Dear Editors,
We would like to thank Yan et al. for providing us with an opportunity to expand on the applicability of our 2-week systematic review (2weekSR) processes [1].
Yan et al. correctly point out that “it is important to recognize not all systematic reviews (SRs) are created equal and complexity of the [SR] topic can greatly impact the time spent on the review.” We state on page 89 of the manuscript that our initial 2weekSR had a narrow PICO question and may not be representative of other SRs. We also note that the narrowness was deliberate—this is because the first 2weekSR was intended to be a ‘proof of concept.’
Yan et al. also note that “the authors did not mention or include the time needed to familiarize researchers with these [SR automation] tools. This may possibly be due to the authors’ [sic] prior experience using such tools.” As we state on p.87 under, facilitators and barriers to completing a 2weekSR, experience with the tools was one of the key facilitators of the 2weekSR: “Pre-existing knowledge and ability to use the automation tools (e.g., SRAHelper, Word Frequency Analyser, etc.) eliminated the ‘learning curve’ and time.”
The point of Yan et al., regarding the need for supplementing simulated decisions (made by the automation tools) with manual independent validation, does not apply to most of the tools we used. However, as the Journal’s readers know, the output of RobotReviewer is only partial (4 of the 7 risk of bias domains) and ipso facto requires a manual check, while the RobotSearch tool has been extensively validated [2].
We appreciate the kind commendation of Yan et al. of our 2weekSR academic exercise. We are happy to share that the 2weekSR process has so far been adopted on 5 SRs, which included a variety of SR types (intervention, prevalence, and adverse events) and study designs (randomized controlled trials, observational studies, both). These SRs involved screening of 638 to 5471 references in the title abstract, 16 to 214 references in the full text, and included from 7 to 37 studies. They have involved teams of 4 to 6, with four of the SR teams involving at least 1 team member with only some SR experience, and 1 team also involving a colleague with no prior SR experience. The duration of the 2weekSRs ranged from 9 to 13 days, that is, from less than 2 weeks to 2.5 weeks. Further 2weekSRs are already in progress or currently in planning stages.
As the 2weekSR methodology is particularly suited to the need for quick but robust and evidence-based decision-making, three of the abovementioned 2weekSRs were on COVID-related topics and were provided as input to decision-makers.
The 2weekSR process is undergoing continual refinement and development, and we are planning to offer workshops to train interested colleagues in the approach.