Year: 2008
Type: article
Study design: Observational Study
Abstract: Neurologists and other clinicians rate of naming, comprehension, prosody, fluency, grammar and articulation in the differential diagnosis of frontotemporal lobar degeneration (FLTD). However, the inter-rater reliability of such assessments is not well studied. We examined intraclass correlations of 4 board-certified behavioral neurologists ratings' of 17 aspects of speech and language (S&L). Language samples were recordings of 13 FTLD patients responses to the Cookie Theft picture of the Boston Diagnostic Aphasia Examination. Thirteen FTLD patients were studied. Three met Neary Consensus criteria for FTD behavioral variant, 4 were classified as progressive non-fluent aphasia, 3 with semantic dementia and 3 with primary progressive aphasia. The behavioral neurologists were masked to the FTLD subtype. They used a 5 point Likert scale ranging from feature absent to severe. Guidelines were provided to the clinicians to aide rating of the following: Reduced prosody, Agrammatism, Anomia, Aphasia, Apraxia of speech, Circumlocution, Dysarthria, Empty content, Hesitations, Neologism, Perseveration, Reduced overall rate, Reduced phrase length, Semantic paraphasias, Phonemic paraphasia, and Telegraphic speech. They also rated the dichotomous diagnostic variable Non-fluent aphasia (yes or no). Following Shrout and Fleiss (1979), we used a two-way ANOVA to determine the intraclass correlations (ICCs) based on absolute agreement. (McGraw and Wong, 1996). ICC's above 0.6 were considered acceptable. Ratings of the following six S&L features showed acceptable ICCs; Agrammatism, Apraxia, Perseveration, Reduced overall rate, Reduced phrase length, and Telegraphic speech. The ICC for the dichotomous diagnostic variable Non-fluent aphasia was .68. Aphasia, Anomia, Reduced prosody, Dysarthria, Hesitation, Neologism, Semantic paraphasias, and Phonemic paraphasia had no-zero agreement that was nevertheless unacceptable. Empty content and Circumlocution showed zero agreement. While S&L features are important components of FTLD differential diagnosis, not all S&L features are rated with acceptable reliability. Some key features like agrammatism and apraxia appear to have good agreement, but other keys features like anomia show unacceptable ICCs. At the level of diagnoses e.g. progressive non-fluent aphasia, ICC is only marginally acceptable. These findings suggest the need for improved methods for assessment of S&L features in FTLD research.
Source: Alzheimer s & Dementia
Language: English
Subfield: Psychiatry and Mental health
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2018604391