Introduction

Postural stability is considered to be an important indicator of musculoskeletal health and therefore could be of importance in view of clinical issues. Postural stability refers to the inherent ability of a person to maintain, achieve or restore a specific state of balance and not to fall [16].

The most frequently used technique to evaluate postural stability, both static and dynamic, is the measurement of the position and displacement of the centre of pressure (COP) using a force plate form. Force plate measurements for postural assessment are widely used in adults and the reliability is well documented in this population [3, 4, 11]. Based on the literature, reliability reports and normative data for bilateral stance assessments in elementary schoolchildren are limited. Children’s development of postural stability using bilateral force plate measurements was earlier described in a number of studies [5, 10, 12, 13, 15, 18, 19]. However, the latter studies did not incorporate reliability reports on the evaluation method of force plate measurements for a childhood population. In contrast, reliability was investigated in childhood and adolescence with respect to other postural assessment techniques. Accordingly, the study of Atwater et al. [1] investigated reliability for the one-leg stance and a tiltboard balance test in 4 to 9 year olds; Gabriel and colleagues [9] evaluated in 5 to 9 year olds test-retest for the Neurocom VSR; Emery et al. [6] evaluated test-retest for the one-leg stance in 14 to 19 year olds and Mc Evoy and Grimmer [14] described repeated testing of upright posture evaluation using saggital plane photography in children aged 5 to 12 years. Further, Baker et al. [2] assessed in 2 to 12 year olds the reliability of two systems assessing static standing balance, but test-retest reliability of the separate systems was not measured. Only when test-retest reliability for postural stability assessments using force plate measurements in elementary school-age children is established, reference data can be determined and possible associations to impairments within the clinical field could be investigated in early stages of the potential problems.

According to Rival et al. [18] investigating developmental changes of postural control in children with respect to standing balance, a transition phase should occur around 7–8 years. In 9 to 10 year olds standing balance appeared to be adult-like [18], however, not fully matured [17]. Therefore, the present study was designed to report test-retest reliability and reference values for postural stability in 9 to 10 years old schoolchildren using the Balance Master system.

Methods

Subjects

The reliability of the Balance Master system was evaluated in 4th and 5th graders from a randomly selected school. The parents of all 153 4th and 5th grade children were notified by a letter and asked for their child’s participation in the reliability assessment for postural stability using the Balance Master system. This invitation was accepted by 47 parents who signed the informed consent form for their child. Out of this group, 20 children aged 9–10 years were selected by simple randomization to participate in the reproducibility study for postural stability (10 boys and 10 girls, mean age 10.1±0.7).

In order to determine reference values for postural stability in elementary schoolchildren, a sample of 4th and 5th graders from 10 simply randomized selected schools was drawn. The parents of all 379 children were contacted to ask for their child’s participation in the reliability assessment of postural stability. A total of 99 parents signed the informed consent form. The study sample in order to determine reference values consisted of 99 children between 9 and 10 years old (41 boys and 58 girls, mean age 9.8±0.5).

Procedure

The reliability study for postural stability using the Balance Master system in 9 to 10 year olds included test and retest measurement with a one week interval. Both the test and retest measurements and the assessments for reference values were performed by the same researcher, according to the following standardized test-setting. Children were barefoot for all measurements. Before performing the balance tests, the children’s age and basic anthropometrical data were registered. Weight was assessed to the nearest 0.1 kg (Seca, max 200 kg). Height to the nearest 1 mm was measured using a wall-mounted stadiometer (Siber Hegner). The balance tests took place in a discrete room free from external distractions. Starting the assessment, the researcher positioned the children’s feet following the appropriate alignments on the force platform for the medial malleolus and the outside border of the heel. All children started with the assessment of static balance, which was followed by the dynamic balance test. For each condition of the static balance test and before the dynamic balance test, one training trial was allowed before data collection. A side-view and frontal positioned camera registered the children performing both the static and dynamic standing balance tests.

The study was approved by the Ethical Committee of the University Hospital of Ghent University.

Instruments

Neurocom - Balance Master system

The Basic Balance Master system (NeuroCom, Clackamas, OR, USA) was used to measure children’s postural stability with respect to standing balance. The Balance Master consisted of a portable force platform connected to a computer including a software program that calculated the centre of pressure relative to the platform coordinates. An estimation of the position and displacement of the centre of gravity (COG) was sampled at 100 Hz, based on a simple inverted pendulum approximation using the sampled centre of pressure data and the subjects’ body height.

Static standing balance

The modified clinical test of sensory interaction on balance (mCTSIB) quantified postural sway velocity of the children standing quietly on the force platform. This test consisted of four different sensory conditions including three consecutive trials lasting a duration of 10 s: (1) standing with eyes open on a firm surface, (2) standing with eyes closed on a firm surface, (3) standing with eyes open on a foam surface, (4) standing with eyes closed on a foam surface. The test sequence of the conditions was identical for all children.

Children were instructed to stand upright as steady as possible with the arms by their sides. In the conditions ‘eyes open’, the children were requested to keep the eyes open and look straight ahead. In the conditions ‘eyes closed’ they were blindfolded and asked to stand upright as steady as possible with eyes closed. The relative absence of sway was a measure for static stability (COG sway velocity).

Dynamic standing balance

The test for the limits of stability (LOS) quantified several movement characteristics associated with the subject‘s ability to voluntary sway towards various locations in space, and briefly maintain stability at those positions. The LOS test measured the child’s volitional (intentional) control of the COG. A limit of stability is the maximum distance a person can lean in a given direction (measured as angular distance from vertical) without losing balance, stepping, or reaching. The limits of stability were calculated individually, based on the children’s body height. Performing the dynamic standing balance task, the location of the child‘s COG was displayed on the computer screen as a cursor providing continuous visual feedback. Cursor control occurred by weight-shifting. The children had to move the cursor (their projection of their centre of gravity) as close as possible to eight targets (their limits of stability). The eight targets were arranged in an ellipse, separated by an angle of 45° (forward, forward-right, right, backward-right, backward, backward-left, left, forward-left). They started at the midline and held the cursor at the target as long as the target remained highlighted. After eight seconds, the cursor disappeared and the child returned to the midline. The same procedure was repeated clockwise for all the targets.

Therefore, the children were instructed to “move as quickly and accurately as possible” to each of the eight targets, without displacing the feet, bending the trunk or moving the arms. Children were instructed to move like a ‘piece of wood’, to emphasize a neutral hip position performing the LOS. When a child lost the correct posture, the test leader stopped the test. Accuracy was indicated by (1) whether or not the subject reached the target (maximal excursion), (2) whether the target was reached on the initial attempt (endpoint excursion), and (3) whether or not progress towards the target was smooth and consistent (directional control).

Data analysis

Outcome measures of standing balance assessment: the standing balance parameters

According to the mCTSIB, the COG sway velocity was calculated as a ratio of distance travelled by the COG (expressed in degrees) to the time of the trial (10 s). The mean COG sway velocity was the average of the COG sway velocity scores from the combined trials of any one condition; the sum of scores divided by the number of trials. The composite mean COG sway velocity was the average of the mean COG sway velocity scores for all conditions; the sum of the four means divided by the number of conditions.

The COG sway velocity scores indicated how well the subject accomplished the objective to stand as still as possible. Small scores reflected little movement, and are “good”. Large scores reflected more movement, and are “worse”.

The measured parameters of the LOS test were: reaction time (RT), movement velocity (sway), endpoint excursion (EXE), maximal endpoint excursion (m-EXE) and directional control (CD). Reaction time (RT) is the time in seconds between the signal to move and the initiation of movement. Movement velocity (MVL) is the average speed of COG movement, expressed in degrees per second, between 5% and 95% of the distance to the primary endpoint. Endpoint excursion (EPE) is the distance travelled by the COG on the primary attempt to reach the target, expressed in % LOS. Maximal endpoint excursion (MXE) is the furthest distance travelled by the COG during the trial. This may be larger than the endpoint excursion if the subject makes additional corrective attempts. Directional control is a comparison of the amount of movement in the intended direction (towards the target) to the amount of extraneous movement (away from the target). In addition to the composite score, the LOS scores from the eight transitions were combined to provide a separate average score for each of the four main directions (forward - backward - right - left).

Statistical analysis

Statistical analyses were conducted using SPSS 11.0 for Windows. The level of statistical significance was set at P<.05. Using single measure intraclass correlation coefficients (ICCs), trial-to-trial (inter-item) and test-retest (inter-session) reliability evaluation for the Balance Master was performed. The ICC values were interpreted according to the general guidelines of Fleiss [8], that is, ICCs>.75 were labelled ‘excellent’, >.40 ‘fair to good’, <.40 ‘poor’.

To determine reference values in 9 to 10 year olds (n=99), independent sample T-tests were executed in order to analyse gender differences for age and anthropometrics (height, weight and BMI). Furthermore, the study sample (n=99) was divided into a group of 9 year olds (n=52) and a group of 10 year olds (n=47). Using univariate analysis of variance, the standing balance parameters were separately analysed as dependent variables. In addition, gender and age were included as between-subjects factors and BMI as a covariate.

Results

Test-retest reliability

The ICCs representing intra-session reliability and inter-session stability for the use of the mCTSIB in the current study sample (n=20) are shown in Table 1. The ICCs for inter-item reliability of the 4 sensory conditions of the mCTSIB showed fair to excellent reliability (ICCs between 0.62 and 0.80). The reproducibility between test and retest was non-significant for the condition ‘firm surface with eyes closed’ (ICC of 0.37), fair to good for the three other sensory conditions (ICCs between 0.59 and 0.68), and excellent for the composite sway velocity (ICC of 0.77). The ICCs representing inter-session stability for the LOS are presented in Table 2. For all composite LOS parameters, the ICCs showed fair to good reproducibility (ICCs between 0.44 and 0.62), with the exception of the non-significant ICC for the composite reaction time (confidence interval: −0.4–0.71). The ICCs for the separate LOS parameters showed fair to good and excellent reliability for nine parameters (ICCs between 0.46 and 0.81), while 11 separate LOS scores did not demonstrate significant ICCs.

Table 1 ICCs for static standing balance (mCTSIB)
Full size table
Table 2 ICCs for dynamic standing balance (LOS)
Full size table

Reference values

The study sample to determine reference values showed no gender differences for age and anthropometrical values (age: t=.447, df=104, P=.656, height: t=.825, df=104, P=.411; weight: t=1.026, df=104, P=.307; BMI: t=1.028, df=104, P=.306).

The reference values for static and dynamic standing balance using the Neurocom Balance Master in boys and girls aged 9–10 are presented in Table 3. Girls performed better on all the composite balance parameters compared to boys, with the exception of reaction time and movement velocity according to the LOS (see Table 3). Girls performed better on all the composite balance parameters compared to boys, with the exception of reaction time and movement velocity according to the LOS (see Table 3). No gender differences were found for the separate dynamic standing balance parameters towards the four main directions, with the exception of a better performance on endpoint excursion backwards in girls. Further, no differences were found on standing balance scores between 9 and 10 year olds (for all parameters: F<2.978, df=1, P>.088). No gender by age interaction effects were found for the balance parameters of both tests revealing that the effect of sex was the same at each age (for all parameters: F<3.150, df=1, P>.081). Consequently, a distinction between 9 and 10 year olds was not made to present the reference values.

Table 3 Reference values for static and dynamic standing balance in girls and boys aged 9–10
Full size table

Measuring static standing balance, BMI was only a significant covariate for the foam conditions (foam eyes open: F=6.581, df=1, P=.012; foam eyes closed: F=11.091, df=1, P=.001) and the composite sway velocity (F=4.860, df=1, P=.030) of the mCTSIB. In the evaluation of dynamic standing balance, BMI was not a confounding factor.

Discussion

The aim of the present study was to examine the reliability of static and dynamic standing balance testing in 9 to 10 year olds and to report reference values in this young population.

Intra-session and test-retest reliability

The intra-session reliability measuring static standing balance in 9 to 10 year olds using the Balance Master demonstrated fair to excellent intra-session reliability for the four conditions of the mCTSIB. Further, mCTSIB stability parameters showed fair to excellent test-retest reliability for three sensory conditions and the composite sway velocity was the most reliable variable considering inter-session stability. However, the level of agreement between the two sessions at one-week interval showed poor reliability for the ‘firm surface eyes closed’ condition. Measuring inter-session reliability of dynamic standing balance, the stability was fair to good for all composite stability parameters, with the exception of the non-significant ‘reaction time’ parameters. Focalizing on children’s performance on the four main transitions of the LOS, none of the ‘reaction time’ parameters were significant. Further, ‘movement velocity’ and ‘maximal endpoint excursion’ parameters showed fair to good reliability for two main directions and excellent reliability for the backward transition. The ‘endpoint excursion’ parameters showed fair to good reliability for transitions backward and left whereas ‘directional control’ parameters demonstrated only for the forward transition fair to good reliability. The present reproducibility data established more significant and in general better inter-session test-retest reliability values for the composite scores compared to the separate scores, both for the static and dynamic standing balance parameters. The study of Lafond et al. [11] pointed out that the reliability of COP measures increased by increasing the trial duration. Along the same lines, based on our study findings, a sufficient number of trials may be an important factor in order to ascertain reliable postural stability assessments.

Like many biological measurements, postural stability has an intrinsic variability influenced by physical, biomechanical, metabolic and psychosocial factors [11]. Consequently, many factors affect the reproducibility of postural outcomes, such as motivation, concentration, fatigue, emotional state, time of the test and relationship with the tester. Therefore, in the current study measurement order, testing sequence, tester and surrounding factors were identical during the two sessions in order to minimize variations between test and retest measurements. So, the modest reproducibility in the current study may be attributed to the inherent variability in children’s balance performances and not to the test protocol. The variable balance performances may be supported by the higher reliability scores for the evaluation of children’s stability recordings at one moment (intra-test reliability) when compared to the evaluation of stability scores at two moments (inter-test reliability). On the other hand, the Neurocom Balance Master’s protocol measuring stability has to deal with the complexity to guarantee the exact same foot positioning on the force plate. Despite the strict prescriptions for foot placement, the chance for possible variation in positioning the feet seems to be a potential danger affecting reproducibility. Finally, the non-significant reliability values of the LOS parameters regarding the separate scores for the main directions may be caused by the calculation of the COG path by the inverted pendulum approximation. Based on the qualitative observation of children’s videotaped stability performances and the standardized test protocol demanding a neutral hip position one may assume that a hip strategy was not used. However, small hip flexion and extension movements may have occurred since minor hip strategy movements are not detectable by eye. Accordingly, a limited part of variation in reliability might also be due to the inverted pendulum principle which doesn’t take into account the possible use of a hip strategy.

Taking the latter into account, it can be concluded that the one-week reproducibility for standing balance assessments using the Balance Master in 9 to 10 year olds pointed out fair to good and even excellent reliability for most parameters, which is in accordance with a study in adults aged 20 to 32 years [4]. However, when interpreting the ‘firm surface with eyes closed’ sensory condition of the mCTSIB and some separate sway parameters for the main directions of the LOS in 9 to 10 year olds, caution is recommended.

Reference values

The current study found significantly lower sway velocities in girls compared to boys for the composite and separate sway parameters of static standing balance, indicating a better postural control in girls at age 9 to 10 years. Analysing the composite scores of dynamic standing balance in 9 to 10 year olds, girls moved at the same velocity towards the targets, but in comparison with boys their initial attempts were better and they hold their COG closer to their LOS in a more consistent progression. None of the separate stability scores for the four transitions towards the main directions of the LOS established significant differences between boys and girls, with the exception of the ‘endpoint excursion’ backwards. The absent gender effect on the separate stability parameters of the LOS could possibly be explained by the lack of reliable measurements for these parameters. The effect of gender on the separate ‘endpoint excursion’ score of the backward direction in relation to the excellent reliability value for this parameter may support this hypothesis. Otherwise, the composite scores are an evaluation of eight trials in comparison with the evaluation of three trials determining the separate stability scores for the main directions. Based on the presented reference data, the composite scores of the LOS seem to include a summation of slightly different separate stability scores between boys and girls. This may explain the significant difference between boys and girls on the total composite stability scores of the LOS.

Further, the current study showed no significant age by gender interactions for postural parameters, which corresponded to the developmental study findings of Figura et al. [7] assessing the static balance in children aged 6 to 10 years old. In this line, the current study findings showed that being 9 or 10 years old did not influence the performance on standing balance measurements. One could suppose that the difference in children’s anthropometrics at the same chronological age might confound the possible effect of age. However, in the current study BMI was only a significant covariate for the measurements on a compliant surface and for composite sway velocity and not for any other parameter. In this line, the reliance of BMI on the balance assessments on a soft surface seems reasonable since BMI directly determines the extent of foam compression. Accordingly, different body weights vary the challenge of the balance task provided by the foam surface.

The present study’s findings about gender differences in standing balance assessment in 9–10 year old children corresponded to the study results of Nolan et al. [15], who examined sex and age differences for postural control in 9 to 10 year olds. This development-orientated investigation for standing posture in children supported our finding with regard to the independency of anthropometrics. Nolan et al. [15] demonstrated that a relation between anthropometric measurements and balance parameters was non-existent, whereas height and weight differed with age. The authors suggested that while postural control may be partly affected by changes in stature as children grow, the development of visual, vestibular, and somatosensory systems may account for age-related changes in balance control to a greater extent. Accordingly, one might presume that boys at age 9–10 years use another control strategy compared to girls of this age group.

Coming to a conclusion, the portable Neurocom Balance Master system is easy to transport and, as such, is extremely useful in the school setting. In the present study, the elementary schoolchildren were volunteers out of a group of randomly selected elementary schools and they did not have any specific background in balance testing or training. Additionally, the present study showed, in general, fair to good reliability for postural measurements using the Balance Master system. Therefore, the postural parameters presented in the current study can be generalized to other elementary schoolchildren aged 9 to 10 years old. However, force plate measurements using the portable Balance Master system in children should be further investigated, as well as the strategies for improving reliability such as by increasing the numbers or durations of the trials. At any rate, the current data on postural control in 9 to 10 year olds are relevant for research in other domains within the clinical field, like obesitas and developmental coordination disorder or in relation to back pain prevalence at early age.