Development of a reliable method to assess footwear comfort during running

Abstract

The purposes of this study were: (a) to determine whether subjects are able to distinguish between differences in footwear with respect to footwear comfort; and (b) to determine how reliably footwear comfort can be assessed using a visual analogue scale (VAS) and a protocol including a control condition during running. Intraclass correlation coefficients (ICCs) between comfort ratings for repeated conditions were high (ICC=0.799). Differences in comfort ratings between the insert conditions were significant. A paired t-test revealed a significant difference in overall comfort ratings for the control insert when tested after the soft insert compared to when tested after the hard insert (P=0.008). The results of this study showed that VASs provide a reliable measure to assess footwear comfort during running under the conditions that: (a) a control condition is included; and (b) the average comfort rating of sessions 4–6 is used.

Introduction

Comfort is an important factor for footwear in recreational physical activities. Most people can quickly identify comfortable or non-comfortable footwear situations. Increasing interest in footwear comfort resulted in several investigations that associated comfort with plantar pressure distribution [1], [2], vertical impact force, rearfoot motion [3], foot and leg shape and alignment [4], and foot sensitivity [5]. It has been speculated that comfort is related to muscle activation and, thus, to fatigue and performance [6], [7]. The specific design, physical properties and construction of footwear have been shown to affect these variables [8], [9] and, thus, seem to be important factors for footwear comfort.

However, comfort is difficult to quantify and it has been suggested that comfort cannot be measured directly [10], [11]. Despite the reported measurements of comfort [1], [2], [3], [4], [5], to date a reliable method to assess footwear comfort has not been developed. All studies investigating comfort in the past had one common shortcoming: The reliability of the comfort measures used was not reported or, if reported, poor. In fact, only two studies reported on the repeatability of the comfort measures used. Miller et al. [4] and Mündermann et al. [5] found rather low intraclass correlation coefficients (ICCs) of comfort ratings for repeated conditions. These results can be in part explained by the following considerations.

Individuals have some feeling regarding the absolute liking of an object, but such sentiments are influenced by the range of objects available. In other words, when expressing sentiments, subjects tend to anchor their responses in terms of: (1) similar stimuli that they have experienced in the past; and (2) the range of stimuli in the set presented [12]. Thus, if someone is required to rate the comfort of a shoe they cannot make such response unless they subjectively compare the features of the shoe within those of the many shoes they have worn before. Therefore, it is speculated that the reliability of a comfort measure can be considerably improved by adding a control condition to the testing protocol such that all conditions to be assessed are tested after and, thus, compared to the same footwear condition.

Since the repeatability of the measures in earlier studies was poor, it must be questioned whether the results of these studies provide sufficient evidence for the existence of a relationship between footwear comfort and biomechanical variable and subject characteristics. A reliable measure for footwear comfort is required to quantify comfort and to determine valid relationships between comfort and shoe constructions, subject characteristics, and biomechanical variables.

Most investigators used ordinal scales to assess footwear comfort [1], [2], [3], [4]. In ordinal scales or ranking scales, objects are ordered from ‘least’ to ‘most’ with respect to an attribute, e.g. comfort. However, in such scales, there is no indication of how much, in an absolute sense, any of the objects possesses the attribute [12]. Furthermore, no information can be obtained as to how different the conditions are with respect to comfort.

One particular kind of ordinal scale is where subjects rate various objects on a 7- or 15-step scale ranging from ‘least’ to ‘most’. For example, the Borg-scale was initially developed to assess perceived exertion [13] and has been used to assess footwear comfort [2], [3], [4]. In general, this scale allows the determination of relative differences between conditions. However, due to the discrete spacing of ratings, very small differences between conditions cannot be detected. Furthermore, non-parametric statistical methods must be used to evaluate ordinal data. Because the sets of possible answers consist of discrete numbers, measuring comfort using an ordinal scale will introduce errors to correlations between comfort and other variables such as biomechanical variables that are measured on continuous scales.

It is proposed that footwear comfort should be measured using a continuous scale. Visual analogue scales (VASs), for example, have been proven to be a reliable measure to assess subjective pain [14], [15]. VAS responses are very easy to obtain from patients and normal volunteers and require little instruction. However, not all pain VASs are bias-free [16].

Studies comparing the use of different types of VASs have shown that the sensitivity and reliability of VASs are somewhat influenced by the words used to anchor the endpoints, by the length of the VAS, and by other factors [17], [18]. Those VASs that most clearly delineate extremes (e.g. the best condition imaginable, the worst condition) and are 100–150 mm in length have been shown to have the greatest sensitivity and are the least vulnerable to distortions or biases in ratings [18]. It has been suggested that giving specific instructions to subjects will increase the reliability of VASs. For example, Price et al. [14] found an inter-test correlation coefficient of 0.970 for a 150 mm VAS and specific instructions.

Therefore, the purposes of this study were: (a) to determine whether subjects are able to distinguish between differences in footwear with respect to footwear comfort; and (b) to determine how reliable footwear comfort can be assessed using a VAS and a protocol including a control condition during running.