Methods: Data were analyzed on 628 randomly selected rheumatoid arthritis (RA) patients in a multicenter, cross-sectional mail survey. We compared patients with and without a history of depression treated with medication in the past 5 years (depression) and their ratings of bother from anti-rheumatic side effects (bother), satisfaction with disease control, self-efficacy to manage arthritis symptoms(SE), and COMRADE confidence in decision (CID). DMARD adherence was estimated with the Medical Adherence Rating Scale (MARS). Between group differences were assessed using the independent samples T test procedure. For estimating the effect the simultaneous influence pathways of the variables and to test study hypothesis, latent variable models (LVM) were fit to the data. The relationships are represented in path diagrams, along with parameter estimates of factor loadings, one way path regression coefficients, and residual covariances. Model fit is summarized by the chi-square and popular index RMSEA.

Results: The response rate was 71%. 20.5% of patients had a history of depression requiring anti-depressants. Patients with depression had statistically significant higher ratings of bother by anti-rheumatic side effects (P<.03), lower levels of self-efficacy (P<.001) satisfaction with disease control (P <.001) and CID (P<.01). Depression was not significantly related to MARS (P<.13). When the data were evaluated with LVM, the results created a meaningful measurement model of patient appraisal of treatment effect relating bother, SE and satisfaction with disease control with an overall chi-square goodness of fit index of .1, for degrees of freedom = 1, and RMSEA = 0 with a 90%-confidence interval (0, .073). CID was related to the latent variable patient appraisal of treatment effect. However when age, low income and RA severity were held constant depression did not significantly relate to the latent variable patient appraisal of treatment effect, nor to CID or MARS. Further depression did not modify the relationship between CID and patient appraisal of treatment effect.

Conclusion: In this large random sample of community RA patients, when major confounders (age, RA severity, low income) were controlled for, depression did not significantly effect patient appraisal of treatment effect or modify the relationship between this appraisal and confidence in decision.