Methods
We conducted a systematic review of all the accessible medical literature through 31 August 2011 that examined the temporal association of depression with RA disease outcomes, the moderating or mediating effects of depression on clinical treatments, and its relation to RA treatment persistency.
Literature Search
The following electronic databases were searched: PubMed, Medline, CINAHL, ISI Web of Science and the Cochrane Central Register of Controlled Trials. The medical subject heading (MeSH) terms used were: depression, rheumatoid arthritis, and depression and rheumatoid arthritis. MeSH terms were combined with certain keywords: disease, health, treatment, outcome and symptoms. The specific search strings used were: depression and rheumatoid arthritis and disease; and depression and rheumatoid arthritis and health; and depression and rheumatoid arthritis and treatment; and depression and rheumatoid arthritis and outcome; depression and rheumatoid arthritis and symptoms. Manual searches of reference lists of the full-text articles that were reviewed were also conducted.
Selection Criteria
Articles had to meet the following criteria: (i) human subjects; (ii) diagnosed RA or a subset of RA patients analysed as a unique study sample; (iii) studies with longitudinal outcome measures in relation to depression; (iv) specific measurement and analysis of depression in relation to downstream RA disease outcomes because due to the possible bidirectional effects studies evaluating both at the same measurement time point cannot provide estimates of temporal associations, or studies where depression served as a mediator or moderator of RA treatment; (v) outcome assessments that constituted a relevant RA disease outcome (e.g. swollen joint counts, DMARD discontinuation); (vi) study data derived from after 1 January 1998 or an interval that overlapped this period; and (vii) longitudinal assessment for at least ≥6 weeks, which was chosen because it was the approximate time period believed to be necessary for any received treatment to take effect. Exclusion criteria were the following: (i) publication before 1 January 1998; (ii) non-English language publication; and (iii) meta-analyses, review articles, case studies, case series and opinion pieces/editorials.
Definitions
Study populations comprised patients with physician-diagnosed RA. Depression was defined using a broad framework and did not imply a clinical major depressive disorder diagnosis. Acceptable measures included physician diagnosis, assessment based on Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (DSM-IV) criteria, validated measurement scales and patient self-report. Study outcome variables encompassed three categories: disease activity, treatment persistence and response to treatment. Pertinent outcomes included the following: clinical disease activity index, 28-joint DAS (DAS28), physician global disease activity using a visual analogue scale (VAS), patient global disease activity VAS, patient pain VAS, HAQ, tender and swollen joint counts (TJCs and SJCs, respectively), and measures indicative of treatment persistency based on medical record review (e.g. DMARD discontinuation).
Data Abstraction and Quality Rating
Articles were first reviewed using their titles and abstracts to determine whether the study included (i) a patient population with diagnosed RA, (ii) longitudinal measures and (iii) measurement and analysis of depression that allowed for a temporal interpretation regarding depression and prospective RA outcomes. Full-text articles were retrieved for the studies meeting these criteria and further reviewed by two study investigators (A.M.R. and G.W.R.). A data abstraction form that was developed and pilot tested prior to data collection was used to gather information, and articles not meeting the more specific inclusion criteria were subsequently excluded. Dual data abstraction was performed in pairs by the primary investigators (A.M.R. and L.R.H. or A.M.R. and G.W.R.) for the included articles. The 26-item Downs and Black quality rating scale was used to evaluate studies included in the review. Pairs of study investigators (A.M.R. and L.R.H. or A.M.R. and G.W.R.) conducted the quality ratings. The results from the data abstractions and quality ratings were then compared and disagreements were discussed among the reviewers until a consensus was reached.
Analyses
The studies were examined regarding their general descriptive characteristics and data such as sample size, outcomes and other related information. The articles were then evaluated for the specific methods utilized, and a simple evaluation of the observed effects regarding depression's association with RA disease activity or treatment was performed. Quantitative analyses were not possible due to the vast heterogeneity of study designs and effect estimates.