Repeatability of the Oxford Foot Model: Comparison of a team of assessors with different backgrounds and no prior experience of the Oxford Foot Model.

Research Question: What is the intra- and inter-assessor error of the Oxford Foot Model (OFM) during healthy adult walking when applied by three assessors with different professional backgrounds and lower limb marker placement experience, not native to the originators of the model and with no prior clinical experience of the model?

Background: No previous OFM studies have examined the repeatability of more than two assessors with different backgrounds, and many of the studies have been conducted by the model originators METHODS: The OFM was applied to ten healthy adults on three separate occasions by three different assessors with varied professional experience and no prior involvement with the OFM (other than local training). Participants walked at self-selected speeds and intra/inter assessor error was calculated using the SEM + 95% upper confidence limit.

Results: Inter-assessor errors ranged from 2.2° to 5.5° whereas intra-assessor errors fell between 1.8° and 5.5°. The error difference between assessors over the same joint angle varied from 0.4° (hindfoot/tibia dorsiflexion) to 1.5° (hindfoot/tibia inversion). The percentage of error to total range of motion varied from 11% (hindfoot/tibia dorsiflexion) to 126% (forefoot/hindfoot adduction).

Significance: Based on commonly used recommendations, the OFM is a largely repeatable tool for measuring foot kinematics during healthy adult walking when applied by assessors with no prior OFM experience, varied experience and not native to the model originators. Intra-assessor error was lower for assessors with prior anatomical knowledge and significant lower limb marker placement experience. The proportion of inter-assessor error to movement exceeded 50% of the total range of motion for four movements, notably forefoot/hindfoot adduction (126%). As such, this movement cannot be recommended as an outcome measure. Inter- and intra-assessor error, specific to each laboratory, should be considered, along with the proportion of error to range of motion when interpreting patient data.