Early identification of toe walking gait in preschool children - Development and application of a quasi-automated video screening procedure

PURPOSE: To develop and test the application of a quasi-automated screening procedure identifying probable toe walking in a large population of preschool children.

METHODS: The proposed screening procedure was designed to identify children exhibiting signs of toe walking in a previously recruited cohort of preschool children (MiPS cohort). The procedure combines parent observation (step 1), objective parameters of foot contact during gait by an automated screening of 3-D video recordings (step 2), and clinical video screening of the children identified in step 1 and/or 2 (step 3).

FINDINGS: From 879 children, gait trials were obtained from 87% (n = 766). Step 1 (parent observation) identified 34 children with potential toe walking, step 2 (automated screening) 122. Fourteen were identified in both step 1 and 2. Thus, 142 children were selected for step 3 (clinical video screening), from which 41 children were classified as showing symmetric signs of toe walking, and five children were identified with asymmetrical signs of toe walking. Of the 41, five had been identified by step 1 only, 32 by step 2 only and four by both steps.

INTERPRETATION: Application of a quasi-automated screening algorithm was feasible and may assist in early detection of toe walking. Disagreements found between parent reported toe walking and video screening, indicate added value in quasi-automated video screening. However, thresholds of heel lift and clinical criteria of toe walking in the algorithm and video screening need to be addressed and validated to confidently identify toe walking gait.