Development of standard computerised adaptive test (CAT) settings for the EORTC CAT Core.

Area of Science:

• Psychometrics and Health Outcomes Measurement
• Clinical Trial Methodology

Background:

• Computerised adaptive testing (CAT) offers individualized patient-reported outcome measurement (PROM) with score comparability across studies.
• Optimal CAT performance depends on well-defined CAT settings, including item selection and termination criteria.
• The European Organisation for Research and Treatment of Cancer (EORTC) CAT Core assesses 14 domains of the EORTC QLQ-C30 questionnaire.

Purpose of the Study:

• To present a general methodology for selecting CAT settings.
• To develop a portfolio of standard CAT settings for the EORTC CAT Core.
• To optimize settings for diverse populations and research objectives.

Main Methods:

• Simulations were used to evaluate measurement properties of CATs with varying lengths and precision levels.
• All CATs were initiated using the most informative EORTC QLQ-C30 item.
• Two fixed-length and two fixed-precision standard CATs were selected for each domain, balancing efficiency (brief) and precision (long).

Main Results:

• Brief fixed-length CATs comprised 3-5 items; longer versions included 5-8 items.
• Fixed-precision CATs targeted reliability levels of 0.65-0.95 (brief) and 0.85-0.98 (long).
• CATs achieved median sample size reductions of 20%-31% compared to EORTC QLQ-C30 scales, with domain-specific variations.

Conclusions:

• The EORTC CAT Core standard settings streamline the selection of appropriate CATs.
• Developed CATs enhance measurement precision and reduce necessary sample sizes.
• These settings can be used directly or adapted for specific research needs.