Article type
Year
Abstract
Background: Conference abstracts provide early results of research, and may be the only information available for systematic reviewers of new health technologies.
Objectives: We compare data presented in conference abstracts of randomised controlled trials (RCTs) of sirolimus-eluting stents to that in subsequent peer-reviewed publications to assess differences in the results.
Methods: Abstracts of three major cardiology conferences between 1999-2003 were handsearched for RCTs of sirolimus-eluting stents. For each RCT, the sample size, endpoints, baseline and follow-up data reported in the earliest abstract was compared with the final publication.
Results: Fourty-six abstracts relating to four RCTs were identified. Three of these RCTs were subsequently published in peer-reviewed journals. The earliest abstract for each trial was compared with the publication.
For all three RCTs, the abstract reported incomplete patient numbers. Two abstracts mis-stated primary and/or secondary endpoints, and these were not described in the third. Baseline data in two abstracts were minimal but consistent with the published report, and were more extensive but differed to the publication in the third. Follow-up results were not described in these early abstracts. Differences between data presented over time will presented.
Conclusions: This analysis suggests that early abstracts will not provide evidence of efficacy, and that information about endpoints and baseline data may change substantially. Implications for those conducting systematic reviews of new health technologies, for whom the systematic identification of RCTs from conference abstracts is a time-consuming task, will be presented.
Objectives: We compare data presented in conference abstracts of randomised controlled trials (RCTs) of sirolimus-eluting stents to that in subsequent peer-reviewed publications to assess differences in the results.
Methods: Abstracts of three major cardiology conferences between 1999-2003 were handsearched for RCTs of sirolimus-eluting stents. For each RCT, the sample size, endpoints, baseline and follow-up data reported in the earliest abstract was compared with the final publication.
Results: Fourty-six abstracts relating to four RCTs were identified. Three of these RCTs were subsequently published in peer-reviewed journals. The earliest abstract for each trial was compared with the publication.
For all three RCTs, the abstract reported incomplete patient numbers. Two abstracts mis-stated primary and/or secondary endpoints, and these were not described in the third. Baseline data in two abstracts were minimal but consistent with the published report, and were more extensive but differed to the publication in the third. Follow-up results were not described in these early abstracts. Differences between data presented over time will presented.
Conclusions: This analysis suggests that early abstracts will not provide evidence of efficacy, and that information about endpoints and baseline data may change substantially. Implications for those conducting systematic reviews of new health technologies, for whom the systematic identification of RCTs from conference abstracts is a time-consuming task, will be presented.