Scripties UMCG - Rijksuniversiteit Groningen
 
English | Nederlands

Structural Validity of the Dutch Version of the Lower Extremity Functional Scale (LEFS-NL) in Patients with Lower Extremity Injuries

(2017) Zwierstra, M. (Myrthe)

Structural Validity of the Dutch Version of the Lower Extremity Functional Scale (LEFS-NL) in Patients with Lower Extremity Injuries
Zwierstra MJ, El Moumni M.
BACKGROUND: The Lower Extremity Functional Scale (LEFS) is developed to assess lower-limb function in a wide variety of patients with lower-extremity orthopedic conditions. Although the assumption is made that the LEFS measures a unidimensional trait, there is a lack of literature supporting this hypothesis.
OBJECTIVE: To evaluate the structural validity of the Dutch version of the LEFS (LEFS-NL).
DESIGN: This was a retrospective analysis of cross-sectional data of 696 patients.
METHODS: Patients aged 16 to 60 with an isolated lower extremity injury sustained in 2014 - 2016 were recruited. Patients treated either conservatively or surgically were included. Patients were excluded if they had multiple injuries or multiple isolated injuries within the researched time frame. They were also excluded if they were unable to speak or read Dutch. The structural validity was analyzed using confirmatory factor analysis (CFA). The internal consistency was examined using Cronbach alpha (α) and omega (ω).
RESULTS: A single-factor model, correlated 2- factor model and a bifactor model all resulted in adequate model fit. Factor loadings in both the single factor model and correlated 2- factor model were high (≥0.72) and statistically significant (p<0.001). The covariance between the two factors in the correlated 2- factor model was remarkably high (0.965, p<0.001). All items in the bifactor model loaded high (at least 0.73) on the general factor, but much lower on the group factors indicating that the contribution of the group factors beyond the general factor is limited.
The reliability estimates of the single factor and the correlated 2- factors were high (>0.96). With respect to the general factor in the bifactor model, coefficient omega hierarchical was high (0.98), but considerably lower for each subscale (<0.09), indicating that it is not reasonable to report subscales.
LIMITATIONS: Patients were predominantly treated conservatively (n = 515) for an ankle distortion (n = 191). In addition, the response rate low (32.4%).
CONCLUSION: The LEFS-NL measures a unidimensional trait. It should hence be reported as a single score.






 
To top