This was an observational, prospective, multicenter study to validate a new HRQOL instrument (PSO-LIFE) for patients with psoriasis. Patients were included in the study between October 2008 and May 2009, and attended a maximum of 3 study visits over the 3 month follow-up period. A total of 39 investigators from centers all over Spain participated in the study.
The study protocol was approved by the Clinical Research Ethics Committee of the Hospital General Universitario de Valencia. All patients included in the study provided their signed, informed consent to participate.
Three types of patients were included in the study. Patients in Group A were patients with a diagnosis of active psoriasis, patients in Group B had inactive psoriasis at the time of inclusion, and patients in Group C were a control group consisting of patients with active atopic dermatitis (AD) or active chronic urticaria (CU). Inactive psoriasis was defined as occurring when plaques remained the same size and no new plaques appeared, even when psoriasis was widespread . Patients in groups A and B completed 3 study visits, at baseline, and at 7 days and 3 months since baseline. Patients in Group C only attended one study visit.
Inclusion criteria for Group A patients (patients with active psoriasis) were: over 18 years of age with a diagnosis of plaque psoriasis for at least 6 months and confirmed by a dermatologist, who was able to complete the study questionnaires, and who provided written, informed consent to participate. Patients were excluded from this group if they had guttate, erythrodermic or pustular psoriasis, psoriatic arthritis, or psoriasis limited to the scalp, nails, palms, or soles. The same inclusion criteria were applied for Group B, although in that case patients were required to have inactive psoriasis. Patients included in the control group were required to have a confirmed diagnosis of AD or CU.
Sample size was set at 325 patients aged over 18 with a diagnosis of active or inactive psoriasis and 66 patients in the control group (33 with AD and 33 with CU). The sample size was calculated so as to allow for testing a full range of psychometric characteristics and was based specifically on the need to test sensitivity to change which was considered to require the largest sample size. Specifically, sample size for patients groups A and B was calculated to be able to detect a change of 0.2 standard deviations on the PSO-LIFE score after 3 months of follow-up, with a p value of 0.01 and a statistical power of 0.80 assuming a 10% loss to follow-up. Applying the same criteria, and in order to detect a between-group difference of 0.45 SDs with a ratio of inactive patients to controls and active patients to controls of 1:2 and 1:3, respectively, a sample size of 65 controls, 130 patients with inactive psoriasis, and 195 patients with active psoriasis was required.
Study instrumentation and variables
The primary study variable was the PSO-LIFE questionnaire. This was a self-administered questionnaire for patients with psoriasis consisting of 20 items measuring a range of HRQOL aspects relevant to individuals with psoriasis (Additional file 1). The items refer to the previous 7 days and each item allows for a 5 point Likert scale (response choices from Always to Never), Each response is given a scoring from 1 (worst HRQOL) to 5 (best HRQOL). The overall questionnaire score is obtained by adding up the 20-item responses and it ranges from 20 to 100 points. In order to simplify its interpretation and to validate the PSO-LIFE questionnaire, the scoring was transformed to a 0 to 100 scale with a higher score indicating better HRQOL.
The PSO-LIFE questionnaire was developed in a previous study phase using questionnaire development techniques; according to FDA guidelines for patient-reported outcome measure , a conceptual framework was established for the questionnaire development. Psoriasis is a dermatological disease which affects physical, emotional and social well-being, these conform the HRQOL in patients with psoriasis. The physical dimension is affected through its symptoms; the psychological dimension is affected through anxiety, depression, lack of concentration and altering self-image; and it also affects the social domain through the limitation on social/leisure activities, disturbing sex life and causing work leave. The questionnaire to be developed pretended to consider the relationship among all these domains and their impact on patient’s life. Considering all these aspects, the initial point was the development of a specific questionnaire that would provide physicians with a tool to assess knowledge on the impact of psoriasis on the patient’s quality of life.
Firstly, a literature review was performed to set the relation between psoriasis and patient’s self-perceived quality of life; then, a qualitative assessment through a focus group with 5 dermatologists was carried out to identify the main domains affecting HRQOL in psoriasis patients, As a second step, a semi-structured interview was conducted with 20 psoriasis patients (10 with active and 10 with non active), from which several items were identified to be related with psoriasis HRQOL; at the end of this stage 139 preliminary statements were identified. Each of the remaining items was subsequently rated by the dermatologist in terms of frequency, importance and clarity and the number of items was reduced to a total of 37 items. The selected items were edited in questionnaire format and administered to a sample of 171 psoriasis patients (52.1% active and 47.9% non active) to allow for a preliminary factor analysis and Rasch analysis in order to obtain the final version of the questionnaire. Factor analysis identified 6 preliminary dimensions (variance explained of 72.4%). The Rasch analysis was used to exclude those items with INFIT or OUTFIT > 1.30 and <0.70 and those which were redundant with other items. This final questionnaire, with a total of 20 items has good preliminary internal consistency (Cronbach’s alpha = 0.94) .
In the present study, socio-demographic variables collected were age, sex, educational level (no education, primary, secondary or tertiary level studies), and employment situation. Clinical variables collected in patients with psoriasis included date of diagnosis, duration of latest episode of active psoriasis, PASI index, current treatment, recurrence or continuation of active episode (in follow-up visits), and co-morbid chronic diseases. In patients with AD, clinical variables included severity of AD using the EASI index, disease intensity (using a 4 point scale from none to severe), and treatment. In patients with CU, variables collected included type of CU, current symptoms, presence of wheals or lesions on inclusion, and treatment.
As well as the PSO-LIFE questionnaire, all patients also completed the Dermatology Life Quality Index (DLQI) and patients in groups A and B also completed the Psoriasis Disability Index (PDI). Both of these have been adapted and validated for use in Spain . The DLQI is a generic questionnaire designed for use in patients with any type of skin disease. It consists of 10 items with a time-frame referring to the last 7 days and measures the impact of disease in terms of symptoms, daily activities, leisure, work/study, personal relationships, and treatment. The score ranges from 0 (minimal impact on HRQOL) to 30 (maximum impact on HRQOL). The PDI is a disease-specific questionnaire consisting of 15 items covering four dimensions of HRQOL: daily activities, work/studies, personal relationships, and leisure, with an additional item on treatment. The overall score ranges from 0 to 45 with a higher score indicating greater impact on HRQOL.
Data were also collected on patient perception of their overall health state using a single item with 7 response options. In the follow-up visits, patients were also asked whether their overall health state had improved, deteriorated or remained the same, again using a 7 option response scale from much improved to much worse (health status transition item).
The data analysis was performed using SPSS 15.0 for Windows. A significance level of 0.05 was used in all between-group comparisons.
The sample’s socio-demographic and clinical characteristics were analyzed using descriptive statistics. Parametric (Student’s t test and ANOVA) and non-parametric (Mann-Whitney U, Kruskal-Wallis, chi-squared) tests were used to test for between-group differences while Friedman and Wilcoxon tests for paired data were used to test for differences over time in the PSO-LIFE index.
The feasibility of the PSO-LIFE questionnaire was tested by analyzing the proportion of patients with missing responses on each item, and the proportion of patients with no missing responses.
Score distributions were evaluated by calculating the observed range of scores and the proportion of patients with the worst and best possible scores (floor and ceiling effects) on each dimension, as an indicator of the extent to which scales capture the range of the underlying dimension. The reliability of the new questionnaire was tested by examining the internal consistency of data in the overall sample of psoriasis patients using Cronbach’s α. Test-retest reliability was examined in patients (group A and B) who reported no change on the health status transition item in the follow-up visit at 7 days and was analyzed using the Intraclass Correlation Coefficient (ICC). For both Cronbach’s α and the ICC, values over 0.7 were considered acceptable .
Construct validity was tested in several ways. In the first place, a principal components factor analysis was performed with Varimax rotation to determine the underlying dimensional structure of the questionnaire. Screen plot analysis (based on eigenvalues –which correspond to the variances of the factors– was also used in deciding the number of factors in the questionnaire .
The questionnaire’s known groups’ validity (or capacity to discriminate between groups expected to have different scores ) was assessed by comparing scores on the PSO-LIFE questionnaire between patients with active and inactive psoriasis, and between those two groups and the control group. We also expected patients with visible lesions to score worse on the PSO-LIFE than those without visible lesions, and we expected those with mild psoriasis (PASI score <10) to score higher on the PSO-LIFE (better HRQOL) than those having more severe psoriasis (PASI score of 10 – 50). Between-group differences were analyzed using linear regression models and controlling for age and level of education (patients vs. controls) and age, level of education, and duration of last crisis (active vs. inactive psoriasis).
Convergent validity was tested by examining the extent to which scores on the PSO-LIFE demonstrated logical relationships with the DLQI and PDI in the baseline visit. We expected moderate to high correlations with the DLQI and PDI overall scores and with most of the individual dimensions on those questionnaires as they measure similar content to the PSO-LIFE. Correlations were calculated using Pearson and Spearman correlation coefficients as appropriate. Correlations under 0.3 were considered weak, correlations between 0.3 and 0.5 were considered moderate, and those over 0.5 were considered strong .
Longitudinal validity and sensitivity to change were examined by comparing changes observed in PSO-LIFE scores with those observed on the PASI Index, psoriasis status, changes in treatment, and presence of adverse effects as well as with changes in scores on the DLQI and PDI, between baseline and the final visit. Changes in PSO-LIFE score were analyzed by sub-groups based on patient perceptions of overall change on completing the study using the health status transition item. Effect sizes were used to estimate the magnitude of changes. Effect size was defined as the difference between baseline and final visit mean scores divided by standard deviation at baseline . An effect size of 0.2 is equivalent to a small effect size, 0.5 equates to a medium effect size, and values of 0.8 or over equate to large effect sizes . For comparison, effect sizes for the DLQI and PDI questionnaires were also calculated.
The minimum important difference (MID) is the smallest difference in scores on a questionnaire that patients perceive as beneficial . The MID was estimated for the PSO-LIFE questionnaire using results from patients who reported having 'slightly improved' health status at 3 months from baseline.