Results All the participants were able to
  • 2. Methods
    Participants were recruited from the student body at Victoria University (Melbourne, image Australia). A total of 83 participants (n = 41 female) were involved in the study. No data that R428 could be used to identify the participants was recorded. All 83 participants were assessed on the same day. The study was approved by the Victoria University Human Research Ethics Committee.
    Participants were required to walk approximately 10 m to the FPI-6 assessment station and stand on a 30 cm high platform whilst two examiners worked through the FPI-6 checklist (Redmond et al., 2006). The examiners assessed the participant at the same time to minimise participant movement and were able to refer to the FPI-6 manual (Redmond, 2005) during the assessment. Each examiner completed a separate FPI-6 sheet per participant and passed these sheets onto a research assistant who coordinated data entry. The only other detail recorded on the sheet was the participant\'s number and gender. One of the authors (BV) oversaw the FPI-6 assessment process to ensure there was no communication between the examiners.
    The two examiners were graduating students, 6 months from completion of their osteopathy program at Victoria University. During their studies these students had extensive training in musculoskeletal assessment but had not previously been exposed to the FPI-6. The two examiners underwent a 1 h training session (Cornwall et al., 2008) with an osteopath experienced in the use of the FPI-6 one week prior to data collection.
    Data were analysed for each foot as raw ordinal scores, Rasch-converted total score (Keenan et al., 2007), and as categorical data according to the FPI-6 total score classification system (Redmond et al., 2008). In the present study an attempt was made to reduce data to analyse the left or right only (Morrison and Ferrari, 2009), but there were significant differences (p < 0.05) between sides for both examiners.
    Inter-examiner reliability was assessed using the intra-class correlation coefficient (ICC2,1) for the Rasch-converted total FPI-6-score (Keenan et al., 2007), weighted Kappa (Kw) for the raw scores, and generalised Kappa for the foot type classification data. Kw was also used to evaluate the examiner agreement between each of the individual criteria on the FPI-6. All coefficients were calculated using the psych package ( Revelle, 2015) in R ( R Core Team, 2015). The ICC and Kw were interpreted according to Landis and Koch (1977): 0 to 0.20 ‘slight’, 0.21 to 0.4 ‘fair’, 0.41 to 0.60 ‘moderate’, 0.61 to 0.8 ‘substantial’, and 0.81 or greater ‘almost perfect’. Three reliability estimates were calculated to evaluate the internal structure of the FPI-6: Cronbach\'s alpha (α) and McDonald\'s omega total (ωt) and hierarchal (ωh) (Zinbarg et al., 2005). McDonald\'s ωh values over 0.7 provide support for the calculation of a total score (Hecimovich et al., 2014).
    3. Results
    Descriptive data and reliability estimates for each examiner are presented in Table 1. The reliability estimates in the present study are acceptable (α & ωt > 0.80) for both examiners and the calculation of a total score for the FPI-6-6 is epiglottis valid (ωh > 0.70). The ωh value obtained in the present study indicates over 70% of the reliable variance in the total FPI-6 score is due to the underlying latent construct (static foot posture). The inter-examiner reliability for the Rasch-converted total FPI-6 score indicates an almost perfect level of agreement (left ICC2,1 = 0.86; 0.80–0.91; right ICC2,1 = 0.85; 0.78–0.90).

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!