OBJECTIVES: To highlight the importance and impact of imputation approach used in reporting rheumatoid arthritis (RA) clinical trial results when data are analyzed using non-responder imputation (NRI) versus last observation carried forward (LOCF). METHODS: Non-responder imputation is a conservative analysis method in which participant dropouts are assumed to be non-responders regardless of actual response status at the time of dropout. Last observation carried forward is an analysis method in which the last measured value of a variable, such as treatment response, is carried forward and assumed to be valid for a future point of analysis. Results at 52 weeks from the PREMIER trial, a double-blind, randomized trial in adult patients with early RA (<3 years) that compared among adalimumab plus methotrexate (ADA+MTX) and monotherapies with either drug were compared using NRI and LOCF analyses. Outcome measures presented here are American College of Rheumatology (ACR) 50%, and 70% responses, and remission based on 28-joint Disease Activity Score (DAS28<2.6). RESULTS: In the ADA+MTX treatment group, outcome measures calculated using NRI and LOCF, respectively, were 62% and 68% for ACR50, 46% and 48% for ACR70, and 43% and 47% for remission. In the MTX-monotherapy group, NRI and LOCF values, respectively, were 46% and 49% for ACR50, 28% and 29% for ACR70, and 21% and 22% for DAS28 <2.6. For all outcome measures, the estimate of drug effect was lesser when using NRI analysis compared with LOCF analysis. CONCLUSIONS: Non-responder imputation  analyses tend to result in more conservative estimates of drug effect on outcome measures than LOCF analyses. In trials in which there are high numbers of participant dropouts, the difference in results using NRI versus LOCF could be substantial. Thus, caution is warranted in comparisons of results across clinical trials using these different imputation methods.