Brazilian Validation of the Alberta Infant Motor Scale

Nadia Cristina Valentini and Raquel Saccani

PHYS THER. 2012; 92:440-447.
Originally published online December 1, 2011
doi: 10.2522/ptj.20110036

The online version of this article, along with updated information and services, can be
found online at: http://ptjournal.apta.org/content/92/3/440

Collections This article, along with others on similar topics, appears

in the following collection(s):
Motor Control and Motor Learning
Motor Development
Tests and Measurements
e-Letters To submit an e-Letter on this article, click here or click on
"Submit a response" in the right-hand menu under
"Responses" in the online version of this article.
E-mail alerts Sign up here to receive free e-mail alerts

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016

 Research Report

Brazilian Validation of the

Alberta Infant Motor Scale
Nadia Cristina Valentini, Raquel Saccani
N.C. Valentini, PhD, Departments
of Physical Therapy and Physical
Education, Universidade Federal
Background. The Alberta Infant Motor Scale (AIMS) is a well-known motor
do Rio Grande do Sul, 750 Feliz- assessment tool used to identify potential delays in infants’ motor development.
ardo Jardim Botanico Porto Alegre, Although Brazilian researchers and practitioners have used the AIMS in laboratories
Rio Grande do Sul 91751-330, and clinical settings, its translation to Portuguese and validation for the Brazilian
Brazil. Address all correspondence population is yet to be investigated.
to Dr Valentini at: nadiacv@esef.
ufrgs.br.
Objective. This study aimed to translate and validate all AIMS items with respect
R. Saccani, Department of Physical to internal consistency and content, criterion, and construct validity.
Therapy, Universidade de Caxias
do Sul, Caxias do Sul, Rio Grande
do Sul, Brazil. Design. A cross-sectional and longitudinal design was used.
[Valentini NC, Saccani R. Brazilian Methods. A cross-cultural translation was used to generate a Brazilian-Portuguese
validation of the Alberta Infant
Motor Scale. Phys Ther. 2012;
version of the AIMS. In addition, a validation process was conducted involving 22
92:440 – 447.] professionals and 766 Brazilian infants (aged 0 –18 months).
© 2012 American Physical Therapy
Association
Results. The results demonstrated language clarity and internal consistency for the
motor criteria (motor development score, ␣⫽.90; prone, ␣⫽.85; supine, ␣⫽.92;
Published Ahead of Print: sitting, ␣⫽.84; and standing, ␣⫽.86). The analysis also revealed high discriminative
December 1, 2011
power to identify typical and atypical development (motor development score,
Accepted: November 7, 2011
Submitted: February 4, 2011 P⬍.001; percentile, P⫽.04; classification criterion, ␹2⫽6.03; P⫽.05). Temporal sta-
bility (P⫽.07) (rho⫽.85, P⬍.001) was observed, and predictive power (P⬍.001) was
limited to the group of infants aged from 3 months to 9 months.

Limitations. Limited predictive validity was observed, which may have been due
to the restricted time that the groups were followed longitudinally.

Conclusions. In sum, the translated version of AIMS presented adequate validity

and reliability.

Post a Rapid Response to

this article at:
ptjournal.apta.org

440 f Physical Therapy Volume 92 Number 3 March 2012

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
 Brazilian Validation of the Alberta Infant Motor Scale

M
otor delays during infancy cally aimed: (1) to translate the AIMS Table 1.
could indicate potential risks to Portuguese-Brazilian, (2) to probe Sample Distribution by Age and Sex
of developmental disorders.1,2 language clarity and pertinence of (N⫽766)
Early assessment of infants’ develop- the Portuguese version of the AIMS, Age (mo) at Female Male
mental status provides valuable and (3) to analyze the temporal sta- Examination 1 n (%) n (%)
information to identify motor deficits bility and construct validity of the 0–3 63 (16.9) 69 (17.5)
and to guide practitioners to prop- AIMS for Brazilian infants. 4–6 68 (18.3) 71 (18.0)
erly define appropriate programs to 7–9 67 (18.0) 78 (19.8)
improve motor competency.3,4 Ade- Method
10–12 73 (19.6) 54 (13.7)
quate assessment of an infant’s Participants
motor development depends on Four bilingual (2 English and 2 Por- 13–15 52 (14.0) 63 (16.0)

the use of reliable and valid instru- tuguese native speakers) individuals 16–18 49 (13.2) 59 (15.0)
ments.5–7 Many well-established participated as translators in this
assessment tools7 have been used to study. In addition, 20 expert health-
identify typical and atypical motor, related professionals (physical thera- in the province of Alberta, Canada.
cognitive, and social development.5,7 pists, physical therapy educators, The scale contains 58 motor items
However, in different cultures, a pediatricians, and nurses) partici- divided into 4 subscales: prone (21
challenge of an early diagnosis is the pated in the process of the AIMS con- items), supine (9 items), sitting (12
absence of translated and valid tent validation. The governmental items), and standing (16 items). Each
instruments to assess motor develop- health and education boards medi- item is described specifically consid-
ment.7–9 The Alberta Infant Motor ated the contact between the ering the weight-bearing surface of
Scale (AIMS) is one example of a researcher and the infants’ families. the body, the posture necessary to
motor assessment tool used broadly A total of 766 preterm and full-term achieve the gross motor skill, and the
in Brazil without specific cultural infants (394 boys and 372 girls), antigravity or voluntary movement
validation.10 –12 from 0 to 18 months of age (Tab. 1), performed by the infant in the posi-
were assessed individually in day tion.26 The assessment can be com-
The AIMS is an observational tool in care centers, basic health govern- pleted typically with 20 minutes of
which the infant’s global develop- mental units, and family homes. behavioral observation. The sum of
ment is assessed based on the inte- Infants with a history of osteomioar- the observed criteria for each sub-
gration of antigravitational muscular ticular affections (fractures, periph- scale comprises the total raw score
control in 4 positions: prone, supine, eral nerve injury, osteomuscular (0 –58 points). The final raw scores
sitting, and standing.13 The scale was infection) or any other neuropathol- can be converted into percentile
designed primarily to screen motor ogy were excluded from the study. ranks and compared with the ranks
development of Canadian full-term Consent was obtained from the cus- of age-matched peers. Classification
and preterm infants. Because of its todial caregivers of each infant as is provided using standard deviations
practicality and psychometric char- well as from each professional par- for each group.13 Infants with scores
acteristics, the AIMS was widely ticipating in the study. All partici- below minus 2 standard deviations
adopted by researchers and health- pants gave informed consent based are referred to as having abnormal
related practitioners working with on procedures approved by the Fed- motor development. Infants scoring
infants in clinical settings.14 –22 eral University of Rio Grande do Sul between minus 2 standard devia-
Although there is clear evidence of Institutional Review Board. tions and minus 1 standard deviation
increased use of the AIMS, estimates are referred to as having suspicious
of its reliability and validity are Instruments motor development. Infants scoring
restricted in several countries. It is AIMS. The AIMS is a behavioral above minus 1 standard deviation are
not known, for example, whether motor assessment tool that requires considered in the range of normal
the current Canadian norms are careful observational techniques and motor development.
appropriate for infants raised with minimal infant handling. It was
different rearing practices or from designed to assess the development Child Behavior Development
different cultural backgrounds. Tests of newborn infants (born preterm Scale (CBDS). The CBDS27 is a
and scales standardized in a certain and full-term) until 18 months of cor- scale standardized for Brazilian chil-
country may not be appropriate for rected age.13 The scale was validated dren that assesses spontaneous and
infants developing under different in 1994. The norms were developed stimulated behaviors. It contains 64
socioeconomic and cultural con- based on data from 2,202 infants items to evaluate cognitive, motor,
straints.18,21,23–25 This study specifi- aged from 1 week to 18 months born and social milestones in infants from

March 2012 Volume 92 Number 3 Physical Therapy f 441

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
Brazilian Validation of the Alberta Infant Motor Scale

1 to 12 months of age. For this study, face validity procedure also was con- session) was examined throughout 6
only the motor subscale was used. ducted. Each evaluator received the months, resulting in a total of 5
All infants were assessed for posture, AIMS-BR and used a 5-point Likert assessments for each infant. Group 2
trunk control, and interlimb coordi- scale to score validity of all motor (13 boys and 19 girls; X⫽5.22,
nation. The scale provides raw, per- items: prone, supine, sitting, and SD⫽4.12, weeks of age at the first
centile, and standardized scores and standing positions. assessment session) was tested and
allows for prediction of expected retested within a 6-month interval.
behaviors for each age. A sociodemo- A sample of 259 infants’ scores were
graphic questionnaire also was used. used to test for internal assessment Data Analysis
The parents or primary caretakers consistency. The infants were A sample size of 600 was needed to
completed questions about infant retested within 1 week (X⫽7.83 determine the validity of the AIMS
sex and chronological and corrected days, SD⫽3.27) by the same evalua- for Brazilian infants within a confi-
ages. tor for the test-retest reliability anal- dence interval of 95% and a 4%
yses. The test-retest reliability was margin of error. A conservative
The test section (AIMS and CBDS) conducted to confirm consistency of approach to prevent missing data
took approximately 20 minutes for the AIMS items and to provide infor- was adopted. Approximately 10
each infant. All test sections were mation about the temporal stability more infants in each age group were
video recorded for further examina- of the scale.31–33 Independent pro- tested. The calculation was obtained
tion. All parents were informed fessionals independently assessed using PEPI-for-Windows: computer
about the results. Information about the videotapes for the interrater and programs for epidemiologists, ver-
public services was provided to the intrarater reliability tests. sion 4.0 (Brixton Health, London,
families of infants identified with low United Kingdom).
motor scores. To verify the discriminant validity of
the AIMS, 124 infants (aged 0 –15 To analyze the clarity and pertinence
Procedure months) were assessed. Sixty-two of the AIMS items, a content validity
A double-back, reverse, indepen- infants with atypical development index (CVI) 30 and the kappa concor-
dent translation procedure was (extremely premature) were paired, dance coefficient31 were used. Per-
adopted.28,29 This procedure involved considering chronological age, with cent agreement was used to assess
4 bilingual professional translators the same number of infants with typ- professional’s face validity. Internal
and required 2 independent transla- ical development (born full term). consistency analysis was verified
tions from English to Portuguese. The premature infants’ age was not using the Cronbach alpha index.34,35
Afterward, 2 independent transla- corrected because the goal was to Test-retest reliability was assessed
tions from Portuguese back to Eng- verify whether the outcomes of the using the Spearman correlation coef-
lish were completed. The translators AIMS assessment allow for discrimi- ficient.33 The interrater and intra-
did not have access to the original nation between infants developing rater scores were analyzed using
English version.13 After independent typically and those developing atyp- intraclass correlation coefficients
completion of 4 versions of the ically. Concurrent validity was tested (ICCs),32 the Friedman test, and the
AIMS, all translators were invited to a using the CBDS and AIMS in a group Wilcoxon test. Discriminant validity
meeting in which all 4 versions were of 40 preterm and full-term infants was tested using the Student t test,
compared with the original version (aged 0 –12 months). This procedure Pearson chi-square test, and Pearson
of the AIMS. The 2 Portuguese ver- was adopted to compare the results correlation coefficient.35 To investi-
sions were revised, and the seman- obtained from the AIMS-BR and the gate concurrent validity, the Spear-
tics were adjusted based on unani- CBDS.27 Predictive validity was used man correlation coefficient and the
mous agreement. A final translated to test the AIMS’ criterion validity. McNemar-Bowker test were used.
and edited scale resulted in the The procedure was conducted using The predictive validity34 of the AIMS
Brazilian-Portuguese version of the 2 groups of infants: group 1, longi- was analyzed using the following:
AIMS (AIMS-BR). tudinally assessed throughout 1 ICC, Friedman test, Pearson chi-
month, and group 2, longitudinally square test, Pearson correlation coef-
A panel of expert health-related pro- assessed within 6 months. This pro- ficient, and analysis of variance
fessionals was assembled to test con- cess verified the instrument’s power (ANOVA) for repeated measures.
tent validity30 of the AIMS-BR. The to accurately predict future develop-
experts used a Likert scale to inde- ment. The infants in group 1 (12 The statistical tests were chosen
pendently assess language clarity and boys and 8 girls; X⫽3.80, SD⫽2.42, according to the data distribution
the pertinence of all motor items. A weeks of age at the first assessment (parametric or nonparametric) and

442 f Physical Therapy Volume 92 Number 3 March 2012

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
 Brazilian Validation of the Alberta Infant Motor Scale

variables (qualitative and quantita- Table 2.

tive) considered for the analyses. Validity Content Index (CVI) and Kappa Concordance Coefficient
Correlation coefficient values over Comparison CVI (%) Kappa (95% CIa) P
.60 were considered strong correla-
Clarity
tion; values between .30 and .60,
Judge 1 ⫻ judge 2 ⫻ judge 3 66.7 ...b ...
moderate correlation; and values
under .30, weak correlation.28,34 Judge 1 ⫻ judge 2 92.8 .87 (.75–.99) ⬍.001
Cronbach alpha values over .80 were Judge 1 ⫻ judge 3 73.2 .75 (.58–.92) .002
considered excellent; values over .70, Judge 2 ⫻ judge 3 67.8 .51 (.28–.74) .028
good; and values between .60 and .70, Pertinence
acceptable.35 Partial ␩2 (eta squared)
Judge 1 ⫻ judge 2 ⫻ judge 3 98.2 ... ...
was presented as an index of effect
Judge 1 ⫻ judge 2 ⬍.001
size (ie, small effect size, ␩2⫽.01; mod- 100 .93 (.84–1.0)

erate effect size, ␩2⫽.06, and large Judge 1 ⫻ judge 3 98.2 .82 (.68–.97) ⬍.001
effect size, ␩2⫽.14).33,36 Significance Judge 2 ⫻ judge 3 98.2 .82 (.68–.97) ⬍.001
level was set at Pⱕ.05. The Statistical a
CI⫽confidence interval.
b
Package for the Social Sciences, ver- Ellipsis indicates statistical test was not conducted.
sion 19.0, (SPSS Inc, Chicago, Illinois)
was used for data analysis.
standing, ␣⫽.86). The results for per- term infants’ results were associated
Role of the Funding Source centile also confirmed the reliability significantly with the typical classifi-
Funding support for this study was for the 4 subscales (␣⫽.73). cation, whereas preterm infants’
provided by Conselho Nacional de results were associated significantly
Desenvolvimento Científico e Tec- Reliability with atypical or at-risk motor perfor-
nológico and Coordenação de Aper- The test-retest reliability results of mance (␹2⫽6.03, P⫽.04).
feiçoamento de Pessoal do Ensino the Spearman test showed a strong
Superior. positive and significant correlation Concurrent Validity
(rho⫽.98, P⬍.0001) for raw scores The results of the Spearman coeffi-
Results and percentile (rho⫽.85, P⬍.001). cient showed a moderate positive
Content and Face Validity The same trend was observed for and significant correlation (rho⫽.34,
The results for content validity are the 4 subscales (rho⫽.93–.98, P⫽.03) between the AIMS and CBDS
presented on Table 2. The concor- P⬍.0001). The ICCs for interrater classification. The McNemar-Bowker
dance results for language clarity of reliability (␣⫽.86 –.99) and the intra- test (7.95, P⫽.047) revealed the
AIMS items ranged from 66.7 to 92.8, rater reliability (␣⫽.91–.99) indi- specific differences between both
and for pertinence, the results were cated strong and congruent results instruments. The AIMS scores
all higher than 98. The CVI for clarity among the experts (Tab. 3). There showed a significantly higher num-
and pertinence among experts was were no significant differences ber of infants as “typical,” and the
strong. The kappa coefficient statis- (P⬎.05) among the evaluators for CBDS classified significantly more
tical test also showed significant raw scores and percentiles. cases as “atypical.”
results (P⬍.05). The apparent or
face validity was evaluated by health Discriminant Validity Predictive Validity
professionals. For this procedure, The statistical analyses showed a The results from the group of 20
the AIMS items received 100% con- significant difference between the infants assessed longitudinally
cordance among the experts. All pro- full-term and preterm groups of (group 1) showed positive, strong,
fessionals scored a high value on the infants for the raw score (t122⫽ and significant correlations for raw
Likert scale for every item. ⫺4.84, P⬍.001) and percentile scores (ICC⫽.97, range⫽.94 –.98,
(t122⫽⫺1.99, P⫽.05). The scores P⬍.0001), percentile (ICC⫽.86,
Internal Consistency among full-term infants were sig- range⫽.72–.93, P⬍.0001), and clas-
The Cronbach alpha coefficient anal- nificantly higher than those of the sification (ICC⫽.87, range ⫽.75–.94,
yses (n⫽766 infants) showed signif- preterm infants. Strong correlation P⬍.0001). Higher scores in the first
icant reliability in the 4 subscales was identified between the raw assessment correlated with higher
(␣⫽.90), as well as for each subscale scores and percentiles for full-term scores in the other assessment times.
independently (prone, ␣⫽.85; infants (r⫽.48, P⬍.0001) and pre- The comparative analysis of raw
supine, ␣⫽.92; sitting, ␣⫽.84; and term infants (r⫽.61, P⬍.0001). Full- scores using ANOVA for repeated

March 2012 Volume 92 Number 3 Physical Therapy f 443

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
Brazilian Validation of the Alberta Infant Motor Scale

Table 3.
Interrater and Intrarater Reliabilitya

Interrater
Interrater Reliability Reliability
Statistical
Position and Total Score Test AⴛB AⴛC BⴛC AⴛBⴛC AⴛA

Prone ICC (␣) .99 .99 .99 .99 .97

Wilcoxon (P) ⫺1.34 (.18) ⫺1.18 (.23) ⫺1.50 (.13) ...b ⫺0.35 (.72)

Friedman (P) ... ... ... 2.24 (.28) ...

Supine ICC (␣) .92 .86 .87 .91 .91

Wilcoxon (P) ⫺0.75 (.45) ⫺1 (.31) ⫺1.26 (.20) ... ⫺1.34 (.18)

Friedman (P) ... ... ... 5.25 (.07) ...

Sitting ICC (␣) .99 .93 .88 .95 .99

Wilcoxon (P) 0.0 (1) ⫺1.55 (.12) ⫺1.03 (.30) ... ⫺1.134 (.25)

Friedman (P) ... ... ... 2.33 (.31) ...

Standing ICC (␣) .99 .98 .97 .98 .98

Wilcoxon (P) ⫺0.46 (.64) ⫺0.36 (.97) ⫺0.49 (.62) ... ⫺0.32 (.74)

Friedman (P) ... ... ... 1.57 (.92) ...

Alberta Infant Motor Scale Score ICC (␣) .99 .96 .97 .98 .98

Wilcoxon (P) ⫺1.47 (.14) ⫺0.49 (.62) ⫺0.57 (.56) ... ⫺0.755 (.45)

Friedman (P) ... ... ... 0.70 (.70) ...

a
b
ICC⫽intraclass correlation coefficient.
Ellipsis indicates statistical test was not conducted.

measures revealed that the observed P⫽.11). The ANOVA for repeated that usually occurs when an instru-
scores increased significantly over measures revealed that raw scores ment is translated for another culture
time (F4,16⫽28.52, P⬍.0001, ␩2⫽88). increased significantly over time by only one translator.28,29
The effect size was large.33,36 (F1,31⫽175.13, P⬍.0001, ␩2⫽85).
Follow-up paired-samples t tests The effect size was large.33,36 The The panel of expert health-related
from time 1 to 2, 2 to 3, 3 to 4, and same tendency was observed for professionals was unanimous in con-
4 to 5 showed a significant increase percentile and classification; a sig- sidering the AIMS-BR content appro-
in scores (P⬍.001). However, the nificant difference was observed priate to evaluate motor develop-
Friedman test showed no significant from assessment 1 to assessment 2 ment of infants in prone, supine,
changes in percentile scores (P⫽.99) (P⫽.02). From the total sample, 69% sitting, and standing positions. The
in the classification of motor devel- of the infants showed higher scores CVI for clarity and pertinence
opment (P⫽.25) over time. Although in the second assessment and 31% showed a strong coherence among
raw scores changed over time, the showed lower scores in the second the experts, as confirmed by the
infants’ percentile and classification assessment. Raw scores, percentile, kappa concordance coefficient,
of motor development remained and classification changed over time which demonstrated similarity in the
similar. (Tab. 4). answers.24,28,30,31,34 These results
indicate that the AIMS showed high
The study also aimed to assess 32 Discussion content validity indexes, with clear
infants who received 2 examinations The purpose of this study was to and pertinent motor criteria.24,28,34
at an interval of 6 months (group 2). translate the AIMS to Portuguese and The results emphasized the proper
A positive and significant correlation validate its norms for Brazilian representation of the items related to
between assessments 1 and 2 were infants. The independent transla- scale concepts and the theoretical
observed for raw score (r⫽.71, tions resulted in a unified and final relevance of the items.
P⬍.0001) and percentile (r⫽.30, Portuguese version, the AIMS-BR.
P⫽.05). However, for the classifica- The double-back and reverse transla- The results highlighted the internal
tion criteria the positive correlation tions as well as the committee meet- consistency of the AIMS. The values
was small and not significant (r⫽.23, ing eliminated the biased translation obtained using the Cronbach alpha

444 f Physical Therapy Volume 92 Number 3 March 2012

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
 Brazilian Validation of the Alberta Infant Motor Scale

Table 4.
Reliability and Validity of Alberta Infant Motor Scale (AIMS) Median (25th–75th Percentile) Scores
AIMS Score
Statistical Procedure
Raw Percentile Classification
for Reliability and
Validity Median (25th, 75th) Median (25th, 75th) Median (25th, 75th)

Reliability (n⫽259)

Test 43 (19, 57) 43 (14, 71) 3 (2, 3)

Retest 53 (36, 58) 61 (23, 71) 3 (2, 3)

Discriminative (n⫽124)

Term-born 32 (22, 48) 52 (23, 71) 3 (2.7, 3)

Preterm-born 18 (11, 28) 36 (11, 63) 3 (2, 3)

Concurrent (n⫽80)

AIMS . . .a ... 2 (2, 3)

b
CBDS ... ... 2 (1, 2)

Predictive, group 1 (n⫽20)

Examination 1 18.50 (8.25, 25) 64 (44.25, 83) 3 (3, 3)

Examination 2 25 (11.50, 34.50) 64.50 (39.75, 87.75) 3 (3, 3)

Examination 3 32.50 (22.25, 45.75) 74 (45.75, 84) 3 (3, 3)

Examination 4 38 (30.50, 51.50) 65 (48.75, 87.25) 3 (3, 3)

Examination 5 45.50 (38.25, 54.25) 68 (58.50, 80.75) 3 (3, 3)

Predictive, group 2 (n⫽32)

Examination 1 23.50 (3.75, 11.50) 49.50 (25, 67) 3 (3, 3)

Examination 2 50.50 (39.75, 57) 67.50 (48, 71) 3 (3, 3)

a
Ellipsis indicates statistical test was not conducted.
b
Child Behavior Development Scale.

index (.73–.90) reflect a high homo- present study are considered appro- case, which indicates that the data
geneity profile among the investi- priate, even when a very conserva- were reliable.34 It is important to
gated variables. The Cronbach alpha tive approach is adopted (ICC⫽.86 – highlight that the percentile correla-
coefficient must be at a minimum of .99). The interrater and intrarater tion values (rho⫽.85) were ade-
.06.35,36 In large samples, concor- reliability values obtained in the Bra- quate. The test-retest reliability
dance among the findings higher zilian validation were similar to the showed good temporal stability for
than .06 is even more difficult to results reported by the AIMS’ author AIMS. It is important to emphasize
achieve.35,36 In the present study, with the Canadian infant sample.13 that one basic measure of reliability
even with the large sample (n⫽766), The author also suggested that for psychometric instruments is a
the results were higher than .07, pre- acceptable values should be higher reasonable level of temporal stability
suming that the items were homoge- than .80.26 that can be related to the defining
neous and encompassed a represen- measures of the constructs.33
tation of a same trace and, therefore, With respect to test-retest reliability,
measure the same construct. high classification and percentile in The results provided evidence of the
the test were associated significantly validity of the AIMS to discriminate
All reliability results were considered with the elevated values at the retest. between typical and nontypical
appropriate.29 Concordance among Comparison tests also showed non- developmental behaviors. The
experts using the same instrument significant differences between the results for percentile and raw score
must be high and positive to guaran- test and retest. These statistical pro- were significantly different in both
tee reliability and validity of the cedures emphasized the temporal of the investigated groups. High
results.36 Values higher than .80 are stability of the AIMS.28,33 The acqui- scores for motor performance were
considered to indicate sufficient con- escence phenomenon (positive or associated significantly with high
cordance.32,36 The results from the negative) did not occur in either percentile, and the same tendency

March 2012 Volume 92 Number 3 Physical Therapy f 445

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
Brazilian Validation of the Alberta Infant Motor Scale

was observed for average and low description of infant motor develop- should be made carefully at ages
scores. The positive correlation ment. However, it is important to when the precision of the scale is
between these 2 measures of the highlight that both scales (AIMS and low.18,21
same construct, also confirming con- CBDS) have advantages and disad-
vergent validity, provided very simi- vantages; professionals must deter- Several procedures adopted in the
lar information about the same con- mine the most appropriate scale to present study emphasize the effi-
struct.35,36 The discriminant validity use by considering the specific ciency of the AIMS to screen motor
was confirmed by the significant dif- needs of each assessment. development of Brazilian infants.
ference between the preterm and Appropriate results were found for
full-term groups. Results from previ- The results underscore the need for reliable, consistent, and discriminat-
ous studies also confirm the power caution in using the AIMS to predict ing power. Limited concurrent and
of the AIMS to screen motor con- future motor scores. The percentiles predictive validity was observed.
straints during the first years of and classifications were similar for The results of the present study may
life.15,16,18,22,37,38 the group of infants followed longi- positively affect the daily practice of
tudinally by monthly examination educators and therapists, as well as
Comparisons between the categori- from 3 to 8 months old. However, researchers, as they can rely on a
zation of AIMS and CBDS revealed a the group of infants followed longi- validated and reliable instrument to
positive moderate correlation (.34) tudinally using 2 examinations assess infants’ development and to
between both assessments. Although within an interval of 6 months (at 5 design an intervention for Brazilian
the present results were satisfactory, and 10 months old) showed differ- children.
concurrent validity should be stron- ences in percentile and classification
ger.35 Contradictory results have and the classification correlation Both authors provided concept/idea/re-
been reported in the literature con- result was not significant. The results search design, writing, data collection and
cerned with concurrent validity. Sim- for group 1 were similar to those analysis, project management, fund pro-
ilar unsatisfactory concurrent valid- obtained in the AIMS validation in curement, participants, facilities/equipment,
ity has been found between AIMS Canada26 and showed a satisfactory institutional liaisons, clerical support, and
consultation (including review of manuscript
and the Test of Infant Motor Perfor- ability of the scale to predict out- before submission).
mance (correlation results averaged comes. However, the results for
from .20 to .67),17 the Bayley Scales group 2 were more closely aligned This study was approved by the Federal Uni-
versity of Rio Grande do Sul Institutional
of Infant Development (correlation with those of a study of Taiwanese Review Board.
results at 5 months⫽.50 and at 10 infants that detected limited predic-
months⫽.21),6 and the Daily Activi- tive validity of the AIMS to evaluate Funding support for this study was provided
by Conselho Nacional de Desenvolvimento
ties of Infants Scale.38 In contrast, preterm infants.21 The satisfactory Científico e Tecnológico and Coordenação
several studies have shown satisfac- predictive results in group 1 (3–9 de Aperfeiçoamento de Pessoal do Ensino
tory results using the Bayley Scales of months old) could be related to Superior.
Infant Development,13,21 Kyoto Scale infant ages and the scale’s greater DOI: 10.2522/ptj.20110036
of Psychological Development,14 sensitivity to assess infants from 3 to
Infant Motor Profile,15 and Peabody 9 months compared with other
Developmental Gross Motor Scale.21 ages.40 The lack of predictability for References
The concurrent validity of the AIMS group 2 (5–12 months old) could be 1 Wijnhoven TMA, Onis M, Onyang AW,
et al. Assessment of gross motor develop-
with the Bayley Scales of Infant related to the limitation of the scale ment in the WHO multicentre growth ref-
Development also was observed to assess older infants. Although the erence study. Food Nutr Bull. 2004;25:
37– 45.
with Brazilian infants.39 The results scale is arranged by increasing order 2 Vaccarino FM, Ment LR. Injury and repair
of the present study may be related of difficulty, a ceiling effect was in the developing brain. Arch Dis Child.
to the fact that the CBDS has few observed, as also reported previ- 2004;89:190 –192.
motor behaviors to be assessed ously.18,40 The few items available 3 Mahoney G, Robinson C, Perales F. Early
motor intervention: the need for new
(n⫽15) compared with the large for discriminating among infants treatment paradigms. Inf Young Child.
number of motor behaviors to be after they pass the controlled lower- 2004;17:291–300.
assessed (n⫽58) with the AIMS. ing through standing position40 4 Hospers CHB, Algra MH. A systematic
review of the effects of early intervention
Even though the CBDS has similar cause a limitation in the interpreta- on motor development. Dev Med Child
purposes regarding motor develop- tion of the results for measures of Neurol. 2005;47:421– 432.
ment, this scale assesses the most infant development. Therefore, the 5 Santos RS, Araújo APQC, Porto MAS. Early
diagnosis of abnormal development of pre-
marked benchmarks, whereas the present results underscore the previ- term newborns: assessment instruments.
AIMS comprises a more detailed ous finding that clinical assumptions J Pediatr. 2008;84:289 –299.

446 f Physical Therapy Volume 92 Number 3 March 2012

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
 Brazilian Validation of the Alberta Infant Motor Scale

6 Campos D, Santos CCD, Gonçalves GMV, 17 Campbell SK, Kolobe HAT, Wright DB, 29 Vallerand RJ. Vers une méthodologie de
et al. Agreement between scales for Linacre MJ. Validity of the Test of Infant validation trans-culturelle de question-
screening and diagnosis of motor develop- Motor Performance for prediction of 6-, 9- naires psychologiques: implications pour
ment at 6 months. J Pediatr. and 12-month scores on the Alberta Infant la recherché en langue française. Can Psy-
2006;82:470 – 474. Motor Scale. Dev Med Child Neurol. chol. 1989;30:662– 680.
2002;44:263–272.
7 Spittle AJ, Doyle LW, Boyd RN. A system- 30 Neuendorf KA. The Content Analysis
atic review of the clinometric properties 18 Fleuren KMW, Smit LS, Stijnen T, Hartman Guidebook. London, United Kingdom:
of neuromotor assessments for preterm A. New reference values for the Alberta Sage Publications; 2002.
infants during the first year of life. Dev Infant Motor Scale need to be established. 31 Sim J, Wright C. Research in Health Care:
Med Child Neurol. 2008;50:254 –266. Acta Pediatrica. 2007;4:424 – 427. Concepts, Designs and Methods. Chelten-
8 Cameron EC, Maehle V, Reid J. The effects 19 Haritou S, Simitsopoulov A, Kontogianni ham, United Kingdom: Nelson Thornes
of an early physical therapy intervention R, Skordilid M. Keskoslasten motorinen Ltd; 2000.
for very preterm, very low birth weight kehitys Alberta Infant Motor Scale (AIMS): 32 Stephen JW. Quality of Life Outcomes in
infants: a randomized controlled clinical testistöllä arvioituna. Inq Sport Phys Ed. Clinical Trials and Health-Care Evalua-
trial. Pediatr Phys Ther. 2005;17:107–119. 2007;5:273–282. tion: A Practical Guide to Analysis and
9 Pin TW, Darrer T, Eldridge B, Galea M. 20 Bartlett DJ, Fanning JE. Use of the Alberta Interpretation. Cheltenham, United King-
Motor development from 4 to 8 months Infant Motor Scale to characterize the dom: Antony Rowe Ltd; 2009.
corrected age in infants born at or less motor development of infants born pre- 33 Cicchetti DV, Rourke BP. Methodological
than 29 weeks’ gestation. Dev Med Child term at eight months corrected age. Phys and Biostatistical Foundations of Clini-
Neurol. 2009;51:739 –745. Occup Ther Pediatr. 2003;23:31– 45. cal Neuropsychology and Medical and
10 Chagas PSC, Mancini MC, Fonseca ST, 21 Jeng SF, Yau Ky, Chen LC, Hsiao SF. Health Disciplines. London, United King-
et al. Neuromuscular mechanisms and Alberta Infant Motor Scale: reliability and dom: Taylor & Francis Group; 2004.
anthropometric modifications in the initial validity when used on preterm infant in 34 Waltz CF, Strickland OL, Lenz ER. Mea-
stages of independent gait. Gait Posture. Taiwan. Phys Ther. 2000;80:168 –178. surement in Nursing and Health
2006;24:375–381. 22 Darrah J, Piper MC, Watt MJ. Assessment Research. New York, NY: Springer Pub-
11 Formiga KMR, Linhares MB. Motor devel- of gross motor skills of at-risk infants: pre- lishing Co; 2010.
opment curve from 0 to 12 months in dictive validity of the Alberta Infant Motor 35 Cronbach LJ, Meehl PE. Construct Valid-
infants born preterm. Acta Paediatr. Scale. Dev Med Child Neurol. ity in Psychological Tests. Minneapolis,
2010;3:1– 6. 1998;40:485– 491. MN: University of Minnesota Press; 1976.
12 Santos DC, Gabbard C, Gonçalves VM. 23 Netelenbos JB. Teachers’ ratings of gross 36 Breakweell GM, Hammond S, Fife-Schaw
Motor development during the first year: a motor skills suffer from low concurrent C, Smith JA. Research Methods in Psychol-
comparative study. J Genet Psychol. validity. Hum Mov Sci. 2005;24:116 –137. ogy. London, United Kingdom: Sage Pub-
2001;162:143–153. 24 Yun J, Ulrich DA. Estimating measurement lications; 2006.
13 Piper MC, Darrah J. Motor Assessment of validity: a tutorial. Adapt Phys Activ Q. 37 Pin T, Eldridge B. Galea MP. Reliability of
the Developing Infant. Philadelphia, PA: 2002;19:32– 47. the Alberta Infant Motor Scale on infants
WB Saunders Co; 1994. 25 Wiart L, Darrah J. Review of four tests of born less than 30 weeks of gestation. Dev
14 Uesugui M, Tokuhisa K, Shimada T. The gross motor development. Dev Med Child Med Child Neurol. 2009;51:71–72.
reliability and validity of de Alberta Infant Neurol. 2001;43:279 –285. 38 Bartlett DJ, Fanning JK, Miller L. Develop-
Motor Scale in Japan. J Phys Ther Sci. 26 Piper MC, Pinnell LE, Darrah J, et al. Con- ment of the Daily Activities of Infant Scale:
2008;20:169 –175. struction and validation of the Alberta a measure supporting early motor devel-
15 Heineman KR, Boss AF, Hadders-Algra M. Infant Motor Scale (AIMS). Can J Public opment. Dev Med Child Neurol.
The Infant Motor Profile: a standardized Health. 1992;83:46 –50. 2008;50:613– 617.
and qualitative method to assess motor 27 Pinto B, Vilanova L, Vieira RM. The Child 39 Almeida KM, Dutra MV, Mello RR, et al.
behavior in infancy. Dev Med Child Neu- Behavior Development scale—CBDS. São Concurrent validity and reliability of the
rol. 2008;50:275–282. Paulo, Brazil: Fundação de Amparo a Pes- Alberta Infant Motor Scale in premature
16 Tse L, Mayson TS, Leo S, et al. Concurrent quisa do Estado de São Paulo; 1997. infants. J Pediatr. 2008;8:442– 448.
validity of the Harris Infant Neuromotor 28 Hernandez-Nieto R. Contributions to Sta- 40 Liao MP, Campbell KS. Examination of the
Teste and the Alberta Infant Motor Scale. tistical Analysis. Mérida, Venezuela: Los item structure of the Alberta Infant Motor
J Pediatr Nurs. 2008;23:28 –36. Andes University Press; 2002. Scale. Pediatr Phys Ther. 2004;16:31–38.

March 2012 Volume 92 Number 3 Physical Therapy f 447

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016
 Brazilian Validation of the Alberta Infant Motor Scale
Nadia Cristina Valentini and Raquel Saccani
PHYS THER. 2012; 92:440-447.
Originally published online December 1, 2011
doi: 10.2522/ptj.20110036

References This article cites 30 articles, 1 of which you can access

for free at:
http://ptjournal.apta.org/content/92/3/440#BIBL
Subscription http://ptjournal.apta.org/subscriptions/
Information
Permissions and Reprints http://ptjournal.apta.org/site/misc/terms.xhtml
Information for Authors http://ptjournal.apta.org/site/misc/ifora.xhtml

Downloaded from http://ptjournal.apta.org/ by guest on February 21, 2016