equivalent to the GDS30.
• The GDS10-IRT outperforms the GDS15 in diagnostic performance and efficiency.
• The GDS10-IRT provides a more accurate and efficient screening tool for geriatric depression by including key discriminating items.