Psychometric properties of the Swedish version of the Falls Efficacy Scale-International for older adults with osteoporosis, self-reported balance deficits and fear of falling.

PURPOSE

Investigate the psychometric properties of the Swedish version of the Falls Efficacy Scale-International (FES-I).

METHOD

Cross-sectional study. Community-dwelling older adults with self-reported balance deficits and fear of falling were recruited from an ongoing randomised controlled study to evaluate the psychometric properties of the FES-I using Rasch model analysis.

RESULTS

The Rasch model analysis revealed good category function, the questionnaire measured one dimension with an explained variance of 68.6% and item goodness-of-fit with mean square values (MnSq) 0.7-1.44. The item map showed that all items are spread over the scale, which indicates different difficulties in the items. Non-satisfactory person goodness-of-fit was shown with seven persons and showed person misfit according to both the MnSq-value and the z-value, 38 persons (40%) showed a person misfit when only following the threshold for MnSq.

CONCLUSIONS

The Swedish version of FES-I shows good psychometric properties with unidimensionality and item goodness-of-fit. Lower person goodness-of-fit was shown probably because of confounding factors that may influence the answers. The transformed values of the FES-I make it possible to use parametric statistics preferable for this population in future research. Implications for rehabilitation The Falls Efficacy Scale-International (FES-I) shows good psychometric properties with unidimensionality, item goodness-of-fit and good item reliability, which means that FES-I is a valuable tool when measuring concerns about falling in an older population with osteoporosis and could be useful in clinical settings. Confounding factors such as pain, high number of falls, low fall self-efficacy, experience of previous falls, and vertigo may influence the answers and result in low person goodness-of-fit.