Validation of the Dermatology Life Quality Index. A Structural Validity, Internal Consistency and Item Response Analysis.

Abstract

Abstract Background: Although the classical test theory (CTT) has been one of the most widely used approaches for QoL scales validation in dermatology, some of its assumptions are problematic as it is sample dependent and does not take into account that scoring can be affected by patients´ variables and the measurement process. Previous results of the Dermatology-Life-Quality-Index validation in Colombia (DLQI-COL) based on the CTT perspective have showed the need to deep into the measurement properties of the instrument through the item response theory (IRT) or the Rasch model. Methods: In this cross-sectional study we aimed to: 1- Assess the dimensionality of the DLQI, to determine its difficulty, discrimination and differential functioning, to evaluate its internal consistency and to assess its discriminative validity between patients with inflammatory and non-inflammatory skin disease. We recruited Colombian patients with any skin disease. Skin diseases were stratified as inflammatory and non-inflammatory. Floor and ceiling effects were examined in addition to the internal-consistency of the instrument. For Rash analysis, a partial credit model was applied. The following model assumptions and measurement properties were tested: dimensionality and differential item functioning (DIF). Results: Data of 461 patients (mean age: 46.8 years) and 56% females, were included. DLQI-COL-scores ranged from 0-29 (median=5). Internal consistency for dichotomous item 7, showed a McDonald’s Omega coefficient of 0.85 (95%CI:0.84-0.87) and a Cronbach's αcoefficient of 0.86 (95%CI:0.84-0.88). For ordinal item 7, McDonald's Omega coefficient was 0.87 (95%IC:0.85-0.89) and Cronbach's αcoefficient of 0.85 (95%CI: 0.82-0.87). Horn’s parallel analysis indicated DLQI unidimensionality with a one-factor solution. Only item 1 fitted the PCM model. The remaining items, including item 9, did not show adequate fit. Differences in scores were found between patients with non-inflammatory and inflammatory skin diseases and between age with no differences between gender. Conclusions: The validity of DLQI-COL based on Rasch analysis proved not to be a suitable instrument to evaluate QoL in our dermatologic population. A refined version of the scale with an adequate cross-cultural adaptation and validation process are needed to obtain an appropriate and reliable version of the instrument for measuring QOL in Colombian patients with skin diseases.