P4‐115: Validation of a new language screening tool with stroke in chinese: The mandarin version of language screening test (MLAST)

• DOI: 10.1016/j.jalz.2015.06.1821
• Publication Date: 2015-12-15T10:34:01Z
• Authors: Hongyan Yang; Shenghua Tian; Constance Flamand...; Ling Gao; Wei Zhang; Yan Li; Jiajia Wang; Zhou Sun; Ying Su; Zhihou Liang
• Keywords: Mandarin Chinese; Stroke

The lack of effective and rapid aphasia screening scales for early post-stroke patients can lead to less rehabilitation gains unquestionable. Language Screening Test (LAST-a LAST-b two parallel versions which were established avoid retest bias) is the first scale that was confirmed patients. We reconstructed Mandarin (MLAST-a MLAST-b) based on primary ones evaluated their validity reliability. Based original edition's design principal combined with characteristics Chinese our culture, 2 (each included 5 subtests a total 16 items, while either editions consists 15 items) revised. administrated them 154 in acute stroke stage 107 chronic phase without Western Aphasia Battery (WAB). statistical index equivalence MLAST, reliability each version, inter-rater After redundancy items deleting, final MLAST 14 perfect score. MLAST-a MLAST-b equivalent an intraclass correlation coefficient 0.991. Internal fine, none showed floor or ceiling effect, good internal consistence (Cronbach alpha 0.956, 0.965) presented, not exist. sensitivity 1 specificity 0.883 at point (total score MLAST-a＜ 14) against 0.981 0.907 same point. Interrater (intraclass 1) median time complete initial 98 seconds (representing least 45 seconds, maximum 196 seconds). Y-axis represents true-positive rate, namely sensitivity; X-axis false-positive 1-specificity. blue curve MLAST-a; green MLAST-b; off-white MLAST-b. Receiver operating characteristic (ROC) showings sensibility (true-positive rate) 1-specificity (false positive (MLAST) comparison battery (WAB) used external validation. optimal cutoff corresponds angle upper left-hand corner diagram When other, improved. At MLAST-b<27). specificity' 0.926 0.981. quotient Scatter between WAB marked by circles: circles. correlation. Histogram matrix version corresponding (p<0.01). color. purple Naming (M-LAST: items/5 points; WAB: 31 items/100 points); Repetition item/2 point; Automatic speech series/1 series/20 Picture recognition 4 items/4 30 items/120 Verbal instructions orders/2 11 orders/80 points). Schematic representation validation process. Each enrolled acute(with new stoke three days) chronic(10 days after onset) completed tests MLAST-b, validity(item analysis, reliability, factor structure) assessed patients; examiners required give marks simultaneously 111 patients, time-taken accomplish test recorded meanwhile, then calculated from respect (enrolled asked followed day, administered different blinded examiners).