Gary SGronseth, Associated Editor Level-of-Evidence Review, University of Kansas Medical Centerggronseth@kumc.edu
Submitted August 26, 2015
The level-of-evidence (LOE) editors of the Constantinescu et al. [1] article agree with Dr. Meilof's assertion that a longer study with more patients would be desirable to establish the tolerability or safety of any intervention. We would like to emphasize the purpose and limitations of the LOE statement. The LOE rating is designed to give the reader a sense of the risk of bias of a study related to study design and execution. The LOE rating does not address issues of generalizability (e.g., short duration of follow up) nor issues related to the imprecision of the result (e.g., small sample size). When possible, we attempt to capture generalizability issues within the wording of the LOE statement and limitations of precision by including confidence intervals. Given practical editorial limitations on the length of the LOE statement, issues related to generalizability and imprecision cannot always be adequately captured. The brief LOE statement cannot convey all of the information needed to assess the quality and impact of a study. For these nuances, one must carefully read the entire article as Dr. Meilof obviously did.
1. Constantinescu CS, Asher A, Fryze W, et al. Randomized phase 1b trial of MOR103, a human antibody to GM-CSF, in multiple sclerosis. Neurol Neuroimmunol Neuroinflamm 2015;2:e117.
For disclosures, please contact the editorial office at nnnjournal@neurology.org.
The level-of-evidence (LOE) editors of the Constantinescu et al. [1] article agree with Dr. Meilof's assertion that a longer study with more patients would be desirable to establish the tolerability or safety of any intervention. We would like to emphasize the purpose and limitations of the LOE statement. The LOE rating is designed to give the reader a sense of the risk of bias of a study related to study design and execution. The LOE rating does not address issues of generalizability (e.g., short duration of follow up) nor issues related to the imprecision of the result (e.g., small sample size). When possible, we attempt to capture generalizability issues within the wording of the LOE statement and limitations of precision by including confidence intervals. Given practical editorial limitations on the length of the LOE statement, issues related to generalizability and imprecision cannot always be adequately captured. The brief LOE statement cannot convey all of the information needed to assess the quality and impact of a study. For these nuances, one must carefully read the entire article as Dr. Meilof obviously did.
1. Constantinescu CS, Asher A, Fryze W, et al. Randomized phase 1b trial of MOR103, a human antibody to GM-CSF, in multiple sclerosis. Neurol Neuroimmunol Neuroinflamm 2015;2:e117.
For disclosures, please contact the editorial office at nnnjournal@neurology.org.