Living Systematic Review with Laser AI

‍Title: Drinking from a firehose: living systematic review with 15,000 new references every month

Date Published: June 2022

German Medical Science: Meeting Abstract

Citation: Nowak , A. et al. (2022) Information retrieval meeting (IRM 2022): Drinking from a Firehose: Living systematic review with 15,000 new references every month, GMS. Available at: https://www.egms.de/static/en/meetings/irm2022/22irm28.shtml (Accessed: 04 November 2024). 

 What is the report about?

In this meeting abstract, the German Institute for Quality and Efficiency in Health Care (IQWiG) discussed a systematic review focused on understanding the baseline risks of thromboembolic complications in hospitalized COVID-19 patients. However, the constant influx of newly published studies demanded a living review process, and the IQWiG team needed to reduce the time spent screening new publications while maintaining accuracy. This abstract looks at how the deduplication and machine learning models helped save time. 

‍How did Laser AI help?

After the initial review, between August 2020 and July 2021, the living review brought in 158,726 new citations to screen. By using automatic deduplication and machine learning features, the authors excluded irrelevant citations and reduced the number to a more manageable number of 37,739.

‍What were the results?

Using machine learning and automatic deduplication results in a more manageable screening workload, which allows vital studies to be identified quickly, contributing to developing living guidelines.