The reliability of a measurement system used clinically, or in research, must be established in order to be confident in achieving reproducible and meaningful results on different testing occasions. In the current study, the system showed good to excellent between-session reliability in assessing barefoot postural control in double-limb quiet standing, in a sample of older people with RA, as evidenced by ICCs ranging from 0.84 to 0.92. However, measurement error, as expressed by the SEM, SEM% and SRD was relatively high compared to mean values.
The reproducibility of the measures may be attributed to the accuracy of the TekScan MatScan® system in capturing the variables of interest. Indeed the system was found to be highly accurate in an independent study which compared several commonly used plantar pressure measurement systems . Further, due to postural sway values being captured by the measuring system and not the examiner, rater error and bias which may be present in non-computerized tools, such as the sway-meter , was minimized.
Measurement error can be due to the precision of the instrument, systematic error introduced by the rater, or the variation in the population being measured . In the current study, measurement error may have occurred as a result of the inherent variability of postural control parameters in the study sample. The wide range in recorded sway values, resulting in a large SD of the mean, supports this possibility. Variation in postural stability parameters within an RA population is to be expected and may be associated with the differing demographic and clinical characteristics of the sample. For example, in a study of 61 patients with RA, Ekdahl  found that age, sex and high C-reactive protein level were related to decreased postural control in quiet standing. Given the relatively broad inclusion criteria in the current study it would be expected that this population would display a broad range of demographic and clinical characteristics and therefore a potentially wide range of postural sway values. Therefore, in the current study, the relatively high SEM, SEM% and SRD values may be indicative of the variability of the population tested rather than the reliability of the equipment used to test the population.
It can be further argued that SEM and SRD values are of more relevance in the analysis of within-subject variability. Indeed in a clinical setting measures of postural stability would be undertaken on individuals not populations and the ability to detect a real change in the variables measured over time is essential. As the study aim was to assess between-session reliability over 1 hour, within-subject variability was not analyzed however it is acknowledged that such analysis would be valuable in determining the potential measurement error of the system in a clinical setting.
Postural stability is controlled by the central nervous system. Afferent input from the somatosensory, visual and vestibular systems combine with coordinated muscle activity to maintain balance in quiet standing. In healthy adults, postural control is maintained through flexible and smooth interaction between these systems in order to maintain a stable equilibrium . This may not be the case in an RA population as the ability to maintain balance in quiet standing has been shown to be decreased compared to healthy controls . Variability in postural control between participants was found in the current study as demonstrated by the relatively high degree of measurement error. For this reason, it was necessary to assess the reliability of the TekScan MatScan® system in measuring postural control in an RA group specifically, as this is a population of particular interest to the researchers.
Postural control is a dynamic phenomenon that changes over time. As such, variability in COP values can be expected within individuals and should be accounted for through repetition of test measurements to obtain a mean value . In the current study, the mean of three test measurements of 30 seconds was taken. This was in accordance with a previous study which found that three measurements were sufficient for obtaining a consistent average for dynamic plantar pressure measurements in patients with foot problems associated with chronic arthritis . Due to the variability of COP variables, comparison of individual measurements, i.e. within-session reliability, was not undertaken in the current study.
The role of vision in postural control is well documented and is of particular importance in older adults . Previous studies have demonstrated that removing visual feedback increases postural sway compared to EO test conditions [32, 37]. Whilst the current study was interested in adults with RA aged 18 years and older, the cohort age range was 60 to 80 years and hence can be defined as older adult. Our results showed an increase in AP and ML postural sway with the eyes closed condition compared to eyes open condition, which is in agreement with previous studies [24, 32, 37, 38]. Further, when assessing AP and ML sway in an RA population compared with healthy controls, Rome et al.  found that, while both groups demonstrated greater sway in eye closed conditions compared to eyes open, the effect was more marked in the RA group. It is important therefore that postural stability in an RA population is assessed with eyes open and eyes closed and hence eyes open and eyes closed test conditions were used to assess the reliability of the equipment in the current study.
The TekScan MatScan® is a portable pressure system commonly used in research and clinical settings to capture and reproduce plantar pressure measures of dynamic foot function. The reliability of the system to accurately and consistently capture dynamic measures has been previously shown [17, 18]. The results of the current study suggest that the TekScan MatScan® is reliable for assessing postural control in double-limb quiet standing in older adults with RA. Research implications include the ability to gain a better understanding of the changes in postural stability that occur with age  and diseases that affect the feet, such as diabetes and RA. Clinical implications include the ability to identify and manage, through podiatric intervention, patients who are at increased risk of falling. The system may also be useful in evaluating the efficacy of clinical interventions, such as pathological callus debridement, foot orthoses and therapeutic footwear, in reducing postural sway in RA patients.
We acknowledge the limitations of this study. The study cohort was not representative of the general RA population, as all participants were over the age of 60 years. RA affects women three times more than men and peak age at onset is most commonly the fifth decade, although a shift towards older age at onset has been seen in recent studies . A sampling frequency of 40 Hz is relatively low, compared to laboratory based force plate technology, however we believe it is acceptable for measuring the postural sway parameters of interest. Inflammatory disease activity was not assessed as part of the study protocol, and therefore it is not possible to assess the impact of disease activity on variability of the instrument. The study does not address the validity of the TekScan MatScan® in assessing postural control in quiet standing. The validity of a measurement tool can be described as its ability to measure what it is supposed to measure . The validity of the TekScan MatScan® system has been reported by the manufacturer  however independent assessment comparing the TekScan MatScan® to force-platform technology would be valuable.
The reliability of the TekScan MatScan® for assessing double-limb quiet standing in healthy adults would be useful. Future investigations should also explore the reliability of the system during more complex dynamic balance tests, in people with RA, as well as other populations of interest such as patients with diabetes or older adults with a history of falls. Testing of the reliability of postural control measures in participants’ usual footwear will also be of interest.