Comparing the performance of narrow vs. broad search strategies when using machine learning-based software for title/abstract screening

Key Points

• This research aims to evaluate the balance between search sensitivity and specificity when using machine learning software for title and abstract screening.
• Comparison of narrow and broad search strategies using machine learning tools.
• Evaluation of time savings achieved through semi-automated screening.
• Consideration of factors affecting sensitivity and specificity in search strategies.
• Broader search strategies may require more screening time despite potential sensitivity benefits.
• Narrow search strategies can be more efficient but may miss important titles/abstracts.

Abstract