Psychometric evaluation of ADAS-Cog and NTB for measuring drug response Karin A, Hannesdottir K, Jaeger J, Annas P, Segerdahl M, Karlsson P, Sj € ogren N, von Rosen T, Miller F. Psychometric evaluation of ADAS-Cog and NTB for measuring drug response. Acta Neurol Scand 2014: 129: 114–122. © 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd. Aims – To conduct a psychometric analysis to determine the adequacy of instruments that measure cognition in Alzheimer’s disease trials. Background – Both the Alzheimer’s Disease Assessment Scale - Cognition (ADAS-Cog) and the Neuropsychological Test Battery (NTB) are validated outcome measures for clinical trials in Alzheimer’s disease and are approved also for regulatory purposes. However, it is not clear how comparable they are in measuring cognitive function. In fact, many recent trials in Alzheimer’s disease patients have failed and it has been questioned if ADAS-Cog still is a sensitive measure. Materials and Methods – The present paper examines the psychometric properties of ADAS-Cog and NTB, based on a post hoc analysis of data from a clinical trial (NCT01024660), which was conducted by AstraZeneca, in mild-to-moderate Alzheimer’s disease (AD) patients, with a Mini Mental State Examination (MMSE) Total score 16-24. Acceptability, reliability, different types of validity and ability to detect change were assessed using relevant statistical methods. Total scores of both tests, as well as separate domains of both tests, including the Wechsler Memory Scale (WMS), Rey Auditory Verbal Learning Test (RAVLT) and Delis- Kaplan Executive Function System (D-KEFS) Verbal Fluency Condition, were analyzed. Results – Overall, NTB performed well, with acceptable reliability and ability to detect change, while ADAS- Cog had insufficient psychometric properties, including ceiling effects in 8 out of a total of 11 ADAS-Cog items in mild AD patients, as well as low test-retest reliability in some of the items. Discussion – Based on a direct comparison on the same patient sample, we see advantages of the NTB compared with the ADAS-Cog for the evaluation of cognitive function in the population of mild-to-moderate AD patients. The results suggest that not all of ADAS-Cog items are relevant for both mild and moderate AD population. Conclusions – This validation study demonstrates satisfactory psychometric properties of the NTB, while ADAS-Cog was found to be psychometrically inadequate. A. Karin 1 , K. Hannesdottir 2 , J. Jaeger 3,4 , P. Annas 1 , M. Segerdahl 1 , P. Karlsson 1 , N. Sj€ ogren 1 , T.von Rosen 5 , F. Miller 1 1 AstraZeneca R&D, S€ odert€ alje, Sweden; 2 AstraZeneca R&D, Cambridge, MA USA; 3 AstraZeneca R&D, Wilmington, DE USA; 4 CogState, New Haven, CT USA; 5 Department of Statistics, Stockholm University, Stockholm, Sweden Key words: acceptability; Alzheimer’ s Disease Assessment Scale-Cog; Alzheimer’ s disease; ceiling effects; cognitive testing; neuropsychological test battery; psychometric properties; validation A. Karin, Q-Med, Seminariegatan 21, SE-75228 Uppsala, Sweden Tel.: +46 708 771 552 Fax: +46 18 474 90 01 e-mail: aleksandra.karin@outlook.com Accepted for publication April 30, 2013 Introduction Proof-of-concept trials to test potential new symp- tomatic treatments for Alzheimer’s disease (AD) typically include an active control arm including a drug with demonstrated benefit (often donepezil) both to demonstrate assay sensitivity and to serve as a comparator to help estimate the effect size of the new drug. However, numerous recent trials have failed to show the benefit of donepezil (1, 2) on regulatory approved cognitive endpoints, mak- ing it difficult to interpret any absence of benefit with the experimental compound. As a result, new scrutiny is being given to methodological factors that might explain this trend. This includes a developing trend toward the inclusion of patients with milder forms of AD and the subsequent chal- lenges in differentiating mild AD from normal 114 Acta Neurol Scand 2014: 129: 114–122 DOI: 10.1111/ane.12153 © 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA NEUROLOGICA SCANDINAVICA