To perform a comparison of randomized controlled trial (RCT) abstracts and manuscripts published in recent spinal literature.Summary of Background Data.
RCTs represent the “gold standard” upon which evidence-based treatment decisions are made. Inconsistencies between an abstract and manuscript can mislead readers' interpretation of findings and conclusions. Abstract findings are often cited without reference to the manuscript itself. In other fields of medicine, studies have shown discrepancies between RCT abstracts and manuscripts.Methods.
A literature search of RCTs published in Spine, The Spine Journal, and Journal of Spinal Disorders and Techniques during a 10-year period (2001–2010) was performed. All manuscripts described as randomized trials concerning lumbar spinal surgery were selected. Manuscripts were analyzed using a standardized 21-item questionnaire to collect data regarding inconsistencies or bias in the abstract compared with the manuscript. Abstracts were considered deficient if they contained data that were either inconsistent with the manuscript or if they failed to include important findings from the manuscript. Four reviewers reported on the 40 manuscripts that met the inclusion criteria. Each manuscript was reviewed by 2 reviewers. In the event of conflicts in analysis, resolution was achieved through discussion between the reviewers.Results.
At least 1 inconsistency was found in 75% of studies. Despite the word “randomized” appearing in 75% of titles and 92.5% of abstracts, the method of randomization was not described in 37.5% of manuscripts and (if described) was considered unacceptable in 28%. The primary outcome of the study was clearly stated in only 22.5% of abstracts and 47.5% of manuscripts. Pertinent negatives were not reported in 40% of the abstracts. Relevant statistically significant results were reported in only 60% of abstracts.Conclusion.
Abstracts are discrepant with full manuscripts in a surprisingly high proportion of manuscripts. Authors, editors, and peer reviewers should strive to ensure that abstracts accurately represent the data in RCT manuscripts.Conclusion.
Level of Evidence: 1