Musculoskeletal Science and Practice 66 (2023) 102809

Contents lists available at ScienceDirect

Musculoskeletal Science and Practice

journal homepage: www.elsevier.com/locate/msksp

Original article

Test-retest reliability of ankle range of motion, proprioception, and balance

for symptom and gender effects in individuals with chronic ankle instability
Xiaojian Shi a, c, *, Charlotte Ganderton a, b, d, Oren Tirosh a, Roger Adams e, Doa EI-Ansary d,
Jia Han b, c, e, **
a
School of Health Science, Swinburne University of Technology, Hawthorn, 3122, Victoria, Australia
b
College of Rehabilitation Sciences, Shanghai University of Medicine and Health Sciences, Shanghai, 201318, China
c
School of Exercise and Health, Shanghai University of Sport, Shanghai, 200090, China
d
School of Biomedical Science and Health, Royal Melbourne Institute of Technology University, Bundoora, 3083, Victoria, Australia
e
Research Institute for Sport and Exercise, University of Canberra, ACT2600, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: Objectives: To determine the reliability of the commonly used musculoskeletal assessments in individuals with
Repeated measurements chronic ankle instability (CAI).
Mobility Design: Within and between-days test-retest reliability.
Ankle inversion discrimination
Setting: University laboratory.
Postural control
Participants: Twenty-four individuals with unilateral CAI.
Method: For both sides, ankle dorsiflexion range of motion (DFROM) was assessed by a goniometer and weight-
bearing lunge test (WBLT), proprioception by the active movement extent discrimination apparatus (AMEDA),
and balance by the Star Excursion Balance Test with anterior (SEBTA), posteromedial (SEBTPM) and postero­
lateral (SEBTPL) components. All measures were taken at enrollment, after 30 min and one week later.
Results: For the asymptomatic side, all assessments demonstrated good to excellent reliability, with ICCs (3,1)
between 0.8 and 0.96. On the symptomatic side, WBLT, SEBTA and SEBTPM showed excellent reliability, with
ICCs (3,1) above 0.90, while SEBTPL, goniometer and AMEDA showed moderate reliability, with the 95% CI of the
ICCs (3,1) crossing 0.5. Three-way repeated measures ANOVA showed a side main effect, with asymptomatic
worse, for WBLT (F = 16.9, p < 0.001) and SEBTA (F = 5.4, p = 0.03); an overall improving time main effect for
SEBTPL (F = 6.9, p = 0.02). Neither a gender main effect nor any interaction effect was found.
Conclusions: WBLT, SEBTA and SEBTPM can be strongly recommended for measuring ankle dorsiflexion mobility
and dynamic balance for both sides of individuals with unilateral CAI, while only WBLT can be used for side-to-
side comparison. The application of a goniometer to measure DFROM, SEBTPL or AMEDA should be done
cautiously for this specific cohort, considering their poor to good reliability for the symptomatic side.

1. Introduction et al., 2019).

When assessing DFROM, both goniometer measurement and the
Chronic ankle instability (CAI) is one the most common lower ex­ weight-bearing lunge test (WBLT) are widely used (Morales et al., 2017).
tremity impairments that usually occurs secondary to poor management Although Powden et al. (2015) concluded that the WBLT had good
of acute ankle sprains (Anandacoomarasamy and Barnsley, 2005; Hertel reliability in assessing DFROM for healthy people, children with cere­
and Corbett, 2019). Individuals with CAI have shown multiple physical bral palsy (Kim et al., 2018) and unspecified ankle injuries (Jones et al.,
and neuromuscular impairments (Thompson et al., 2018), such as 2005), the reliability and validity in assessing DFROM impairment in
decreased ankle dorsiflexion range of motion (DFROM), diminished CAI has not yet been reported. In addition, an ankle proprioception
ankle proprioception, and reduced balance ability (Han et al., 2022, Shi deficit can be quantified by different psychophysiological methods,

* Corresponding author. School of Health Science, Swinburne University of Technology, Hawthorn, 3122, Victoria, Australia.
** Corresponding author. College of Rehabilitation Sciences, Shanghai University of Medicine and Health Sciences, Shanghai, 201318, China.
E-mail addresses: [email protected] (X. Shi), [email protected] (J. Han).

https://doi.org/10.1016/j.msksp.2023.102809
Received 7 January 2023; Received in revised form 15 May 2023; Accepted 15 June 2023
Available online 16 June 2023
2468-7812/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
X. Shi et al. Musculoskeletal Science and Practice 66 (2023) 102809

based on joint position reproduction, threshold to detection of passive volunteered to participate in this study (Table 1). The inclusion criteria
motion and active movement extent discrimination apparatus (AMEDA) were: adults aged over the age of 18 years old; participants had to have
(Han et al., 2016). Among these, the AMEDA has been argued to have at least one significant ankle sprain that occurred in the past year
superior ecological validity to the former two testing methods (Han (Gribble et al., 2014); participants had to score less than 21 on the
et al., 2016, Horváth et al., 2022). Although the AMEDA has been re­ Taiwan-Chinese version of the Cumberland Ankle Instability Tool
ported to have good test-retest reliability for individuals with ankle in­ (CAIT) on the symptomatic side, given that this is an instrument which
juries (Witchalls et al., 2014), it was unclear what the status of the can provide 87% sensitivity and 85% specificity (Lin et al., 2022a). In­
participants was with respect to CAI, as per the inclusion criteria stated dividuals were excluded if they were undergoing treatment, had any
by the International Ankle Consortium (Gribble et al., 2014). Therefore, residual inflammatory symptoms such as pain or swelling, any surgical
the reliability of the AMEDA has not been clearly established for history that affected lower extremity functioning, or reported significant
assessing ankle proprioception in a clearly defined CAI population. medical episodes including cancer, abnormal sensation, or paralysis in
In terms of balance, the Star Excursion Balance Test (SEBT) has the lower extremity.
demonstrated good discriminative capacity between CAI and non-CAI
participants and has become a common tool for assessing dynamic 2.2. Procedure
balance performance. Recently, Picot et al. (2021) systematically
reviewed the application of SEBT and found that the evidence to support This was a prospective, observational study with repeated mea­
reliability mostly came from studies involving healthy populations. surements. Demographic information, and CAIT scores for asymptom­
Despite clinical popularity, the relevant reports regarding the reliability atic and symptomatic ankles were collected (detailed in Table 1).
of tests employed in assessing CAI were scarce. Physical assessment of both sides was undertaken at enrollment and all
To date, performance on ankle DFROM, dynamic balance and pro­ participants were re-assessed after 30-min and again one week later. The
prioceptive acuity has been assessed and compared between individuals 30-min initial follow-up was a pragmatic decision chosen to reflect a
with and without CAI, but their differences between the symptomatic typical physiotherapy appointment length. Physical assessments were
and asymptomatic side in individuals with unilateral CAI was not fully chosen to capture ankle mobility, ankle proprioception and dynamic
determined. In addition, there has been an increasing number of studies balance performance. One assessor with 9 years of physiotherapy
suggesting the existence of gender differences in functional performance practice experience and expertise in musculoskeletal and orthopedic
in CAI individuals (Lin et al., 2021, Lin et al., 2022b). However, it is rehabilitation was allocated to perform or supervise all assessments. To
unknown whether such disparity would also be seen between male and avoid subjective bias, the assessor did not get access to the previous
female participants with CAI. results, while reassessing the participants in the following sessions.
Previous work has suggested that individuals with CAI demonstrate Details regarding assessment are in the following section.
greater variation in outcome measures during multiple proprioceptive
tasks, a feature which has been considered to be related to altered 2.3. Ankle dorsiflexion range of motion
learning efficacy due to ankle pathologies (Witchalls et al., 2014). We
assumed that it would be likely that such dysfunction would affect the Ankle dorsiflexion range of motion (DFROM) was measured by
test-retest reliability of different commonly used assessment tools in goniometer and WBLT successively. To measure DFROM using a goni­
individuals with CAI. Accordingly, this study was to primarily investi­ ometer (Fig. 1a), participants were instructed to lay supine and with the
gate the reliability of ankle dorsiflexion mobility measured by a goni­ knee straight, and actively dorsiflex their ankles to end of range by
ometer and WBLT, ankle proprioception based on AMEDA, and balance bringing their toes up towards their head. For angle measurement, the
performance based on three components of SEBT, in individuals with
unilateral CAI, and then establish the minimal important differences for
this specific cohort. Secondly, this study was conducted to perform
exploratory analysis to compare difference in the measures between the
asymptomatic and symptomatic sides, and between genders. We hy­
pothesized that the reliability of DFROM measured by a goniometer and
WBLT, balance performance based on SEBT with three components, and
ankle proprioceptive performance based on AMEDA would vary from
the previous reports regarding nonspecific lower limb pathologies.

2. Material and method

2.1. Participants

A convenience sample of twenty-four individuals with unilateral CAI

Table 1
Demographic information of participants (Mean ± SD).
Participants 24(12 males, 12 females)

Involved side 10 Left, 14 Right

Leg dominance 24 Right
Age (yrs.) 26.0 ± 5.0
Height (cm) 168.8 ± 10.4
Weight (kg) 68.0 ± 14.8
CAIT_AS 27.1 ± 4.2
CAIT_SS 15.3 ± 4.2*

* Indicative of statistical significance.

CAIT: Cumberland ankle instability tool, AS-asymptomatic side, SS- Fig. 1. Ankle dorsiflexion range of motion Assessment
symptomatic side. (a. goniometer; b. Weight-bearing lunge test).

2
X. Shi et al. Musculoskeletal Science and Practice 66 (2023) 102809

axis center of the goniometer was placed on the most prominent point of proprioceptive discrimination sensitivity score. AUC values range from
the lateral malleolus, with the fixed arm of the goniometer along with 0.5, equivalent to chance responding, up to 1.0, representing perfect
the ipsilateral shank, and the movable arm parallel with the sole of the discrimination.
foot. Following this procedure, WBLT (knee-to-wall test) was assessed
(Wyndow et al., 2018). To measure DFROM in the WBLT, participants 2.5. Dynamic balance
were instructed to lunge forward to touch the wall using the knee, with
both heels in neutral, in line with the big toe, and consistently planted on SEBT was used to assess dynamic balance, using the anterior
the ground (Fig. 1b). In this position, participants attempted to shuffle (SEBTA), posteromedial (SEBTPM) and posterolateral (SEBTPL) testing
the foot away from the wall as far as possible, while keeping the knee on directions in a random sequence (Fig. 3). Prior to conduct of the SEBT,
the wall, until the final maximum distance was reached and measured as participants placed the hallux at the crossroad of the lines along the
that between the hallux and the wall. For both methods, three trials were three directions, with hands akimbo to control trunk displacement.
performed, and the values were averaged. Thereafter, they were instructed to keep unipedal balance on the
crossroad, while using the contralateral leg to reach along one of the
directions as far as possible, then return and land on the near side of the
2.4. Ankle proprioception crossroad. Three direction reaches were performed in a randomized
order, with a 1-min rest given in between. Each direction was performed
The AMEDA was employed to provide ankle movement discrimina­ for three trials, with an up to 10-s interval for readiness between trials;
tion scores that represent acuity of ankle proprioception(Waddington the corresponding reach distances were then averaged (Picot et al.,
and Adams, 1999). This device consists of a fixed wooden platform 2021). Due to inherent body shape variability, the performance of each
housing a swinging square wooden plate that rotates around an axle direction was represented by the reaching distance being normalized
aligned to the long axis of the foot to be tested. Prior to the test, par­ with leg length (Picot et al., 2021), which was measured from the
ticipants were instructed to look ahead and stand astride the device, anterior superior iliac spine to the ipsilateral medial malleolus, in supine
with their tested foot centered over the axle of the movable plate and the (seen in the following equation):
untested one on the fixed platform. There was an array of wooden stops
with variable heights that could be placed underneath the movable 1∑ 3
SEBT(%) − A, PM or PL = D ÷ Leg length × 100
plate, which were used to restrict the movement extent of ankle inver­ 3 i=1
sion and to generate four variable positions, from the shallowest to the
deepest (positions 1, 2,3 and 4, representing 10, 12, 14 and 16 degrees of D: reaching distance(cm) in anterior(A), posteromedial (PM) or
inversion, respectively) (Fig. 2). posterolateral (PL) directions.
A complete AMEDA assessment consisted of familiarization and data
collection sessions. In the familiarization session of 12 trials, the par­ 2.6. Statistical analysis
ticipants were offered three rounds of experience of the four variable
movement extents. They were instructed to swing the movable plate, Data analysis was performed using SPSS version 28.0.1.1 (IBM
with feedback provided when the movement was terminated by one of Corporation, Route 100, NY105089), with a p value of 0.05 or less
the four physical stops that were manipulated in order, and thereafter considered as statistical significance. Reliability analysis was performed
the foot was actively returned to the start position to prepare for the next based on a two-way mixed effect model to test the consistency of results
trial. In the data collection session of 40 trials, each of the four physical between the two testing sessions. Intraclass correlation coefficients (ICC
stops was randomly presented 10 times. The participants moved the (3,1)) were used to quantify intra-rater or test-retest reliability, where
plate as before, without feedback provided, and they needed to orally values less than 0.5 are indicative of poor reliability, values between 0.5
report the perceived position using numbers 1,2,3 or 4, when the
movement was terminated. All participants were informed that they
could take a rest if needed during the assessment of 40 trials.
The ability of participants to discriminate between the variable ankle
positions was analyzed based on non-parametric signal detection theory,
wherein the perceived positions and actual ones were paired and used to
produce Receiver Operating Characteristic (ROC) Curves. Thereafter,
the mean Area under the curve (AUC) was calculated as the ankle

Fig. 2. Active movement extent discrimination assessment. Fig. 3. Three components of star excursion balance test.

3
X. Shi et al. Musculoskeletal Science and Practice 66 (2023) 102809

and 0.75 moderate reliability, values between 0.75 and 0.9 indicate (F = 6.9, p = 0.02), with the performance at the third session signifi­
good reliability, and values greater than 0.90 indicate excellent reli­ cantly better than the first session (MD(SE) = 10.1(3.8), 95%CI (0.23,
ability (Weir, 2005). Thereafter, the standard error of measurements 20.0), p = 0.04). Among these assessments, neither any gender main
(SEM) and the minimal detectable change with 90% confidence effect nor interaction effect was found (Table 4).
(MDC90), were calculated and reported in this study (Lin et al., 2009),
through the formulations that SEM equals the pooled standard deviation 4. Discussion
(SDpooled) multiplied by the square root of 1-ICC, and MDC90 equals
1.65 multiplied by the square root of 2 × SEM. Paired sample t tests were For the purpose of this study, we identified and applied criteria to
used to compare CAIT scoring between sides; three-way repeated mea­ ensure we assessed a cohort with unilateral CAI, and then investigated
sures analysis of variance (ANOVA) was utilized to investigate the ef­ the intra-rater reliability for a goniometer and WBLT used to measure
fects of symptom, sex and testing times on assessments, as well as any ankle dorsiflexion mobility, SEBT with three reaching components to
underlying interactions. Post-hoc analysis was then performed using the assess dynamic balance, and test-retest reliability for the AMEDA test of
Bonferroni method when any statistically significant difference was ankle proprioception. The most important finding here was that all as­
detected, wherein the mean difference (MD) and standard error (SE) sessments showed good-to-excellent reliability for the asymptomatic
were calculated. side, but were relatively less reliable for the symptomatic side.
Measurement of DFROM with a goniometer demonstrated good
3. Results reliability for the asymptomatic side, a finding which is comparable with
previous research involving healthy cohorts (Konor et al., 2012), while
Results demonstrate that CAIT scoring varied between sides, with the this instrument was only moderately reliable for the symptomatic side. It
score for the symptomatic side significantly lower than the asymptom­ should be noted that ankle DFROM can be measured using a goniometer
atic (mean difference (MD) = 11.8 ± 6.2, t = 9.3, p < 0.001). Perfor­ at different knee flexions (Alawna et al., 2019, Baumbach et al., 2016,
mance of assessments over the test series, and test-retest reliability, SEM, Searle et al., 2018). The reliability of different positions may vary, which
and MDC90 were calculated and are presented in Tables 2 and 3. For the warrants further investigation. Nevertheless, the current study showed
asymptomatic side, WBLT and the three components of SEBT demon­ that the reliability for ankle range measured using a goniometer with the
strated excellent reliability, with the intraclass correlation coefficients knee in extension was impacted in CAI cohorts. In contrast, WBLT
(ICCs (3,1)) between 0.94 and 0.96; and goniometer measured DFROM demonstrated consistently excellent reliability for both sides. The
and AMEDA proprioception demonstrated good reliability, with both MDC90 values of WBLT were 1.09 cm for the symptomatic side, and 1.70
ICCs (3,1) exceeding 0.80. For the symptomatic side, WBLT, SEBTA and cm for asymptomatic side, slightly lower than the 1.90 cm that was
SEBTPM all had excellent reliability, with ICC (3,1) between 0.90 and previously reported through pooling date from diverse populations
0.95; however, the reliability for goniometer-measured DFROM and (Powden et al., 2015). These findings underpin the clinical application
AMEDA were all moderate, with ICCs (3,1) between 0.6 and 0.67, but of WBLT as a method of high reliability for assessing DFROM deficits in
their 95% CI boundaries cross 0.5, the cutoff of poor reliability. CAI populations.
In the three-way (side*gender*time) repeated measures ANOVA, This study established the test-retest reliability of the AMEDA for this
DFROM by WBLT showed a significant side main effect (F = 16.91, p < specific population, via adoption of a strict inclusion criterion for CAI
0.001), and the post-hoc analysis using the Bonferroni method indicated (Lin, Mayer, 2022a). The present results showed that reliability was
that the asymptomatic side significantly outperformed the symptomatic good for the asymptomatic side but moderate for the symptomatic side,
side (MD(SE) = 1.4 (0.33) cm, 95%CI (0.67, 2.0),p = 0.00), while from indicating that the reliability of AMEDA scores was also impacted by
using the goniometer the symptom difference was not statistically sig­ ankle dysfunction. This suggests that the proprioceptive performance of
nificant (F = 2.0, p = 0.17). There was neither a main effect nor inter­ the injured ankle may be not as consistent as for the uninjured side. In
action for the ankle proprioception tests conducted by the AMEDA. In other work involving the AMEDA, the testing protocol was modified to
term of balance tests using the SEBT, SEBTA showed a significant side optimize the interaction with daily or sport-specific movement, e.g.
main effect (F = 5.38, p = 0.03), with the asymptomatic side signifi­ stepping (Witchalls et al., 2012) or landing (Han et al., 2021), giving
cantly outperforming the symptomatic side (MD(SE) = 2.1(0.92),95%CI good to excellent reliability. These authors argued that participants
(0.23,4.0), p = 0.03). There was no significant finding for SEBTPM over might be able to source more information regarding position and joint
test series, while SEBTPL showed an overall significant time main effect orientation in the course of interacting with the movement. Cumula­
tively, these findings imply that assessing ankle proprioception should
be task-specific, and possibly side-specific.
Table 2 Over the measurements of SEBT for CAI, performances in the three
Assessment performance over test series (Mean ± SD). directions all showed excellent test-retest reliability for the asymptom­
Side Variables Testing sessions atic side. This result is consistent with a previous report where analysis
First Second Third was performed by pooling studies involving healthy cohorts of varying
ages (Picot et al., 2021). For the symptomatic side, SEBTA and SEBTPM
AS GM (deg) 15.2 ± 6.2 15.1 ± 4.7 15.0 ± 4.9
WBLT (cm) 12.9 ± 2.0 12.7 ± 1.9 13.2 ± 1.8 consistently showed excellent reliability; however, SEBTPL only showed
AMEDA 0.67 ± 0.09 0.68 ± 0.07 0.69 ± 0.08 poor to moderate reliability. This disparity may be ascribed to a series of
SEBTA (%) 81.3 ± 8.1 78.5 ± 10.7 80.9 ± 8.3 improvements among the recruited participants over the multiple exe­
SEBTPM (%) 93.7 ± 11.1 92.5 ± 13.8 94.9 ± 12.2 cutions of SEBTPL. As identified by previous researchers, SEBTPL re­
SEBTPL (%) 82.2 ± 18.2 82.4 ± 19.0 87.8 ± 16.4
quires more muscle co-activation (Earl and Hertel, 2001), and may be
SS GM (deg) 12.0 ± 5.6 14.2 ± 6.7 13.5 ± 5.0 more difficult to perform than the other two directions. Indeed, as is
WBLT (cm) 10.9 ± 2.1 11.4 ± 2.5 11.0 ± 2.4
evident from Table 2, the values SEBTPL vs SEBTPM for the first session,
AMEDA 0.67 ± 0.09 0.65 ± 0.09 0.68 ± 0.06
SEBTA (%) 78.1 ± 9.1 77.6 ± 9.1 78.9 ± 8.5 i.e., 82.18% vs 93.66% for the asymptomatic side, and 79.75% vs
SEBTPM (%) 93.1 ± 13.7 92.7 ± 10.8 95.6 ± 12.0 93.08% for the symptomatic side, do suggest that the initial poorer
SEBTPL (%) 79.8 ± 18.1 83.4 ± 17.8 94.7 ± 32.1 performance in the SEBTPL may be improved through repeated expo­
AS: asymptomatic side, SS: symptomatic side, GM: goniometer, WBLT: weight- sure. In this current CAI cohort, only part improved in SEBTPL perfor­
bearing lunge test, AMEDA: active movement extent discrimination assess­ mance during the assessment procedure, which subsequently affected
ment, SEBT (A, PM, PL): anterior, posteromedial, and posterolateral components the previous performance rating within this cohort. Given the restriction
of the star excursion balance test. of sample size, the current study was unable to further explore this

4
X. Shi et al. Musculoskeletal Science and Practice 66 (2023) 102809

Table 3
Test-retest reliability for both sides.
Variable ICC (3,1) AS MDC90 ICC (3,1) SS MDC90

95%CI SEM 95%CI SEM

GM 0.80 (0.60 0.91) 2.4 5.6 0.67 (0.34 0.85) 3.00 7.0
WBLT 0.94 (0.88 0.97) 0.47 1.1 0.90 (0.81 0.95) 0.74 1.7
AMEDA 0.80 (0.60 0.91) 0.04 0.09 0.60 (0.22 0.82) 0.05 0.15
SEBTA 0.94 (0.88 0.97) 2.2 5.2 0.95 (0.89 0.98) 2.1 4.5
SEBTPM 0.95 (0.90 0.98) 2.8 6.8 0.95 (0.90 0.98) 2.7 5.6
SEBTPL 0.96 (0.92 0.98) 3.6 8.2 0.51 (0.02 0.78) 16.6 47.3

AS: asymptomatic side, SS: symptomatic side; GM: goniometer, WBLT: weight-bearing lunge test, AMEDA: active movement extent discrimination assessment, SEBT
(A, PM, PL): anterior, posteromedial, and posterolateral components of the star excursion balance test. ICC (3,1): Intraclass Correlation Coefficient. All participants were
assessed by the same tester, and the reliability index was calculated from three times of measurement. ICC less than 0.5 representing poor, between 0.5 and 0.75 -
√̅̅̅̅̅
moderate, between 0.75 and 0.9 – good, and >0.9 excellent. SEM: standard error of measures, SEM = SDpooled* (1 − ICC); MDC90: minimal detectable change given
√̅̅̅
90% confidence, MDC90 = 1.65* 2*SEM.

Table 4
Three-way (Side* Gender*Time) repeated measures ANOVA.
Side Gender Time Side*Time Gender* Side Time* Sex Side* Gender*Time

GM F = 2.0, p = 0.17 F= 0.25, p = 0.63 F = 0.08, p = 0.79 F= 1.3, p = 0.26 F = 0.89, p = 0.17 F = 0.91, p = 0.35 F= 1.1, p = 0.31
WBLT F = 16.9, p < 0.001* F= 3.9, p = 0.06 F = 3.1, p = 0.09 F= 0.00, p = 0.96 F = 0.69, p = 0.42 F = 2.6, p = 0.13 F= 1.6, p = 0.22
AMEDA F = 1.30, p = 0.27 F= 0.34, P = 0.56 F = 0.14, P = 0.71 F= 0.78, p = 0.39 F = 0.44, p = 0.52 F = 0.18, p = 0.67 F= 0.23, p = 0.63
SEBT A F = 5.4, p = 0.03* F= 0.44,p = 0.51 F = 0.09, p = 0.77 F= 1.7, p = 0.20 F = 1.5, p = 0.23 F = 2.2, p = 0.16 F= 0.001, p = 0.98
SEBT PM F = 0.34, p = 0.57 F= 1.31, p = 0.27 F = 2.3, p = 0.14 F= 2.7, p = 0.12 F = 1.46, p = 0.24 F = 0.31, p = 0.58 F= 1.04, p = 0.32
SEBT PL F = 0.22, p = 0.64 F= 3.1, p = 0.09 F = 6.9, p = 0.02* F= 2.0, p = 0.17 F = 0.18, p = 0.67 F = 0.75, p = 0.38 F= 1.6, F = 0.23

* Indicative of statistical significance.

GM: goniometer, WBLT: weight-bearing lunge test, AMEDA: active movement extent discrimination assessment, SEBT (A, PM, PL): anterior, posteromedial, and
posterolateral components of the star excursion balance test.

phenomenon via subgroup analysis, and it would be necessary to employ performance between the genders (Lu et al., 2022). Together with the
a larger sample to identify the underlying mechanisms in the future. current findings, a gender effect on motor control may present implicitly
Conversely, neither SEBTA nor SEBTPL showed learning effects over the through internal modulation, rather than explicitly in functional
test series, which may indicate their stability in application. Cumula­ performance.
tively, SEBTA and SEBTPM may be more reliable than SEBTPL in There were some limitations to this study. First, we only investigated
assessing dynamic balance for individuals with unilateral CAI. test-retest and intra-rater reliability, not inter-rater reliability. Secondly,
This study was also powered to investigate the effects of symptom a Chinese-version of the CAIT with a cutoff of 21 for CAI was utilized at
and gender on impairment outcome measures. The current findings enrollment, therefore the current findings might be only applied to
suggested WBLT demonstrated better discriminative validity than Chinese-speaking individuals with CAI. Thirdly, the study may not have
DFROM measurements made using a goniometer, with the symptomatic had sufficient power to detect differences regarding gender differences
side consistently being underscored by 1.25–1.55 cm relative to the in the assessments due to the relatively small sample size. Additionally,
asymptomatic side, and higher than the currently calculated MDC90 (1.1 this study cohort did not include older adults. Considering that aging
cm) for the asymptomatic side, suggestive of a unilateral impairment in may affect balance, proprioception, and joint mobility, in that older
terms of ankle mobility. In terms of balance, SEBT with its three individuals tend to underperform consistently relative to the young, the
reaching components did not show a symptomatic effect. Specifically, present data avoided the potential confounding due to age heteroge­
the asymptomatic side significantly outperformed the symptomatic side neity; nevertheless, the findings may not be generalizable to older
on the SEBTA, reflecting the established relationship with WBLT (Hoch populations. Future work is expected to explore the differences in
et al., 2012), while the detected side-to-side difference of between 1.93 neuromuscular control between young and older individual with CAI.
and 2.19 was less than the calculated MDC90 for both sides. This may
reflect the contribution of other elements in this reaching performance, 5. Conclusion
e.g., lower limb strength, or proprioception (Picot et al., 2021). No side
difference was identified in SEBTPM or SEBTPL. Furthermore, over the Our results suggest that WBLT, SEBTA and SEBTPM can be strongly
test series with the AMEDA, there was also no significant difference recommended for detecting ankle dorsiflexion mobility deficit, and dy­
between asymptomatic and symptomatic sides. The mean AUCs over namic balance deficit for both sides of individuals with unilateral CAI,
three sessions all were between 0.65 and 0.69, compared to scores all while only WBLT can be used for side-to-side comparison. The inter­
above 0.71 in peers and senior populations investigated in preceding pretation of goniometer results taken to measure DFROM in non-weight
studies (Han et al., 2014, Witchalls et al., 2012, Yang et al., 2022). bearing position, and SEBTPL, and AMEDA should be made cautiously
Cumulatively, the current findings confirm the bilateral impairments in for this specific cohort, considering their poor to good reliability for the
ankle proprioception and balance control identified by previous studies symptomatic side. Taken together, the findings here suggest that re­
(Sousa et al., 2017, Wikstrom et al., 2010, Witchalls et al., 2012). On the searchers and clinicians should be cautious when using these measures
other hand, the current results did not show any gender effect, and over on the symptomatic side in individuals with CAI.
test series the DFROM, ankle proprioception and balance performance
scores were comparable between male and females with unilateral CAI. Funding
Nevertheless, our previous work has suggested varying reliance on hip
abductor strength in perception of ankle instability and functional This research did not receive any specific grant from funding

5
X. Shi et al. Musculoskeletal Science and Practice 66 (2023) 102809

agencies in the public, commercial, or not-for-profit sectors. Lin, C.I., Houtenbos, S., Lu, Y.H., Mayer, F., Wippert, P.M., 2021. The epidemiology of
chronic ankle instability with perceived ankle instability- a systematic review.
J. Foot Ankle Res. 14. https://doi.org/10.1186/s13047-021-00480-w.
Lin, C.I., Mayer, F., Wippert, P.M., 2022a. Cross-cultural adaptation, reliability, and
Declaration of competing interest validation of the taiwan-Chinese version of Cumberland ankle instability tool.
Disabil. Rehabil. 44, 781–787. https://doi.org/10.1080/09638288.2020.1774928.
We declare that there was no conflict of interest in this study. Lin, C.I., Mayer, F., Wippert, P.M., 2022b. The prevalence of chronic ankle instability in
basketball athletes: a cross-sectional study. Bmc Sports Sci Med Rehabili 14. https://
doi.org/10.1186/s13102-022-00418-0.
References Lin, K.C., Hsieh, Y.W., Wu, C.Y., Chen, C.L., Jang, Y., Liu, J.S., 2009. Minimal detectable
change and clinically important difference of the Wolf Motor Function Test in stroke
Alawna, M.A., Unver, B.H., Yuksel, E.O., 2019. The reliability of a smartphone patients. Neurorehabilitation Neural Repair 23, 429–434. https://doi.org/10.1177/
goniometer application compared with a traditional goniometer for measuring ankle 1545968308331144.
joint range of motion. J. Am. Podiatr. Med. Assoc. 109, 22–29. https://doi.org/ Lu, J.L., Wu, Z.G., Adams, R., Han, J., Cai, B., 2022. Sex differences in the relationship of
10.7547/16-128. hip strength and functional performance to chronic ankle instability scores.
Anandacoomarasamy, A., Barnsley, L., 2005. Long term outcomes of inversion ankle J. Orthop. Surg. Res. 17 https://doi.org/10.1186/s13018-022-03061-0.
injuries. Br. J. Sports Med. 39, e14. https://doi.org/10.1136/bjsm.2004.011676 Morales, C.R., Lobo, C.C., Rodriguez Sanz, D., Corbalan, I.S., Ruiz, B.R., Lopez, D.L.,
discussion e. 2017. The concurrent validity and reliability of the Leg Motion system for measuring
Baumbach, S.F., Braunstein, M., Seeliger, F., Borgmann, L., Böcker, W., Polzer, H., 2016. ankle dorsiflexion range of motion in older adults. PeerJ 5. https://doi.org/10.7717/
Ankle dorsiflexion: what is normal? Development of a decision pathway for peerj.2820.
diagnosing impaired ankle dorsiflexion and M. gastrocnemius tightness. Arch. Picot, B., Terrier, R., Forestier, N., Fourchet, F., McKeon, P.O., 2021. The star excursion
Orthop. Trauma Surg. 136, 1203–1211. https://doi.org/10.1007/s00402-016-2513- balance test: an update review and practical guidelines. Int. J. Athl. Ther. Train. 26,
x. 285–293. https://doi.org/10.1123/ijatt.2020-0106.
Earl, J.E., Hertel, J., 2001. Lower-extremity muscle activation during the star excursion Powden, C.J., Hoch, J.M., Hoch, M.C., 2015. Reliability and minimal detectable change
balance tests. J. Sport Rehabil. 10, 93–104. https://doi.org/10.1123/jsr.10.2.93. of the weight-bearing lunge test: a systematic review. Man. Ther. 20, 524–532.
Gribble, P.A., Delahunt, E., Bleakley, C., Caulfield, B., Docherty, C., Fourchet, F., et al., https://doi.org/10.1016/j.math.2015.01.004.
2014. Selection criteria for patients with chronic ankle instability in controlled Searle, A., Spink, M.J., Chuter, V.H., 2018. Weight bearing versus non-weight bearing
research: a position statement of the International Ankle Consortium. Br. J. Sports ankle dorsiflexion measurement in people with diabetes: a cross sectional study.
Med. 48, 1014. https://doi.org/10.1136/bjsports-2013-093175. BMC Muscoskel. Disord. 19, 183. https://doi.org/10.1186/s12891-018-2113-8.
Han, J., Anson, J., Waddington, G., Adams, R., 2014. Sport attainment and Shi, X., Han, J., Witchalls, J., Waddington, G., Adams, R., 2019. Does treatment duration
proprioception. Int. J. Sports Sci. Coach. 9, 159–170. https://doi.org/10.1260/1747- of manual therapy influence functional outcomes for individuals with chronic ankle
9541.9.1.159. instability: a systematic review with meta-analysis? Musculoskelet Sci Pract 40,
Han, J., Luan, L., Adams, R., Witchalls, J., Newman, P., Tirosh, O., et al., 2022. Can 87–95. https://doi.org/10.1016/j.msksp.2019.01.015.
therapeutic exercises improve proprioception in chronic ankle instability? A Sousa, A.S.P., Leite, J., Costa, B., Santos, R., 2017. Bilateral proprioceptive evaluation in
systematic review and network meta-analysis. Arch Phys Medicine rehabili. https:// individuals with unilateral chronic ankle instability. J. Athl. Train. 52, 360–367.
doi.org/10.1016/j.apmr.2022.04.007. https://doi.org/10.4085/1062-6050-52.2.08.
Han, J., Waddington, G., Adams, R., Anson, J., Liu, Y., 2016. Assessing proprioception: a Thompson, C., Schabrun, S., Romero, R., Bialocerkowski, A., van Dieen, J., Marshall, P.,
critical review of methods. J Sport Health Sci 5, 80–90. https://doi.org/10.1016/j. 2018. Factors contributing to chronic ankle instability: a systematic review and
jshs.2014.10.004. meta-analysis of systematic reviews. Sports Med. 48, 189–205. https://doi.org/
Han, J., Yang, Z.H., Adams, R., Ganderton, C., Witchalls, J., Waddington, G., 2021. Ankle 10.1007/s40279-017-0781-4.
inversion proprioception measured during landing in individuals with and without Waddington, G., Adams, R., 1999. Discrimination of active plantarflexion and inversion
chronic ankle instability. J. Sci. Med. Sport 24, 665–669, 10.101/j. movements after ankle injury. Aust. J. Physiother. 45, 7–13. https://doi.org/
jsams.2021.02.004. 10.1016/s0004-9514(14)60335-4.
Hertel, J., Corbett, R.O., 2019. An updated model of chronic ankle instability. J. Athl. Weir, J.P., 2005. Quantifying test-retest reliability using the intraclass correlation
Train. 54, 572–588. https://doi.org/10.4085/1062-6050-344-18. coefficient and the SEM. J. Strength Condit Res. 19, 231–240. https://doi.org/
Hoch, M.C., Staton, G.S., Medina McKeon, J.M., Mattacola, C.G., McKeon, P.O., 2012. 10.1519/15184.1.
Dorsiflexion and dynamic postural control deficits are present in those with chronic Witchalls, J., Waddington, G., Blanch, P., Adams, R., 2012. Ankle instability effects on
ankle instability. J. Sci. Med. Sport 15, 574–579. https://doi.org/10.1016/j. joint position sense when stepping across the active movement extent discrimination
jsams.2012.02.009. apparatus. J. Athl. Train. 47, 627–634. https://doi.org/10.4085/1062-6050-
Horváth, Á., Ferentzi, E., Schwartz, K., Jacobs, N., Meyns, P., Köteles, F., 2022. The 47.6.12.
measurement of proprioceptive accuracy: a systematic literature review. J Sport Witchalls, J.B., Waddington, G., Adams, R., Blanch, P., 2014. Chronic ankle instability
Health Sci. https://doi.org/10.1016/j.jshs.2022.04.001. affects learning rate during repeated proprioception testing. Phys. Ther. Sport 15,
Jones, R., Carter, J., Moore, P., Wills, A., 2005. A study to determine the reliability of an 106–111. https://doi.org/10.1016/j.ptsp.2013.04.002.
ankle dorsiflexion weight-bearing device. Physiotherapy 91, 242–249. https://doi. Wyndow, N., Collins, N.J., Vicenzino, B., Tucker, K., Crossley, K.M., 2018. Foot and ankle
org/10.1016/j.physio.2005.04.005. characteristics and dynamic knee valgus in individuals with patellofemoral
Kim, D.H., An, D.H., Yoo, W.G., 2018. Validity and reliability of ankle dorsiflexion osteoarthritis. J. Foot Ankle Res. 11, 1–6. https://doi.org/10.1186/s13047-018-
measures in children with cerebral palsy. J. Back Musculoskelet. Rehabil. 31, 0310-1.
465–468. https://doi.org/10.3233/bmr-170862. Yang, N., Adams, R., Waddington, G., Han, J., 2022. Ankle complex proprioception and
Konor, M.M., Morton, S., Eckerson, J.M., Grindstaff, T.L., 2012. Reliability of three plantar cutaneous sensation in older women with different physical activity levels.
measures of ankle dorsiflexion range of motion. Int J Sports Phys Ther 7, 279–287. Exp. Brain Res. 240, 981–989. https://doi.org/10.1007/s00221-021-06273-8.