COMMENTS FROM THE EDITORS HAS THE PENDULUM SWUNG TOO FAR? I n this issue of the Journal there are new guidelines for the use of statistics in prospective manuscripts submitted to the Journal. They are in agreement with the “Uniform Requirements for Manuscripts Submitted to Biomedical Journals” published in the Annals of Internal Medicine. Biostatistics has become an ever more important component of biomedical research and, in the past, statisticians have criticized journals for lack of statistical rigor. Thus, in January, the Journal initiated guidelines that were to aid the author and expedite the review process. Some authors have perceived them as too rigid and coercive. In light of these comments, we have rewritten the guidelines with the hope that they will be more “user friendly,” still maintaining statistical standards. We also recognize the unique aspect of surgical research in which randomization is difficult and obviously nonblinded. Numbers are often small both in clinical trials and in animal experiments where availability or cost can become prohibitive when larger numbers are de- manded in order to meet the stringent requirements of some biostatisticians. Biostatistical review remains essential to elicit the true differences in results between two or more groups. It will be our aim to find the pathway between the extremes so that authors and readers will have reasonable confidence in the conclusions presented without going through a biostatistical “inquisition.” We will try to bring the pendulum back to center. Eugene H. Blackstone, MD John A. Waldhausen, MD Copyright © 1998 by Mosby, Inc. 0022-5223/98 $5.00 1 0 12/1/91665 186

Has The Pendulum Swung Too Far?

Download PDF Report

Upload
eugene-h-blackstone
View
212
Download
0

Embed Size (px)

Citation preview

COMMENTS FROM THE EDITORS

HAS THE PENDULUM SWUNG TOO FAR?

In this issue of the Journal there are new guidelinesfor the use of statistics in prospective manuscripts

submitted to the Journal. They are in agreementwith the “Uniform Requirements for ManuscriptsSubmitted to Biomedical Journals” published in theAnnals of Internal Medicine.

Biostatistics has become an ever more importantcomponent of biomedical research and, in the past,statisticians have criticized journals for lack of sta-tistical rigor. Thus, in January, the Journal initiatedguidelines that were to aid the author and expeditethe review process. Some authors have perceivedthem as too rigid and coercive. In light of thesecomments, we have rewritten the guidelines with thehope that they will be more “user friendly,” still

maintaining statistical standards. We also recognizethe unique aspect of surgical research in whichrandomization is difficult and obviously nonblinded.Numbers are often small both in clinical trials and inanimal experiments where availability or cost canbecome prohibitive when larger numbers are de-manded in order to meet the stringent requirementsof some biostatisticians.

Biostatistical review remains essential to elicit thetrue differences in results between two or moregroups. It will be our aim to find the pathwaybetween the extremes so that authors and readerswill have reasonable confidence in the conclusionspresented without going through a biostatistical“inquisition.” We will try to bring the pendulumback to center.

Eugene H. Blackstone, MDJohn A. Waldhausen, MD

1 8 6