Reporting quality of abstracts presented at the European Association of Urology meeting: a critical assessment

Abstract

Purpose: We assessed the reporting quality of randomized and nonrandomized, controlled trials presented in abstract form at the European Association of Urology annual meeting in a 10-year period and determined the impact on subsequent publication.

Materials and methods: Abstracts presented at the European Association of Urology annual meetings in 1998, 1999, 2008 and 2009 were retrieved and included in analysis. Two 2-year meeting periods were considered, including 1) 1998 and 1999, and 2) 2008 and 2009. Two standardized forms were constructed based on the CONSORT (Consolidated Standards of Reporting Trials) and STROBE (Strengthening the Reporting of Observational studies in Epidemiology) guidelines, each including 15 and 16 items for randomized and nonrandomized, controlled trials, respectively. Reporting quality was assessed by measuring the proportion of items respected by authors when preparing the abstract, defined as the score ratio. Subsequent full-length publication within 2 years after the meeting was also determined by a PubMed® search. Differences between the 2 periods were analyzed by the chi-square and simple t tests. Predictors of subsequent full-length publication were evaluated by multiple logistic regression using meeting period, topic, country of origin, design, multi-institutional study and the proportion of reported items (score ratio).

Results: A total of 3,139 abstracts were included in analysis, of which 375 (11.9%) were randomized, controlled trials. Overall oncology represented the main topic (49.2% of all abstracts). The score ratio (proportion of adequately reported items in each abstract) was better for period 1 than 2 for randomized, controlled trials (63% vs 57%) but better for period 2 than 1 for nonrandomized, controlled trials (55.4% vs 53.2%, each p <0.001). Abstracts describing multi-institutional studies were more likely to be followed by full-length publication (OR 1.82, 95% CI 1.44-2.30). Other features, including reporting quality (score ratio), did not predict subsequent publication.

Conclusions: The reporting quality of European Association of Urology meeting abstracts did not improve in a decade. Nevertheless, this factor did not impact subsequent full-length publication. Ultimately, the reporting quality of abstracts remains to be improved by following currently available guidelines.