The present study was based on a Master's degree thesis and was approved by the Research Ethics Committee, according to Resolution N. 510/2016 of the National Health Council, under Opinion n. 2.572.678.

This cross-sectional descriptive study identified 13 children with significant reading difficulties. The study was based on a population of 301 children enrolled in the 4th and 5th years of elementary school in one of the best public schools in the ranking of the state of São Paulo, Brazil, with a Basic Education Development Index — IDEB12 of 8.1 in 2017. Of these, 166 were included in the study and met the criteria: age between 8 years and 12 years and 11 months, normal otorhinolaryngological and audiological evaluation, normal visual acuity, consent form signed by parents or guardians and assent form signed by the child and absence of previously diagnosed diseases, such as genetic syndromes, intellectual disability, autism spectrum disorder and attention-deficit/hyperactivity disorder. Subsequently, the teachers of the 166 children answered a questionnaire (Table 1) to identify those who had great difficulty in reading to the point of compromising their school performance, despite the teacher’s dedication, so that a professional evaluation by a specialist in phoniatrics could be indicated.

Finally, based on the questionnaire answered by the teacher, only 13 children with great difficulty in reading were selected for the study, which corresponded to 7.8% of the total population that was initially analyzed and included in the proposed criteria; there were 5 females and 8 males, with a mean age of 9 years and 9 months; 12 children were attending the 4th year of elementary school and one child was attending the 5th year of elementary school.

The 13 children selected for the study were evaluated during a phoniatric consultation and then were classified into groups, according to reading skills and retelling of the text “The Bundle of Sticks” (Fig. 1). The reading was performed aloud by the child, during which fluency and decoding of graphic symbols were observed, while text comprehension was assessed through retelling. The examiner then read the text aloud and again assessed text comprehension through the child's retelling.

Fig. 1.

Text for reading assessment — “The Bundle of Sticks”. The children in GR1 did not read it because they were not literate yet.

(0.41MB).

Fluency and decoding were considered very altered when the reading was performed with difficulty, showing problems in precision and speed, prolonged pauses and several changes and (or) omissions of phonemes. Fluency and decoding were considered little altered when the child showed some changes or omissions of phonemes that showed little impact on the speed and accuracy of reading. Fluency and decoding were considered adequate when the child showed good accuracy and reading speed and did not show phoneme changes and (or) omissions. This analysis was considered subjective and depended on the examiner's observation and judgment.

Based on the reading and retelling task, the children were grouped into 6 groups (GR):

GR1 — children who could not read the text. They were not literate yet. They did not retell the text after the examiner's reading.

GR2 — children who read the text but did not retell anything after their own reading or after the examiner’s reading.

GR3 — children who read the text and did not retell after their own reading; however, after the examiner’s reading, they retold the overall content of the text.

GR4 — children who read the text and were not able to retell any of the text after their own reading; however, after the reading was performed by the examiner, they were able to retell the entire text in details.

GR5 — children who read the text and retold the overall content of the text after their own reading.

GR 6 — children who read the text and retold the text in detail after their own reading.

Table 2 depicts the children who belong to each group. None of the children retold the text in details after reading it themselves, a necessary criterion to be included in GR6.

Subsequently, the 13 selected children were evaluated by standardized and non-standardized tests used in the phoniatric assessment according to parameters published by Dualibi et al.13 The assessment of auditory and visual perceptual skills was analyzed through the application of 22 tests:

Figure naming test for children — this test is a short version of the 60 figure naming test by Seabra et al.14 The variable below the mean for age was used for the analysis.

Auditory discrimination test — assessed using a standardized technique for children aged 5–9 years, consisting of 30 pairs of syllables, with 10 equal pairs and 20 different pairs by Rodrigues.15 The altered variable was used for the analysis.

Syllable synthesis, phonemic synthesis, alliteration, syllable segmentation, phonemic segmentation, rhyme, syllabic manipulation, phonemic manipulation, syllabic transposition and phonemic transposition — standardized oral production phonological awareness skills tests for children aged 3–14 years by Seabra and Capovilla.16 The variable below average for age was used for the analysis.

Repetition of numbers in random order and repetition of numbers in reverse order — auditory working memory tests, carried out through the repetition of digits in random order and repetition of digits in reverse order, by Capellini et al.17 the altered variable was used for the analysis.

Pseudoword repetition — word and pseudoword repetition test by Seabra and Capovilla.16 The variable below average for age was used for the analysis.

Visual discrimination of letters and visual discrimination of words — evaluation of the visual discrimination of letters and words, using exercises of discrimination in letters and words proposed by Myklebust and Johnson.18 The altered variable was used for the analysis.

Visual memory — the standardized test from the book “Cognitive perceptual motor disfunction”, by Rubin et al.19 for children aged 6 years 2 months to 9 years and 7 months was performed. In children above this age, visual memory was considered impaired when it was below the highest age range standardized in the test. The child received 10 cards with one, three or four geometric figures of varying complexity. One card at a time was presented for 10 s; then, the examiner removed the form and the child had to draw the corresponding geometric shapes. The altered variable was used for the analysis.

Rapid figure naming — a test standardized by level of schooling by Capellini et al. was used.17 Naming tests require quick, successive and sequential evocation of symbols, as well as in reading, where there is rapid and successive decoding of symbols. In this test, for children attending the 4th and 5th years of elementary school, it is expected that the time spent to name the figures in sequence do not exceed 40″, being considered under attention when this time exceeds 50″. The figures used in this test are simple and common to children. It was considered adequate when the obtained result was within the expected level for the level of schooling; it was considered altered when the obtained result was “under attention” The altered variable was used for the analysis.

Visual synthesis with words — a non-standardized test proposed by the examiner was used. In this test, the child has to identify six words through scattered letters: plate, door, car, blue, plane, soap and pencil. The child was considered as having an adequate performance in the visual synthesis task when they were able to write most of the presented words (50% + 1); and altered when they could not analyze and write any of the words or less than 50% of the presented words. The altered variable was used for the analysis.

Figure copying — a test that assesses spatial organization in the graphic plane through the copying of geometric figures.20 The altered variable was used for the analysis.

The data were initially descriptively analyzed due to the small sample size. Subsequently, the dendrogram statistical method was used, which is a quantitative method that organizes in a diagram the groupings comprising the variables and their levels of similarity,21 aiming to assess the correlation between altered or below-average results in auditory and visual perception tests. In the word repetition test, none of the children obtained below-average values; therefore, this variable was not considered in the remainder of the analysis and, therefore, 21 tests were considered in this analysis of groupings. Using this technique, the test results were grouped in such a way that those in the same group are more correlated with each other than with the tests in another group. The phi correlation coefficient was considered a measure of similarity and the grouping method adopted was the mean of distances.22 The phi correlation coefficient was also used to identify tests of which results are highly correlated.

Finally, the results obtained in the dendrogram with the grouping of tests were compared with the classification of the six groups according to reading and retelling skills. The discriminatory capacity of the groups was assessed individually, as the sample size does not allow the use of multivariate techniques.

Table 3 shows anamnesis data related to the subjects’ personal background, by groups, namely: speech delay; persistent speech disorders at six years of age, perinatal diseases, previous history of recurrent otitis, prematurity; previous speech therapy and previous psychological therapy. Other background data were not reported by the subjects during the phoniatric consultation and therefore are not included in the table.

There were no children with previous speech disorders in GR5 and children who had a history of persistent speech disorder at six years of age belonged to GR4, children with considerable decoding difficulty. It is also observed, in relation to personal history, that the two subjects who belonged to GR3 had perinatal diseases and mild prematurity.

Regarding therapy, it is shown in Table 3 that five children from different groups had never undergone any type of therapy, such as treatment for complaints of school difficulties, while eight children were undergoing or had undergone therapy, such as speech therapy and psychological therapy; among them was subject 12 from GR5, who also had a positive personal history of recurrent otitis.

As for family history, four children had a positive family history of language or learning problems; of these, two belonged to GR1, one to GR2 (subject 3) and another to GR5 (subject 10). Two children had a positive family history of psychological disorders, one from GR1 (subject 2) and the other from GR2 (subject 4).

Table 4 represents the percentage of individuals in each group with alterations in fluency and decoding. This table shows that the three children from GR4 read with very altered fluency and decoding and this was the group with the worst performance among those who participated in the reading fluency and decoding assessment. The GR5 was the group with the best performance, and subjects 11 and 12, from this group, showed little altered fluency. It is also observed that all subjects in GR3 showed alterations in reading fluency. The two subjects from GR1 could not read.

The two children in GR1 did not participate in the reading test, as they were not yet literate, and the individual totals were divided by the number of tests in which each child participated, thus obtaining, for each child, the proportion of altered or below-average results. Fig. 2 shows the proportion of the median of the altered test results in which each child participated; the proportion of altered results decreases from GR1 to GR3 and the only group with a median greater than GR4 is GR1 and the lowest median was observed in GR5. It is also observed that the medians in GR3 and GR5 are close. The points are identified by the child’s number in each group.

Fig. 2.

Individual and median proportions of altered or below average results in tests of auditory and visual perceptions and reading in the GR1 to GR5 groups. *The numbers in each point correspond to the children’s identification. **GR6 — none of the children retold the text in details after their own reading.

(0.11MB).

Fig. 2 shows that subject 12, despite belonging to GR5, that is, having better reading comprehension in comparison to GR4, has perceptual characteristics that are similar to those of GR4.

To identify the variables among the 21 tests, the results of those with similar behavior were grouped using the group analysis technique, which generated a dendrogram (Fig. 3) with seven groups of variables.

Fig. 3.

Dendrogram obtained from the analysis of the grouping of 21 variables related to tests of auditory and visual perceptions in which all subjects participated. *Grouping 1: figure naming, phonemic synthesis, rhyme, phonemic manipulation; grouping 2: auditory discrimination; grouping 3: syllable synthesis; grouping 4: alliteration, phonemic segmentation, syllable transposition, reverse order repetition, visual memory, visual synthesis with words, figure copying, rapid figure naming; grouping 5: syllable segmentation, visual letter discrimination; grouping 6: syllable manipulation, random number repetition, visual word discrimination; grouping 7: phonemic transposition, pseudoword repetition.

(0.32MB).

The discriminatory capacity of the tests was individually assessed considering the groupings. The percentages of individuals with an altered or below-average results in each of the variables in the different groupings are shown in Tables 5, 6, 8, 9, 11, 12 and 14. The phi correlation coefficient values ​​of variables in the same grouping are shown in Tables 7 and 10. The performance in the auditory discrimination and syllabic synthesis tests showed a poor correlation with the other tests. These results suggest that:

In grouping 1: the figure naming, phonemic synthesis, rhyme and phonemic manipulation tests discriminated the groups GR1, GR2 and GR4 from the others (Table 5). There is a strong correlation between the phonemic synthesis and phonemic manipulation tests, and between phonemic manipulation and rhyme (Table 7); they were the ones with the highest phi correlation coefficient value, with both being significant. The figure naming test showed a moderate correlation with phonemic synthesis and phonemic manipulation (Table 7). Most of the children from GR1, GR2 and GR4 showed alterations in phonemic synthesis, phonemic manipulation, rhyme and figure naming. These children represent those who had difficulty retelling the text after their own reading. They are the ones that showed the worst performance in reading, either due to decoding difficulty, as those in GR4, comprehension difficulty, as those in GR2 or due to both difficulties, as those in GR1. Although the existing difficulties among the children are diverse, this set of tests showed sensitivity to identify children with the greatest difficulties in reading.

In grouping 2: the auditory discrimination test discriminated the GR4 from the others (Table 6). The auditory discrimination, although it did not show a significant correlation with the other tests, showed good sensitivity to identify children in GR4, being altered in all of them.

In grouping 3: the syllabic synthesis test discriminated GR1 and GR4 from the others (Table 8 ). In this case, an alteration was observed only in the children from GR1 and GR4, the groups to which the children with greater difficulties in decoding belong.

In grouping 4: the alliteration, phonemic segmentation, syllable transposition, repetition of numbers in reverse order, visual memory, visual synthesis with words, figure copying, and rapid figure naming tests discriminated GR1, GR2 and GR4 (Table 9). The high correlations between the repetition of numbers in reverse order (working memory) and figure copying with most variables in this grouping stand out (Table 10).

In grouping 5: the results shown in Table 11 suggest that the syllable segmentation and visual letter discrimination tests did not adequately discriminate the groups.

In grouping 6: the results shown in Table 12 suggest that the syllable manipulation, random number repetition and visual word discrimination tests discriminated GR1 from the others. The phi correlation coefficient value between the random number repetition and the visual word discrimination tests is equal to 1 (Table 13), that is, the individuals in the sample showed the same response in both tests.

In grouping 7: The results shown in Table 14 suggest that the phonemic transposition and pseudoword repetition tests did not discriminate between the groups.