Bells Test Cancellation Tools Data 311023

Published: 31 October 2023| Version 1 | DOI: 10.17632/9rfxfhtfmd.1
Jamie Berry


Background: Whilst the outcome variables on the Bells test (e.g., total omissions) are easily quantified, process data (e.g., search strategy) have been more difficult to quantify objectively. CancellationTools is a software package that calculates cancellation task-based process variables following input of cancellation target identification order. The primary aim of the current study was to examine the psychometric properties of several CancellationTools process variables for the Bells test. Method: The CancellationTools process variables: mean distance, standardized mean distance, standardized angle, best r, and intersections rate, were calculated for the Bells Test in a diverse neurological sample (n=101) and a healthy Australian sample (n=57). Ratings of cancellation path organization using an ordinal categorical variable (systematic, disorganized, indeterminate) was completed by two experienced clinicians. Construct validity, criterion validity, known groups validity, test-retest reliability and test operating characteristics were examined. Results: Mean distance, standardized angle, best r and intersections rate, but not standardized mean distance, showed good construct (convergent) validity with clinician ratings. Standardized angle, best r and intersections rate showed good divergent validity when compared with outcome variables. Criterion validity was established for best r and intersections rate. The CancellationTools measures did not show known groups validity, although clinician ratings did. Good test-retest reliability was demonstrated for best r (ICC = .79) and clinician-rated search strategy (ICC = .75). Best r explained 88% of the area under the curve when classifying disorganized vs other search strategies based on clinician ratings. Conclusion: Best r emerged as the most psychometrically robust of the CancellationTools measures.



Macquarie University