Validity and Diagnostic Accuracy of a Clinical Trial Battery for Primary Brain Tumor Patients
Objective: Neurocognitive function (NCF) is a relevant endpoint in clinical trials of primary brain tumor patients. The Clinical Trial Battery (CTB) consisting of the Hopkins Verbal Learning Test-Revised, Trail Making Test, and Controlled Oral Word Association is currently in use for multinational / multisite clinical trials of primary brain tumor patients. Although the CTB is brief, useful, and consists of standardized measures, validity and diagnostic accuracy of the CTB are unknown and the measures are primarily verbally based. The CTB generates multiple outcome variables for cognitive domains of attention, executive function, learning, memory, and processing speed, and can therefore be analytically challenging. The present study aimed to investigate the validity and diagnostic accuracy of the CTB and to evaluate the added value of including a test requiring predominantly visuospatial function.
Methods: To integrate and interpret the six major outcome variables generated by the CTB, an unweighted average standard composite score (NCF6Z) was calculated. Block Design (BD) was added to the CTB in order to examine the impact of including a test of visuospatial function, which yielded another composite score (NCF7Z) for the extended CTB (CTBE). Validity and diagnostic accuracy of NCF6Z and NCF7Z of untreated primary brain tumor patients (n = 260) were assessed against the criteria: performance on comprehensive clinical assessment, clinician ratings of NCF impairment status, tumor characteristics, and clinician report measures of functional status. Results: NCF6Z was found to be adequately valid. Assessed against clinician ratings of NCF impairment status, at cutpoint -0.50, NCF6Z demonstrated sensitivity = 0.86, specificity = 0.78, PPV = 0.80, and NPV = 0.85. At cutpoint -0.6257, NCF6Z displayed optimal diagnostic accuracy (sensitivity = 0.85, specificity = 0.83, PPV = 0.84, NPV = 0.84). Addition of BD did not augment validity or diagnostic accuracy of the original CTB, even when the scores were weighted to offset the purported verbal bias of the CTB.
Conclusion: The present investigation demonstrated validity and diagnostic accuracy related evidence for NCF6Z. The CTB, which is a brief, useful, valid, and diagnostically accurate measure of NCF among primary brain tumor patients, is suitable for use in large scale research studies involving primary brain tumor patients.