Objective. To determine the comparability of a text to a mannequin format for the assessment of joint counts (JC) among patients with rheumatoid arthritis (RA) participating in a randomized clinical trial (RCT). Methods. A subgroup of patients participating in the MIRA (Minocycline in RA) RCT completed self-administered JC and joint scores (JS), which were compared to those of a trained assessor. Results. JC and JS data were consistently higher for the patient than for the assessor. Higher correlations were obtained for JC than for JS. Conclusion. Our data suggest JC can be used in the context of clinical trials or in the clinical setting, but are not interchangeable with trained assessor JC.