Further Validation of the Dermatology Life Quality Index (DLQI) Using a 13 European Country Dataset

Speaker(s)

Salek SS1, Johns J2, Ali FM3, Finlay AY2, Dalgard F4, Kupfer J5
1University of Hertfordshire, Hatfield, UK, 2Cardiff University, Cardiff, CRF, UK, 3Cardiff University, Cardiff, UK, 4National Center for Dual Diagnosis, Brumundal, Norway, 5Institute of Medical Psychology, Justus Liebig University, Giessen, Germany

Presentation Documents

OBJECTIVES: To further validate the DLQI using the “Psychological Burden of Skin Diseases” European Society of Dermatology and Psychiatry dataset.

METHODS: Data from an European multicentre observational cross-sectional study conducted in 13 countries were analysed. In each dermatology clinic, 250 consecutive adult out-patients were recruited.

RESULTS: From 3,635 patients, 3408 patients completed the DLQI questionnaire with no missing data. 55.8% of patients were female and mean age was 46.6 years (SD 17.82). The commonest conditions were: psoriasis (17.4%), non-melanoma skin cancer (10.9%), pyoderma gangrenosum (9.5%), recurrent herpes simplex (6.7%), eczema (6.2%), acne (6.2%), nevi (5.0%), atopic dermatitis (4.5%), epidermal cyst (4.2%), eczema (contact dermatitis) (4.1%) and leg ulcers (2.8%). Using the DLQI score meaning banding, their disease had no effect (n=962, 28.2%), small effect (912, 26.8%), moderate effect (674, 19.8%), very large effect (691, 20.3%) and extremely large effect (169, 5.0%) on their quality of life. Psychometric analyses were conducted and Eigenvalues, screen plots, item response theory (IRT) and Very Simple Structure (VSS) indicated unidimensionality of the measure. Cumulative explained common variance did not significantly increase from 0.479 with addition of further factors. Confirmatory factor analysis fit statistics for the 1-factor solution were acceptable, and BIC was optimum for 1-factor model. IRT indicated good fit statistics, no misfitting items (infit and outfit), and no local dependence (LG2 and Q3). Reliability by Cronbach’s alpha was 0.900. Cohen’s d effect sizes for total DLQI score between different EQ-5D item levels were moderate (>0.5) or large (>0.8), and large between different levels of disease severity. Good correlation was found between EQ-5D VAS scale and DLQI total score (Spearman’s r =-0.409, p<0.001). Known group analysis of DLQI total score by severity was significant between groups (Kruskal-Wallis Test p<0.001).

CONCLUSIONS: he DLQI showed good psychometric properties in this dataset using both classical test theory as well as IRT.

Code

PCR63

Topic

Patient-Centered Research

Topic Subcategory

Instrument Development, Validation, & Translation, Patient-reported Outcomes & Quality of Life Outcomes

Disease

Sensory System Disorders (Ear, Eye, Dental, Skin)