Abstract 11938: Citation Bias Favoring Studies With Higher Success Rates for Catheter Ablation of Atrial Fibrillation
Background: Studies of catheter ablation for the treatment of atrial fibrillation (AF) are numerous and report widely varying success rates. We sought to determine if citation bias is present, based on the over-citation of studies reporting high success rates.
Methods: We systematically identified observational studies and clinical trials of catheter ablation for atrial fibrillation indexed in PubMed between 1990 and 2012 and obtained manuscript citation counts from the Web of Science citation index (Thomson Reuters, New York, NY). Poisson regression was used to estimate association between study success rate (freedom from AF recurrence) and total citation count, adjusting for sample size, mean duration of follow-up, journal impact factor, and time since publication. We performed separate analyses for studies reporting success rates after single and multiple ablation procedures.
Results: We identified 180 articles meeting our inclusion criteria, consisting of 36,549 participants. Reported success rates ranged from 10% - 92% after single ablations and 31% - 95% after multiple ablations. In unadjusted analysis, a 1-point increase above the mean in reported success rate was associated with 2.6 additional citations for studies of single ablation and 2.5 additional citations for studies of multiple ablations (P <0.0001 for both). After adjustment for covariates, every 10-point increase above the mean in the reported success rate was associated with 4.1 additional citations (2.4-5.7, P <0.0001) among single ablation studies and 1.7 citations (0.03-3.4, P <0.05) among multiple ablation studies. Sample size, journal impact factor, and the time since publication were significant covariates.
Conclusion: We found citation bias present among studies of catheter ablation of AF, with higher success rates associated with a greater number of citations. To readers of the medical literature, AF ablation may seem more effective than the data supports.
- © 2013 by American Heart Association, Inc.