R E S E A R C H A R T I C L E

Reliability and Diagnostic Accuracy

of Clinical Tests of Vestibular
Function for Children
Jennifer B. Christy, PT, PhD; JoAnne Payne, MA, CCC-A; Andres Azuero, PhD, MBA; Craig Formby, PhD
Department of Physical Therapy, School of Health Professions (Dr Christy), and Department of Community Health,
Outcomes and Systems, School of Nursing (Dr Azuero), The University of Alabama at Birmingham, Birmingham,
Alabama; Department of Communicative Disorders (Ms Payne and Dr Formby), College of Arts and Sciences, The
University of Alabama, Tuscaloosa, Alabama.

Purpose: To determine reliability, diagnostic values, and minimal detectable change scores, 90% confidence
(MDC90 ) of pediatric clinical tests of vestibular function. Methods: Twenty children with severe to profound
bilateral sensorineural hearing loss and 23 children with typical development, aged 6 to 12 years, participated.
The Head Thrust Test, Emory Clinical Vestibular Chair Test, Bucket Test, Dynamic Visual Acuity, Modified Clinical
Test of Sensory Interaction on Balance, and Sensory Organization Test were completed twice for reliability.
Reference standard diagnostic tests were rotary chair and vestibular evoked myogenic potential. Reliability,
sensitivity, specificity, predictive values, likelihood ratios, and MDC90 scores were calculated. Results: Reliability
ranged from an intraclass correlation coefficient of 0.73 to 0.95. Sensitivity, specificity, and predictive values,
using cutoff scores for each test representing the largest area under the curve, ranged from 63% to 100%. The
MDC90 for Dynamic Visual Acuity and Modified Clinical Test of Sensory Interaction on Balance were 8 optotypes
and 16.75 seconds, respectively. Conclusions: Clinical tests can be used accurately to identify children with
vestibular hypofunction. (Pediatr Phys Ther 2014;26:180–190) Key words: child, diagnostic tests/standards,
postural balance, reproducibility of results, ROC curves, sensitivity and specificity, sensorineural hearing loss,
vestibular function tests

INTRODUCTION Although studies demonstrate the effectiveness of vestibu-

The reported incidence of vestibular-related deficits lar exercise in children with hypofunction,3,12,13 vestibular
in children has grown.1,2 Impairments identified in this function is rarely tested because simple inexpensive clini-
group are poor gaze stability with head movement, as cal tests of vestibular function have not been developed for
well as delayed development of postural control and motor children. Because of the known adverse effect of vestibu-
skills.3-6 These impairments persist and progress through- lar impairment on motor function, postural control, and
out childhood, adversely affecting high-level gross mo- stability of gaze, there is a pressing need for valid, reliable,
tor function, reading ability, and school performance.7-11 and simple tests to assess children’s peripheral vestibular
system integrity.
The development of valid and reliable testing tools
0898-5669/110/2602-0180
for pediatric assessment is warranted because of increas-
Pediatric Physical Therapy ing reports of vestibular dysfunction in children, including
Copyright C 2014 Wolters Kluwer Health | Lippincott Williams &
benign positional vertigo,14 vestibular neuronitis,15,16 oti-
Wilkins and Section on Pediatrics of the American Physical Therapy
Association
tis media,17 and drug-induced vestibulopathy.18 Variabil-
ity of reported incidence of vestibular hypofunction (VH)
Correspondence: Jennifer B. Christy, PT, PhD, Department of Physical (30%-100%) in children with sensorineural hearing loss
Therapy, The University of Alabama at Birmingham, 1720 2nd Ave S, (SNHL) exists because of differences in study populations
SHPB 360X, Birmingham, AL 35294 ([email protected]).
and vestibular testing.19-23
Grant Support: This study was funded by a research grant from the
Section on Pediatrics, awarded to Dr Christy.
Cushing et al20,21 measured vestibular function by
The authors declare no conflicts of interest.
using reference standard tests of caloric, rotary chair and
vestibular evoked myogenic potential (VEMP).24 Caloric
DOI: 10.1097/PEP.0000000000000039
testing detects unilateral vestibular hypofunction (UVH) of

180 Christy et al Pediatric Physical Therapy

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 the horizontal canal. Caloric testing is a potentially fright- been used in studies of postural control in adults40,41 and
ening experience for children since cold and warm water children.42,43 The original test had 6 conditions to mimic
must be placed in the ear, which elicits vertigo.25 The rotary SOT. A dome was used for sway-referenced visual condi-
chair test also evaluates horizontal canal function. It uses tions and thick foam for sway-referenced somatosensory
natural rotational stimulation to detect bilateral vestibu- conditions. The pediatric version (P-CTSIB) also included
lar hypofunction (BVH) and is tolerated by most children. tandem and single-legged stance conditions.42,43 The
Vestibular evoked myogenic potential tests the saccule, CTSIB was modified, dome conditions were removed, and
which contributes to the vestibulospinal system and pos- the test was validated for adults at risk for falls.41,44 Good
tural control.24 Pediatric normative data are available for test-retest and interrater reliability (r ≥ 0.75), and moder-
the rotary chair and VEMP tests.26–28 Cushing et al20 found ate correlations with SOT, were reported for adults.41,45,46
that 9 of 9 children with profound deafness following bac- Reliability and validity of the Modified Clinical Test of
terial meningitis had VH. Their balance scores, measured Sensory Interaction on Balance (MCTSIB), without extra
using the balance subscale of the Bruininks-Oseretsky Test conditions used in the P-CTSIB, have not been established
of Motor Proficiency, were less than that of peers who were for children.
not hearing impaired and children with SNHL but normal The purpose of this preliminary study was to deter-
vestibular function. In a separate study, Cushing et al21 mine reliability, sensitivity, specificity, predictive values,
tested 153 children with congenital and acquired SNHL of likelihood ratios, and cutoff scores for clinical tests of
varying etiologies and found that 50% had measurable VH; vestibular function. The clinical tests (ie, HTT, Bucket
37% of these children had severe VH or areflexia. Test, DVA, MCTSIB, and ECVCT) were compared with
Clinical tests of the vestibulo-ocular reflex or hori- reference standard tests (ie, VEMP and rotary chair) in a
zontal canal function include the Head Thrust Test (HTT; small cohort of children with severe to profound SNHL.
response to head turns right/left), the Dynamic Visual Acu-
ity test (DVA; acuity difference with head still and moving
at 2 Hz), and the Emory Clinical Vestibular Chair Test METHODS
(ECVCT; time of nystagmus following rotations in the A convenience sample was recruited from the Birm-
dark). A clinical test of utricular function is the Bucket ingham, Alabama, community. To be included, subjects
Test of subjective visual vertical (SVV): ability to perceive had to be between 6 and 12 years of age and diagnosed
a line as vertical in the absence of visual cues. Although with severe to profound SNHL as determined by audio-
reliability and validity of these tests have been reported for metric testing. Subjects were excluded if they showed ev-
adults,29-32 the only test with reported reliability, sensitiv- idence of neurological, central visual, or musculoskeletal
ity, and specificity for children is the DVA test.33,34 The abnormalities, fear of darkness, motion sensitivity, or a
DVA test had good inter- and intratester reliability (intr- history of neck trauma. Inclusion and exclusion criteria
aclass correlation coefficient [ICC] = 0.94 and 0.84, re- were determined via phone interview with a parent be-
spectively) and 100% sensitivity/specificity for identifying fore the first testing session and via neuromuscular screen.
children with confirmed BVH.33 The HTT had good sen- Twenty children with severe to profound SNHL (mean age
sitivity/specificity for adults with unilateral and bilateral = 8.9 ± 1.8 years) and 23 children with typical develop-
caloric weakness.29,30 The ECVCT had moderate to good ment (mean age = 9.5 ± 2.9 years) participated (Table 1).
test-retest, intrarater, and interrater reliability for adults The Institutional Review Boards at The University of Al-
(r > 0.70). Sensitivity was 78%, specificity was 92% for abama at Birmingham and The University of Alabama,
identifying 5 adult subjects with VH.35 The Bucket Test is Tuscaloosa, approved this study. Parents and children pro-
a low-technical version of the laser SVV that tests utricu- vided informed consent.
lar function. Zwergal et al31 reported that healthy adults
were able to perceive true vertical within 0.9 ± 0.7◦ . Sub-
jects with vestibular deficits perceived vertical within 8.9 Clinical Tests
± 5.4◦ . Inter- and intratester reliability of the Bucket Test After obtaining informed consent, subjects completed
were r = 0.90 and r = 0.92, respectively. 3 testing sessions. The first session began with a neuromus-
Postural control involves input from the vestibular, vi- cular screen composed of oculomotor tests (ie, smooth
sual, and somatosensory systems. Studies have shown that pursuit, saccades, vergence, ocular range of motion); cere-
children with VH also have poor postural control.3,22 A bellar tests (ie, rapid alternating movements, heel to shin);
clinical test of postural control is the Sensory Organization manual muscle testing and range of motion. Children with
Test (SOT). The SOT is not a direct measure of vestibu- abnormal results were excluded.
lospinal function but determines how vestibular informa- The same examiner, with 20 years of pediatric physi-
tion is used to control posture.36 The ICC demonstrating cal therapy experience and advanced training in vestibular
reliability of the SOT was greater than 0.75 for children rehabilitation, completed clinical tests (J.B.C.). Published
with and without VH.37,38 However, many clinics do not methods29 were used for HTT. Facing the subject at eye
have access to the equipment required to complete the level, the examiner flexed the neck 30◦ to maximally stimu-
SOT. The Clinical Test of Sensory Interaction on Balance late the horizontal canal. This was done using an imaginary
(CTSIB)39 was designed as a low-technical SOT and has line between the lateral canthus of the eye and the external

Pediatric Physical Therapy Clinical Tests of Vestibular Function for Children 181
Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 TABLE 1 on the subject’s face. The other examiner timed duration
Subject Characteristics of nystagmus. The examiners were trained by the primary
SNHL TD investigator (J.B.C.) and practiced before testing. To deter-
(n = 20) (n = 23) mine interrater reliability, nystagmus was videorecorded
so that raters could later watch and time the nystagmus.
Age in years, mean (SD) 8.9 (1.8) 9.5 (2.9)
To determine whether or not goggles were necessary, the
Gender, n
Male 14 10 test was repeated without goggles. After rotating with eyes
Female 6 13 closed, the subject looked at a white sheet and nystagmus
Ethnicity, n was timed. Nystagmus was again videotaped so that in-
White 16 22 terrater reliability could be determined later. Four trained
African American 1 1
examiners scored the videos, and scored the same videos
Hispanic 1 0
Other 2 0 1 week later. For intra- and interrater reliability of the m-
Hearing level, n ECVCT, fixation removed, 2 examiners scored the videos.
Bilateral profound 7 For the Bucket Test, a straight line was drawn into
Bilateral severe/profound 8 the bottom of an opaque bucket (23.5-cm diameter and
Bilateral moderate/severe 5
24-cm long). An angle finder was placed on the bottom of
Etiology of hearing loss, n
Conexin 26 4 the bucket in the same plane as the line.31 Each subject
Enlarged vestibular aqueduct 2 practiced with the bucket held away from the face (ie,
syndrome available visual cues) until the examiner was certain that
Pendred syndrome 1 the subject understood the task. Subjects also confirmed
Goldenhar syndrome 1
that the only object they could see when the bucket was
Heredity 5
Unknown 7 held over the face was the line. After training, the subject
Cochlear implants, n was seated with eyes closed. The mouth of the bucket was
Bilateral 10 placed around the subject’s face and the examiner turned
Unilateral right 3 the bucket to set the line off vertical. The subject then
None 7
opened the eyes. As the examiner slowly turned the bucket,
Abbreviations: SNHL, sensorineural hearing loss; TD, typical develop- the subject said “now” when the line reached vertical. The
ment. angle of degrees and direction off 0◦ was recorded for 10
trials of clockwise and counterclockwise rotations. The
acoustic meatus. The subject attempted to keep the eyes on mean degrees off 0◦ and mode of direction tilted (ie, left,
the examiner’s nose, which was decorated with a sticker. right, or straight) were calculated.
The subject’s head was unpredictably and quickly turned For the DVA test, the subject sat 10 away from the
to the right and left from center at an amplitude of 5◦ to 10◦ . Lea Symbols (ie, house, circle, heart, and square) chart. The
The examiner watched for a corrective saccade following chart had a total of 15 lines of 5 optotypes, ranging from
each HTT, and then the head was returned to center. This Snellen acuity levels of 20/200 to 20/8. The subject began
maneuver was randomly repeated 3 times in each direc- at an acuity level where all symbols on a line could be cor-
tion. The HTT was positive if at least 2 corrective saccades rectly identified, and continued to identify progressively
were observed to the right and/or the left. smaller symbols until no symbol could be identified.33 The
Modified methods were used for the modified ECVCT number of optotypes unable to be identified was static vi-
(m-ECVCT).35 The subject sat in a rotating office chair sual acuity. For the DVA, the neck was flexed 30◦ by using
with the head centered and slightly flexed and eyes closed. the same anatomical landmarks as for the HTT, the head
The chair was rotated right for 30 seconds at 0.5 Hz, us- was moved at 2 Hz (120◦ per second) to a metronome in
ing a metronome. This timing differed from the original the yaw plane, and the number of unidentified optotypes
test,35 which rotated adult subjects for 60 seconds. During was recorded. The DVA was completed twice, averaged,
development of the protocol before data collection began, and scored as the difference in optotypes missed between
it was determined that younger children became restless the DVA and static visual acuity tests.
and tried to open their eyes after 30 seconds. Therefore, For the SOT, the subject wore a safety harness and
after 30 seconds of rotation, the chair was stopped and the stood on the SMART EquiTest platform (NeuroCom, a
timer started. The subject did not open the eyes until in- division of Natus, Clackamas, Oregon). The subject stood
frared camera goggles were placed over the eyes, blocking still during 6 conditions: (1) stable platform, (2) stable
fixation. The goggles were not placed on the eyes during platform eyes closed, (3) sway-referenced visual surround,
rotation to avoid damage to the goggle cables. Nystagmus (4) sway-referenced platform, (5) sway-referenced plat-
was observed on the monitor and timed until it subsided. form eyes closed, and (6) sway-referenced visual surround
The subject rested for 2 minutes or double the duration and platform. A blindfold was used for eyes closed condi-
of nystagmus to dampen the effect of the first rotation.35 tions. Each condition lasted 20 seconds, and 3 trials were
The test was then repeated to the left. All subjects received completed. On the basis of reports that children with
rightward and then leftward rotations. We used 2 examin- hypofunction have low visual, vestibular, and somatosen-
ers. One examiner rotated the chair and held the goggles sory effectiveness ratios,3,47 these ratios were calculated as

182 Christy et al Pediatric Physical Therapy

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 follows: somatosensory = 3/1; visual = 4/1; and vestibular cameras on I-Portal 100-Hz binocular video oculography
= 5/1.48 goggles. Subjects were seated in a chair with a computer-
For the MCTSIB, the subject stood barefoot with arms controlled motor within a closed booth in complete dark-
across the chest and feet together for 30 seconds during 4 ness. If necessary, small children sat in the lap of an adult.
conditions: (1) floor eyes opened; (2) floor eyes closed; (3) Calibration of eye movement was performed before test-
Neurocom foam eyes opened; and (4) Neurocom foam eyes ing. An adult investigator sat in the darkened room and
closed. A blindfold was used for eyes closed conditions. On used a radio for communication with the audiologist, who
eyes opened conditions, subjects faced a white sheet. Three visually monitored the test via infrared cameras. Alerting
trials of each condition were completed only if the subject tasks were used.12,27,28
was unable to complete the entire 30 seconds on the first or Oculomotor tests (ie, smooth pursuit, saccades, and
second trial. If the child completed 30 seconds on the first optokinetic nystagmus) were completed to rule out central
trial, a score of 30 seconds was given for that condition. visual problems. The gain and accuracy of eye movements
The mean of the 3 trials was calculated for each condition were compared with pediatric normative data (Neuro Ki-
and then added for a total score (maximum = 120 seconds, netics, Inc). The SVV paradigm was based on previously
30 seconds for each of the 4 conditions). reported methods.53 The subject practiced the technique
The HTT, Bucket Test, DVA, MCTSIB, and m-ECVCT in room light until he or she understood. In complete
were completed a second time, 4 hours to 7 days later, darkness, the subject set the off-vertical laser line to per-
for test-retest reliability. Some children were unable to ceived vertical, using buttons that rotated the line clock-
be tested on 2 separate days, whereas others preferred to wise or counterclockwise in 0.1◦ increments. The subject
return on a different day. Given the nature of the tests, said “now” when he or she perceived that the line was
we felt that 4 hours was sufficient time to negate effects vertical. The test was repeated 5 times to each side and
of the first testing session, and 7 days was short enough responses were averaged.
to negate effects of maturation. Reference standard tests Rotary chair tests included the SHA and step rota-
were completed within 1 month after the clinical tests tions to measure horizontal canal function. The subject’s
by an audiologist with 24 years of experience, blinded to head was pitched forward 30◦ and gently immobilized.
clinical test results. Reference standard test results were For the SHA test, 5 frequencies were examined (ie, 0.01,
not expected to change in 1 month. 0.04, 0.08, 0.16, and 0.64 Hz) at 60◦ per second over 3
to 9 cycles. Data included phase, gain, and asymmetry of
the vestibulo-ocular reflex response. Step rotation testing
Reference Standard Tests included a series of complete chair rotations to both the
The cervical VEMP (cVEMP) assessed the function of right and left (ie, 100◦ per second for 60 seconds). Pre- and
the saccule and inferior vestibular nerve following proto- postrotary time constants (ie, the amount of time required
cols previously reported.49,50 Before testing, subjects were for nystagmus to decrease to 37% of its original strength)
examined with otoscopy and tympanometry to ensure the were measured. Results were compared with normative
ears were free of obstruction. Subjects sat in a semireclined data and scored as follows: (1) negative, within 1.5 SDs of
position during the test. Following electrode placement, the comparative sample mean; (2) bilateral loss (positive),
subjects turned the head to activate the sternocleidomas- reduced gain values outside the comparative range; and (3)
toid. Electromyography with video feedback assured that unilateral loss (also positive), abnormal phase and asym-
baseline sternocleidomastoid muscle activity was between metry values in the presence of normal gain values.12,27,28
35 and 50 μV. Tone burst stimuli (ie, 500 Hz with 2-0-2 If either the rotary chair (ie, SHA, step rotation) or cVEMP
rise-plateau-fall) were presented at 97 dB normal hearing tests were positive, the vestibular result was classified as a
level (nHL) or 105 dB nHL at a rate of 5 per second with rar- “positive” diagnosis. Subjective visual vertical was not used
efaction polarity. One subject had complete atresia of both for classification due to lack of pediatric normative data.
ear canals, requiring protocol modification. For this sub-
ject, stimuli were presented via B71-10 bone conduction
vibrator placed on the mastoid bone.51 Recordings were Data Analysis
amplified with a gain of 5000×, for a bandpass setting Descriptive statistics were computed for participant
of 2 to 250 Hz, and time window of 30 millisecond. The characteristics and clinical tests scores. For each clinical
initial positive and negative peaks were marked, and the test, a receiver operating characteristic curve determined
P1-N1 amplitude asymmetry ratio was calculated. Subject optimal cutoff scores for classification of VH as diagnosed
responses were compared with normative data.27,52 Re- by reference standard tests. The area under the ROC
sponses greater than 2 SDs from the normative mean were curve (AUC) was tabulated for comparison of accuracy
considered positive. among clinical tests. The closer an AUC to the maximum
Oculomotor tests, laser SVV, and sinusoidal harmonic value of 1, the more accurate the test. After cutoff scores
acceleration (SHA) were completed as was step rotation were determined, cross-tabulations between hypofunction
testing using a Neuro Kinetics, Inc (Pittsburgh, Pennsyl- status and clinical test results were conducted to estimate
vania) Rotary Chair with VEST 6.10 software. Videonys- the frequencies of true positive (TP), false positive (FP),
tagmography measured eye movements through infrared true negative (TN), and false negative (FN) classifications.

Pediatric Physical Therapy Clinical Tests of Vestibular Function for Children 183
Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 Measures of sensitivity (TP/(TP+FN)), specificity RESULTS
(TN/(TN+FP)), positive predictive value (TP/(TP+FP)), Nineteen of 20 subjects with SNHL completed ref-
negative predictive value (TN/(TN+FN)), positive like- erence standard tests. All subjects had normal results on
lihood ratio (sensitivity/(1 − specificity)) and negative the neuromuscular screen and oculomotor tests. Of this
likelihood ratio ((1 − sensitivity)/specificity), and 95% group, 3 subjects had BVH, 5 had UVH, and 11 had nor-
confidence intervals (CIs) were computed. The Delta logit mal vestibular function (NVF). One child, aged 6 years,
method54 and the method by Simel et al55 were used to refused reference standard testing but completed clinical
estimate proportion and likelihood ratio CIs, respectively. tests. Two subjects with typical development (TD) com-
Repeated assessments estimated test-retest reliability and pleted reference standard and clinical tests. Twenty-three
interrater reliability with the ICC;56 95% CIs for these subjects with TD completed some or all of the clinical tests
measures were computed with the Fisher z transforma- (Figure 1).
tion. Pearson correlations were used to examine linear Test-retest reliability was good (ICC ≥ 0.73) for
associations. To determine the minimal detectable change all clinical tests except for condition 4 of the MCTSIB
(90% confidence) (MDC90 ), the following √ formula was (Table 2). Two tests that can be used to detect change
applied:57 MDC90 = 1.65 × pooled SD × (2[1 − ICC]). due to intervention are DVA and MCTSIB. The MDC90 for
Analyses were conducted using SAS version 9.3 (SAS these outcomes were 8 optotypes for DVA and 16.75 sec-
Institute, Cary, North Carolina). onds for the MCTSIB total score. Intrarater reliability for

Eligible
n=42
Excluded (n=0)

Index Tests:
HTT (n=42); MCTSIB (n=42), DVA (n=42),
m-ECVCT (n=24); SOT-VR (n=41)

Abnormal Result: Normal Result: Inconclusive

HTT (n=8) HTT (n=34) SOT-VR (n=1)
MCTSIB (n=9) MCTSIB (n=33) (refused)
DVA (n=9) DVA (n=33)
m-ECVCT (n=6) m-ECVCT (n=18)
SOT-VR (n=7) SOT-VR(n =33)

No VFT No VFT (n=21);

n=1 Reason: time to travel for
(refused) testing (all TD)

VFT VFT
n=8 n=13

Inconclusive Inconclusive
n=0 n=0

VH + VH - VH + VH -
HTT (n=6) HTT (n=2) HTT (n=3) HTT (n=10)
MCTSIB (n=7) MCTSIB (n=1) MCTSIB (n=2) MCTSIB (n=11)
DVA (n=7) DVA (n=1) DVA (n=4) DVA (n=9)
m-ECVCT (n=5) m-ECVCT (n=3) m-ECVCT (n=0) m-ECVCT (n=13)
SOT-VR (n=6) SOT-VR (n=2) SOT-VR (n=1) SOT-VR (n=12)

Important Note: See Erratum to this Figure pasted at end of this PDF.
Fig. 1. Flow diagram of subjects completing testing. DVA indicates Dynamic Visual Acuity; HTT, Head Thrust Test; MCTSIB, Modified
Clinical Test of Sensory Interaction on Balance; m-ECVCT, Modified Emory Clinical Vestibular Chair Test; SOT-VR, Sensory Organization
Test, Vestibular Ratio; VFT, Vestibular Function Tests; TD, typical development; VH, vestibular hypofunction.

184 Christy et al Pediatric Physical Therapy

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 TABLE 2
Test-Retest Reliability

Name of Test n (TD) n (SNHL) ICC (95% CI)

HTT, +/− 17 20 0.73 (0.53-0.85)

DVA score, # optotypes 15 20 0.81 (0.66-0.90)
m-ECVCT room light, total seconds 4 18 0.88 (0.75-0.95)
left + right
m-ECVCT fixation removed, total 5 15 0.95 (0.88-0.98)
seconds left + right
MCTSIB total score, s 18 20 0.74 (0.55-0.86)
MCTSIB condition 4, s 18 20 0.56 (0.15-0.77)
Bucket Test, mean degrees off 15 20 0.74 (0.49-0.87)
center
Bucket Test Direction, most often 15 20 κ: 0.24 (0.00-0.49)
tilted: right, left, straight

Abbreviations: CI, confidence interval; DVA, Dynamic Visual Acuity; HTT, Head Thrust Test; ICC, intraclass correlation coefficient; κ, kappa; MCTSIB,
modified clinical test of sensory interaction for balance; m-ECVCT, modified Emory Clinical Vestibular Chair Test; SNHL, sensorineural hearing loss;
TD, typical development.

the m-ECVCT in room light was good for the 4 raters. The 69%). Likelihood ratio CIs were wide because of the low
ICC ranged from 0.76 (95% CI, 0.51-0.89) to 0.97 (95% prevalence of hypofunction in the sample.
CI, 0.92-0.99). However, interrater reliability was poor for All 5 subjects with UVH had a laser SVV result greater
the 4 raters on the m-ECVCT room light. The ICC ranged than 2◦ , tilted to the lesioned side in all but 1 subject. One
from 0.37 (95% CI, 0.10-0.62) for left rotations to 0.40 subject with BVH had an abnormal laser SVV. Forty-five
(95% CI, 0.13-0.65) for right rotations. Intra- and inter- percent of subjects with normal rotary chair and cVEMP
rater reliability was high for 2 raters on m-ECVCT with results had abnormal laser SVV results. Both subjects with
fixation removed; ICC = 0.86 (95% CI, 0.67-0.93) for left TD who completed testing had laser SVV results less than
+ right rotations. 2◦ . Laser SVV results agreed with the Bucket Test results
Moderate to good correlations were found between in only 38% of cases.
the m-ECVCT fixation removed, the m-ECVCT room light,
and the rotary chair time constant. The m-ECVCT room
light total score correlated with the m-ECVCT fixation re- DISCUSSION
moved total score (r = 0.59 [95% CI, 0.22-0.80]). The This is the first study to determine reliability and di-
m-ECVCT fixation removed after spinning right/left cor- agnostic accuracy for pediatric clinical tests of vestibular
related moderately with the rotary chair time constant after function. The HTT had good test-retest reliability and cor-
spinning right/left—right: r = 0.66 (95% CI, 0.31-0.84); rectly predicted vestibular function scores with 88% agree-
left: r = 0.81 (95% CI, 0.55-0.92). A fair correlation was ment. Schubert et al29 tested subjects with UVH/BVH and
found between SOT composite scores and the MCTSIB to- compared the HTT with caloric tests. Sensitivities for sub-
tal scores (r = 0.37; P = .02). However, a moderate to good jects with UVH and BVH were 71% and 84%, respectively.
correlation was found between the MCTSIB total scores Specificity was 82%. In the current study, we combined
and SOT vestibular ratios (r = 0.58; P < .001). The cor- subjects with UVH (n = 5) and BVH (n = 3) because of
relation between the Bucket Test mean degrees off center low prevalence. Even so, the HTT was reliable (ICC =
and laser SVV mean scores was not statistically significant 0.73), sensitive (75%), and specific (91%). The HTT was
(r = −0.34; P = .13). simple and required no special equipment. However, clin-
Means and SDs for all clinical tests are provided in icians should practice correct technique.
Table 3. The optimal cutoff scores to predict hypofunction Similar to adult performance,41 the MCTSIB total
for clinical tests, on the basis of the AUC, are summarized score was reliable. The sensitivity, specificity, and pre-
in Table 4. Positive and negative likelihood ratios with CIs dictive values were 78% or greater for a cutoff score of
are provided in Table 5. Except for the Bucket Test, all 110 total seconds. However, if only doing condition 4
clinical tests had an AUC ranging from 0.64 to 0.89. The of the MCTSIB, test-retest reliability decreased (ICC =
Bucket Test had an AUC of 0.55, indicating slightly better 0.56). Therefore, we recommend that all conditions of the
than chance prediction of hypofunction. The highest MCTSIB be completed. If using the MCTSIB as an outcome
overall values were obtained with the HTT (sensitivity tool to measure improvement, then the MDC90 score of
= 75%; specificity = 91%), the MCTSIB total score 16.76 seconds should be used. The moderate to good cor-
(sensitivity = 88%; specificity = 85%), the m-ECVCT relation between the SOT vestibular ratio and the MCTSIB
fixation removed (sensitivity = 63%; specificity = 100%), score indicated that the latter provided information about
the SOT vestibular ratio (sensitivity = 75%; specificity vestibular input to postural control. This differed from the
= 92%), and the DVA (sensitivity = 88%; specificity = findings of Gagnon et al,58 who reported that the Pediatric

Pediatric Physical Therapy Clinical Tests of Vestibular Function for Children 185
Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 TABLE 3
Means and SDs

Clinical Test TD SNHL NVF SNHL BVH SNHL UVH

HTT R n = 23 n = 11 n=3 n=5

# positive results 0 2 1 4
HTT L n = 23 n = 11 n=3 n=5
# positive results 0 3 1 2
DVA score n = 23 n = 11 n=3 n=5
Mean optotypes (SD) 11.6 (8.7) 8.3 (3.4) 19.3 (13.5) 15 (5.6)
m-ECVCT room light n=6 n = 10 n=3 n=5
Mean total seconds (SD) 19.3 (7.6) 17.9 (9.3) 2.7 (2.5) 15.3 (10.8)
m-ECVCT fixation removed n=5 n = 11 n=3 n=5
Mean total seconds (SD) 52.1 (7.5) 48.3 (9.5) 9.7 (16.9) 41.9 (17.7)
Bucket Test n = 21 n = 11 n=3 n=5
Mean degrees off center (SD) 1.8 (0.44) 1.7 (0.47) 1.7 (0.58) 1.6 (0.54)
Laser SVV score n=2 n = 11 n=3 n=5
Mean degrees off center (SD) 0.49 (0.62) 2.15 (1.5) 2.7 (2.2) 5.0 (2.7)
MCTSIB total score n = 23 n = 11 n=3 n=5
Mean seconds (SD) 118.6 (4.3) 116.0 (6.3) 84.4 (23.3) 98.4 (20.4)
SOT visual ratio n = 22 n = 11 n=3 n=5
Mean ratio (SD) 0.76 (0.18) 0.59 (0.18) 0.64 (0.29) 0.69 (0.15)
SOT vestibular ratio n = 22 n = 11 n=3 n=5
Mean ratio (SD) 0.57 (0.17) 0.43 (0.17) 0.00 0.22 (0.16)
SOT somatosensory ratio n = 22 n = 11 n=3 n=5
Mean ratio (SD) 0.96 (0.05) 0.95 (0.67) 0.95 (0.05) 0.99 (0.05)

Abbreviations: BVH, bilateral vestibular hypofunction; DVA, Dynamic Visual Acuity; HTT, Head Thrust Test; L, left; m-ECVCT, modified Emory Clinical
Vestibular Chair Test; MCTSIB, modified clinical test of sensory interaction for balance; NVF, normal vestibular function; R, right; SNHL, sensorineural
hearing loss; SOT, sensory organization test; SVV, subjective visual vertical; TD, typical development; UVH, unilateral vestibular hypofunction.

CTSIB and SOT did not correlate and concluded that the used to determine whether or not gaze stability exercises
2 tests measured sensory organization differently. In the are working.59 A change in the DVA score of greater than 8
current study, we did not do tandem or single-legged optotypes (or approximately 1.6 lines) can be considered
stance conditions. We also used the vestibular ratio rather a change that is greater than error.
than the SOT stability scores. The clinical DVA test was The m-ECVCT with fixation removed was reliable
reliable and predicted vestibular function test results with and predicted vestibular function with 86% accuracy, us-
a 76% success rate, using a cutoff score of 10 optotypes ing a cutoff of 29.2 seconds (ie, following 30-second left
(ie, 2 lines). This differed from a previous study that re- + right rotations). It correlated moderately with the ro-
ported a 100% success rate for predicting hypofunction.33 tary chair time constant. The m-ECVCT in room light had
Unlike original study methods, we continued testing until good test-retest and poor interrater reliability. The cutoff
the subject missed all optotypes on a line and then counted score of 15.3 seconds yielded only 65% correct prediction.
the total number of missed optotypes, pushing subjects to The 4 raters who participated in the interrater reliability
their limit of the DVA. The original test33 ended when sub- study commented that it was difficult to determine when
jects missed 3 optotypes on a line. The DVA test can be nystagmus stopped in room light. Some children fixated
TABLE 4
Diagnostic Values and Cutoff Scoresa

Optimal Sensitivity Specificity PPV NPV

Clinical Test AUC Cutoff Score % Agree (95% CI) (95% CI) (95% CI) (95% CI)

HTT (pos or neg) NA NA 88 0.75 (0.25-0.89) 0.77 (0.32-0.86) 0.67 (0.25-0.84) 0.83 (0.31-0.90)
MCTSIB condition 4 only 0.89 20 s 90 0.88 (0.22-0.95) 0.92 (0.28-0.96) 0.88 (0.22-0.95) 0.92 (0.28-0.96)
SOT vestibular ratio 0.88 0.20 (ratio) 86 0.75 (0.25-0.88) 0.92 (0.28-0.96) 0.86 (0.21-0.95) 0.86 (0.33-0.91)
MCTSIB total 0.88 110 s 86 0.88 (0.22-0.95) 0.85 (0.32-0.90) 0.78 (0.26-0.89) 0.92 (0.27-0.96)
DVA score 0.85 10 optotypes 76 0.88 (0.22-0.95) 0.69 (0.30-0.82) 0.64 (0.27-0.81) 0.90 (0.25-0.95)
m-ECVCT, room light 0.74 15.3 s (total) 65 0.75 (0.25-0.89) 0.58 (0.27-0.78) 0.55 (0.25-0.78) 0.78 (0.26-0.89)
m-ECVCT, fixation removed 0.74 29.2 s (total) 86 0.63 (0.23-0.84) 1.00 (NA) 1.00 (NA) 0.81 (0.35-0.87)
SOT visual ratio 0.67 0.76 (ratio) 76 0.63 (0.23-0.84) 0.85 (0.32-0.90) 0.71 (0.22-0.88) 0.79 (0.33-0.86)
SOT somatosensory ratio 0.64 0.98 (ratio) 67 0.50 (0.20-0.80) 0.77 (0.32-0.86) 0.57 (0.20-0.84) 0.71 (0.32-0.81)

Abbreviations: AUC, area under the curve; CI, confidence interval; cVEMP, cervical vestibular evoked myogenic potential; DVA, Dynamic Visual Acuity;
HTT, Head Thrust Test; MCTSIB, Modified Clinical Test of Sensory Interaction on Balance; m-ECVCT, modified Emory Clinical Vestibular Chair Test;
NA, not applicable; neg, negative; NPV, negative predictive value; pos, positive; PPV, positive predictive value; SOT, Sensory Organization Test.
a Positive cases tested positive on the cVEMP or rotary chair test or both.

186 Christy et al Pediatric Physical Therapy

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 TABLE 5
Likelihood Ratios

Clinical Test LR+ (95% CI) LR− (95% CI)

HTT (pos or neg) 3.25 (1.11-9.48) 0.33 (0.09-1.12)

MCTSIB condition 4 only 11.38 (1.70-76.14) 0.13 (0.02-0.85)
SOT vestibular ratio 9.75 (1.42-66.85) 0.27 (0.08-0.91)
MCTSIB total 5.87 (1.54-20.90) 0.14 (0.02-0.94)
DVA score 2.84 (1.20-6.70) 0.18 (0.03-1.17)
m-ECVCT, room light 1.80 (0.83-3.93) 0.43 (0.12-1.56)
m-ECVCT, fixation removed NA 0.38 (0.15-0.92)
SOT visual ratio 4.06 (1.02-16.20) 0.44 (0.18-1.12)
SOT somatosensory ratio 2.17 (0.65-7.27) 0.65 (0.31-1.38)

Abbreviations: CI, confidence interval; DVA, Dynamic Visual Acuity; HTT, Head Thrust Test; LR+, positive likelihood ratio; LR−, negative likelihood
ratio; MCTSIB, Modified Clinical Test of Sensory Interaction on Balance; m-ECVCT, modified Emory Clinical Vestibular Chair Test; NA, not applicable;
SOT, Sensory Organization Test.

and stopped nystagmus immediately. Therefore, this test with a functioning vestibular system. The current study
should be done with fixation removed. included subjects with chronic VH. It is possible that test
The Bucket Test had good test-retest reliability of results might differ in subjects with acute lesions.
mean degrees off vertical, poor reliability for direction,
and did not correlate with the laser SVV results. A cut- Clinical Implications
off score could not be determined because the AUC was
only slightly better than chance. These results differed from To apply the results of this study in a clinical
those obtained by Zwergal et al,31 who found good reliabil- context, we will consider the subject who refused ref-
ity, sensitivity, and specificity of the Bucket Test. In adults, erence standard tests. This was a 6-year-old girl with
SVV in the acute stage following UVH tilts toward the side SNHL, of unknown origin. She had bilateral cochlear
of the lesion. Studies differ as to when or if SVV ever fully implant surgery at the age of 4 years. She walked in-
compensates.60 We do not know how SVV compensates dependently at 15 months of age and had a normal
in children with VH because adult paradigms of utricular neurological and oculomotor screen. She presented with
testing have not been tested in children.61,62 To our knowl- positive HTT bilaterally and a static visual acuity and
edge, this is the first study to report SVV results in children DVA score difference of 34 (>10 optotypes cutoff). Her
with VH. Given the results of the laser SVV in this study, MCTSIB total score was 95.57 seconds (<110 cutoff). All
it is tempting to hypothesize that the children with UVH conditions on the MCTSIB were normal except for the
had static uncompensated deficits, whereas the 5 subjects foam eyes closed condition (5.57 seconds). She refused the
with normal vestibular function tests but abnormal perfor- SOT. The m-ECVCT score was 23.3 seconds (<cutoff of
mance for laser SVV had a utricular deficit. Except for 1, 29.2 seconds). According to clinical tests, this child likely
subjects with BVH did not have asymmetric SVV, which has BVH affecting gaze stability and balance. This child
was expected. Utricular function should be tested in larger could potentially benefit from vestibular exercises.3,59
numbers of children with and without VH using the ocular
VEMP, a laboratory test of utricular function.63 LIMITATIONS
Cochlear implantation is increasingly being offered First, only 8 of 19 children with SNHL had confirmed
for individuals with severe to profound SNHL.64 The VH. This contributed to the wide 95% CIs for diagnostic
surgery involves an array of electrodes inserted into the values and likelihood ratios. The low number of subjects
cochlea to send electrical signals to the auditory nerve.65 may have also contributed to the low diagnostic capacities
Histological studies revealed saccular damage in some chil- of the Bucket Test and condition 4 of the MCTSIB. More
dren who received cochlear implantation, ostensibly due subjects with confirmed pathology are needed to add to
to trauma during electrode insertion into the cochlea, this preliminary data. Second, the reliability of laser SVV
which lies in close proximity to the saccule.66,67 In these in children has not been established. We do not know
cases, persisting vertigo occurred when the device was how development or the presence of a vestibular deficit
activated.66 In contrast, some subjects demonstrated bet- affects SVV in children. Third, this study can only be
ter balance with the cochlear implant turned on than generalized to children aged 6 to 12 years with severe to
off.19,68,69 Importantly, children undergoing cochlear im- profound SNHL and without other neurologic problems.
plant surgery should be screened for vestibular deficits. This battery of tests should be completed on other groups
The timing of vestibular injury is important. The of children with VH.
vestibulo-ocular reflex develops rapidly during the first
2 years of life.70 The use of vestibular input for postural
control does not become adult-like until after the age of 15 CONCLUSION
years.71 Therefore, an adult who developed typically and The best tests to determine whether a child with an
acquired VH will differ from a child who did not develop otherwise normal neurological system has VH include (1)

Pediatric Physical Therapy Clinical Tests of Vestibular Function for Children 187
Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 the HTT, (2) the m-ECVCT fixation removed, (3) the DVA, cochlear implantation in children correlate poorly with functional
(4) the MCTSIB, and (5) SOT vestibular ratio. The MDC90 outcome. Otol Neurotol. 2009;30(4):488-495.
21. Cushing SL, Gordon KA, Rutka JA, James AL, Papsin BC. Vestibular
scores should be considered if using the DVA and MCTSIB end-organ dysfunction in children with sensorineural hearing loss
as outcome tools to detect improvement due to interven- and cochlear implants: an expanded cohort and etiologic assessment.
tion. Tests of utricular function require further study in Otol Neurotol. 2013;34(3):422-428.
pediatric populations. 22. Suarez H, Angeli S, Suarez A, Rosales B, Carrera X, Alonso R. Balance
sensory organization in children with profound hearing loss and
cochlear implants. Int J Pediatr Otorhinolaryngol. 2007;71(4):629-
REFERENCES 637.
23. Tribukait A, Brantberg K, Bergenius J. Function of semicircular
1. Rine RM. Growing evidence for balance and vestibular problems in canals, utricles and saccules in deaf children. Acta Otolaryngol.
children. Audiol Med. 2009;99999:1. 2004;124:41-48.
2. O’Reilly RC, Morlet T, Nicholas BD, et al. Prevalence of vestibular and 24. Sheykholesami K, Kaga K, Megerian CA, Arnold JE. Vestibular-
balance disorders in children. Otol Neurotol. 2010;31(9):1441-1444. evoked myogenic potentials in infancy and early childhood. Laryn-
3. Rine RM, Braswell J, Fisher D, Joyce K, Kalar K, Shaffer M. Improve- goscope. 2005;115:1440-1444.
ment of motor development and postural control following inter- 25. Wuyts FL, Furman J, Vanspauwen R, Van de HP. Vestibular function
vention in children with sensorineural hearing loss and vestibular testing. Curr Opin Neurol. 2007;20(1):19-24.
impairment. Int J Pediatr Otorhinolaryngol. 2004;68(9):1141-1148. 26. Eviatar L, Eviatar A. The normal nystagmic response of infants to
4. Rine RM, Spielholz NI, Buchman C. Postural control in children caloric and perrotatory stimulation. Laryngoscope. 1979;89:1036-
with sensorineural hearing loss and vestibular hypofunction: deficits 1045.
in sensory system effectiveness and vestibulospinal function. In: Duy- 27. Valente M. Maturational effects of the vestibular system: a study
sens j, Smits-Engelsman BCM, Kingma H, eds. Control of Posture and of rotary chair, computerized dynamic posturography, and vestibu-
Gait. Amsterdam, the Netherlands: Springer-Verlag; 2001:40-45. lar evoked myogenic potentials with children. J Am Acad Audiol.
5. Inoue A, Iwasaki S, Ushio M, et al. Effect of vestibular dysfunction on 2007;18(6):461-481.
the development of gross motor function in children with profound 28. Staller SJ, Goin DW, Hildebrandt M. Pediatric vestibular evalu-
hearing loss. Audiol Neurootol. 2013;18(3):143-151. ation with harmonic acceleration. Otolaryngol Head Neck Surg.
6. Shall MS. The importance of saccular function to motor development 1986;95(4):471-476.
in children with hearing impairments. Int J Otolaryngol. 2009;2009:1- 29. Schubert MC, Tusa RJ, Grine LE, Herdman SJ. Optimizing the sensi-
5. tivity of the Head Thrust Test for identifying vestibular hypofunction.
7. Braswell J, Rine RM. Evidence that vestibular hypofunction af- Phys Ther. 2004;84(2):151-158.
fects reading acuity in children. Int J Pediatr Otorhinolaryngol. 30. Halmagyi GM, Curthoys LS, Cremer PD, et al. The human horizon-
2006;70:1957-1965. tal vestibulo-ocular reflex in response to high acceleration stimula-
8. Kaga K. Vestibular compensation in infants and children with con- tion before and after unilateral vestibular neurectomy. Exp Brain Res.
genital and acquired vestibular loss in both ears. Otorhinolaryngology. 1990;81:479-490.
1999;49:215-224. 31. Zwergal A, Rettinger N, Frenzel C, Dieterich M, Brandt T, Strupp M.
9. Rine RM, Cornwall G, Gan K, et al. Evidence of progressive delay of A bucket of static vestibular function. Neurology. 2009;72(19):1689-
motor development in children with sensorineural hearing loss and 1692.
concurrent vestibular dysfunction. Percept Mot Skills. 2000;90:1101- 32. Hall CD, Hoover J, Jacobs M, et al. A new measurement tool for
1112. vestibular hypofunction: validity of the Emory Clinical Vestibular
10. Hlavacka F, Mergner T, Krizkova M. Control of the body ver- Chair Test. J Neurol Phys Ther. 2009;33(4):232.
tical by vestibular and proprioceptive inputs. Brain Res Bull. 33. Rine RM, Braswell J. A clinical test of Dynamic Visual Acuity for
1999;40(5/6):431-435. children. Int J Pediatr Otorhinolaryngol. 2003;69(11):1195-1201.
11. Franco ES, Panboca I. Vestibular function in children underperform- 34. Rine RM, Roberts D, Corbin BA, et al. A new portable tool to screen
ing in school. Braz J Otorhinolaryngol. 2008;74(6):815-825. vestibular and visual function: National Institutes of Health Toolbox
12. Casselbrant ML, Mandel EM, Sparto PJ, et al. Longitudinal postur- Initiative. J Rehabil Res Dev. 2012;49(2):209-220.
ography and rotational testing in children three to nine years of age: 35. Hall CD, Abbott KN, Lane EC, et al. Measurement of vestibular
normative data. Otolaryngol Head Neck Surg. 2010;142(5):708-714. hypofunction: validity of the Emory Clinical Vestibular Chair Test
13. Rine RM, Wiener-Vacher S. Evaluation and treatment of vestibular (ECVCT). J Neurol Phys Ther. 2007;31(4):210-211.
dysfunction in children. NeuroRehabilitation. 2013;32(3):507-518. 36. Evans MK, Krebs DE. Posturography does not test vestibulospinal
14. Saka N, Imai T, Seo T, et al. Analysis of benign paroxysmal function. Otolaryngol Head Neck Surg. 1999;120(2):164-173.
positional nystagmus in children. Int J Pediatr Otorhinolaryngol. 37. Gabriel LS, Mu K. Computerized platform posturography for chil-
2013;77(2):233-236. dren: test-retest reliability of the sensory test of the VSR System. Phys
15. Monobe H, Murofushi T. Vestibular neuritis in a child with otitis me- Occup Ther Pediatr. 2002;22(3/4):101-117.
dia with effusion; clinical application of vestibular evoked myogenic 38. Rine RM, Rubish K, Feeney C. Measurement of sensory system ef-
potential by bond-conducted sound. Int J Pediatr Otorhinolaryngol. fectiveness and maturational changes in postural control in young
2004;68:1455-1458. children. Pediatr Phys Ther. 1998;10:16-22.
16. Zannolli R, Zazzi M, Muraca MC, Macucci F, Buoni S, Nuti D. A 39. Shumway-Cook A, Horak FB. Assessing the influence of sensory in-
child with vestibular neuritis. Is adenovirus implicated? Brain Dev. teraction on balance. Phys Ther. 1986;66:1548-1550.
2006;28(6):410-412. 40. Horak FB. Clinical measurement of postural control in adults. Phys
17. Casselbrant ML, Villardo RJ, Mandel EM. Balance and otitis media Ther. 1987;67(12):1881-1885.
with effusion. Int J Audiol. 2008;47(9):584-589. 41. Cohen H, Blatchly CA, Gombash LL. A study of the clinical test of
18. Contopoulos-Ioannidis DG, Giotis ND, Baliatsa DV, Ioannidis JP. sensory interaction and balance. Phys Ther. 1993;73(6):346-351.
Extended-interval aminoglycoside administration for children: a 42. Crowe TK, Deitz JC, Richardson PK, Atwater SW. Interrater reliability
meta-analysis. Pediatrics. 2004;114(1):e111-e118. of the pediatric clinical test of sensory interaction for balance. Phys
19. Buchman CA, Joy J, Hodges A, Telischi FF, Balkany TJ. Vestibular Occup Ther Pediatr. 1990;10(4):1-27.
effects of cochlear implantation. Laryngoscope. 2004;114:1-22. 43. Westcott SL, Lowes L, Richardson PK. Evaluation of postural sta-
20. Cushing SL, Papsin BC, Rutka JA, James AL, Blaser SL, Gordon bility in children: current theories and assessment tools. Phys Ther.
KA. Vestibular end-organ and balance deficits after meningitis and 1997;77(6):629-645.

188 Christy et al Pediatric Physical Therapy

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 44. Wrisley DM, Whitney SL. The effect of foot position on the modified 59. Braswell J, Rine RM. Preliminary evidence of improved gaze stability
clinical test of sensory interaction and balance. Arch Phys Med Rehabil. following exercise in two children with vestibular hypofunction. Int
2004;85(2):335-338. J Pediatr Otorhinolaryngol. 2006;70:1967-1973.
45. Weber PC, Cass SP. Clinical assessment of postural stability. Am J 60. Hafstrom A, Fransson PA, Karlberg M, Magnusson M. Subjective
Otol. 1993;14(6):566-569. visual tilt and lateral instability after vestibular deafferentation. Acta
46. El-Kashlan HK, Shepard NT, Asher AM, Smith-Wheelock M, Telian Otolaryngol. 2006;126(11):1176-1181.
S. Evaluation of clinical measures of equilibrium. Laryngoscope. 61. Clarke AH, Schonfeld U, Helling K. Unilateral examination
1998;108(March):311-319. of utricle and saccule function. J Vestib Res. 2003;13(4-6):
47. Enbom H, Magnusson M, Pyyko I. Postural compensation in children 215-225.
with congenital or early acquired bilateral vestibular loss. Ann Otol 62. Bohmer A. The subjective visual vertical as a clinical parameter for
Rhinol Laryngol. 1991;100:472-478. acute and chronic vestibular (otolith) disorders. Acta Otolaryngol.
48. Charpiot A, Tringali S, Iionescu E, Vital-Durand F, Ferber-Viart C. 1999;119:126-127.
Vestibulo-ocular reflex and balance maturation in healthy children 63. Iwasaki S, Smulders YE, Burgess AM, et al. Ocular vestibular evoked
aged from six to twelve years. Audiol Neurootol. 2010;15(4):203-210. myogenic potentials in response to bone-conducted vibration of the
49. Akin FW, Murnane OD, Panus PC, Caruthers SK, Wilkinson midline forehead at Fz. A new indicator of unilateral otolithic loss.
AE, Proffitt TM. The influence of voluntary tonic EMG level Audiol Neurootol. 2008;13(6):396-404.
on the vestibular-evoked myogenic potential. J Rehabil Res Dev. 64. Rubin LG, Papsin B. Cochlear implants in children: surgical site
2004;41(3B):473-480. infections and prevention and treatment of acute otitis media and
50. Halmagyi GM, Curthoys IS. Clinical testing of otolith function. Ann meningitis. Pediatrics. 2010;126(2):381-391.
N Y Acad Sci. 1999;871:195-204. 65. James AL, Papsin BC. Cochlear implant surgery at 12
51. McNerney KM, Burkard RF. The vestibular evoked myogenic po- months of age or younger. Laryngoscope. 2004;114(12):
tential (VEMP): air- versus bone-conducted stimuli. Ear Hear. 2191-2195.
2011;32(6):e6-e15. 66. Basta D, Todt I, Goepel F, Ernst A. Loss of saccular function after
52. Zhou G, Kenna MA, Stevens K, Licamelli G. Assessment of saccular cochlear implantation: the diagnostic impact of intracochlear electri-
function in children with sensorineural hearing loss. Arch Otolaryngol cally elicited vestibular evoked myogenic potentials. Audiol Neurootol.
Head Neck Surg. 2009;135(1):40-44. 2008;13(3):187-192.
53. Bohmer A, Mast F. Assessing otolith function by the subjective visual 67. Tien HC, Linthicum FH Jr. Histopathologic changes in the
vertical. Ann N Y Acad Sci. 1999;871:221-231. vestibule after cochlear implantation. Otolaryngol Head Neck Surg.
54. Agresti A. Categorical Data Analysis. 2nd ed. Hoboken, NJ: Wiley; 2002;127(4):260-264.
2002. 68. Ribari O, Kustel M, Szirmai A, Repassy G. Cochlear implantation
55. Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence: influences contralateral hearing and vestibular responsiveness. Acta
sample size estimation for diagnostic test studies. J Clin Epidemiol. Otolaryngol. 1999;119(2):225-228.
1991;44(8):763-770. 69. Bonucci AS, Costa Filho OA, Mariotto LD, Amantini RC, Alvarenga
56. Rousson V, Gasser T, Seifert B. Assessing intrarater, interrater KF. Vestibular function in cochlear implant users. Braz J Otorhino-
and test-retest reliability of continuous measurements. Stat Med. laryngol. 2008;74(2):273-278.
2002;21(22):3431-3446. 70. Fife TD, Tusa RJ, Furman JM, et al. Assessment: vestibular testing
57. Haley SM, Fragala-Pinkham MA. Interpreting change scores of tests techniques in adults and children: report of the Therapeutics and
and measures used in physical therapy. Phys Ther. 2006;86(5):735- Technology Assessment Subcommittee of the American Academy of
743. Neurology. Neurology. 2000;55:1431-1441.
58. Gagnon I, Swaine B, Forget R. Exploring the comparability of the 71. Ferber-Viart C, Onescu E, Orlet T, Roehlich P, Ubreuil C. Balance in
Sensory Organization Test and the Pediatric Clinical Test of Sen- healthy individuals assessed with Equitest: maturation and normative
sory Interaction for Balance in children. Phys Occup Ther Pediatr. data for children and young adults. Int J Pediatr Otorhinolaryngol.
2006;26(1/2):23-41. 2007;71(7):1041-1046.

CLINICAL BOTTOM LINE

Commentary on “Reliability and Diagnostic Accuracy of Clinical Tests of Vestibular Function for Children”

“How could I apply this information?”

Clinicians can use the positive and negative likelihood ratios (LR+, LR−) provided in Table 5 to determine
the likelihood that a specific child diagnosed with severe to profound sensorineural hearing loss (SNHL) has
vestibular hypofunction (VH). We use the example of the 6-year-old child with SNHL in this study who did not
undergo reference standard Vestibular Function Testing (VFT) yet had positive results on 4 clinical tests: (1) Head
Thrust Test, (2) Dynamic Visual Acuity Test, (3) Modified Clinical Test of Sensory Interaction on Balance total
score, and (4) Modified Emory Clinical Vestibular Chair Test with fixation removed. Her pretest probability for
VH was 42%, on the basis of the fact that 8 of 19 children with SNHL in this study had VH as diagnosed by VFT.
Using a nomogram1 and the LR+ from each clinical test (range, 2.84-5.87), her posttest probability of having VH
as diagnosed by VFT would be approximately 65% to 82% (Figure 1). If each clinical test had been negative (LR−
range, 0.14-0.38), her posttest probability of having VH as diagnosed by VFT would be approximately 8% to 20%
(Figure 2). In this case, the use of a nomogram and the LR of each clinical test provided critical quantitative
information that increased certainty that the child had VH to make appropriate referrals for reference standard
VFT and determine the most appropriate intervention plan.

Pediatric Physical Therapy Clinical Tests of Vestibular Function for Children 189
Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
 “What should I be mindful about when applying this information?”
Clinicians must be mindful that only 8 participants had VH as diagnosed by reference standard VFT and
that the psychometric properties of the 5 clinical tests of vestibular function can only be generalized to 6- to
12-year-old children with severe to profound SNHL from chronic lesions. Furthermore, 13 of 20 participants with
SNHL used cochlear implants, yet no mention was made as to whether these children had a higher incidence of
VH or whether their implants were on or off during the clinical tests of vestibular function.

Fig. 2. Using a nomogram to find the posttest probabil-

ity based on negative test results. Pretest probability =
Fig. 1. Using a nomogram to find the posttest probabil- 42%. m-ECVCT-fixation removed − LR = 0.38, posttest
ity of VH based on positive test results. Pretest probabil- probability = 20%. HTT − LR = 0.33, posttest proba-
ity = 42%. MCTSIB total +LR = 5.87, posttest probabil- bility = 18%. DVA − LR = 0.18, posttest probability =
ity = 82%. HTT +LR = 3.25, posttest probability = 70%. 10%. MCTSIB total − LR = 0.14, posttest probability =
DVA +LR = 2.84, posttest probability = 65%. A positive 8%. A negative result on each of these 4 clinical tests
result on each of these 3 clinical tests changes the prob- changes the probability of the child having VH from
ability of the child having VH from 42% (pretest) to 65% 42% (pretest) to 8% to 20% (posttest), depending on
to 82% (posttest), depending on the test. DVA indicates the test. DVA indicates Dynamic Visual Acuity; HTT, Head
Dynamic Visual Acuity; HTT, Head Thrust Test; LR, like- Thrust Test; LR, likelihood ratio; MCTSIB, Modified Clin-
lihood ratio; MCTSIB, Modified Clinical Test of Sensory ical Test of Sensory Interaction on Balance; m-ECVCT,
Interaction on Balance. Modified Emory Clinical Vestibular Chair Test.

REFERENCE
1. Fagan T. Nomogram for Bayes’s theorem. N Engl J Med. 1975;293:257.

Barabara Sargent, PT, PhD, PCS

University of Southern California
Los Angeles, California
Jennifer Pate, PT, DPT
Los Angeles Children’s Hospital
Los Angeles, California
The authors declare no conflicts of interest.
DOI: 10.1097/PEP.0000000000000040

190 Christy et al Pediatric Physical Therapy

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.
The following correspondence resulted in a revised Flow Diagram (Fig. 1) that the primary author
submitted to the journal as an erratum. The revised figure is included below.

From: Steve Allison [email protected] 07 October 2014

To: [email protected]

Hello Dr. Christy:

Thank you for your recent publication: Christy JB, Payne J, Azuero A, Formby C. Reliability and diagnostic accuracy
of clinical tests of vestibular function for children. Pediatr Phys Ther. 2014 Summer;26(2):180-9.
I have selected this article as an assignment for critical appraisal in an EBP course I teach at RMUoHP, and am in
the process of preparing some supporting materials to help students through the process. As I attempted to replicate
results for the diagnostic accuracy statistics in Tables 4 & 5, and using the frequency counts from Figure 1, I
encountered difficulties. I hope you can clarify for me.

I'm using the counts for true positives, false positives, false negatives, and true negatives from the bottom 4 boxes in
Figure 1. Taking the first reported result (HTT) as an example, I see there were:

6 true positives (abnormal result on HTT and with VH);

2 false positives (abnormal result on HTT but without VH);
3 false negatives (normal result on HTT but with VH);
10 true negatives (normal result on HTT and without VH).

Entering these values into a common calculator (I used the PEDro Confidence Interval Calculator spreadsheet), I get
the following results:

Sensitivity = .67 (disagrees with reported Sensitivity = .75 in Table 4)

Specificity = .83 (disagrees with reported Specificity = .77 in Table 4)
PLR = 4.00 (disagrees with reported PLR = 3.25 in Table 5)
NLR = 0.40 (disagrees with reported NLR = 0.33 in Table 5)

I've pasted results from the PEDro spreadsheet into the attached MS Word file. Similar discrepancies were found for
the other index tests in Tables 4 & 5 as I worked through computations for a couple of them. I suspect that I must be
misinterpreting the counts reported in the Figure.

Thanks in advance for any help resolving the discrepancies.

All the best,

SA

From: Jennifer Braswell Christy [email protected] 10 October 2014

Dear Steve,

First of all, I am SO GLAD that you put together this assignment for your students. You found a MAJOR ERROR in
the bottom part of my figure that wasn’t previously noticed. Somehow, the numbers in the middle 2 boxes at the
bottom were flipped so that the false positives/ false negatives were in the wrong box. I immediately revised the
figure and sent it to Ann Van Sant so that an erratum could be published. This is my first erratum (and hopefully
my last). I attached the new (and correct) figure here, for you to use with your students. I am grateful to you for
pointing it out to me!!! I am hoping that Ann will publish the erratum soon so that others don’t have the same
confusion. THANKS!!!!

Sincerely,
Jennifer
 ERRATUM
Reliability and Diagnostic Accuracy of Clinical Tests of Vestibular Function for Children: Erratum

In the article cited above, Figure 1 on page 184 of the Summer 2014 issue of Pediatric Physical Therapy included
errors. The corrected figure appears below, and the error has been noted in the online version of the article, which
is available at www.pedpt.com.

Eligible
n=42
Excluded (n=0)

Index Tests:
HTT (n=42); MCTSIB (n=42), DVA (n=42),
m-ECVCT (n=24); SOT-VR (n=41)

Abnormal Result: Normal Result: Inconclusive

HTT (n=8) HTT (n=34) SOT-VR (n=1)
MCTSIB (n=9) MCTSIB (n=33) (refused)
DVA (n=9) DVA (n=33)
m-ECVCT (n=6) m-ECVCT (n=18)
SOT-VR (n=7) SOT-VR (n=33)

No VFT No VFT (n=21);

n=1 Reason: time to travel for
(refused) testing (all TD)

VFT (VH +) VFT (VH -)

n=8 n=13

Inconclusive Inconclusive
n=0 n=0

True + False - False + True -

HTT (n=6) HTT (n=2) HTT (n=3) HTT (n=10)
MCTSIB (n=7) MCTSIB (n=1) MCTSIB (n=2) MCTSIB (n=11)
DVA (n=7) DVA (n=1) DVA (n=4) DVA (n=9)
m-ECVCT (n=5) m-ECVCT (n=3) m-ECVCT (n=0) m-ECVCT (n=13)
SOT-VR (n=6) SOT-VR (n=2) SOT-VR (n=1) SOT-VR (n=12)

Fig. 1. Flow diagram of subjects completing testing. DVA indicates Dynamic Visual Acuity; HTT, Head Thrust Test; MCTSIB, Mod-
ified Clinical Test of Sensory Interaction on Balance; m-ECVCT, Modified Emory Clinical Vestibular Chair Test; SOT-VR, Sensory
Organization Test, Vestibular Ratio; VFT, Vestibular Function Tests; TD, typical development; VH, vestibular hypofunction.

Reference
Christy JB, Payne J, Azuero A, Formby C. Reliability and diagnostic accuracy of clinical tests of vestibular function for children. Pediatr Phys
Ther. 2014;26(2):180–190.

Pediatric Physical Therapy Erratum 1

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins and the Section on Pediatrics of the American Physical Therapy
Association. Unauthorized reproduction of this article is prohibited.