This analysis used data retrieved from the Adelphi Disease Specific Programme, a large, syndicated, retrospective, multinational surveys of physicians and patients in a real-world clinical setting for a range of common diseases . The Disease Specific Programme collects quantitative and qualitative survey data and provides a comprehensive overview of a given disease and treatment of that disease from the perspective of both physician and patient . Two Rheumatology Disease Specific Programme surveys conducted in the USA between January and March 2011 and over a similar time period in 2014, were used for this study. The Disease Specific Programme included a geographically diverse sample of US rheumatologists and their respective patients with PsA. The Rheumatology Disease Specific Programme was conducted in accordance with the US Health Insurance Portability and Accountability Act 1996 (HIPAA; www.hhs.gov/ocr/privacy/) and the Health Information Technology for Economic and Clinical Health legislation (2014 only as this legislation was not present in 2011). The Disease Specific Programme is a market research project and complies with all relevant market research guidelines and legal obligations. The research methodology and nature of the collected data make submission to national and/or local ethics committees and regulatory bodies unnecessary. Namely, the Disease Specific Programme is non-interventional and employs solely a retrospective data collection, and both physician and patient data are collected anonymously and independently.
Physicians were identified from public lists of healthcare professions. The physician sample included 200 US rheumatologists (100 sampled in each year) responsible for managing patients with PsA. Eligible physicians had to meet the following pre-specified criteria: primary specialty was rheumatology, currently treating rheumatoid arthritis (RA), PsA, and spondyloarthropathy; typical monthly workload involved consultations with three or more patients with PsA; and qualification as a physician between 3 and 40 years prior to initiation of the survey.
Each physician completed a response form for three consecutive, consulting, adult patients with PsA, generating 600 forms across the two surveys. Eligible patients had to be ≥18 years of age and have a diagnosis of PsA on or before the day of consultation. Patients were excluded if they were involved in a clinical trial. All patients gave their informed consent. Subjects had the right to opt-out of the survey at any time.
The Rheumatology Disease Specific Programme was developed by Adelphi Real World (Adelphi Real World, Cheshire, UK). All physician-completed patient record form answers were confidential and maintained physician and patient anonymity; data were fully de-identified prior to receipt by the research team. The physician-completed patient record form provided information on a wide a range of patient and disease characteristics, including demographics, comorbidities, symptomatology, and satisfaction with PsA control. All data from the physician-completed patient record form were based on evidence available to the physician at the time of the consultation; no tests or investigations were performed as part of this research.
Patients were asked to fill out patient self-completed questionnaires on a voluntary basis. To preserve anonymity, patients were asked to complete the form independently of the physician and return the patient self-completed questionnaires in a sealed envelope. To ensure the physician did not see any patient responses (including patient-reported outcome measures and symptom assessments), the patient put their responses in an envelope and sealed this prior to return. Each pair of forms (i.e., the physician-completed patient record form and the patient self-completed questionnaire) was linked during data processing using non-identifying unique identification numbers. All eligible pairs of linked physician-completed patient report forms and the patient self-completed questionnaires were included for analysis.
Patients reported their satisfaction with PsA control; in 2011, this was in response to a categorical question, and in 2014, this was captured as a response to a Likert scale (Fig. 1).
The validated Work Productivity Activity Impairment (WPAI)  and Health Assessment Questionnaire-Disability Index (HAQ-DI) (HAQ-DI max = 3.0)  questionnaires were included in the patient self-completed questionnaires for completion by patients, allowing scores to be derived for both measures. The overall percentage of “work impairment” as well as the percentage of “presenteeism” and “absenteeism” was calculated for patients who were employed at the time of the survey. However, the percentage of “activity impairment” derived from the WPAI responses was calculated for the whole sample (both employed and unemployed patients).
Determination of misalignment on satisfaction
The responses to the satisfaction questions from each pair of linked forms (physician-completed patient record form and patient self-completed questionnaire) were compared to determine if the physician and patient pair were ‘aligned’ or ‘misaligned’ in terms of their satisfaction with PsA control. Pairs were classified as ‘aligned’ when both the patient and physician felt satisfaction or dissatisfaction in terms of PsA control or as ‘misaligned’ when the physician felt satisfied, but the patient was dissatisfied with PsA control or vice versa.
Study variables of interest included physician and patient reported satisfaction with PsA control, demographics characteristics, disease characteristics, and burden of disease. The following variables were captured from physicians in the physician-completed patient record form: satisfaction with PsA control (categorized as satisfied or dissatisfied), patient age, sex, comorbidities (listed below), time since diagnosis measured in years, current treatment (e.g., topical agents, nbDMARDs, bDMARDs), tender joint count (TJC), swollen joint count (SJC), percent body surface area (BSA) affected by psoriatic skin lesions, and number of PsA symptoms currently present including joint symptoms (tenderness, swelling, stiffness, etc.) and skin symptoms for those with psoriatic skin lesions (itching, pain, scaling, etc.). Comorbidities included anxiety, depression, type 2 diabetes, hyperlipidemia/elevated cholesterol, gastric condition, hypertension, liver impairment, malignancy, obesity, renal impairment, osteoporosis, respiratory conditions, tuberculosis, and vasculitis. The patient self-completed questionnaire captured satisfaction with PsA control (categorized as satisfied or dissatisfied), WPAI (a percentage of overall work impairment scored as 0 to 100% impairment and impairment of “presenteeism”, “absenteeism”, and activity impairment) , and HAQ-DI (continuous variables of scores ranging from 0 to 3 with a score of 0 indicating performance without any difficulty and up to score 3 meaning performance cannot be done at all) .
Data were reported descriptively for each variable (i.e., patient age, sex, comorbidities, time since diagnosis, current treatment, TJC, SJC, percent BSA affected by psoriatic skin lesions, number of PsA symptoms currently present, HAQ-DI scores, and WPAI responses). Categorical variables were summarized using frequency counts and percentages. Continuous variables were summarized by the number of observations, their mean, and standard deviation (SD).
Bivariate statistical comparisons were made between the aligned and misaligned groups for each variable. P values were obtained using the Wilcoxon rank-sum test for continuous variables, and Fisher’s exact test for categorical variables.
For the primary objective, multivariate logistic regression analyses evaluated what factors may be associated with patient and physician misalignment. Two analyses were performed as follows: the first included all patient-physician pairs and the second included the subgroup of patients who were employed and had completed the WPAI. The dependent variable was whether patients were aligned with their physicians in regard to satisfaction of PsA control. Independent variables included age, current bDMARD treatment, SJC, percent BSA affected by psoriatic skin lesions, and HAQ-DI. For the multivariate analysis, TJC was omitted from the model because there was multicollinearity with SJC. In addition, TJC is confounded by other diseases, such as osteoarthritis, and is an indirect measure of inflammation . SJC, on the other hand, is a good measure of inflammation . All variables were included in the models at the same time. Standard errors were adjusted to allow for possible intragroup correlation within the reporting physician. The multivariate logistic regression was repeated on the subpopulation of patients who had completed the WPAI using WPAI as an independent variable to identify independent predictors of misalignment of employed patients.
Sub-analysis of satisfaction with PsA control in patients with active joint disease
An exploratory analysis was performed in two groups of patients with active joint disease (>3 TJC): satisfied and not satisfied with PsA control. Data were reported descriptively for each variable. Categorical variables were summarized using frequency counts and percentages. Continuous variables were summarized by the number of observations, the mean, and SD.
Bivariate statistical comparisons were made between the satisfied and not satisfied groups for each variable. P values were obtained using the Wilcoxon rank-sum test for continuous variables, and Fisher’s exact test for categorical variables.