Enhancing the Efficiency of Confrontation Naming Assessment for Aphasia Using Computer Adaptive Testing

Area of Science:

• Neuroscience
• Speech and Language Pathology
• Psychometrics

Background:

• Anomia, a common symptom of aphasia, significantly impacts communication abilities.
• The 175-item Philadelphia Naming Test (PNT) is a standard tool for assessing anomia.
• Developing shorter, efficient assessments is crucial for clinical and research applications.

Purpose of the Study:

• To evaluate the psychometric agreement between the full Philadelphia Naming Test (PNT) and a newly developed 30-item computer adaptive version (PNT-CAT).
• To determine the suitability of the PNT-CAT for measuring anomia in individuals with aphasia.
• To assess the precision of score estimates derived from the PNT-CAT using item response theory (IRT).

Main Methods:

• Administered the full PNT and the PNT-CAT to 47 participants with aphasia in a counterbalanced order.
• Analyzed latent trait-naming ability estimates using a Bayesian framework.
• Evaluated agreement through correlation, bias, variable error, and total error measures.

Main Results:

• Demonstrated strong agreement between the PNT and PNT-CAT, evidenced by a high correlation (r = .95).
• Observed negligible bias and low variable and total error, indicating consistent measurement.
• Found that the IRT model adequately predicted the precision of individual score estimates.

Conclusions:

• The PNT-CAT is a psychometrically sound and efficient alternative to the full PNT for assessing anomia in group studies.
• The PNT-CAT's robust score precision supports its utility in individual clinical assessments.
• IRT methods provide a foundation for further refinement of the PNT-CAT to minimize measurement error.