1. Understanding Systematic Literature Reviews (SLRs) and Their Steps [0]

thumbnail A Systematic Literature Review (SLR) is a structured and transparent method for finding, selecting, evaluating, and combining research evidence, which differs significantly from a normal literature review.

1.1. What is an SLR? [0]

  1. Performing an SLR manually can be overwhelming and time-consuming, especially when dealing with hundreds of papers for screening and data extraction. [0]
  2. A normal literature review involves steps like searching Google Scholar, shortlisting relevant studies, reading them, organizing findings, and explaining the research. [8]
    1. However, this approach can lead to questions about the selection process, such as "why did you choose these 30 papers?" or "are you sure you did not miss any important studies?". [11]
  3. An SLR defines a systematic and transparent process for finding, selecting, evaluating, and synthesizing research evidence. [12]
    1. This ensures that any researcher reviewing the SLR can understand exactly how the final studies were chosen. [13]
    2. This transparency is the fundamental difference between a normal literature review and an SLR. [14]

1.2. Six Steps to Conduct an SLR [15]

  1. Define your research question. [17]
    1. This step clarifies what the SLR aims to answer. [18]
    2. A broad topic, such as "treatment of anemia," needs to be narrowed down to a focused research question. [19]
    3. Frameworks like PICO (Population, Intervention, Comparison, Outcome) can help refine a broad topic into a specific research question. [21]
      1. P stands for Population: what is being studied (e.g., "adults with eczema"). [24]
      2. I stands for Intervention: the treatment or technique given to the population (e.g., "biological therapies"). [25]
      3. C stands for Comparison: what the intervention is compared against (e.g., "standard care"). [26]
      4. O stands for Outcome: what is being measured (e.g., "symptom control"). [27]
      5. An example of a focused research question using PICO is: "In adults with eczema, do biological therapies improve symptom control compared with standard care?" [28]
  2. Define the review protocol. [29]
    1. A protocol is a complete plan for conducting the SLR. [30]
    2. It includes decisions on: [31]
      1. Which databases to search (e.g., PubMed, Scopus, Semantic Scholar).
      2. What keywords and search strings to use.
      3. Inclusion and exclusion criteria for studies.
      4. How papers will be screened.
      5. How data will be extracted.
      6. How the quality or risk of bias of studies will be assessed.
      7. How findings will be combined or synthesized (e.g., narrative review, meta-analysis).
    3. These decisions are crucial for the systematic review process and become part of the protocol. [33]
  3. Register the protocol. [35]
    1. Registering the protocol on a public registry, such as PROSPERO (for health and social care) or OSF Registries (broader platforms), "seals" the finalized plan. [36]
    2. This prevents changes to the protocol later based on the results obtained during the review. [36]

1.3. Using AnswerThis AI Tool for SLR Steps [38]

  1. The AnswerThis AI tool can assist with the first two steps: defining the research question and developing the review protocol. [38]
  2. Developing a search strategy and finding studies is where AnswerThis becomes even more relevant. [40]
    1. Users log in, go to "Systematic Reviews," enter their research question, and start the review. [41]
    2. The tool displays eligibility criteria based on the PICO framework, which users fill in with labels, descriptions, and inclusion/exclusion criteria. [42]
    3. Users can also add custom criteria, choose the number of screening reviewers (AI or human), and select databases (e.g., PubMed, Semantic Scholar, ClinicalTrials.gov). [44]
  3. Forming search strings is the next part on AnswerThis. [48]
    1. Researchers need to define search terms and their synonyms because different researchers might use different terminology (e.g., "biologic therapies," "biologic treatment," "biologic agents"). [49]
    2. Synonyms and alternative terms are combined with Boolean operators: [51]
      1. "Or" is used between similar terms (e.g., "biologic therapy or biologic treatment or biologic agent or biologic drug"). [52]
      2. "And" is used between different concepts (e.g., "amna or bronchila AND biologic therapy or biologic treatment or biologic agent AND amna control or symptom control or symptom improvement"). [53]
    3. Traditionally, identifying and refining these search strings is a manual process involving extensive literature review. [55]
    4. AnswerThis automatically generates search strings based on the research question and criteria, which users can review and modify. [57]
  4. Fetching the papers is done by AnswerThis based on the research question, search strings, and selected databases. [59]
    1. Multiple databases are used because no single database contains all papers. [63]
    2. Using multiple databases can result in duplicate papers. [64]
  5. Duplication removal and mechanical filtering is the next step. [65]
    1. AnswerThis removes duplicates by DOI and title. [66]
    2. Papers are also filtered based on publication date, citation count, and missing abstracts. [67]
    3. This process significantly reduces the number of papers; for example, from 1,876 to 532 papers. [68]

1.4. Screening, Data Extraction, and Quality Assessment [69]

  1. Title and abstract screening is the next step to remove irrelevant studies. [70]
    1. For example, a paper titled "Safety of Biologics for the Treatment of ADHD in Children and Adolescents" would be excluded if the research question focuses on adults. [73]
    2. If a paper seems relevant after title and abstract screening, it moves to full-text screening rather than being immediately included. [75]
    3. AnswerThis can reduce the number of studies significantly at this stage (e.g., from 532 to 25 studies). [77]
    4. The tool clearly shows which studies are included or excluded and provides reasons for exclusion (e.g., "Ama Condition," "Adults," "Biologic Therapies"). [79]
  2. Full-text screening is the subsequent step. [81]
    1. Users need to upload the full text of papers, especially those behind a paywall, as AnswerThis might not have access to all of them. [82]
    2. While AnswerThis can perform full-text screening, it is highly recommended that researchers also review the papers themselves to ensure they meet eligibility criteria and to clearly document reasons for exclusion. [85]
    3. The entire process, from initial records to final included papers, is illustrated through a PRISMA flow diagram, which AnswerThis also provides. [88]
  3. Data extraction and quality assessment are performed after full-text screening, once a final set of studies (e.g., 12 studies) is identified. [91]
    1. Data extraction involves deciding what information is needed from the papers and extracting it into structured columns, rather than making random notes. [93]
      1. On AnswerThis, users can define their own variables or choose from recommended extraction variables based on selected papers. [94]
      2. The tool then extracts this information neatly into structured columns. [96]
    2. Quality assessment (or risk of bias assessment) determines if all studies are equally reliable. [97]
      1. For example, a well-designed randomized controlled trial with 500 participants (Study A) is more reliable than a study with 20 participants, unclear patient assignment, and missing data (Study B). [98]
      2. AnswerThis recommends assessment frameworks, such as Cochrane RoB 2.0, based on the research question. [102]
      3. The tool adds columns indicating which studies meet the assessment criteria. [104]

2. Evidence Synthesis and AI Tool Capabilities [106]

thumbnail

Evidence synthesis is the final step in an SLR, where findings from selected studies are combined to answer the research question, and AI tools like AnswerThis can assist throughout the entire process.

2.1. Evidence Synthesis and Reporting [106]

  1. Evidence synthesis is the final step to answer the research question (e.g., "in adults with rheumatoid arthritis, do biologic therapies improve symptom control compared with standard care?"). [106]
  2. There are two main ways to synthesize evidence: [109]
    1. Narrative synthesis: This involves looking at all studies together to find common patterns, similarities, and differences, rather than listing findings one by one. [110]
    2. Meta-analysis: This is a statistical technique that quantitatively combines the results of multiple studies to obtain an overall or pooled effect, if the studies allow for it. [112]
  3. After synthesis, conclusions are drawn based on the quality, consistency, and limitations of the reviewed studies. [114]
  4. AnswerThis provides additional features for synthesis: [116]
    1. Evidence grading: Outcomes are presented in a comparison table called the "Summary of Findings," which combines the main information extracted from the papers. [117]
    2. Meta-analysis: The tool indicates which studies are not pooled and the outcomes found across all studies, considering factors like outcome, comparison, and time point. [120]
    3. Users can export files from AnswerThis at any stage, including screening, extraction, risk of bias assessment, meta-analysis, and forest plots. [123]
  5. AnswerThis can also help write the systematic literature review report. [124]
    1. While generating AI content for writing is generally not recommended, users can generate a report for reference and then write their own. [125]

2.2. Benefits of Using AI Tools for SLR [126]

  1. The SLR process involves significant manual work, including: [127]
    1. Building complex search strings.
    2. Searching across multiple databases.
    3. Screening hundreds or thousands of papers.
    4. Documenting reasons for excluding each study.
    5. Extracting information from each paper.
    6. Assessing the risk of bias.
    7. Synthesizing all findings.
  2. Platforms like AnswerThis are helpful because they consolidate all these tasks into a single platform. [128]
    1. This eliminates the need for multiple tools. [129]
    2. The majority of the manual tasks are handled by the tool, simplifying the SLR process step by step. [129]
  3. The video concludes by offering courses on research writing and mastering AI tools for research, including a discount coupon for AnswerThis. [130]