What is Sysrev?
Sysrev is a web application that assists users in conducting a review of available research on a topic. Once a user has defined a question they want to answer, Sysrev facilitates the search for relevant data and the extraction of information from identified literature. It does this by providing an interface that enables users to perform reviews more efficiently, by using AI to optimize the identification and tagging of relevant data, and by providing tools that drastically improve the ability to collaborate with others.
Users create sysrev projects to upload and extract information from articles. Public projects are visible to the world, google and other search engines can discover public projects and anybody on the internet can download the information in a public project. Private and public projects have the same workflows, but private projects cannot be accessed by non-members.
After a researcher has defined a question that they want to investigate, they have to identify relevant research. Traditionally, this was performed by searching through journals and the citations/sources in other relevant articles; scanning through thousands of articles to find a few relevant ones, while often missing many. Today, researchers often use keyword searches in online databases (similar to a google search). This is much a more efficient way to search, both reducing the need to search through many articles that are very unlikely to be relevant, and reducing the likelihood of missing relevant articles.
Sysrev is a next iteration. It allows users to conduct a keyword search within the web application or to import a body of literature they have gathered on their own. Users then identify the relevant data within the platform, electronically tagging an article as relevant or not, and immediately having the next article displayed for review. This system allows users to review articles much more quickly. Additionally, Sysrev's machine learning algorith will use these responses to build a prediction model that assigns a score to all the literature within the project, predicting the likelhood it will be relevant to the researcher. The researcher can use the prediction model to prioritize reviewing articles scored as most likely to be relevant. Many researchers will choose not to review articles with low scores. In some instances, this may allow researchers to reduce their review by as much as 90%.
Sysrev users have the ability to apply labels, annotations, and notes (together referred to as tags) to every article as they conduct a review. These tags can be customized for the specific project. For example, you may create a label for study type, allowing each article to be labeled as a case-control, cohort, or systematic review. Or, you may want to label an article as showing a positive, negative, or neutral outcome to a drug being tested.
Applying these tags during the review process helps users more quickly and accurately extract data from the relevant literature. By creating tags at the project level, you can standardize how tags are applied throughout the review, greatly reducing the risk of errors or omission. Throughout the review, metrics on the usage of the tags will be provided on the overview page, allowing for an effortless overview of the outcomes of the relevant literature throughout the process. And finally, use of the tags makes it much easier to find specific literature when conducting your analyses. Finally, the standardized application of the tags facilitates the .
Sysrev automatically creates visualizations for summary metrics, including any labels or annotations being applied to articles. This allows users to effortlessly monitor the progress of their project while conducting for the review. For example, by seeing the relative frequency of different labels that have been applied so far.
Sysrev was built to enable more effective collaboration between researchers. When starting a project, it is easy to invite colleagues to view or participate in the project. Having multiple reviewers can drastically increase the speed of reviewing literature and applying tags. Some users will set up a project to have each article be reviewed only once, dramatically increasing speed. Other reviewers set up their project to have articles reviewed by multiple people to ensure accuracy. In the case that multiple reviewers disagree on a inclusion or a tag, Sysrev has a conflict resolution system that highlights the conflict to enable the team to reach a resolution
In addition to using collaborators as additional reviewers on a project, some teams have opted to leverage public reviewers. These are people who users can leverage to facilitate the review process. Users can source their own public reviewers, or they can leverage the public reviewers who have signed up with Sysrev to review article on a pay-per-article basis.
Finally, users have the option of making their project public. This will make your project publicly available to view. It is a great way to share your data, increase transparency in your research, and allow other future collaborators to benefit from your work.