Blog post written by Dr. Andrew Booth, University of Sheffield, September 2021
When a review of research is urgently needed to inform decision-making, and time and resources are scarce, the rapid review methodology can fit the bill.
Transform Nutrition West Africa has applied rapid reviews to synthesize the evidence on nutrition in the region. In collaboration with methodological experts, the project developed and compiled the procedures, templates, and relevant resources required to undertake a rapid review within a resource bank. Here, Dr. Andrew Booth, a senior methodology expert from the University of Sheffield, who contributed to this publication, introduces what you need to know to carry out a rapid review.
A 7 min read
What is a rapid review?
“A rapid review is a form of knowledge synthesis that accelerates the process of conducting a traditional (i.e., conventional) systematic review through streamlining or omitting specific methods to produce evidence for stakeholders in a resource-efficient manner” (Hamel et al. 2020).
Rapid review methods follow those for a systematic review, but to speed a typically lengthy process, these reviews rely on predetermined procedures, standardised descriptions, and reusable templates. Rapid reviews often employ labour-saving software and previous review documents, adapted for each review. In short, much detailed thinking on the design and conduct of the review is done beforehand, so that a review team “hits the ground running.”
Rapid reviews, however, are not characterised by their speed of production alone. Equally important is the degree of interaction with stakeholders, which results in an output well-tailored to the needs of the commissioners and intended users of the review. Because rapid reviews will streamline or omit some methods used by other reviews in order to speed the process or reduce the costs, there must be trust and communication between the review team and those commissioning the review. To ensure transparency, these changes and omissions should be thoroughly documented (Hamel et al. 2021) and the reviewers must outline any implications for confidence in the final review product. It is important to that, although their names are often used interchangeably, scoping reviews and mapping reviews are not the same as rapid reviews. Scoping and mapping reviews may be conducted over extended periods of time and serve as the basis for extended follow-up research. Rapid reviews, however, often employ scoping and mapping processes to target limited time and people resources efficiently.
How do rapid reviews support decision-makers in low- and middle-income countries?
Rapid reviews are useful when (Tricco et al. 2017):
- Decision-makers require a solution to an urgent problem, or they need to make a decision within a timeframe shorter than the 12–14 months required by a typical systematic review.
- Resources are not available to support a full systematic review, yet a limited evidence-based response is required.
- Decision-makers need a provisional answer to inform action that will take place while a full systematic review is being produced.
- Decision-making takes place against the background of a rapidly unfolding situation, such as an emergency or pandemic, and/or where the volume or content of the available evidence is likely to change rapidly.
- Decision-makers want to address areas of uncertainty rapidly before proceeding with subsequent research or policy action.
- Decision-makers are unsure whether available evidence on a topic is sufficient to sustain a systematic review.
What is the difference between a rapid review and a systematic review?
Rapid review refers to a rigorous and targeted review methodology in its own right. Rapid reviews typically start from a systematic review template, adapted to meet a specific decision-making window. Rapid reviews should employ the same knowledge synthesis skills required by a systematic review. Typically, they are conducted by teams of experienced reviewers who are experienced and knowledgeable about shortcuts and their implications. Rapid reviews are frequently produced by academic centres, development programmes, and government units that either specialise in these reviews or regularly move within a mixed portfolio.
The “rapid review” label should not be used to justify a poorly or carelessly performed review. Nor should a rapid review be conducted if time and resources are available to perform a full systematic review.
How to conduct and report a rapid review
In conducting a rapid review, you will produce two key documents: (1) the review protocol and (2) the review report itself. Reporting standards exist for both these documents, based on systematic review procedures. PRISMA-P (http://www.prisma-statement.org/Extensions/Protocols.aspx) lists requirements for producing a protocol, along with an explanation, examples, and a checklist to monitor and document adherence. PRISMA (http://www.prisma-statement.org/PRISMAStatement/) provides the same three components for a final review. Because the protocol will be the methods section of the final review, some review teams use a combined protocol/review template for consistency and to fast-track the writing process.
Plans are currently underway for a reporting guideline for rapid reviews (PRISMA-RR) but, for the moment, the best approach is to follow PRISMA-P/PRISMA reporting guidelines and include a description of any limitations from streamlined or omitted methods. If the primary purpose of the rapid review is scoping or mapping, you could follow specific PRISMA-ScR (scoping review) guidelines (http://www.prisma-statement.org/Extensions/ScopingReviews).
The Transform Nutrition West Africa Rapid Reviews Resource Bank offers an easily navigable ten-step process to guide researchers and decision-makers in conducting a rapid review. More details on these steps can be found in the TNWA Rapid Review Resource Bank. It also provides a set of eight useful templates to guide different steps of the process.
In addition, the Cochrane Rapid Reviews Methods Group recently produced 26 recommendations (R1–26) organised around the steps of a rapid review (Garrity et al. 2021). This source offers additional, detailed information on the methods summarised below.
Stepwise approach to conducting a rapid review (TNWA Rapid Review Resource Bank)
[Read More]
Step 1: Topic selection and refinement and Step 2: Setting the Research question
A successful rapid review will involve key stakeholders (e.g., review users such as consumers, health professionals, policymakers, and decision-makers) to set and refine the review question, eligibility criteria, and outcomes of interest. Consultation with stakeholders should continue throughout the review process, particularly when changes are proposed (R1).
Step 3: Protocol
The protocol should document review questions, question components, and inclusion and exclusion criteria. You should be able to amend and adapt the published protocol (R23; R24) in response to all post hoc changes (R25).
Step 4: Setting eligibility criteria
Key stakeholders should help the review team decide what to include and what is a manageable number of interventions (R2) and comparators (R3). Similarly, the review should focus on the outcomes most important for decision-making (R4). Date restrictions should ideally have a clinical or methodological justification (R5), as should any restrictions by setting (R6). Reviews should usually default to English language only, the primary language of scientific communication, extending only to other languages if justified (R7). Repurposing data collected for existing systematic reviews can save time (R8), as can focusing on research with higher-quality study designs (e.g., systematic reviews or RCTs). Any extra workloads associated with additional study designs should be discussed with the team and the commissioners (R9).
Step 5: Searching for evidence
An information specialist should be involved to support a rigorous and timely response. The number of main databases (e.g., Cochrane CENTRAL, MEDLINE [e.g., via PubMed], and Embase, if available, for clinical trials) (R10) can be limited, and the team can opt to search just one or two specialized databases (e.g., PsycInfo and CINAHL) for particular topics or omit these if time and resources are limited (R11). At least one search strategy (e.g., MEDLINE) should be peer-reviewed to highlight problems in terminology or search logic (R12). Grey literature and supplemental searching should be used only if you are looking for elusive types of literature (R13). Online systematic review software (e.g., Covidence, DistillerSR, or EPPI-Reviewer) can be used to streamline processes (R26). If there is time left after screening abstracts and full texts, the team should search study registries or scan reference lists of other systematic reviews or included studies.
Step 6: Screening and study selection
Your standardized form for title and abstract screening should be piloted by the entire screening team across the same 30–50 abstracts. Two reviewers then dual screen a minimum of 20 percent of abstracts, resolving all conflicts. One reviewer then screens the remaining abstracts, while a second reviewer screens all excluded abstracts to identify any that have been mistakenly discarded (R14).
Full-text screening also uses a standardized screening form, piloted on five to ten full-text articles by the entire screening team. One reviewer should screen all included full-text articles, while a second reviewer screens all excluded full-text articles (R15).
Step 7: Data extraction
A single reviewer pilots the data extraction form, with a second reviewer to check for correctness and completeness (R16). Data extraction can be limited to a minimal set of data items (R17). Data previously extracted into existing systematic reviews can be repurposed to reduce time spent on data extraction (R18).
Step 8: Quality assessment
A valid risk of bias tool should be used to assess quality, if available. One reviewer should rate risk of bias, and have it verified by a second reviewer (R19). Risk of bias ratings can be focused on outcomes most important to decision-making (R20).
Step 9: Knowledge synthesis
The default method of synthesis will be narrative synthesis (that is, using text, tables, and graphics for presentation). Accepted methods of meta-analysis may also be appropriate (R21). A single reviewer can grade the certainty of evidence, provided that all judgments (and footnoted rationales) are verified by a second reviewer (R22).
Step 10: Report production and dissemination
Although you should be thinking about your audience and purpose throughout your rapid review, this is particularly important for the production of the Rapid Review report to identify the end users and the most appropriate channels for dissemination. While there may be may have opportunities to publish the rapid review as a formally peer-reviewed journal article, it is equally likely that the focus will be on producing a tailored report that is more readily, and more speedily, available to those working at the front line.
What else is required
Of course, it is not enough to simply follow the steps outlined within a rapid review “cookbook” – sufficient time, research skills, and project management expertise are equally necessary. Within Cochrane, we have crafted the acronym RETREAT as a memory aid for what is needed (Noyes et al. 2021). Into the ‘bodywork’ of:
Research question-Epistemology (knowledge view)-Audience and purpose-Type of data
we insert the “engine” of: Time-Resources-Expertise
Over several years, we have worked with researchers across Africa to build up review capacity (time, resources, expertise) using multiple training techniques. Techniques have included attendance at our Sheffield courses, e-learning via virtual classroom sessions, mentoring, and tailored problem-solving clinics. An experienced facilitator is essential for working with the local teams and then brokering their learning and problem-solving needs with a review methodologist through “pop-up” consultancy sessions. Over time, the teams have become more proficient and self-sufficient, the sessions have become less frequent, the production more rapid, and the content of discussions more sophisticated and advanced.
Top Tips
Top tip 1: Keep the review commissioner informed of key decisions and changes to the review plan.
Top tip 2: If publishing a rapid review as a journal article, only use that label in the title if processes have been streamlined or omitted. If a full scoping, mapping or even systematic review has been produced rapidly, then use that specific label in your title.
Top tip 3: Follow the Cochrane Rapid Review Methods Group recommendations and stay up-to-date with emerging methods via their web pages: https://methods.cochrane.org/rapidreviews/cochrane-rr-methods
Top tip 4: Use PRISMA-P and PRISMA documents as templates to structure your rapid review.
Dr. Andrew Booth, Cochrane Qualitative and Implementation Methods Group, University of Sheffield: School of Health and Related Research (ScHARR) and Course Co-Lead ScHARR Rapid Reviews Course
References
Garritty, C., G. Gartlehner, B. Nussbaumer-Streit, V.J. King, C. Hamel, C. Kamel, L. Affengruber, and A. Stevens. 2021. “Cochrane Rapid Reviews Methods Group Offers Evidence-Informed Guidance to Conduct Rapid Reviews.” Journal of Clinical Epidemiology 130: 13–22.
Hamel, C., A. Michaud, M. Thuku, L. Affengruber, B. Skidmore, B. Nussbaumer-Streit, A. Stevens, and C. Garritty. 2020. “Few Evaluative Studies Exist Examining Rapid Review Methodology across Stages of Conduct: A Systematic Scoping Review.” Journal of Clinical Epidemiology 126: 131–40.
Hamel, C., A. Michaud, M. Thuku, B. Skidmore, A. Stevens, B. Nussbaumer-Streit, and C. Garritty. 2021. “Defining Rapid Reviews: A Systematic Scoping Review and Thematic Analysis of Definitions and Defining Characteristics of Rapid Reviews.” Journal of Clinical Epidemiology 129: 74–85.
Noyes, J., A. Booth, M. Cargo, K. Flemming, A. Harden, J. Harris, R. Garside, K. Hannes, T. Pantoja, and J. Thomas. 2021 [update]. “Chapter 21: Qualitative Evidence.” In Cochrane Handbook for Systematic Reviews of Interventions, Version 6.2., 525–545. https://training.cochrane.org/handbook/current/chapter-21 .
Tricco, A.C., E.V. Langlois, and S.E. Straus. 2017. “Rapid Reviews to Strengthen Health Policy and Systems: A Practical Guide.” World Health Organization. Alliance for Health Policy and Systems Research. http://www.who.int/alliance-hpsr/resources/publications/rapid-review-guide/en/.
Verstraeten, R., L. Salm, and A. Booth, 2021. Transform Nutrition West Africa Rapid Reviews: A Resource Bank [interactive web page]. Dakar: International Food Policy Research Institute (IFPRI). https://www.ifpri.org/interactive/resourcebank
