Capture-recapture is a potentially useful method for assessing publication bias

Area of Science:

• Gerontology
• Exercise Science
• Biostatistics

Background:

• Publication bias is a significant concern in systematic reviews of randomized controlled trials.
• It can distort the perceived effectiveness of interventions, particularly in fields like exercise science.
• Assessing publication bias is crucial for the reliability of systematic review conclusions.

Purpose of the Study:

• To evaluate the influence of publication bias on a systematic review examining Progressive Resistance Training (PRT) effectiveness in older adults.
• To compare different methods for detecting and adjusting for publication bias in this context.

Main Methods:

• Systematic review methodology was employed, identifying studies through electronic databases, expert consultation, and handsearching.
• Methods used to assess publication bias included capture-recapture, funnel plot inspection, Egger's regression test, Trim and Fill, and a selection model.
• These techniques were applied to evaluate the robustness of findings regarding PRT's effectiveness in older populations.

Main Results:

• Multiple methods indicated the presence of publication bias.
• Capture-recapture suggested 3 missed studies, while Trim and Fill estimated 16 missed studies.
• Egger's test and funnel plot regression confirmed bias, though a selection model suggested asymmetry might not solely stem from publication bias.

Conclusions:

• The capture-recapture method shows promise for assessing publication bias in systematic reviews.
• Further simulation studies are needed to validate the accuracy of various bias detection methods across different scenarios.
• Consistent results across methods enhance confidence in the assessment of publication bias.