A ‘systematic literature review’ (SLR) attempts to answer big questions, and the explanation lies in the name: systematic. Like any good system, the process for conducting a literature review should be sturdy and robust, it should consider the quality and limitations of the literature, and give thought to the cost constraints of time and labour. But it’s not just assessing words. The objectives of the review, the methods used to search for literature, the outcomes of the search results, and your conclusions for it’s potential in answering your big question, must all be examined. Set criteria is crucial for reducing the risk of bias.
Often, authors consider their reviews to be systematic literature reviews, but they don’t always meet the minimum criteria as set out by PRISMA. To summarize, once your scientific or working hypothesis has been determined, a literature search should be conducted in three stages across two verticals. The three stages are: identification, screening and inclusion. These should be considered when identifying studies via databases and registers, and when using other methods like websites and organisations.
In the case of databases and registers, once records are identified, duplicate records should be removed, appropriate records marked as eligible and then included in the review. In the case of websites and organisations, eligible reports should be clearly marked, reasons provided for excluded reports. PRISMA is clear in stating that excluded records should identify those removed by humans, and those removed by automation tools.
Using a methodical approach
Using a checklist like the PRISMA checklist effectively provides a template for a robust review. It’ll help you cover title, abstract and introduction (including the rationale for review and statement of objective) and then naturally move into appropriate questioning for the method used, including all inclusion and exclusion criteria. Search strategies and limits should be identified and made clear, plus risk and bias assessed and defined. Equally important is clear coverage of results, particularly the reasoning for inclusion and exclusion of records, and a detailed view of all possible outcomes, statistical syntheses, sensitivity analyses and certainty of the evidence.
Types of literature reviews
All literature reviews should be systematic in their approach to finding answers or solutions. They should also maintain high standards. Three kinds of literature reviews are Narrative, Umbrella and Comprehensive.
Four common narrative reviews are general literature, theoretical literature, methodological literature and historical literature. Each has a different purpose with a narrowed focus on specific criteria. They’re an important part of the research process and help to establish a framework from which to identify patterns and trends.
Essentially, this is a review of previous reviews. It examines existing reviews and outlines a top-level summary and helps define research and evidence already known. Umbrella reviews are helpful for highlighting areas where research is still required. The search process used in umbrella reviews is vitally important to ensure the inclusion of correct literature, and to reduce bias.
Accurate, ethical and evidence-heavy, a comprehensive literature review is effective in its use of critical thinking. They search and evaluate extensive results across multiple databases and registers along with websites and organisations. Evaluations are rich in context and show relationships between previous studies and theories. They explore the available literature and then organise the information to allow for quantitative and qualitative analysis. In many cases, the comprehensive review is the first stage of a larger dissertation or project.
Without a clear objective and method-control for your review, there are problems that can rear up with a negative impact. A few of them are:
- Lack of relevance: This can render a review almost useless unless appropriate expert feedback is included.
- Lack of a defined priori protocol: It should be registered somewhere like PROSPERO. Without it, it’s harder to reign in your criteria and objectives, resulting in a total blowout of the review.
- Lack of transparency (for replication purposes): the review should be clear enough to be repeated, if necessary. CEE Guidance, PRISMA and ROSES are helpful.
- Selection bias: caused by a search that wasn’t comprehensive enough to start with.
- Short, rapid-search reviews: these still require defined search strings, appropriate database use, documented methods, analysis and outcomes.
Consider the cost component
Literature reviews can take anywhere from six months to almost two years and be costly. Time, labour and dollars spent all add up quickly, with one SLR on the topic of SLR’s, determined the cost of a single SLR to be close to $150,000.
We have previously outlined the Understanding the Literature Search Process.
Here at Rx we use explicit and reproducible methods to systematically search, critically appraise, and synthesize on a specific issue. Our aim is to answer focused questions to: inform professionals and patients of the best available evidence when making healthcare decisions; influence policy; and identify future research priorities. Get in touch. Our team at Rx are ready to answer all your queries.