OBJECTIVE: The demand for comparative effectiveness research (CER) data from payers, physicians, and patients is significant, but the cost and time associated with prospective randomized trials is a barrier to rapid decision-making.  Use of patient-reported outcomes (PROs) collected via on-line patient communities provides one channel for rapid data collection, particularly in conditions such as rheumatoid arthritis (RA), where validated PRO instruments are available. METHODS: A random sample of iGuard.org members in the US treated with non-steroidal anti-inflammatory drug (NSAIDs), oral disease-modifying antirheumatic drugs (DMARDs), or biologics for RA completed the Rheumatoid Arthritis Disease Activity Index (RADAI) and a series of other questions related to their disease.  iGuard.org is a free medication monitoring service that is introduced to patients through multiple sources including physician, pharmacy and online referrals.  For this study, we report pilot baseline data on patient-reported RADAI, pain, and joint counts across the three treatment groups to demonstrate use of on-line communities in supporting CER. RESULTS: A total of 153 RA patients completed the study: 49 treated with NSAIDs only, 51 exposed to oral DMARDs, and 53 exposed to biologics.  The mean (SD) RADAI score was 4.59 (2.16).  Adjusting for age and gender and multiple comparisons, there were significant differences between the three treatment groups on RADAI scores (p=0.0045) and patient global assessment of pain (p=0.0357) but not on the number of painful joints (p=0.3512).  The trend was towards patients on NSAIDs only having worse outcomes compared to patients on biologics or patients on oral DMARDs. CONCLUSIONS: This pilot study demonstrates the possibility of collecting baseline disease severity data directly from patients using the RADAI, which is sensitive to detect differences by treatment modality.  The next step in the pilot program will be to investigate the potential for capturing longitudinal disease progress information amongst patients in on-line communities.