How to conduct a systematic literature review (SLR)

What is a systematic literature review (SLR)? How is it different from other kinds of reviews?
A systematic review is a meta-analysis of previous research within the scope of a particular research topic. It makes an effort to gather all available information on a certain subject in order to address a specific research issue. Before beginning the systematic review, the authors develop criteria for determining whether evidence should be included or eliminated. This helps mitigate the possibility of bias and increases the reliability of the study’s results.

Systematic reviews should have the following characteristics:

• they should clearly state their objectives and employ an explicit and reproducible methodology
• they should attempt to define all studies that fulfilled the inclusion criteria through a detailed search strategy
• they should assess the validity of the included studies’ findings
• they should synthesize the studies’ findings in a systematic manner

An overview of a systematic literature review

Conducting a systematic review is a significant undertaking. It may even appear daunting at first as you gain a sense for the road ahead. However, the difficulty is rewarding. We conduct systematic reviews in order to ascertain the answer to a critical clinical issue. Additionally, because high-quality evidence may make a significant impact in the lives of real people.

Assistance is on the way. A systematic review is frequently done in collaboration with others due to the breadth of information, skills, and experience required to accomplish it effectively. The second good news is that a variety of practical tools have been created over time to streamline and simplify the review process.

Systematic reviews are highly structured and adhere to a consistent methodology. The procedure may be divided into smaller, more doable stages. So pull up a chair, put yourself at ease, and let us go through the major steps. The steps are detailed hereafter.

Formulate the research questions

The first stage is to formulate a research topic. All subsequent choices will be made in light of the research topic, therefore it is critical to consider carefully, consult extensively, and clarify precisely.

The review should address an area of knowledge deficiency. Check the PROSPERO database to see what systematic reviews are currently available. Consider who will read the review. Discuss with these individuals what info they require and why. This ensures the review is useful.

Consider the context. Choosing between a wide and a tight emphasis will be determined by a variety of variables. A wide scope enables a thorough piece of writing but necessitates considerable resources. While a limited focus may be more manageable, it runs the danger of obtaining insufficient data to create a meaningful synthesis. If the review is expected to generate a significant quantity of evidence, consider utilizing a platform to organize the study’s data, record decision-making, and monitor progress.

Question frameworks may be very helpful in providing direction. The framework you choose will be determined by the area of study, the kind of the issue, and the kind of data you anticipate to discover. While the acronym PICO (patient, intervention, comparison, outcome) is often used in clinical medicine, the acronym SPIDER (sample, phenomena of interest, design, evaluation, and research type) may be used for qualitative evidence evaluations. But there are many others!

Establish criteria for inclusion and rejection

The review question determines the inclusion and exclusion criteria. Here are the most critical aspects of your review.  If your inquiry is a PICO, you will now identify the patients, treatments, and contrasts you will examine. Typically, outcome reporting is not utilized to determine research eligibility. This is because doing so would exclude research that measured the result but did not disclose it. The research authors’ selective reporting may add biases into the review.

Additionally, these criteria will be dependent on how you want to organize the data for subsequent synthesis throughout the evaluation process (the final step). This is to ensure that the search returns the data necessary for doing the synthesis. Your choices will be documented in a protocol or study strategy, which will also include your data synthesis procedures. In certain applications, you can store your screening criteria and see them beside the search results for easy reference during the screening process.

Conduct a literature search

Conducting literature searches is a specialized skill. Yes, it is complex. However, you may create an efficient search strategy with the assistance of a library or information expert. How are you going to compile a list of all possibly relevant studies? How many papers are sufficient? Which software applications are included in your institutional registration? Your librarian will have the answers to all of these questions and more. Proof (if there was any doubt) that not all heroes do not wear capes.

The search has to be be very precise. It has to get back the maximum number of possibly relevant research. Regrettably, increased sensitivity comes with the cost of overall accuracy. Your task is to identify an suitable tradeoff that captures the information you require, eliminates the information you do not require, and spares you from the destiny of endlessly searching through irrelevant references.

Extract research publications

Are you curious as to what info the search uncovered? While this is an exciting period, it can also be very time demanding. You may be feeling concerned about the volume of data you currently have to go through. Not to fret: there is a simple two-step procedure for this.

To begin, you will read the titles and abstracts of each citation and determine whether or not the research presented is possibly admissible. If you are uncertain about a specific research due to a lack of information, you will retain it in the blend for the time being.

Second, you will examine each research in more detail by screening the entire text. This extra information will assist in determining eligibility for those who were unsure during the first screen.
These stages may be readily handled with special software, which records decisions and facilitates team communication. For instance, when team members make incompatible judgments regarding the inclusion or removal of research, some software can record the disagreement and generates an alert, allowing the issue to be examined and addressed promptly and openly.

Data extraction

Until you can work on these, the information from the various studies must be extracted in a consistent and trustworthy manner. Here, the data extraction form is your buddy. The precise format will vary according on the data type and nature of the inquiry. The Cochrane Handbook states that the key to success is “creating simple-to-use forms and collecting adequate and clear data that faithfully reflect the source in a structured and organized way.”

It is not uncommon to come across several published reports on the same research. Multiple reports must be connected in order to prevent counting the very same data again. You may quickly and simply combine reports into a single study as part of a data extraction process with automated software.

You will gather data in accordance with the procedure or study plan. Additionally, the extraction procedure shall adhere to applicable reporting standards and the PRISMA declaration. You can simply build, customize, and fill data extraction tables. Having the data and tables in about the same tool simplifies and speeds up the process.

Evaluate the quality

Let us discuss bias. Certain aspects of randomised trials’ design (insufficient blinding, for instance) have been shown to skew their findings uncertain instances. This is problematic because skewed findings may underestimate or exaggerate the intervention’s actual impact. To minimize this and other kinds of bias, reviewers have to thoroughly evaluate the quality of each trial included in the review and, in specific, the possibility of bias in the trial’s findings.

Whether you utilize Cochrane’s risk of bias instrument or create your own evaluation template, special software can help with this. You may create a collection of bias domains that enables a systematic evaluation of the planning, implementation, and presentation of the included research. You may modify these areas as you go, store work in progress and return to it later, and simply verify that review team members have fulfilled their assigned responsibilities.

Perform a synthesis and present all of the results obtained

We are almost there. It is now time to provide a summary of the study’s findings. Here, you will summarize the features of the research you have included and combine the data from the studies to create something that is more valuable as a whole than sum of the parts.

The manner wherein you report your findings will vary depending on the kind of data, the type of analysis, and the relevant reporting requirements. This will have been stated in the procedure or study plan, but there will be unanticipated problems that you will have to address, justify, and explain in the review.

You might do a meta-analysis, a statistical method for combining the findings of several studies. Meta-analysis may be beneficial in terms of increasing overall accuracy. However, it may be unreliable if done without appropriate regard for, yes, bias. At step 6, we discussed the possibility of bias in research outcomes. There is now an added danger of bias being introduced into the synthesis’s findings.