Developing Nested Knowledge: 2022 and Beyond

2022 is going to be the year of the living systematic review.

Software Design and Development:

In the early spring of 2020, the two-fold vision of Nested Knowledge became clear: systematic review workflow software, married to a set of visualization tools. 

The workflow software (dubbed AutoLit) was built to reflect tried and true manual review methods, but with a twist. As a user, you could enter our software, create an account, enter a search query, and begin screening studies. In other words, you could accomplish the first two steps in the systematic review process, and you could do so more quickly and more accurately by using our tool than doing the same steps by hand or in excel. 

However, AutoLit also introduced a step that manual best practice did not call for, which turned out to have several, distinct advantages. By tagging concepts of interest in each study, you create a flexible hierarchy of key ideas, which could in turn become an intelligent way to assimilate study outcome data. As you read and tag studies, your hierarchy continues to grow. In that way, each reference in your review becomes multiple qualitative data points, all of which are easily visualized, enabling point and click analysis. Better yet, as new studies are published, the data and insights from those studies can be integrated into your review, and by extension, your visuals. The perfect tool to conduct continuously updatable or “living” systematic reviews was born.

Still, the Nested Knowledge platform had a lot of growing up to do. For one, there was no support for data extraction and the associated visual outputs necessary for a proper meta-analysis, two features that were added by early fall of 2020. That same year, Nested Knowledge entered into a mutually beneficial partnership with ECRI institute: they received early access to our software, and Nested Knowledge gained experienced early users. Armed with additional insights into the workflows of ECRI’s experts, Nested Knowledge pushed many new features with an eye towards speed and accuracy. By the end of that year, RoboPICO automatically highlighted key terms in the screening and tagging stages, dual screening allowed for two passes by two separate reviewers, and Study Inspector allowed users to bypass the sequential process in order to spot-fix issues. Also added: automatic compliance with the PRISMA process, with an automatically updating flow diagram which could be published along with the other visuals. More features were on the horizon, and the utility of the software was clear, but it still had one, glaring issue: it was kinda ugly.

2021 saw increased professionalization across the board, with a standardized UI, and several key new features to ensure the user could conduct every aspect of their review in one platform. Like two bookends, the revised Search Exploration and Manuscript tools allowed users to stay in the Nested Knowledge ecosystem as they completed the beginning and final stages of their reviews. Instead of forcing users to start with an already perfect search query, Search Exploration allows users to enter just a few PICO terms relevant to their research question and explore an expanded set of related terms. By adding related terms of interest, users could generate a usable query automatically, which could then be later refined in Query Builder. On the other end of the process, the Manuscript tool allows users to write and cite on the site. Instead of generating a static PDF with static tables, Manuscript is a cloud-based document, with custom tables, visualizations, and citations that update as the underlying review is updated. 2021 also saw the addition of a Risk of Bias module and set of visuals, including a traffic light diagram.

Looking ahead to 2022, the Nested Knowledge team is excited to bring additional features to the platform over the coming year, namely dual extraction, dual risk of bias, more quality control options, and an advanced organization-level library (including the very exciting “review inspector”). As always, if you have ideas for how we can improve, please leave us feedback.