Background: A problem encountered when analyzing long-term efficacy is that the number of patients in follow-up decreases with time for different reasons. The method used to account for missing observations for the therapy under analysis has a great influence on the inference of efficacy.
Objective: To describe the long-term efficacy of etanercept for psoriasis in daily practice using 3 analytical approaches.
Methods: Prospective data from a cohort of patients with psoriasis treated with etanercept for at least 24 weeks were analyzed using 3 analytical approaches: as treated analysis, intention-to-treat analysis (ITT) with last observation carried forward (LOCF) and intention-to-treat analysis with modified nonresponder imputation (modified NRI).
Results: One hundred thirty-one patients were treated with etanercept during 134 treatment episodes with a mean treatment duration of 2.7 years. The maximum follow-up was 6.0 years. The methodological approach chosen had a great influence. Psoriasis Area and Severity Index (PASI) 75 response rates varied from 60% in the as-treated approach to 34% in LOCF and to 29% in modified NRI at week 264.
Limitations: All analytical methods applied have limitations. Other outcome measures could be used to overcome the bias introduced by each method of analysis, such as drug survival.
Conclusions: The methodological approach chosen to analyze long-term efficacy data has a great influence on the inferences that may be drawn regarding the degree of efficacy. Therefore we support the use of different methods to present long-term efficacy data.
Copyright © 2012 American Academy of Dermatology, Inc. Published by Mosby, Inc. All rights reserved.