The BREV neuropsychological test: Part II. Results of validation in children with epilepsy C Billard* MD, Department of Neuropaedatrics, Bicêtre Hospital, Le Kremlin Bicêtre; J Motte MD; M Farmer MD, Department of Neuropaedatrics, American Hospital, Reims; M O Livet MD, Department of Neuropaedatrics, Regional Hospital, Aix-en-Provence; L Vallée MD, Department of Neuropaedatrics, Roger Salengro Hospital, Lille; P Gillet PhD, Department of Neuropaedatrics, Clocheville Hospital, Tours; S Vol MSc, Institute of Health (IRSA), La Riche, France. The ‘BREV’ group: MA Barthez, N Bednarek, N Badinand, Y Chaix, R Cheminal, JC Cuvelier, JM Cuisset, A Dusser, B Echenne, S Gonzales, J Mancini, ML Moutard, S Nguyen, D Parain, MJ Peniello, S Peudenier, C Rousselle, P Sabouraud, A de Saint Martin, D Sotto, M Tardieu. *Correspondence to first author at Unité de rééducation Neuropédiatrique Service de Neuropédiatrie, Bicêtre, 78 rue du Général Leclerc, 94275 Le Kremlin Bicêtre, France. E-mail:
[email protected] The Battery for Rapid Evaluation of Cognitive Functions (Batterie Rapide d’Evaluation des Fonctions Cognitives: BREV) is a quick test to screen children with higher-functioning disorders and to define the patterns of their disorders. After standardization tests in 500 normally developing children aged 4 to 8 years, validation consisted of comparative evaluation of the specificity and sensitivity of the BREV with a wide reference battery in 202 children with epilepsy (108 males, 94 females; mean age 6 years 6 months, SD 1 year 8 months). Children were divided into 10 age groups from 4 to 8 years of age and represented eight epileptic syndromes. The reference battery included verbal and non-verbal intelligence assessment using the Wechsler scale, oral language assessment with a French battery for oral language study, drawing with the Rey figure, verbal and visuo-spatial memory with the McCarthy scale subtest and the Rey figure recall, and educational achievement with the Kaufman subtests. Every function evaluated with the BREV was significantly correlated with the reference battery testing a similar function (p=0.01 to 0.001). Specificity and sensitivity of the BREV verbal and non-verbal scores were correlated with those of the Wechsler scale in more than 75% of children. The BREV, therefore, appears to be a reliable test which has been carefully standardized and validated and is valuable in screening for cognitive impairment in children.
398
Developmental Medicine & Child Neurology 2002, 44: 398–404
Early screening of children for impairment of cognitive and verbal and non-verbal functions is essential for effective treatment. Many of such screening tests have been developed in English (see Law et al. 1998 for a systematic review of the literature). A comparative review of 51 tests in English (Sturner et al. 1994) emphasizes the importance of batteries that explore both oral language and non-verbal functions to guide appropriate choices of education and treatment. The current screening batteries standardized in French involve only verbal functions. Each battery is standardized for a single age group: the épreuves pour le repérage des troubles du langage chez l’enfant de 4 ans (ERTL 4; Alla et al. 1998 ) for 4-yearolds; the épreuves pour le repérage des troubles du langage et des apprentissages pour les enfants de 6 ans (ERTLA 6; Roy et al.1999) for 5- to 6-year-olds; and the battery designed by Zorman and Jacquier-Roux (1999) also for 5- to 6-year olds. A paediatric version of the Minimum Mental State examination (MMS) has been validated recently (Ouvrier et al. 1999), but it only provides an overall score and not the pattern of possible deficits and has not been standardized in French. The BREV is a neuropsychological clinical examination tool which in 30 minutes provides objective information concerning expressive and receptive verbal language, nonverbal and executive functions, attention and verbal and visuo-spatial memory, and educational achievement (reading, spelling, and numeracy). As with every screening battery, any abnormal score should be investigated with a more detailed classical test. When it was conceived the BREV took into consideration normal cognitive development and learning. The first validation phase comprised battery standardization to define reference values and reproducibility in a population of 500 children, divided into 10 age groups from 4 to 8 years of age, attending mainstream school and free of any neurological disorder (see Part I: Billard et al. 2002). The second validation phase, reported here, estimated the sensitivity and specificity of the BREV for each of the functions explored in comparison with a reference battery in 202 children with epilepsy. Method PARTICIPANTS
The validation population of 202 children with epilepsy (108 males, 94 females; mean age 6 years 6 months, SD 1 tear 8 months) was chosen from 15 French neuropaediatric centres according to the following criteria: (1) attendance at an ordinary school (specialized class or not); (2) regular distribution in 10 × 6-month age groups between the ages of 4 and 8 years (mean 6.9 years), and according to sex distribution; and (3) diagnosed with one of the eight generalized or partial epileptic syndromes (Table I). Children with untestable intellectual competence due to a major deficit or behaviour problems were excluded. PROCEDURE
BREV scores were evaluated by comparing the results with those of a large battery of classical cognitive function tests. The protocol included a neuropsychological evaluation in three sessions of one hour to one hour 15 minutes over a single day. The order of the three sessions was random: 1, 2 then 3; or 2, 3 then 1, or 3, 1 then 2). Tests were performed by a single psychologist at each centre who was experienced in carrying out such tests.
Table I: Characteristics of population according to epilepsy type Epilepsy type
n
Petit mal absences Grand mal Doose syndrome Generalized (other) Benign epilepsy with centrotemporal spikes Benign partial (other) Partial symptomatic Partial cryptogenic
33 19 9 24 18 21 40 38
Age, mean (range) y
VIQ (SD)
6 y PIQ (SD)
7 (4–9) 6.5 (4.4–8.7) 6.6 (4.6–8.9) 6.2 (4–8.6) 6.7 (4.1–8.8) 6.7 (4–9) 6.7 (4–9) 6.7 (4.2–8.8)
103 (15) 101 (25) 78 (26) 80 (20) 102 (25) 93 (12) 75 (21) 104 (13)
93 (10) 97 (14) 70 (21) 74 (15) 96 (17) 90 (14) 74 (14) 87 (15)
93 (16) 89 (20) 56 (15) 84 (17) 96 (24) 92 (16) 79 (16) 81 (17)
94 (22) 90 (18) 59 (15) 84 (25) 100 (27) 96 (14) 76 (20) 79 (17)
VIQ, verbal intelligence quotient; PIQ, performance intelligence quotient.
Table II: Children with epilepsy (E) and reference population (R): verbal scores for BREVa and WPPSI or WISC–III subtests (mean, SD), for 6 month age groups Subtest n with epilepsy
(4–4.6) 11
Verbal functions Phonology Mean (E) SD Mean (R) Naming Mean (E) SD Mean (R) Repetition of Sentences Mean (E) SD Mean (R) Sentence in response to pictures Mean (E) SD Mean (R) Fluency Mean (E) SD Mean (R) Grammar comprehension Mean (E) SD Mean (R) Syllable subtraction Mean (E) SD Mean (R) Phoneme subtraction Mean (E) SD Mean (R) BREV Verbal score Mean (E) SD Mean (R) VIQ (WPPSI-R/WISC-III) Mean (E) SD
(4.6–5) 13
(5–5.6) 25
(5.5–6) 21
Age group (y) (6–6.6) (6.6–7) 23 24
(7–7.6) 23
(7.6-–8) 19
(8–8.6) 19
(8.6–9) 24
15.4 4.2 18.4
15.6 3 18.6
17.3 3 18.8
16.6 2.8 19.2
18.5 2.8 19.2
18.6 1.8 19.5
16.8 2.7 19.5
19 1.7 19.8
18.7 2.4 19.7
19 2.1 19.8
14.5 4.5 18.2
17.5 2.7 19.1
17.8 3.4 19.3
11.6 5.4 14.6
12.3 4 15.8
13.7 5.1 16.3
13.8 6.1 16.3
13.1 6.5 16.6
13.9 4.3 17.2
14.2 5.9 17.5
10.7 6.1 15.5
14.5 5.1 17
15.1 4.5 17.1
14.5 5.6 18.1
15.7 3.8 18.8
17 3.5 19.3
15.3 4.3 19.3
17.3 3.4 19.5
18.5 2.9 19.8
18 2.7 19.9
13.6 4.7 15.3
16 1.9 15.5
15.9 1.7 16.1
14.2 4.8 16
16 1.9 16.5
16.7 1.7 16.6
16.4 1.8 16.9
17.5 1.7 16.8
17.6 1.4 17
17.9 1.9 17.3
2.5 1.6 4.5
4.7 1.8 4.5
3.8 1.7 5.3
4.4 2.1 5.8
5 1.6 5.7
4.6 2 6.1
4.3 2.1 6.3
5.3 1.7 6.9
5.4 1.7 7.8
6 2.5 7.7
7.4 6.2 11.5
10.3 4.4 14.3
10.9 5.3 15.0
10.6 5.7 16.2
13.5 3.3 17.4
13.8 2.7 17.5
10.4 5.3 17.1
14.8 4.8 17.8
16 3.7 17.8
16.6 3.4 18.3
7.2 8.8 7.6
5.2 7.5 12
8.5 8.9 16.5
12.5 8.7 18.2
9.8 9.4 18.9
14.7 8.2 19.7
17 .3 6.5 19.3
16.7 6.7 19.7
1.4 5 8.7
4.5 6.1 11.2
3.1 6 14.1
10.5 8 16.8
11.1 8.9 16.9
9.9 8 18.5
10.7 2.9 14
13.1 2.4 14.9
13.4 2.2 15.4
12 3.1 15.1
13.2 1.6 15.4
14.1 1.8 15.9
12.7 3 15.8
14.5 2.6 16.2
15 1.6 16.5
15.3 2.3 16.7
92.5 26.4
90.2 16.8
91.7 21.8
86.1 23.2
89.3 16.1
87.3 14.8
76.5 19.8
86.4 21.2
86.7 20.1
87.5 21.4
aMean scores recalculated over 20. WPPS, Wechsler Preschool and Primary Scale of Intelligence (Wechsler 1995); WISC–III, Wechsler
Intelligence Scale (Wechsler 1996); –, test not applicable for these ages.
BREV Neuropsychological Test. Part II C Billard et al.
399
Session 1: BREV (for precise description of method, see Billard et al. 2002). Six normal development-inspired subtests (Chevrie Müller et al. 1981, De Agostini et al. 1998) tested oral language: phonology, naming skills, fluency, grammatical production (repetition of two sentences and one sentence in response to a picture), grammar comprehension, and phonological awareness. Five subtests evaluated non-verbal functions: drawing, visual discrimination, visuo-spatial planning based on Raven’s matrices (Raven 1976), and resolving mazes. Seven subtests evaluated attention and memory: visual attention, selective attention (based on the ‘go-not-go’ paradigm; Passler et al. 1985), digit span, verbal memory (recall of words named in pictures and recall of two repeated sentences), visuo-spatial memory (recall of the positions of the pictures named and recall of the figures copied). Three subtests based on the K-ABC battery (Kaufman and Kaufman 1993) evaluated educational achievement. Reading was tested by non-words, regular words, irregular or orthographically ambiguous words, and sentences. Spelling was tested by dictating sentences chosen from school books and non-words of growing difficulty. Numeracy was evaluated with different counting tests, mental or written calculation, and problems chosen according to arithmetic level. Verbal score was calculated by adding together the scores for phonology, naming, grammar production, grammatical comprehension, and fluency. Non-verbal score was
calculated by summing the scores for drawing, visual discrimination, visuo-spatial planning, labyrinth, and visual attention. Session 2: Wechsler Preschool and Primary Scale of Intelligence–Revised (WPPSI-R; Wechsler 1995) for children under 6 years of age or the Wechsler Intelligence Scale for Children-III (WISC-III; Wechsler 1996) for children from 6 to 8 years of age was used to evaluate verbal intelligence (VIQ) and performance quotients (PIQ). Five of the six verbal subtests (Information, Similarities, Vocabulary, Comprehension, Arithmetic, Digit, or Sentence memory) were used to establish VIQ, and five of the six non-verbal subtests (Picture Completion, Block Design, Object Assembly, Geometric Drawing, Labyrinths, Picture Arrangement, Animal Chequers, or Coding) were used to establish PIQ. Session 3: Multiple trials. Phonological production and naming skills were evaluated with two subtests (word repetition and picture naming) from the French battery ‘Epreuves pour l’examen du Langage’ (Chevrie Müller et al. 1981), verbal fluency with the corresponding McCarthy scale subtest (McCarthy et al. 1972), and syntactic comprehension with the French version of the North Syntax Screening Test (Weil Harpen et al. 1983). Drawing ability was evaluated with the Rey figure (Rey 1959) and verbal memory was evaluated with the story recall subtest of the McCarthy scale (McCarthy et al. 1972). Visuo-spatial memory was evaluated with Rey figure
Table III: Children with epilepsy (E) and reference population (R): non-verbal scores for BREV. and WPPSI-R or WISC–III (mean, SD) subtests, for 6 month age groups Subtest n with epilepsy Non-verbal functions Drawing Mean (E) SD Mean (R) Visual discrimination Mean (E) SD Mean (R) Visuo-spatial planning Mean (E) SD Mean (R) Maze Mean (E) SD Mean (R) Visual attention Mean (E) SD Mean (R) Non-verbal Score Mean (E) SD Mean (R) PIQ (WPPSI / WISC) Mean (E) SD
Age group (y) (6–6.6) (6.6–7) 23 24
(4–4.6) 11
(4.6–5) 13
(5–5.6) 25
(5.5–6) 21
(7–7.6) 23
(7.6-–8) 19
(8–8.6) 19
(8.6–9) 24
1.3 1.3 3.6
2.2 2.6 4.1
2.8 2.1 6.4
4.8 3.4 8.1
3.9 2.8 6.1
5.3 2.6 6.7
2.8 2.8 7.8
6.5 3.9 9.3
7.9 4 9.4
8.2 4.5 10.9
5.3 3.5 8.9
7.4 2.7 8.8
7.5 3.8 9.1
8.1 4.0 10.7
9.7 2.6 12.1
9.8 2.6 12.8
8.6 4 12.8
10.8 4 13.8
10.4 2.6 14.5
12.3 3 14.7
3.1 3.1 5.3
4.5 4.2 6.4
7.7 3.4 9.5
7.2 4.1 9.9
9.2 4.7 12.5
9.7 5 12.9
8.2 4.8 12.7
13.8 4.4 14
13.2 5.4 14.9
14 4.5 15.8
4.2 6 13.8
11.6 7.2 15.4
12.4 5.4 16.8
10 5.6 14
11.6 5.4 15.2
11.8 4.8 15.8
8 6.8 15.6
13.8 5.2 16.6
12.4 5.6 16.8
13.2 5.8 16.8
–3.6 5.6 7.8
6.6 4.5 8.4
8.1 3.4 10.2
9.5 4.8 11.8
11 3.5 13
12 3.8 13.7
9.1 5.9 15.5
13.5 3.9 16.1
13.9 4.1 17.4
15.4 4.2 18.3
2.8 2.7 8.2
6.5 3.3 8.7
7.7 2.5 10.5
7.9 3.4 10.8
9 2.4 11.9
9.7 2.5 12.4
7.2 4.1 13
11.7 3 14
11.4 3.1 14.5
12.6 3.4 15.3
82.7 16
91 19.5
82.3 14.7
82.5 17.8
89 15.9
88.3 22.5
74.4 25.3
86 21.8
83.2 19.9
91.8 23.3
a Scores recalculated over 20. WPPSI-R, Wechsler Preschool and Primary Scale of Intelligence(Wechsler 1995); WISC-III, Wechsler Intelligence
Scale (Wechsler 1996).
400
Developmental Medicine & Child Neurology 2002, 44: 398–404
coefficient between each BREV subtest and the corresponding reference battery subtest. The ability of the BREV to detect children with learning disorders was estimated using two levels of screening for PIQ and VIQ: a mean threshold of 85 (nearly the median) and a low threshold of 70 (the first quintile). Optimization of corresponding thresholds for the BREV was obtained by the ROC curves, after score adjustment according to the mean for the group.
recall (Rey 1959). Reading and mathematical achievement were evaluated with the deciphering and calculation subtests of the K-ABC battery (Kaufman and Kaufman 1993). Reliability, sensitivity, and specificity were appraised by comparing verbal and non-verbal BREV data, restored to the reference population mean, with the VIQ and PIQ of the Wechsler scales (WPPSI-R and WISC-III). STATISTICAL ANALYSIS
Statistical analysis was performed with Number Cruncher Statistical Systems 60 program. Scores were all recalculated over 20, except for the verbal fluency score which is an unrefined score. Results for children with epilepsy are shown independently of the type of epilepsy, by age group or school level, with group size (mean and SD) in comparison with similar age and sex samples of the reference population. BREV performance was analyzed by Spearman’s correlation
Results The population with epilepsy comprised 202 children (108 males, 94 females; mean age 6 years 11 months) of whom 24.9% were left handed. Eighty-five children had generalized epilepsy (petit mal epilepsy, tonic–clonic generalized epilepsy, Doose syndrome, or other generalized epilepsies), partial idiopathic epilepsy (n=39), and partial non-idiopathic epilepsy (n=78; partial symptomatic or cryptogenic epilepsy). Age and sex distributions were similar for each epileptic
Table IV: Children with epilepsy and reference population: memory and attention tests for BREV subtests (mean SD)a, for 6-month age groups Subtest n with epilepsy
Age group (y) (6–6.6) (6.6–7) 23 24
(4–4.6) 11
(4.6–5) 13
(5–5.6) 25
(5.5–6) 21
16.2 2.9 18
17.1 2.8 18.2
16.8 3.3 19.3
17.1 6.1 19.5
18.4 2.1 19.8
14.2 7.9 18.3
15.7 5.3 18.7
Selective Attention Control Test Mean (E) SD Mean (R) Selective Attention Conflict Test Mean (E) SD Mean (R) Digit Span Mean (E) SD Mean (R) Word Recall Mean (E) SD Mean (R) Sentence Recall Mean (E) SD Mean (R) Verbal memory Mean (E) SD Mean (R) Position Recall Mean (E) SD Mean (R) Drawing Recall Mean (E) SD Mean (R) Visuo-spatial memory Mean (E) SD Mean (R)
(7–7.6) 23
(7.6-–8) 19
(8–8.6) 19
(8.6–9) 24
18.8 1.6 19.9
17 3.2 19.8
18.8 2.2 19.9
19.4 1.6 19.9
20 0 19.9
16.1 7 19.5
16.7 4.4 18.7
15.3 7 19.6
16.6 5.7 19.5
18.2 3.9 19.5
18.3 4.2 19.8
6.7 2.9 8.3
7.7 2.9 9.8
9.3 4 10.7
8.7 2.9 12
8.9 2.8 13.5
10.5 3.2 13.7
8.3 3.4 14
12.1 4.1 15
11.7 4.4 15
13.3 4.4 15.9
4 5.4 10.4
9.8 6.9 12.3
10.7 3.8 12.8
9.3 4.4 13.4
10.3 5.1 12.9
11.3 5 12.7
9.7 6.6 13.5
10.7 5.7 15.2
11.4 6.4 13.5
12.7 5.4 14.9
1.5 3.2 1.6
1 2.2 2
1.5 2.8 1.8
2.7 3.9 3.5
3 3.4 4.1
5.4 5 4.4
2.5 3.2 5.6
4.2 5.7 6
5.8 4.8 8.1
6.5 5.9 7.2
2.1 2.7 3.9
3.1 2.5 4.5
3.7 2.2 4.5
4.3 3.4 5.9
4.7 3.1 6.2
6.8 4 6.5
4.2 3 7.5
5.7 4.7 8.2
7.1 4.4 9.5
7.9 5.2 9.1
8.4 9 18
17.8 3.9 17.7
15.4 5.9 18.4
13.7 7.2 18.6
16.9 5.9 18.5
17.8 3.3 18
16.5 5 17.7
17.3 5.7 19.1
16.4 6.2 18.2
17.8 4.2 18.9
0.5 0.6 5.8
1.7 1.6 5.6
2.2 2.3 8.4
3.8 2.7 12.1
3.8 1.8 7.3
3.2 2.5 7.9
2.7 2.5 7.6
4.5 2.9 9.1
4.9 2.7 10
5.3 3.5 9.6
4.4 4.4 11.8
9.8 2.3 11.7
8.8 3.5 13.3
8.8 4.2 15.4
10.3 3.3 13
10.5 2.2 13
9.6 3.1 12.9
10.9 3.8 14.4
10.7 3.3 14.4
11.6 3.4 14.3
aMean scores recalculated over 20.
BREV Neuropsychological Test. Part II C Billard et al.
401
syndrome (see Table I). Intelligence scores with the Wechsler scales for the children with epilepsy are given in Tables II and III. Means and SDs of the BREV subtests in the group with epilepsy are shown in Tables II and III for verbal and non-verbal data, in Table IV for attention and memory tests, and in Table V for educational achievement in comparison with 202 children
Table V: Children with epilepsy (E) and reference population (R): learning scores (meana, SD) according to educational level Subtest
Reading Mean (E) SD Mean (R) Spelling Mean (E) SD Mean (R) Numeracy Mean (E) SD Mean (R)
Level 1b n=41
10.7 7.3 14.3
Level 2c n=52
Level 3 n=46
Level 4 n=28
Level 5 n=12
6.2 5.4 10.5
9 7.6 16.5
15.2 5.5 17.5
18.1 2 17.6
7.5 5.8 10.5
9.3 6.8 16.5
11.7 6.2 17.5
11.5 3.4 17.6
14.4 5.1 17.4
9.4 5.8 13
14.9 4.3 18.2
15.5 5.6 16.1
aMean scores were recalculated over 20. Level 1, second preschool year; Level 2, last preschool year; Level 3, first school year; Level 4, second school year; Level 5, third year school. For reading and spelling: b 3rd year of kindergarten from Christmas or first year of primary school before Christmas; c First year of primary school from Christmas.
from the 500 reference population (Billard et al. 2002), matched for age and sex. Scores for this population with epilepsy according to epilepsy type will be the subject of a future publication. Each of the BREV subtests was compared with a similar subtest in the reference population (Table VI). For some of the subtests, the functions evaluated were similar (e.g. phonology or educational achievement). For some other subtests, functions were similar but test conditions were different (e.g. Wechsler scale picture completion test: not penalized by time or visual discrimination; BREV scale: limited to 20 seconds). Other subtests were close in so far as they were testing a verbal or non-verbal cognitive function, even when the test principles were different (e.g. visuo-spatial planning [BREV] compared with PIQ, but the tests were different). The correlations between the BREV and reference batteries are detailed in Table VI. Each BREV subtest and comparable cognitive function item in the reference battery scores was significantly correlated, with correlation coefficients ranging from 0.32 (visual discrimination; p
