Machine Learning Methods for Systematic Reviews:: A Rapid Scoping Review

Area of Science:

• Medical research
• Computer Science
• Data Science

Background:

• Natural language processing (NLP), a subfield of machine learning, analyzes textual data.
• NLP has significantly contributed to synthesizing data from diverse biomedical studies.
• This review explores NLP's role in optimizing systematic reviews within medical research.

Purpose of the Study:

• To investigate the impact of machine learning on systematic review methodologies.
• To assess improvements in search strategies, article selection, and data abstraction using ML.

Main Methods:

• A rapid scoping review methodology was employed.
• A comprehensive literature search was conducted across multiple databases, including grey literature and handsearching.
• The search concluded on April 14, 2023.

Main Results:

• The initial search identified 23,190 studies.
• After removing duplicates, 117 studies (1.70%) met the inclusion criteria for the review.

Conclusions:

• Machine learning methods are being developed to assist all stages of systematic reviews.
• Integrating ML tools with human expertise can enhance efficiency and speed up evidence generation.
• Faster evidence dissemination through improved systematic reviews supports timely decision-making.