A cross-sectional study was conducted that included patients with SLE who were attached to the External Consultation centre of the Rheumatology Unit of the “Ruíz y Páez” University Hospital Complex and the Centro Clínico Universitario de Oriente in Ciudad Bolívar, Venezuela, between April and July 2018, with at least 4 classification criteria from the 1982 American College of Rheumatology21. Patients with additional diagnoses of other different autoimmune diseases to SLE were excluded, save those patients with a diagnosis of anti-phospholipid syndrome.

The study was conducted in keeping with the ethical principles for medical research in human beings from the Declaration of Helsinki22, with the corresponding informed consent signed by all patients.

The LupusQoL10 questionnaire, consolidated with a valid, reliable HRQoL from the patient and specific to the disease for adults with a diagnosis of SLE, contains eight domains and a total of 34 items which are responded to using a five-point Likert scale. It gives a score by domains which ranges from 0 (worse HRQoL) to 100 (best HRQoL), which may be obtained if the following formula is followed: the responses by domain are added up and divided by the total number of items of this domain. The resulting score is divided by 4 and then multiplied by 100.The version translated into Spanish of the LupusQoL was used; usage licence was requested through the online system of RWS Life Sciences, Inc.

The “Generalitat de Catalunya”23 (GENCAT) scale form — an objective tool designed in keeping with advances made on the multidimensional general quality of life model proposed by Schalock in 2002 and Verdugo in 2003 was used. It contains eight dimensions and a total of 69 items which are responded to using a four-option scale of frequency. The GENCAT scale is a valid and reliable tool23 which has been used in national24,25 and international studies26,27 to measure the quality of life in patients with SLE.

The accumulated disease activity and damage were calculated through the Systemic Lupus Erythematosus Disease Activity Index28 (SLEDAI) and Systemic Lupus International Collaborating Clinics29 (SLICC), respectively. The SLEDAI assesses disease activity in the last 10 days and includes 24 items which determine specific symptoms in 9 organs or systems, with a maximum score of 10528. The SLICC assesses irreversible damage of the disease during the last 6 months and includes 42 items to measure involvement of the 12 domains with a maximum score of 46 points29. Both models have been proven to be valid and reliable for assessing activity (SLEDAI) and accumulated damage (SLICC) in patients with SLE28,29.

Patients were approached during their routine rheumatology consultation. Those who agreed to participate in the study received an envelope with an informed consent form, a sociodemographic data form, a copy of the LupusQoL and a copy of the GENCAT scale for them to complete in the medical centre or take home to complete and deliver within a week. The patients who could not read or write received help to complete the documents, with by the authors in the medical centre or their family members when they were at home.

The SLEDAI and SLICC indices of the medical records for the previous 6 months were later obtained.

Reliability was determined by internal consistency using Cronbach’s alpha coefficient. Cut-off points were generated for the LupusQoL and the GENCAT scale from an analysis of latent classes which highlighted one or more non observed (latent) classes with respect to a variable. By applying an analysis of the Bayesian information criterion values, we confirmed that two classes were optimum for both questionnaires. Means and marginal probabilities were then studied for the latent classes, grouping each patient into one of the two classes. From these, using Receiver Operating Characteristics curves (ROC), the cut-off points for the LupusQoL were determined (for each domain and the total) and for the GENCAT scale. These cut-off points led to stratifying the HRQoL of the patients into “better” or “worse”. Convergent and discriminating validity was analysed using Cohen’s kappa coefficient and the Spearman correlation coefficient, respectively. Finally, the predictive validity was analysed using COR curves with SLEDAI and SLICC values grouped into: <4 or ≥4 points and 0 or ≥1 point, respectively; these cut-off points were taken because they had been satisfactorily used in previous validation studies of LupusQoL11–16,19. The STATA version 16 and the SPSS version 25 were used for statistical analysis. Statistical significance was set at p < .05.

A total of 100 patients with SLE completed the questionnaires; 93% were women, with mean age at 42 years (SD: 13) and mean disease duration of 11 years (SD: 9). SLEDAI and SLICC means were 3 and 1, respectively. The other sociodemographic characteristics are contained in Table 1.

Internal consistency was .96, which proves LupusQoL’s high reliability.

The highest correctly classified patient percentage (99%) was obtained with a value of 64.55 points in the LupusQoL (area below the curve = .99; sensitivity: 100%; specificity: 97.56%), suggesting that the patients with 64.55 points or more presented with a good HRQOL (Fig. 1). Furthermore, cut-off points for each LupusQoL domain (Table 2) domain were obtained.

Fig. 1.

ROC curve of the LupusQoL in Venezuelan patients with SLE.

(0.09MB).

For the GENCAT scale, the cut-off points which correctly classified most patients (95%) was 67 points (area under the curve = .98; sensitivity: 92.73%; specificity: 97.78%), and patients with the same or higher score to this sum therefore presented with a better quality of life (Fig. 2). After grouping the patient in accordance with the LupusQoL cut-off points and the GENCAT scale, a moderate convergent validity which was statistically significant was obtained (Cohen’s kappa coefficient = .556; p = .000).

Fig. 2.

ROC curve of the GENCAT scale in Venezuelan patients with SLE.

(0.09MB).

The LupusQoL was inversely correlated with the SLEDAI (rho = –3.27; p = .001) (Fig. 3A) and with the SLICC (rho = –.246; p = .014) (Fig. 3B) significantly.

Fig. 3.

Correlation between the SLEDAI and the SLICC with LupusQoL in Venezuelan patients with SLE. (A) Correlation between the SLEDAI and the LupusQoL. (B) Correlation between the SLICC and the LupusQoL.

(0.14MB).

The LupusQoL discreetly predicted disease activity (SLEDAI) for scores ≥4 (area under the curve = .704; sensitivity: 74.24%; specificity: 67.65%) (Fig. 4A), classifying 72% of patients. The LupusQoL gave a poor prediction of accumulated damage (SLICC) for scores ≥1 (area under the curve = .642; sensitivity: 66.67%; specificity: 52.94%) (Fig. 4B), classifying only 62% of patients.

Fig. 4.

ROC curves of the LupusQoL with the SLEDAI and the SLICC in Venezuelan patients with SLE. (A) ROC curves of LupusQoL with the SLEDAI. (B) ROC curves of the LupusQoL with the SLICC.

(0.21MB).

Average scores for each domain of the LupusQoL were compared in patients with and without disease activity (SLEDAI: <4 or ≥4 points), and also in patients with or without accumulated damage (SLICC: 0 or ≥1 points) (Table 3), with significant differences being found between both groups in all domains, except in “sexual relations” for the SLEDAI and SLICC scales and “burden for others” for the SLICC scale.