Adaptação do protocolo M1-Alpha para a avaliação da linguagem à beira do leito

Abstract

The objective of this study was to adapt the M1-Alpha and to search for content validity evidence. The process of adaptation of the instrument was divided into four steps. Step 1: Three language and speech therapists, two of whom were experts on aphasiology and one, a master's student in human communication disorders, analyzed the instruments M1-Alpha (only available for research and for some clinicians), the Montreal-Toulouse Language Assessment Battery for Aphasia-BR (MTL-BR) and MTL-BR version B (not marketed), to check for possible needs for modifications and select the stimuli to compose the instrument being adapted. Step 2: Twenty-eight non-expert judges analyzed the representativeness of 65 drawings. When necessary, the judges could make suggestions for improvements. Step 3: Five judges, who are experts on assessment of aphasia, determined whether the pictorial stimuli were adequate while four other judges, experts on linguistics and phonological acquisition, determined if the words, pseudowords and sentences were appropriate. As in Step 2, the expert judges were instructed to suggest modifications when necessary Step 4: This step was divided into pilot study 1 and pilot study 2, both with healthy individuals. After modifications suggested by the judges, the first draft of the instrument was designed and then applied to the sample of pilot study 1, composed of seven individuals. Based on pilot study 1, punctuation errors and inadequacies of some stimuli could be identified. After the adjustments required, the latest version of the protocol M1-Alpha was designed and then called MTL-BR Brief. The instrument was applied to 63 individuals from pilot study 2. Analysis of data from steps 1 and 4 was descriptive. For step 2, index of agreement among judges was determined using simple percentage agreement. The results of the expert judges were analyzed by calculating the content validity ratio while the first-order agreement coefficient (AC1 statistic) of Gwet was used to check for the correlation between expert judges. After analysis of the instruments (Step 1), 65 figures were designed and redesigned and then analyzed by non-expert raters (Step 2), and 60 of them showed more than 80% agreement. After adjustments, they were analyzed by the expert judges (Step 3); all 23 words, pseudo words and phrases were considered to be appropriate, while 22 out of the 66 pictorial stimuli required changes. Subsequently, they were analyzed by the same expert judges, and there was high agreement AC1= 0,98 [IC=0,96-1]. On the basis of the pilot studies (Step 4), some errors were found and then corrected to produce the latest version of the instrument. In conclusion, satisfactory content validity could be achieved in the adaptation of the M1-Alpha test, originating the MTL-BR Brief Battery.