Manual and Visual Clinical Assessment of Foot Type and Posture Correlate Poorly with Radiographic Assessment

Abstract

Category: Midfoot/Forefoot; Hindfoot Introduction/Purpose: Accurate characterization of foot type can facilitate identification of patients at risk for different foot disorders and direct operative treatment. Foot type can be determined qualitatively with visual assessment and quantitatively using arch index (AI), calcaneal pitch (CP), Meary's angle (MA), and foot posture index (FPI). Agreement between various foot typing measures is unclear. The purpose of this study was to assess the agreement between different measurements of foot type. We hypothesized that quantitative measures would have better interrater reliability than qualitative measures, and that visual assessment of foot type would be strongly correlated with radiographic assessment. Methods: Bilateral foot photographs were acquired from 92 asymptomatic participants (50 men, 42 women, mean age of 43.3+- 18.6 years) to assess photographic arch type (PAT) and photographic foot posture (PFP). Hindfoot alignment and lateral radiographs were acquired of each foot to assess calcaneal pitch (CP), Meary's angle (MA), x-ray foot posture (XFP), and x-ray arch type (XAT). Photographs and radiographs were assessed by five reviewers in random order. Foot Posture Index (FPI) was determined through manual exam. Arch Index (AI) was measured on lightbox images. Interrater reliability was assessed using Intraclass Correlation Coefficient (ICC) for continuous variables and Kendall's W for categorical variables. Correlations between measures and sensitivity/specificity were assessed using calcaneal pitch as the reference standard with cavus feet >= 26 degrees and planus feet measuring <= 19 degrees. Results: Interrater reliability of XFP (W=0.522), XAT (W=0.676), PAT (W=0.831), PFP (W=0.612) ranged from strong to extremely strong. Interrater reliability of MA (ICC=0.711) and CP (ICC=0.804) were moderate and good. Correlations between foot arch measurements ranged from low to high (Table 1). Fair correlations were identified between foot posture measurements (r=0.356 to r=0.442). Overall, XAT had the greatest sensitivity and specificity for planus (0.68, 0.93) and normal feet (0.92, 0.62), and the greatest specificity for cavus feet (0.98). AI demonstrated the greatest sensitivity (0.47) for cavus feet. Compared to AI, photographs had superior sensitivity for identifying normal feet (0.73 vs. 0.43), the greatest sensitivity for identifying planus feet (0.54 vs. 0.06), and the greatest specificity for identifying cavus feet (0.95 vs. 0.71). Conclusion: Agreement between methods varied widely, suggesting that classification is dependent upon methodology and cutoff values selected. PAT had the highest interrater reliability, whereas correlation between XAT and PAT was moderate, suggesting that radiographs are needed to accurately assess the longitudinal arch.