AD8 revisited: what’s the optimal cutoff value?

Abstract

AbstractBackgroundThe 8‐item Informant Interview to Differentiate Aging and Dementia (AD8) was developed as a screening tool for dementia with a cutoff of 2 suggested by the initial study. However, various studies found different cutoff values, and many suggested a cutoff of 2 might result in a high false positive rate (up to > 80%). A higher false positive rate in Taiwan was repeatedly shown when AD8 was self‐administered in local government screening programs. This study aimed to examine the performance of AD8, determine its best cutoff value, and discuss factors that may affect the performance of AD8.MethodWe recruited 119 participant‐informant dyads from a psychiatric outpatient clinic of a university medical center. For each dyad, the AD8 was administered first, and then Mini‐Mental Status Examination (MMSE) and the Clinical Dementia Rating (CDR) to minimize contamination effect. Two geriatric psychiatrists made a consensus clinical diagnosis for each subject based on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM‐5) using all available clinical information, including MMSE, CDR, clinical history, other past neuropsychological tests, and neuroimaging examinations (if available).ResultThere were 37 subjects with normal cognition, 44 with mild neurocognitive disorder (mild NCD), and 28 with major neurocognitive disorder (major NCD). When the cutoff value was 2 for the differentiation between major NCD and no‐dementia subjects (including normal cognition and mild NCD), the AD8 had an AUC of 0.756, a sensitivity of 0.929, and a specificity of 0.582. But the optimal cutoff value was 4 according to the Youden index, with an AUC of 0.888, a sensitivity of 0.786, and a specificity of 0.835.ConclusionAlthough many studies recommended 2 as the best cutoff value for screening dementia, our findings and those of others suggest a cutoff of 2 may result in a high false positive rate. Many factors could affect the performance of AD8, including sample characteristics (e.g. education level), region, conduct of the test, flow of the administration (whether AD8 is administered first or contaminated by other tests), etc. Details will be discussed during the AAIC meeting.