Reliability of a two-dimensional video analysis protocol to assess forward head posture during walking

OBJECTIVE To develop a protocol to investigate the dynamic forward head posture during walking, using 2D video analysis, and to examine its reliability. METHOD Thirty-six healthy volunteers were recruited (15 men and 21 women; age 23.6 +/- 7.65 years). A video camera of a mobile phone was used for recording the forward head posture of the participants during walking in two sessions. Three recordings were obtained at each one of the two sessions. The second session was performed 8 days after the first session. Videos were analyzed by three raters with a different level of experience. Forward head posture was described based on craniovertebral angle. RESULTS Intra-rater (intraclass correlation coefficient [ICC]: 0.90-0.92, standard error of measurement [SEM]: 1.44-1.84), inter-rater (ICC: 0.98-0.99, SEM: 0.79-1.08) and test-retest reliability (ICC: 0.95-0.98, SEM: 1.01-1.52) were found to be excellent for the three raters. The test-retest bias (rater A:-0.3 +/- 1.97, rater B:-0.74 +/- 2.14, rater C:-0.31 +/- 1.43) and the inter-rater bias between the two more experienced raters (0.25 +/- 1.53) were very low. CONCLUSIONS The assessment of dynamic forward head posture by using 2D video-based motion analysis is a feasible and reliable method with small measurement error. A minimal training of raters is sufficient for providing reliable estimates of dynamic forward head posture.