Are published manuscripts representative of the surgical meeting abstracts? An objective appraisal.

Abstracts accepted for presentation at a pediatric surgical meeting usually result in publication. This study attempts to evaluate the relationship of abstracts to the finished, published manuscript. Thirty-three papers published in the Journal of Pediatric Surgery over a 2-year period were randomly selected and compared with their original program abstracts. Only 10 of 33 (30%) had the same title and authors as their abstract. Nine manuscripts underwent revisions of their title and two had different first authors. Nine (27%) had one more author, two had 2 more authors, and 3 had one less author. Only 11 of 33 (33%) had the same data in the paper and in the abstract. Fifteen (45%) had numbers that were mathematically inconsistent with their abstracts. (Range was from 75% fewer cases to 210% more patients.) Ten papers had smaller numbers of patients or experiments while nine had larger numbers than their respective abstracts. Of the seven experimental papers, four had smaller numbers than their original abstract. The conclusions stated in the papers were similar to the program abstracts in 23 of 33 comparisons. In the remaining ten, the conclusions were not only different but rather routinely weaker than in the abstract. In no instance were the data or conclusions stronger. Two papers bore little relationship to their abstract. These facts suggest a need for change in the current evaluation process wherein a highly competitive abstract leads to a presentation, which leads to publication. At the very least, the need is shown for authors to evaluate data completely before submitting an abstract, so that the program committee actually reads the true data.(ABSTRACT TRUNCATED AT 250 WORDS)