Psychometric properties of the Clinical Dementia Rating-sum of Boxes and other cognitive and functional outcomes in a prodromal Alzheimer’s disease population

December 21, 2020

Authors: F. McDougall, C. Edgar, M. Mertes, P.Delmar, P. Fontoura, D. Abi-Saab, C.J. Lansdall, M. Boada, R. Doody

Journal: The Journal of Prevention of Alzheimer's Disease

DOI: 10.14283/jpad.2020.73

Year Published: 2020


The Clinical Dementia Rating–Sum of Boxes (CDR-SB) has been proposed as a primary outcome for use in prodromal AD trials. However, the psychometric properties of this, and of other commonly used measures, have not been well-established in this patient population.


To describe the psychometric properties of commonly used efficacy measures in a clinical trial of prodromal AD.


Data were gathered as part of a two-year clinical trial.


Patients had biomarker confirmed prodromal AD.


Clinical Dementia Rating (CDR), Functional Activities Questionnaire (FAQ), Alzheimer’s Disease Assessment Scale – Cognition Subscale 11 and 13 (ADAS-Cog), Mini Mental State Exam (MMSE), and Free and Cued Selective Reminding Test (FCSRT-IR [words]). Assessments were conducted at least every 24 weeks.


For the CDR-SB, test-retest reliability was good (intra-class correlation coefficient [ICC]=0.83); internal consistency was 0.65 at baseline but above 0.8 at later assessments. Relationships between the CDR-SB and other measures were as expected (higher correlations with more closely related constructs), and the CDR-SB differentiated between patients with different severities of dementia (-2.9 points difference between CDR-Global Score 0.5 and 1, P<.0001). Floor and ceiling effects on the CDR-SB total score were minimal; however, at baseline there were ceiling effects in the personal care domain. Further detail is provided on the psychometric properties of ADAS-Cog, MMSE, FCSRT-IR and FAQ in this population.


The psychometric properties of the CDR-SB are adequate in prodromal AD and continued use is warranted in clinical trials. However, there remains scope for improvement in the assessment of functional constructs and development of novel measures should continue.

Back to Publications