In this cross-sectional study, recruited patients attended the dermatological consultation of private practice offices, health provider centers and hospitals of 6 different Colombian regions: North-East; North-West Central area; and South-West. This study was conducted in full compliance with the Declaration of Helsinki of 1975, as revised in 1983 and the ethical standards of the School of Medicine Bio-Ethics Committee and Fundacion Hospital San Vicente Research and Ethics Committee (Medellin, Colombia). Informed consent was obtained from all adult patients and adolescents between 16 and <18 years-old, as well as a signed assent form from their parents or guardians. Individuals older than 16 years-old of any gender and with any skin disease and who signed informed consent were included too. Patients with mental disabilities were excluded.

The DLQI in Colombian Spanish applied to all patients is a 10-item instrument divided into six aspects: activities of daily living, work/school, personal relationships, leisure, symptoms and feelings and therapy. Its scoring ranges from 0 up to 30: 0–1=no impact; 2–5=mild impact; 6–10=moderate impact; 11–20=substantial impact and 21–30=extreme impact on the patient's QoL. The higher the score, the worse the QoL. Each DLQI item is answered on a 4-point scale scored as follows: ‘not at all’=0, ‘a little’=1, ‘a lot’=2 and ‘very much’=3. Non-relevant response (NRR) options are available for items 3 through 10.

Quantitative variables were expressed as means and standard deviations or median and interquartile range (Quartile 1 and Quartile 3) while qualitative variables were expressed as absolute numbers and their relative frequencies in percentages. Skin diseases were stratified as inflammatory and non-inflammatory. Response options are expressed as percentages for each item. Floor and ceiling effects were examined by the proportion of patients who achieved the DLQI best and worst scores, with cutoff values >15%, as described by McHorney et al.12

Item #7 was assessed either as a dichotomous variable (“Within the last week, has your skin problem prevented you from working or studying?”) and as an ordinal variable (“If “Yes”, over the last week how much has your skin been a problem at work or school?”). For all items except for item #7, the matrix of polychoric correlations across items was determined, and for item #7, as well, in its dichotomous version, the tetrachoric relationship with the other items was calculated. Internal-consistency was determined using McDonald's Omega coefficient and Cronbach's alpha (α) coefficient with their respective 95% confidence intervals (CI). Values>0.70 were considered adequate.13,14

For Rash analysis, we used Partial Credit Model (PCM) for polytomic items. The assessment of the assumption of unidimensionality and dimensionality of the instrument among non-inflammatory and inflammatory skin diseases was conducted using Horn's parallel.15 Then we assessed the structural validity of the one-dimensional model with a confirmatory factor analysis while the goodness-of-fit measures was evaluated using the root mean square error of approximation (RMSEA) with values <0.05 in addition to the comparative fit index (CFI) and the Tucker–Lewis index (TLI) with values>0.95.16

We examined differential item functioning in items across subgroups of patients with non-inflammatory and inflammatory diseases and between gender and age ranges categorized either as: “Younger”≤45 and “Older”>45 years-old and into 3 groups: <30, 30–60, >60 years-old) through the Hedges G test based on the standardized mean differences. Fritz et al.17 classification was used with 0.20, 0.40 and 0.80 considered as a low, medium, and high differences,17,18 respectively. In addition, differential item functioning was evaluated using Mantel–Haenszel model19 for both genders, inflammatory and non-inflammatory diseases. Differential functioning was stablished with p values<0.05 as reported by Holland et al.20

All analyzes were performed using the R program version 4.3.0 and the mirt, psych, lavaan and MBESS packages.21–25

Overall, 865 patients with a mean age of 49.3 years (SD, 21; 61% females) were included. Around 52% of patients were not on any therapies, 54% of patients suffered from comorbidities and most patients included had non-inflammatory skin diseases (n=482) (Table 1).

Dermatology-Life-Quality-Index scores ranged from 0 up to 30, with a median of 5 (percentile25=1; percentile75=13) The distribution of item scores was asymmetrical with positive skewness for total score and response distribution for each item (Figs. 1 and 2 of the supplementary data). Floor effect12 was present in all items (percent of response in option “not at all”>29%). Ceiling effect was not present in any of the items (Table 2).

Internal consistency for the version with dichotomous item #7, showed a McDonald's Omega coefficient of 0.85 (95%CI, 0.84–0.87) and a Cronbach's α coefficient of 0.86 (95%CI, 0.84–0.88). For the version with item #7 as ordinal, McDonald's Omega coefficient was 0.87 (95%CI, 0.85–0.89) and Cronbach's α coefficient of 0.85 (95%CI, 0.82–0.87). Correlation matrix is presented in Supplement 3.

Horn's parallel analysis indicated DLQI unidimensionality with a one factor solution. A two-factor model alternative only included item #7 in the second factor. Therefore, we chose the one factor solution model (Fig. 4 of the supplementary data). In addition, when analyzing the structural validity under a one-dimensional model, the goodness-of-fit indices showed better results for the model that had an ordinal item #7 with CFI=0.999, TLI=0.998 and RMSEA=0.012 (90%CI, 0.000–0.028) compared to the dichotomous item #7 which showed CFI=0.999, TLI=0.999 and RMSEA=0.000 (90%CI, 0.000–0.024).

When Horn's parallel analysis was performed stratifying by inflammatory and non-inflammatory skin disease, we found that in the former, the first dimension would include item #2, item #3, item #4, item #5, item #6, item #7, item #8, item #9, item #10 while the second dimension would only include item #1 as seen in Fig. 1. The internal consistency by Cronbach's Alpha and McDonald's Omega was 0.83 (95%CI, 0.80–0.87) and 0.86 (95%CI, 0.84–0.89), and 0.58 and 0.58 for the first and second dimensions, respectively. Regarding Inflammatory skin diseases, the single dimension included all items, except for item #7, as seen in Fig. 2. The internal consistency by Cronbach's Alpha and McDonald's Omega was 0.73 (95%CI, 0.66–0.81) and 0.80 (95%CI, 0.75–0.84), respectively.

Figure 1.

DLQI dimensionality in inflammatory disease.

(0.11MB).

Figure 2.

DLQI dimensionality in non-inflammatory disease.

(0.11MB).

Regarding non-inflammatory skin diseases, the first dimension would include item #2, item #5, item #6 and item #8 and the second, item #1, item #3, item #4, item #7, item #9 and item #10, as seen in Fig. 2. The internal consistency by Cronbach's Alpha and McDonald's Omega was 0.71 (95%CI, 0.65–0.77) and 0.61 (95%CI, 0.53–0.69) and 0.72 (95%CI, 0.66–0.78) and 0.73 (95%CI, 0.68–0.78) for the first and second dimensions, respectively.

Figs. 3 and 4 show the characteristic curves of the responses of items #1 and #9, respectively. Only item #1 fitted the PCM model.

Figure 3.

Option characteristic curves for item #1 fitting a partial credit model. P1: “Not at all”, P2: “A little”, P3: “A lot”, P4: “Very much”.

(0.2MB).

Figure 4.

Option characteristic curves for item #9 fitting a partial credit model. P1: “Not at all”, P2: “A little”, P3: “A lot”, P4: “Very much”.

(0.19MB).

Fig. 5 of supplementary data describes the characteristic curves of the items using the Partial Credit Model (PCM) for polytomic items. On the other hand, Table 3 shows the discrimination and difficulty parameters for the response options of each item where it is observed that the options “A little” and “A lot” do not preserve the order of the responses in any of the items except for item #1.

Furthermore, in the items information functions, a slight shift to the right was observed in all items toward individuals with impaired QOL close to +1 standard deviation (Fig. 6 of the supplementary data).

Differences between non-inflammatory and inflammatory patients was found with G Hedges=0.89 (95%CI, 0.73–1.04), but no gender differences were ever found: 0.07 (95%CI, 0.13–0.15) (Table 4).

Differential item function (DIF) analysis showed that items #2, #5, #8 and #10 had large DIF in non-inflammatory and inflammatory patients (Table 4). Item #2 had a large DIF between male and female patients (Supplement 7). In addition, when age was categorized in patients younger and older than 50 years (median), items #3 and #6 presented large DIFs (p<0.05), but when age was categorized into 3 groups (<30, between 30 and 60 and >60 years-old) we saw that only item #3 presented a large DIF (p<0.05) (Table 5).