Abstract TP148: The Montreal Assessment of Connected Speech Offers Good Psychometric Properties to Monitor Ecological Language Recovery in Post-stroke Aphasia

    loading  Checking for direct PDF access through Ovid


Introduction: Clinical trials in post-stroke aphasia measure language outcomes with available standardized tests. Most of these are primarily diagnostic tools and without evaluated test-retest reliability and responsiveness to change. They do not have parallel versions to prevent learning bias, and their language tasks (e.g., naming) lack ecological validity. The Montreal Assessment of Connected Speech (MACS) was designed to address these issues. Patients are asked to freely describe five pictures illustrating scenes of daily life. Speech samples are scored for their efficiency in transmitting correct information (% correct content units out of the total number of words). Inter-rater reliability (r = .84; R2 = 0.70) and three parallel versions were validated with 105 healthy young and older francophone adults.

Hypothesis: The MACS has good test-retest reliability (Spearman’s coefficient > .8; Wilcoxon signed-rank test p > .05) and responsiveness to change (discriminates between improved and unimproved patients with area under the receiver operating characteristic curve [AUC] >.8).

Methods: Patients with subacute or chronic aphasia undergoing intensive speech therapy at the Jewish General Hospital in Montreal with improvement in the Boston Naming Test were included in the improved group (n = 7; 3 females; age mean = 64.1 [SD = 10.8]). Patients with chronic post-stroke aphasia receiving no speech therapy were recruited from an association in the Montreal area for the unimproved group (n = 12; 7 females; age mean 62.4 [SD = 11.6]). All participants were French-speaking, had good or corrected visual acuity and underwent pseudorandomized parallel versions of the MACS at T1 and T2 with a two-week interval between them.

Results: Correlation between scores at T1 and T2 (rs(12) = .97, p < .001) and stability (Z = -1.38, p = .17) were found in unimproved patients. Score changes showed good discriminative ability between groups (AUC = .81, p = .03).

Conclusion: Good test-retest reliability and responsiveness to change make the MACS a promising standardized tool for aphasia clinical trials in stroke patients.

Related Topics

    loading  Loading Related Articles