Psychometric evaluation of separate anxiety and depression item banks - potential supplements to the EORTC CAT Core

The European Organisation for Research and Treatment of Cancer's computerized adaptive test instrument, the EORTC CAT Core, provides dynamic, individualised assessment across 14 health-related quality of life domains, including emotional functioning (EF), which encompasses anxiety and depression. When one of these symptoms is of primary interest, separate scores may better support clinical decision-making and increase measurement sensitivity. The aim here was to evaluate separate item banks for anxiety and depression as potential supplements to the EORTC CAT Core.

Responses to 33 candidate EF items from 1,023 cancer patients were available from the original EF item bank project. Expert-based content classification was followed by psychometric analyses, including confirmatory factor analysis (CFA), item response theory (IRT) model calibration and evaluation, differential item functioning (DIF) analyses, and CAT simulations to assess measurement properties of the two item banks.

Experts allocated nine items to anxiety and 17 to depression while seven items were judged to assess something else. CFA supported unidimensionality for each item bank. All items demonstrated acceptable IRT model fit, and no DIF of practical relevance was identified. CAT simulations indicated greater measurement precision than fixed 1- and 2-item scales, with potential sample size savings of up to 30% for anxiety and 50% for depression.

Well-functioning and psychometrically sound item banks can be formed for anxiety and depression, respectively. These findings support the use of separate anxiety and depression CATs as potential supplements to the EORTC CAT Core for more targeted assessment. Further studies are needed to confirm the content validity of the two item banks.