# Confirmation of unidimensionality of the Dermatology Life Quality Index (DLQI) using a multinational 3,408 patient dataset

## TL;DR

This study confirms that the Dermatology Life Quality Index (DLQI) is a reliable and unidimensional tool for measuring the impact of skin diseases on patients' quality of life.

## Contribution

The study provides empirical validation of DLQI's unidimensionality using a large multinational dataset of 3,408 patients.

## Key findings

- DLQI demonstrated unidimensionality through psychometric analyses including eigenvalues, item-response theory, and confirmatory factor analysis.
- DLQI scores showed strong correlations with disease severity and quality of life measures like EQ-5D.
- Cronbach’s alpha of 0.90 and good fit statistics confirm the reliability and validity of DLQI.

## Abstract

The Dermatology Life Quality Index (DLQI) was designed to be a simple, practical questionnaire for routine clinical use and is the most widely used tool to measure the burden of skin diseases and assess effectiveness of interventions based on patients’ perspective. The aim of this study was to further validate the DLQI using a multinational European study dataset.

Data from a cross-sectional study conducted in 13 European countries were analysed. In each dermatology clinic, 250 consecutive adult out-patients were recruited. A wide range of classical test theory and IRT tools were used to investigate unidimensionality and known-group properties.

From 3,635 patients, 3408 completed the DLQI with no missing data. 55.8% of patients were female, mean age 46.6 years (SD = 17.82). The commonest conditions were psoriasis (17.4%), non-melanoma skin cancer (10.8%), infection of the skin (6.7%), hand eczema (6.2%) and acne (6.2%). Using DLQI score banding, the patients’ disease had no effect (n = 962, 28.2%), small (912, 26.8%), moderate (674, 19.8%), very large (691, 20.3%) or extremely large effect (169, 5.0%) on their quality of life. Psychometric analyses employing eigenvalues, scree plots, item-response theory (IRT) and very simple structure indicated unidimensionality. Cumulative explained common variance did not significantly increase from 0.618 with additional factors. Confirmatory factor analysis fits for a 1-factor solution were very good and a 1-factor model was optimum. IRT indicated good fit statistics, no misfitting items (infit and outfit), and no local dependencies (LG2 and Q3). Cronbach’s alpha was 0.90. Cohen’s-d effect sizes for DLQI scores between different EQ-5D item levels were moderate (> 0.5) or large (> 0.8), and large between different disease severity levels. Good correlation was found between EQ-5D VAS scale and DLQI score (r2 =-0.409, p < 0.001). Known group analysis of DLQI score by severity was significant between groups (p < 0.001).

Analysis showed good psychometric properties confirming a 1-factor unidimensional model.

The online version contains supplementary material available at 10.1186/s41687-026-01025-0.