Related Experiment Video
Updated: Aug 28, 2026
Development of a Virtual Reality Assessment of Everyday Living Skills
Published on: April 23, 2014
Derivation of score interpretation thresholds in clinical outcome assessments measuring cognitive impairment
Gill Worthy1, Kim Cocks1, Ben Pascoe1
1Adelphi Values Ltd, Bollington, Cheshire, UK.
Background:
Cognitive impairment is a major factor impacting the functioning and quality of life of individuals with schizophrenia. Several clinical outcome assessment (COA) measures have been developed to measure various aspects of cognition. This study aimed to identify thresholds for interpreting changes that would be meaningful to patients.
Methods:
The study utilised COA data from a multinational program of randomised controlled trials evaluating iclepertin for the treatment of cognitive impairment associated with schizophrenia (CIAS). Endpoints included the MATRICS Consensus Cognitive Battery (MCCB), Schizophrenia Cognition Rating Scale (SCoRS), Virtual Reality Functional Capacity Assessment Tool (VRFCAT) and Patient Reported Experience of Cognitive Impairment in Schizophrenia (PRECIS). Analysis utilised anchor-based methods supported by distribution-based methods to derive interpretation thresholds. Triangulation resulted in a single value or small range of values for between-group difference and meaningful within-patient change (MWPC) thresholds.
Results:
MWPC thresholds for improvement were increases of 4 to 7 points for MCCB overall composite T-score and 10 points for VRFCAT overall T-score and decreases of 5 to 7 points for SCoRS interviewer total score and 0.4 to 0.6 points for PRECIS total score. MCCB and VRFCAT thresholds were estimated using distribution-based methods due to insufficient correlation between anchors and the target scores.
Conclusion:
The proposed score interpretation thresholds may be valuable for assessing improvement in CIAS in future clinical research. Anchor-based analyses may be limited in this patient population as patient, caregiver and clinician-based global ratings of severity tend to have low correlations with multi-item COA scores, particularly with performance-based measures.
