Oral presentation bias: a retrospective cohort study

Area of Science:

• Biomedical Research Ethics
• Scientific Communication
• Research Methodology

Background:

• Oral presentations at scientific meetings are a key method for disseminating research findings.
• The potential for bias in the presentation of research results, similar to publication bias, is a growing concern.
• Understanding oral presentation bias is crucial for accurately interpreting research landscapes.

Purpose of the Study:

• To assess the extent and characteristics of oral presentation bias in biomedical research at a national level.
• To compare the dissemination strategies of research findings, including oral presentations and publications.
• To identify factors associated with the oral presentation of research results.

Main Methods:

• A retrospective cohort study analyzing 248 completed biomedical research protocols approved by French research ethics committees in 1994.
• Data collection involved extracting initial protocol characteristics and follow-up information on results dissemination via questionnaires to principal investigators.
• Statistical analysis, including adjusted odds ratios, was used to identify variables associated with oral presentations.

Main Results:

• Approximately half (49%) of the completed studies were orally presented.
• Confirmatory results showed a significantly higher likelihood of oral presentation (adjusted OR = 6.4).
• Factors such as study scope, investigator status, adverse events, and interim analyses were also associated with oral presentation rates.

Conclusions:

• Scientific meetings may present a biased representation of research, with a tendency to favor confirmatory findings.
• Oral presentation bias appears to be a significant issue in scientific communication, potentially exceeding publication bias.
• Strategies for disseminating research results should consider the potential for oral presentation bias.