Introduction: Systematic reviews are considered the gold standard in collating available evidence related to a specific question and are used to inform policy for health care public health. They are considered to be essential in producing trustworthy guidelines. However, they are time- and resource-intensive undertakings which may not meet the timeline of stakeholders and policy-makers when urgent answers are required. The aims of rapid reviews are to produce evidence reviews in a timely manner, while maintaining rigorous and robust methods. However, to date, the only consensus around a definition of a rapid review is that a formal definition does not exist. Additionally, there is no standardized set of methods for rapid reviews, nor is there a comprehensive review which has compiled empirically evaluated rapid review methods and evaluated the impact of these abbreviated methods. The aim of this doctoral dissertation was to: (i) identify how rapid reviews have been defined in the literature and perform a thematic analysis of these definitions to identify the key themes; (ii) identify and create a repository of empirically evaluated methods abbreviations, and identify any gaps in the research; and (iii) evaluate the reduction in the screening burden and perform of an artificial intelligence and active machine-learning tool in an online systematic review software. Methods: RR definitions: A systematic scoping review identifying rapid reviews published between 2017 and January 2019 was performed. Definitions of rapid reviews were extracted verbatim from these rapid reviews and a thematic analysis was performed to identify the key themes which should be included when defining a rapid review. RR methods: A systematic scoping review identifying formally evaluated rapid review methods abbreviations published from 1997 onward was performed. In order to create a comprehensive repository of rapid review documents, additional studies (e.g., around guidance on conducting rapid reviews, discussing terminology) were identified. All publications were divided into one of four main categories based on the purpose of the publication. Those that formally evaluated rapid review methods abbreviations were mapped to the Methodological Expectations of Cochrane Intervention Reviews (MECIR) to determine if they met these criteria. Lastly, an experimental evaluation was conducted in DistillerSR ® on 10 completed systematic reviews, using the artificial intelligence simulation tool, to measure the reduction in the screening burden and accuracy (i.e., how many relevant records were missed) when prioritized screening using active machine-learning was employed. 26 Results: RR definitions: A total of 204 definitions that could be thematically analyzed were identified in 216 rapid reviews and 90 rapid review methods papers. Eight major themes were identified, with four themes found in 48.5% or more of the definitions: Theme 4: Compare and contrast to SRs (68.1%; 139/204), Theme 2: Variation in shortcut methods (54.9%; 112/204), with Theme 1: Accelerated/rapid process and Theme 6: Resource efficiency rationale tied (48.5%; 99/204 each). This lead to a suggested definition of “A rapid review is a form of knowledge synthesis that accelerates the process of conducting a traditional systematic review through streamlining or omitting a variety of methods to produce evidence in a resourceefficient manner.” RR methods: Ninety rapid review methods papers were identified, of which 14 formally evaluated rapid review methods abbreviations addressing several, but not all, key dimensions related to the conduct of a review. Only a cursory mapping to MECIR criteria was possible, as insufficient information impeded the ability to determine if criteria were met. Active machine-learning prioritization tool: The active machine-learning tool, employing prioritized screening, greatly reduced the screening burden of the 10 systematic reviews that were evaluated. The median percentage of studies required to be screened to identify 95% of the records included at the title and abstract level (true recall @ 95%) was 47.1% (IQR: 37.5 to 58.0%). Among the 5% that were not yet identified as included (i.e., title and abstract false negatives), none were included in the final review, resulting in 100% accuracy. Conclusion: The emergence of rapid reviews, highlighted by the ongoing COVID-19 pandemic, requires consistency in how they are defined, in order to identify and produce a homogenous set of products regardless of the term used to identify them. Producers of rapid reviews also need guidance on which abbreviated methods may be used to keep potential bias minimized. Lastly, active machine-learning is a viable method to reduce the screening burden and was shown to be very accurate.