Visual categorisation of the arch index: a simplified measure of foot posture in older people

Abstract

Background: Foot posture is considered to be an important component of musculoskeletal assessment in clinical practice and research. However, many measurement approaches are not suitable for routine use as they are time-consuming or require specialised equipment and/or clinical expertise. The objective of this study was therefore to develop and evaluate a simple visual tool for foot posture assessment based on the Arch Index (AI) that could be used in clinical and research settings.

Methods: Fully weightbearing footprints from 602 people aged 62 to 96 years were obtained using a carbon paper imprint material, and cut-off AI scores dividing participants into three categories (high, normal and low) were determined using the central limit theorem (i.e. normal = +/- 1 standard deviation from the mean). A visual tool was then created using representative examples for the boundaries of each category. Two examiners were then asked to use the tool to independently grade the footprints of 60 participants (20 for each of the three categories, randomly presented), and then repeat the process two weeks later. Inter- and intra-tester reliability was determined using Spearman's rho, percentage agreement and weighted kappa statistics. The validity of the examiner's assessments was evaluated by comparing their categorisations to the actual AI score using Spearman's rho and analysis of variance (ANOVA), and to the actual AI category using percentage agreement, Spearman's rho and weighted kappa.

Results: Inter- and intra-tester reliability of the examiners was almost perfect (percentage agreement = 93 to 97%; Spearman's rho = 0.91 to 0.95, and weighted kappas = 0.85 to 0.93). Examiner's scores were strongly correlated with actual AI values (Spearman's rho = 0.91 to 0.94 and significant differences between all categories with ANOVA; p < 0.001) and AI categories (percentage agreement = 95 to 98%; Spearman's rho = 0.89 to 0.94, and weighted kappas = 0.87 to 0.94). There was a slight tendency for examiners to categorise participants as having higher arches than their AI scores indicated.

Conclusions: Foot posture can be quickly and reliably categorised as high, normal or low in older people using a simplified visual categorisation tool based on the AI.

Figure 1

Footprint obtained using carbon paper imprint material.

Figure 2

Calculation of the AI. The length of the footprint excluding the toes (L) is divided into equal thirds. The AI is then calculated as the area of the middle third of the footprint divided by the entire footprint area (AI = B/A + B + C).

Figure 3

Histogram of AI scores obtained from 602 people aged 62 to 96 years. Cut-off scores defining high and low-ached feet based on ± 1 SD are indicated.

Figure 4

The visual AI categorisation tool.

Figure 5

Mean (± SD) AI scores obtained by the computer graphics tablet according to each examiners’ categorisations in session 1 and 2.