Concurrent validity,test-retest reliability,and sensitivity of a PostureRite system measurement on dynamic postural sway and risk of fall in cerebral palsy

Abstract

BACKGROUND:

Accurately diagnosing dynamic postural sway (DPS) is essential for effective and sustainable intervention in children with cerebral palsy (CP). We developed an accurate, inexpensive, and wearable DPS measurement system to measure DPS accurately and consistently during walking and functional activities of daily living.

OBJECTIVE:

We investigated the validity and reliability of this PostureRite system in children with CP, and the link between PostureRite and clinical measures including gross motor function measure (GMFM), pediatric balance scale (PBS), and fall efficacy scale (FES).

METHODS:

Twenty-one participants were categorized as follows: 11 healthy adults (3 females, mean age, 25.00±1.00 years) and 10 children with CP (mean age, 11.10±6.28 years). We determined the concurrent validity of PostureRite by comparing DPS data to the gold standard accelerometer measurement results. We determined test-retest reliability by measuring DPS data on three occasions at 2-h intervals. We assessed PostureRite measurement sensitivity to ascertain differences between healthy children and children with CP DPS measurements.

RESULTS:

Random and mixed intraclass correlation coefficients (ICC_2,k and ICC_3,k) were obtained; an independent T-test was performed (P < 0.05). Concurrent validity analysis showed a good relationship between the gold standard accelerometer and PostureRite (ICC_2,k = 0.973, P < 0.05). Test-retest reliability demonstrated a good relationship across the three repeated measures of the DPS data (ICC_3,k = 0.816–0.924, P < 0.05). Independent T-test revealed a significant difference in DPS data between healthy adults and children with CP (P < 0.05).

CONCLUSIONS:

Keywords

Validity reliability sensitivity dynamic postural sway cerebral palsy

1 Introduction

Dynamic postural sway (DPS) is a cardinal sign of neuromuscular impairment in children with cerebral palsy (CP), which often predisposes them to a higher risk of falls during ambulation (Tracy et al., 2019; Morgan & Mcginley, 2013). Conventionally, DPS is measured by a variety of clinical measurement tools such as the pediatric balance scale (PBS) (intraclass correlation coefficient [ICC], > 0.9), fall efficacy scale (FES) (ICC = 0.96), gross motor function measure (GMFM)-D (ICC = 0.53), and GMFM-E (ICC = 0.67); these tests vary in their degree of validity, reliability, and specificity in children with CP (Franjoine et al., 2010; Delbeare et al., 2010; Bania et al., 2014; Begnoche et al., 2016). Furthermore, based on an extensive review of the literature, the most common clinical rating scales used to evaluate postural control are limited by bias of the clinicians, insensitivity to mild impairments (ceiling effects), and poor reliability (ICC_1,1 = 0.55 0.84) (Mancini et al., 2012; Horak et al., 2009). Based on the current conceptual framework of the precision medicine model, such limitations in the current clinical measures affect” precision” clinical decision-making to accurately diagnose a disease, determine the prognosis of it, determine intervention efficacy, or treat people with mild balance deficits (Mancini et al., 2012; Horak et al., 2009). Consequently, a more sophisticated, objective approach for DPS measurement using a force plate system has been utilized. However, measuring postural sways during functional activity and participation in children with CP has significant accessibility and portability limitations in the clinical setting (Clifford & Holder-Powell, 2010; Prado et al., 2007). Thus, there is a clear need for a more accurate, portable, comfortable, and inexpensive measurement system for DPS in children with CP.

To mitigate the shortcomings of the current DPS measurement systems, we recently developed a portable, low-cost, wearable DPS measurement device (PostureRite; Mezoo, Wonju, Republic of Korea). PostureRite was designed to measure, analyze, and provide accurate biofeedback on dynamic postural control and gait deviations across the impairment, functional activity, and participation domains as defined in the World Health Organization International Classification of Functioning, Disability and Health (McDougall et al., 2010). While PostureRite is highly capable of detecting static and postural sway during functional activities such as walking, stair climbing, reaching, grasping, and associated participation at home and school settings, the measurement system’s validity and reliability have yet to be determined.

In the present study, we aimed to determine (1) the validity, reliability, and sensitivity of PostureRite to detect falls and the DPS measurement device in healthy adults and children with CP and (2) the correlation between PostureRite and conventional clinical measures such as GMFM-88, PBS, and FES in children with CP. We hypothesized that PostureRite would demonstrate good validity and reliability in DPS measurement, as well as strong relationships between the extent of DPS and clinical measures in children with CP.

2 Materials and methods

2.1 Participants

A convenience sample of 11 healthy adults (mean age, 25.00±1.00 years) and 10 children with CP (mean age, 11.10±6.28 years) was recruited from a major university and local community rehabilitation center, respectively. The study was approved by the Yonsei Institutional Review Board and the Ethics Committee (IRB no. 1041849-202106-BM-086-02). All participants provided informed consent prior to participation. All healthy participants had no known medical conditions. Inclusion criteria for children with CP were (1) diagnosis of CP, (2) age between 13 and 20 years, and (3) levels of I, II, and III as determined by the Gross Motor Function Classification System. Exclusion criteria were the presence of epilepsy and severe cognitive impairment. Participants completed a demographic and health questionnaire on medical history, sensation, range of motion, muscle strength, and balance to ensure safe participation in the present study.

2.2 Instrumentation

We developed a wireless, portable, 3-dimensional PostureRite sensor (Mezoo, Wonju, Republic of Korea) consisting of a three-axis acceleration sensor (MPU-9250, intensity) and a low-power Bluetooth System-on-Chip (SoC) (NRF52832, a nodic semiconductor). The low-power Bluetooth SoC is responsible for acquiring and preprocessing three-axis acceleration signals through inter-integrated circuit communication and transmitting them to the monitoring software of the laptop via Bluetooth (Fig. 1). The laptop monitoring software stores the post-processing acceleration signals received via Bluetooth locally. The monitoring software was created using LabVIEW software (LabVIEW 2018). The sensor is designed to detect DPS using three-axis data (anterior–posterior [AP], medial–lateral [ML], and vertical [V] directions). The sensor recorded 3-D linear accelerations and angular velocity, while the controller continuously. PostureRite is small (40×30×5 mm) and weighs 8 g; therefore, it has the advantage of being easy to move. The sensor recorded 3-D linear acceleration and angular velocity continuously. As shown in Fig. 2, PostureRite (indicated by a red cube) was consistently superimposed at approximately 2/3^rds of the line between the xiphoid process and jugular notch and slightly to the left, indicating three-direction acceleration.

Fig. 1

PostureRite sensor system block diagram.

Fig. 2

PostureRite location and 3-axis orientation.

2.3 Experimental procedure

All testing conditions were maintained as consistently as possible, using the same testers and procedures (e.g., consistent instruction, calibration, foot position, and testing sequence) throughout the experiment. This study consisted of two phases. The first phase was performed in the motion analysis laboratory to determine the concurrent validity of new accelerometers and compare them to the Vernier 3-D accelerometer (Vernier, OR, USA) for 11 healthy adults. The second phase was conducted in a local community rehabilitation center to determine the test-retest reliability and clinical sensitivity of PostureRite system for healthy adults and children with CP. Clinical sensitivity was determined by comparing the composite score of the DPS data (in the AP, ML, and V directions) with standardized clinical tests including GMFM-88, PBS, and FES. DPS was defined as the body sway performed in the highest GMFM-88 level possible as measured with an accelerometer device (in the AP, ML, and V directions).

2.3.1 GMFM-88

The GMFM-88 is a standard measure of quantitative changes in gross motor function in children with CP. The GMFM-88 measures the following five gross motor domains: (1) lying, (2) crawling and kneeling, (3) sitting, (4) standing, and (5) walking, running, and jumping. The scoring was graded on a 4-point scale (0 = unable to initiate the task; 1 = able to initiate the task; 2 = able to perform the task partially; and 3 = able to perform the task completely) (Alotaibi et al., 2014). The validity and reliability of the GMFM-88 were reported to be r = 0.99 and ICC = 0.95, respectively (Ko & Kim, 2013; Park, E.Y. & Park, S.Y., 2010).

2.3.2 PBS

The PBS is a revision of the Berg balance scale (BBS) and utilized in a pediatric population to determine functional balance in preschoolers and school-aged children with CP. The PBS has 14 items with maximum item score of 5 and assesses the following functional activities that a child must safely and independently perform at home or school, or in the community: sitting balance, standing balance, sitting to standing, standing to sitting, transfers, stepping, reaching forward, reaching to the floor, turning, and stepping on and off of an elevated surface (Franjoine et al., 2010; Yi et al., 2012). Scores range from 0 (unable to perform) to 4 (able to perform the task as instructed without difficulty) for each item. The validity and reliability of PBS were reported to be ICC_3,1 = 0.90 and r = 0.579, respectively (Chen et al., 2013; Yi et al., 2012).

2.3.3 FES

The FES is used to evaluate the scale of recognized efficacy (i.e., self-confidence) in avoiding a fall during each of the 10 relatively harmless activities of daily living. Scores range from 1 (not at all concerned) to 4 (very concerned) for each item that assesses the level of concern about falling when carrying out each activity on a four-point scale (Delbaere et al., 2010). The validity and reliability of the FES were reported to be ICC_3,1 = 0.70 and Cronbach’s alpha = 0.90, respectively (Jung et al., 2015; Delbaere et al., 2010).

2.4 Concurrent validity

Concurrent validity was determined by comparing the AP-, ML-, and V-axis DPS measurement data that were concomitantly recorded by the gold standard Vernier accelerometer in 11 healthy adults. For data acquisition, as illustrated in Fig. 3 (concurrent validity test procedure), each participant was initially positioned to be standing. The PostureRite and Vernier accelerometer sensors were consistently superimposed at approximately 2/3^rds of the line between the xiphoid process and jugular notch and slightly to the left. The participant was instructed to walk barefoot on a 7-m pathway at a self-selected speed.

Fig. 3

Concurrent validity test procedure (Red cube: PostureRite; blue cube: Vernier).

2.5 Test-retest reliability

Test-retest reliability was established by determining an instrument’s capability of measuring the DPS measurement data (by the AP-, ML-, and V-axis accelerometers) with consistency. To examine the test-retest reliability, 10 children with CP and 11 healthy adults were tested at the community rehabilitation center by two investigators. As with the validity test procedure, each participant was initially positioned to be standing, and the PostureRite sensor was attached. The participant was then instructed to walk. All participants in this study were given three to five practice trials, and data were collected three times over 10-min intervals.

2.6 Data processing

The data processing flow is shown in Fig. 4. Data were obtained by setting the scale of the acceleration sensor to±8 G, the cutoff frequency of the digital low-pass filter to 44.8 Hz, and the sampling frequency to 50 Hz. During pre-processing, the output of the acceleration sensor was down-sampled to 25 Hz by taking a two-point moving average. For post-treatment, the tilt angle (theta), tilt angle (phi), and tilt angle (psi) were determined using the tilt angle calculation as the three axes of acceleration. The root mean square (RMS) calculation was used to determine the RMS (AP direction), RMS (ML direction), and RMS (V direction) for each acceleration axis every 5 s. Acceleration 3-directional data, tilt angle, and RMS were stored automatically in local storage.

Fig. 4

Data processing flow. RMS, root mean square; DLPF, digital low pass filter.

Details on the post-processing method are provided as follows. The tilt angle was calculated every 25 Hz with an arctan formula for a three-axis component of 1 G of gravitational acceleration. The RMS was calculated for each axis of acceleration every 5 s.

Phi has a value of –90° to 90° and is positive when inclined to the left and negative when inclined to the right. $phi (degree) = {tan}^{- 1} (\frac{- Acc . Y}{\sqrt{Acc . X^{2} + Acc . Z^{2}}}) * \frac{180}{π}$

Theta has a value of –90° to 90°, positive when tilted forward and negative when tilted backward. $theta (degree) = {tan}^{- 1} (\frac{- Acc . X}{\sqrt{Acc . Y^{2} + Acc . Z^{2}}}) * \frac{180}{π}$

Psi has a value of 0° to 180° and indicates that it is inclined in any direction, left/right, and front/back. $psi (degree) = {tan}^{- 1} (\frac{\sqrt{Acc . Y^{2} + Acc . Z^{2}}}{Acc {. X}^{2}}) * \frac{180}{π}$ If psi < 0, psi = psi + 180

2.7 Statistical analysis

Descriptive statistics are expressed in mean and standard deviation. The minimum sample size was determined to be 20, with an effect size (eta squared, η²) of 1.02 and power (1-β) of 0.8 based on our previous pilot data obtained using the GMFM and PBS variables. Baseline demographic characteristics between groups were compared using the independent t-test for continuous variables and χ2 test for categorical variables. ICC_2,k tests were used to determine the validity of the DPS system in measuring body sway and test-retest reliability. All continuous variables were analyzed using the Kolmogorov-Smirnov test, assuming a normal distribution. Between-group (healthy adults and children with CP) differences in sensitivity were determined using an analysis of covariance (ANCOVA) because the baseline was statistically different between groups. A Pearson correlation coefficient was used to determine relationships between clinical outcome measures (GMFM, PBS, FES, and DPS). SPSS for Windows (version 25.0, SPSS, Chicago, IL, USA) was used to conduct statistical analyses. The alpha level was set at 0.05.

3 Results

All participants who successfully completed the test were included in the analysis. Table 1 summarizes the demographic and clinical characteristics of the participants.

Table 1
Demographic characteristics of the participants (N = 21)

Variable Healthy adults (n = 11) Children with CP (n = 10) P-value

Age (years), mean±SD 25.00±1.00a 11.1±6.28 0.412

Height (cm), mean±SD 176.00±5.78 131.2±29.23 0.661

Weight (kg), mean±SD 82.55±13.17 35.9±19.62 0.573

CP type (%) N/A Spastic Diplegia 70 –

Ataxia 10

Spastic Hemi 10

Athetoid 10

Male (%) 100 70 0.234

Female (%) 0 30

Variable	Healthy adults (n = 11)	Children with CP (n = 10)	P-value
Age (years), mean±SD	25.00±1.00a	11.1±6.28	0.412
Height (cm), mean±SD	176.00±5.78	131.2±29.23	0.661
Weight (kg), mean±SD	82.55±13.17	35.9±19.62	0.573
CP type (%)	N/A	Spastic Diplegia 70	–
		Ataxia 10
		Spastic Hemi 10
		Athetoid 10
Male (%)	100	70	0.234
Female (%)	0	30

a, mean±standard deviation; CP, cerebral palsy; N/A, not applicable; SD, standard deviation.

3.1 Concurrent validity

ICC analysis between DPS and Vernier accelerometer measures showed excellent similarity among 3-axis body sway measures (ICC_2,k = 0.973, P < 0.05). Figure 5 shows the linearity between the two measures, indicating that the DPS measurements obtained were valid to assess body sway. Concurrent validity analysis indicated a statistically significant difference in DPS between the two groups (P < 0.05), suggesting that PostureRite could be used instead of the gold standard accelerometer.

Fig. 5

A graph showing linearity in body sway in the AP, ML, and V directions measures between the DPS and Vernier methods. DPS, dynamic postural sway.

3.2 Test-retest reliability

Table 2 shows the results of the test-retest reliability of the participants’ repeated measures using the DPS measurement data. ICC_3,k was measured in the three gait training tests. The correlation between the repeated measures was high, with ICC_3,k ranging from 0.816–0.924 (P = 0.05).

Table 2
Test-retest reliability (N = 21)

AP direction ML direction V direction Mean

Healthy adults 0.838^* 0.816^* 0.924^* 0.859^*

Children with CP 0.832^* 0.850^* 0.911^* 0.864^*

	AP direction	ML direction	V direction	Mean
Healthy adults	0.838^*	0.816^*	0.924^*	0.859^*
Children with CP	0.832^*	0.850^*	0.911^*	0.864^*

^*P < 0.05; AP, anterior–posterior; ML, medial–lateral; V, vertical.

3.3 PostureRite DPS measurement sensitivity

Table 3 shows the comparisons of DPS among the three directions. The mean DPS differences among the groups were examined by using ANCOVA, which showed a significant difference among the groups (P = 0.01). The mean difference of the healthy adults’ group was –0.15±0.39, whereas that of the children with CP group was 5.75±12.43.

Table 3
DPS Measurement sensitivity (N = 21)

Healthy adults Children with CP P-value

AP direction –0.58±0.76 –1.17±6.18 0.001^*

ML direction –0.09±0.19 –4.52±21.3 0.001^*

V direction 0.22±0.23 22.93±9.80 0.001^*

	Healthy adults	Children with CP	P-value
AP direction	–0.58±0.76	–1.17±6.18	0.001^*
ML direction	–0.09±0.19	–4.52±21.3	0.001^*
V direction	0.22±0.23	22.93±9.80	0.001^*

^*P < 0.05; AP, anterior–posterior; ML, medial–lateral; V, vertical.

3.4 Correlation statistics between the clinical measures with PostureRite DPS Data

Table 4 shows the correlation of the GMFM-88, PBS, and FES scores with PostureRite DPS measurement data. An inverse relationship was observed between GMFM-88 and the AP direction (r = –0.593, P < 0.05). PBS scores were inversely correlated with the AP direction (r = –0.669, P < 0.05) and V direction (r = –0.541, P < 0.05). The correlation between FES scores and the AP direction was moderate (r = 0.623, P < 0.05). The ML direction did not correlate significantly with GMFM-88, PBS, and FES scores. The average of the 3 directions was inversely correlated with the GMFM and PBS (r = –0.410 and –0.526, respectively; P < 0.05).

Table 4
The correlation results of GMFM-88, PBS, and FES with PostureRite (N = 11)

AP direction ML direction V direction Mean of 3 direction

GMFM-88 –0.593^* –0.276 –0.455 –0.410^*

PBS –0.669^* –0.160 –0.541^* –0.526^*

FES 0.623^* –0.069 0.488 0.279

	AP direction	ML direction	V direction	Mean of 3 direction
GMFM-88	–0.593^*	–0.276	–0.455	–0.410^*
PBS	–0.669^*	–0.160	–0.541^*	–0.526^*
FES	0.623^*	–0.069	0.488	0.279

^*P < 0.05; PBS, pediatric Balance Scale; FES, fall efficacy scale; GMFM, gross motor function measure.

4 Discussion

The present investigation evaluated the concurrent validity, test-retest reliability, and clinical sensitivity of a novel PostureRite DPS measurement system in healthy adults and children with CP. As hypothesized, the PostureRite measurements demonstrated a high level of validity, reliability, and sensitivity in discriminating pathological postural sway in children with CP, as well as strong relationships between PostureRite and standardized clinical measurements including scores from GMFM-88, PBS, and FES.

4.1 Concurrent validity

The concurrent validity analysis revealed high correlations (ICC_2,k = 0.973) for all three directions of dynamic postural sway values during gait, suggesting that the accuracy of the PostureRite DPS measurement is compatible with the gold standard method of measurement. This finding was consistent with previous validity studies (Ahlrichs et al., 2016; Alessandrini et al., 2017; Mancini et al., 2012; O’Neil et al., 2016) that compared the 3-axis acceleration and center of pressure (COP) measurements during gait, which are considered the gold standard approach for DPS. Mancini and colleagues (2012) determined the concurrent validity by comparing acceleration and COP data recorded from an accelerometer and force plate in 13 healthy adults and 13 participants with idiopathic Parkinson’s disease, respectively (Mancini et al., 2012). Correlation data between COP and acceleration postural sway data ranged from r = 0.64 to 0.74, supporting moderate to good measurement accuracy. Alessandrini and colleagues (2017) investigated the concurrent validity between the acceleration and force plate measurements in 13 healthy adults and 13 participants with vestibular disorder during gait and reported a correlation of r = 0.90, indicating excellent postural sway measurement accuracy (Alessandrini et al., 2017). O’Neil and colleagues (2016) reported high concurrent validity (rho = 0.70–0.85) of accelerometer-based motion sensors with indirect calorimetry during physical activity including resting, writing, active video games, and walking in 57 youth with CP (O’Neil et al., 2016). It is important to examine postural stability in children with CP as most falls occur during ambulation. However, no previous validity study correlating COP and acceleration postural sway data has been reported in children with CP.

4.2 Test-retest reliability

Test-retest reliability analysis showed excellent reliability (ICC_3,k=0.816–0.924) of the “trunk” acceleration measurement of postural sway during gait in healthy adults and children with CP. This result was in line with that of previous reliability studies on postural sway measurements (O’Neil et al., 2016; You, 2005). O’Neil and colleagues (2016) examined the test-retest reliability of the “trunk” acceleration measurement of postural sway during gait children with CP and found an excellent measurement consistency (ICC = 0.98, 95% confidence interval: 0.92–0.99), which demonstrated excellent inter-instrument reliability (ICC = 0.94–0.99) of accelerometer-based motion sensors during physical activity in 57 young individuals with CP (O’Neil et al., 2016). You (2005) reported high test-rest reliability (ICC_2,k = 0.88–0.99) of the SENSERite system measurement of five ankle position sense tests using a potentiometer in 10 healthy adults, 22 healthy older adults, and 14 older adults with a history of falls (You, 2005). Previous studies suggest that, regardless of sensor location and subject conditions, the test-retest reliability of postural sway during gait was very consistent when measured with the inertial measurement unit (IMU) or acceleration sensor devices.

4.3 PostureRite DPS measurement sensitivity

DPS measurement sensitivity analysis showed significant differences in the ML, V, and AP directions between the healthy adults’ group and the children with CP group (P = 0.01), indicating the measurement system’s capability to discriminate differences in DPS between healthy adults and children with CP. Specifically, children with CP showed greater DPS deviations in the ML (99%), V (98%), and AP (50%) directions than healthy adults. The sensitivity of PostureRite was comparable to that of previous studies (Ahlrichs et al., 2016; Shany et al., 2011; Muir et al., 2013). You (2005) evaluated the sensitivity of the SENSERite or position sensor in healthy older adults and older adults with a history of falls (You, 2005). Muir and colleagues (2013) reported that the balance score of elderly participants with a history of recent falls was found to be approximately two times greater than the balance score of elderly non-fallers (Muir et al., 2013). Cho and colleagues (1995) evaluated the sensitivity of the two linear single-plane accelerometers in those with a history of falls and those without and reported that older adults without a history of falls showed greater Romberg’s test time (20 seconds) than older adults with a history of falls (Cho et al., 1995).

Interestingly, the ML-direction DPS deviation was consistently the largest among the V and AP directions, supporting neuromechanical evidence that the base of support (BOS) in the ML direction is relatively smaller than that in the AP direction to maintain the COP of the inverted body pendulum in the BOS during ambulation (Saxena et al., 2014). Winter (2009) observed greater ML and AP postural sway deviations in children with CP than in healthy children (Winter, 2009). Postural controls may stem from a lack of selective ankle joint neuromuscular control as a strategy for compensatory anticipatory postural control (Girolami et al., 2011; Winter, 2009).

4.4 Correlation of clinical measures with PostureRite

Correlation analyses showed a relationship be-tween PBS and the AP axis (r = –0.669, P < 0.05) as well as a positive relationship between FES and the AP axis (r = 0.623; P < 0.05), indicating that the DPS recorded from PostureRite was closely related to the clinical measure scores. Shahzad and colleagues (2017) studied the relationship between BBS scores and acceleration data recorded from the accelerometer in 23 older adults and reported a strong correlation (r = 0.86), suggesting an excellent relationship between balance and postural sway (Shahzad et al., 2017). Senden and colleagues (2012) evaluated Tinetti’s fall risk measurement scores and acceleration RMS data obtained from the triaxial accelerometer measurement in 100 older adults during walking and reported a Pearson correlation (r = 0.60), representing a moderate relationship between fall risk and postural sway (Senden et al., 2012). These findings suggest that the moderate to excellent relationships between the triaxial accelerometer measurement and clinical outcomes substantiate that our accelerometer measurement can be utilized to estimate the static and dynamic balance control and associated risk of falls in children with CP. Similarly, moderate relationships were noted between GMFM-88 and AP-axis (r = –0.593, P < 0.05), and between the V-axis and PBS (r = –0.541, P < 0.05), as well as between the mean of the three axes and PBS (r = 0.526, P < 0.05). In the current literature, only a single study by Abaid and colleagues (2013) compared GMFM scores and acceleration data recorded from IMU measurement in 10 children with CP and reported a Pearson correlation of r = 0.92, indicating a good relationship between gross motor function and postural sway (Abaid et al., 2013).

Taken together, the present results indicate that PostureRite is a wearable and real-time measurement sensor, which can be used for monitoring and quantifying key postures and movement patterns in children with and without the postural and movement impairments that are commonly observed in CP during standing and walking. Furthermore, the PostureRite system measurements can help clinicians to differentiate abnormal posture and movement patterns from normal control data, thus providing important feedback information for the accurate diagnosis and implementation of effective intervention strategies.

A couple of study limitations should be considered in future research. One main limitation is that the age-matched comparison was not performed between individuals with and those without CP in the present study; however, the independent T-test statistical analysis failed to show a significant difference between groups, indicating relative group homogeneity. The other limitation is that the sensitivity was only established in children with spastic diplegic CP and ataxic CP. Therefore, generalizing other types of postural and movement impairments associated with balance and falls in children with CP is difficult. Our current findings should thus be interpreted with caution for the clinical management of children with CP.

5 Conclusions

We developed a portable, wireless, and affordable PostureRite system to measure DPS during gross motor function associated with daily activity and participation, and established the concurrent validity, test-retest reliability as sensitivity, and clinical relevance by comparing the DPS obtained from the participants with and without CP. The PostureRite system measurement was accurate, reliable, sensitive, and highly correlated with the clinical balance test and fall risk measures. The PostureRite system measurement can be used as an alternative DPS measurement tool to measure DPS in individuals with CP when designing accurate assessment and effective and sustainable intervention strategies to improve balance and gross motor function and reduce the risk of falling during daily activities and participation. We believe our results have great therapeutic value and can aid clinicians in prescribing even more effective CP rehabilitation programs.

Footnotes

Acknowledgments

This research received financial and administrative support from the Korea Health Industry Development Institute (grant no. HI18C1687000020).

Conflict of interest

None to report.

References

Abaid,

, Cappa,

, Palermo,

, Petrarca,

, & Porfiri,

(2013). Gait detection in children with and without hemiplegia using single-axis wearable gyroscopes. PloS One, 8(9) e73152.

Ahlrichs,

, Samà,

, Lawo,

, Cabestany,

, Rodríguez-Martín,

, Pérez-López,

, Sweeney,

, Quinlan,

L. R.

, Laighin,

G. Ò.

, & Counihan,

(2016). Detecting freezing of gait with a tri-axial accelerometer in Parkinson’s disease patients. Medical & Biological Engineering & Computing, 54(1), 223–233.

Alessandrini,

, Micarelli,

, Viziano,

, Pavone,

, Costantini,

, Casali,

, Paolizzo,

, & Saggio,

(2017). Body-worn triaxial accelerometer coherence and reliability related to static posturography in unilateral vestibular failure. Acta Otorhinolaryngologica Italica, 37(3), 231.

Alotaibi,

, Long,

, Kennedy,

, & Bavishi,

(2014). The efficacy of GMFM-88 and GMFM-66 to detect changes in gross motor function in children with cerebral palsy (CP): a literature review. Disability and Rehabilitation, 36(8), 617–627.

Chen,

C.L.

, Shen,

I.H.

, Chen,

C.Y.

, Wu,

C.Y.

, Liu,

W.Y.

, & Chung,

C.Y.

(2013). Validity, responsiveness, minimal detectable change, and minimal clinically important change of Pediatric Balance Scale in children with cerebral palsy. Research in Developmental Disabilities, 34(3), 916–922.

Cho,

C. Y.

, Kamen,

, & Holt,

(1995). Using accelerometry for assessing balance performance in normal elderly and frequent fallers. In Washington, DC: World Confederation of Physical Therapy Conference (Vol. 995).

Clifford,

A. M.

, & Holder-Powell,

(2010). Postural control in healthy individuals. Clinical Biomechanics, 25(6), 546–551.

Delbaere,

, Close,

J. C.

, Mikolaizak,

A. S.

, Sachdev,

P. S.

, Brodaty,

, & Lord,

S. R.

(2010). The falls efficacy scale international (FES-I). A comprehensive longitudinal validation study. Age and Ageing, 39(2), 210–216.

Franjoine,

M. R.

, Darr,

, Held,

S. L.

, Kott,

, & Young,

B. L.

(2010). The performance of children developing typically on the pediatric balance scale. Pediatric Physical Therapy, 22(4), 350–359.

10.

Girolami,

G. L.

, Shiratori,

, & Aruin,

A. S.

(2011). Anticipatory postural adjustments in children with hemiplegia and diplegia. Journal of Electromyography and Kinesiology, 21(6), 988–997.

11.

Jung,

, Shin,

H. H.

, Choi,

Y. W.

, & Kim,

K. M.

(2015). The relationship between fall efficacy and activity level in older adults. Physical & Occupational Therapy in Geriatrics, 33(1), 53–63.

12.

Ko,

, & Kim,

(2013). Reliability and responsiveness of the gross motor function measure-88 in children with cerebral palsy. Physical Therapy, 93(3), 393–400.

13.

Mancini,

, Salarian,

, Carlson-Kuhta,

, Zampieri,

, King,

, Chiari,

, & Horak,

F. B.

(2012). ISway: a sensitive, valid and reliable measure of postural control. Journal of Neuroengineering and Rehabilitation, 9(1), 1–8.

14.

McDougall,

, Wright,

, & Rosenbaum,

(2010). The ICF model of functioning and disability: incorporating quality of life and human development. Developmental Neurorehabilitation, 13(3), 204–211.

15.

Morgan,

, & McGinley,

(2013). Performance of adults with cerebral palsy related to falls, balance and function: a preliminary report. Developmental Neurorehabilitation, 16(2), 113–1203.

16.

Muir,

J. W.

, Kiel,

D. P.

, Hannan,

, Magaziner,

, & Rubin,

C. T.

(2013). Dynamic parameters of balance which correlate to elderly persons with a history of falls. PloS One, 8(8), e70566.

17.

O’Neil,

M. E.

, Fragala-Pinkham,

, Lennon,

, George,

, Forman,

, & Trost,

S. G.

(2016). Reliability and validity of objective measures of physical activity in youth with cerebral palsy who are ambulatory. Physical Therapy, 96(1), 37–45.

18.

Park,

E.Y.

, & Park,

S.Y.

(2010). Comparison of reliability and validity between GMFM-88 and GMFM-66 in children with cerebral palsy. Physical Therapy Korea, 17(3), 40–47.

19.

Prado,

J. M.

, Stoffregen,

T. A.

, & Duarte,

(2007). Postural sway during dual tasks in young and elderly adults. Gerontology, 53(5), 274–281.

20.

Saxena,

, Rao,

B. K.

, & Kumaran,

(2014). Analysis of postural stability in children with cerebral palsy and children with typical development: an observational study. PediatricPhysical Therapy, 26(3), 325–330.

21.

Senden,

, Savelberg,

, Grimm,

, Heyligers,

, & Meijer,

(2012). Accelerometry-based gait analysis, an additional objective approach to screen subjects at risk for falling. Gait & Posture, 36(2), 296–300.

22.

Shahzad,

, Ko,

, Lee,

J.A

, & Kim,

(2017). Quantitative assessment of balance impairment for fall-risk estimation using wearable triaxial accelerometer. IEEE Sensors Journal, 17(20), 6743–6751.

23.

Shany,

, Redmond,

S. J.

, Narayanan,

M. R.

, & Lovell,

N. H.

(2011). Sensors-based wearable systems for monitoring of human movement and falls. IEEE Sensors Journal, 12(3), 658–670.

24.

Tracy, James

, Drew Petersen,

, Jamie Pigman,

Benjamin Conner,

Henry Wright,

, Christopher Modlesky,

, Freeman Miller, , Curtis Johnson,

, & Jeremy Crenshaw,

(2019). Dynamic Stability During Walking in Children with and without Cerebral Palsy. Gait & Posture, 72, 182–187.

25.

Winter,

D. A.

(2009). Biomechanics and motor control of human movement. John Wiley & Sons.

26.

Yi,

S.H.

, Hwang,

J. H.

, Kim,

S. J.

, & Kwon,

J.Y.

(2012). Validity of pediatric balance scales in children with spastic cerebral palsy. Neuropediatrics, 43(06), 307–313.

27.

You,

S. H.

(2005). Joint position sense in elderly fallers: a preliminary investigation of the validity and reliability of the SENSERite measure. Archives of Physical Medicine and Rehabilitation, 86(2), 346–352.

Concurrent validity,test-retest reliability,and sensitivity of a PostureRite system measurement on dynamic postural sway and risk of fall in cerebral palsy

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1 Introduction

2 Materials and methods

2.1 Participants

2.2 Instrumentation

2.3.1 GMFM-88

2.3.2 PBS

2.3.3 FES

2.4 Concurrent validity

2.6 Data processing

3 Results

Table 2 Test-retest reliability (N = 21) AP direction ML direction V direction Mean Healthy adults 0.838* 0.816* 0.924* 0.859* Children with CP 0.832* 0.850* 0.911* 0.864*

Table 3 DPS Measurement sensitivity (N = 21) Healthy adults Children with CP P-value AP direction –0.58±0.76 –1.17±6.18 0.001* ML direction –0.09±0.19 –4.52±21.3 0.001* V direction 0.22±0.23 22.93±9.80 0.001*

Table 4 The correlation results of GMFM-88, PBS, and FES with PostureRite (N = 11) AP direction ML direction V direction Mean of 3 direction GMFM-88 –0.593* –0.276 –0.455 –0.410* PBS –0.669* –0.160 –0.541* –0.526* FES 0.623* –0.069 0.488 0.279

4.1 Concurrent validity

4.2 Test-retest reliability

4.3 PostureRite DPS measurement sensitivity

4.4 Correlation of clinical measures with PostureRite

5 Conclusions

Footnotes

Acknowledgments

Conflict of interest

References

Table 2
Test-retest reliability (N = 21)

AP direction ML direction V direction Mean

Healthy adults 0.838^* 0.816^* 0.924^* 0.859^*

Children with CP 0.832^* 0.850^* 0.911^* 0.864^*

Table 3
DPS Measurement sensitivity (N = 21)

Healthy adults Children with CP P-value

AP direction –0.58±0.76 –1.17±6.18 0.001^*

ML direction –0.09±0.19 –4.52±21.3 0.001^*

V direction 0.22±0.23 22.93±9.80 0.001^*

Table 4
The correlation results of GMFM-88, PBS, and FES with PostureRite (N = 11)

AP direction ML direction V direction Mean of 3 direction

GMFM-88 –0.593^* –0.276 –0.455 –0.410^*

PBS –0.669^* –0.160 –0.541^* –0.526^*

FES 0.623^* –0.069 0.488 0.279