REPORTING OF RANDOMIZED CLINICAL TRIAL DESCRIPTORS AND USE OF STRUCTURED ABSTRACTS
Roberta W Scherer and Kay Dickersin
University of Maryland School of Medicine, 506 W Fayette St, Century Building, Baltimore, MD 21201, USA
Objective: To assess whether the use of structured abstracts has changed the overall reporting of design
and operational characteristics (descriptors) of randomized clinical trials (RCTs).
Design: We hand-searched 2 ophthalmology journals for reports of RCTs published in the year preceding,
the year of, and the year following the requirement for structured abstracts (1992, 1993, and 1994 for
Archives of Ophthalmology; 1991,1992, and 1993 for Ophthalmology). We excluded reports stating that no design
characteristics were reported because they had previously been published. We extracted RCT descriptors, as described
in the Consort Statement (JAMA. 1996;276:637-639) from each report, and compared reporting of these descriptors in RCTs
using structured abstracts to RCTs not using structured abstracts.
Results: We found 97 published reports of RCTs fulfilling our criteria, 46% (45/97) with a structured abstract
and 54% (52/97) with a nonstructured abstract. Those with structured abstracts appeared only in the year of (n=17)
or following (n=28) the journal requirement. Reporting of descriptors did not generally vary by abstract type,
except in a few cases. Reports with structured abstracts more frequently included either "random" or "trial" in the
title: 64% (29/45) vs 19% (10/52)); described their statistical methods more often: 87% (39/45) vs 67% (35/52)); and
used intention-to-treat analyses more frequently: 22% (10/45) vs 8% (4/52).
Conclusions: The presence of a structured abstract is associated with better reporting of a few RCT descriptors.
Return to Session Information
