University Tutorial: Advanced Literature Review Methodology and PRISMA Implementation

Incorporating Feedback from Assessment $1$

• The primary goal of current tutorials is to facilitate the evolution of research projects using feedback from peers and instructors regarding assessment item $1$.

• Feedback is intended to refine the research question, reshape specific objectives, or strengthen the overall search strategy.

• Adjustments made to the question or strategy following assessment $1$ are expected and encouraged; students will not be penalized for changes made to their original proposal when submitting assessments $2$ and $3$.

The PRISMA Framework and Protocol Development

• The PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) provides a step-by-step checklist to ensure methodological rigor.

• Different PRISMA checklists exist for various review types, specifically scoping reviews and systematic reviews.

• A standard title for a review must explicitly indicate the review type. For example: "Features of Neurodiversity: A Systematic Review."

• The Abstract section is typically for written journal articles and may be bypassed if the final submission is solely a PowerPoint presentation.

Formulating the Mockup Research Question and Rationale

• Mockup Research Question: "What are the documented features of neurodivergence and their diverse presentations?"

• Rationale for the Review:    - Neurodivergent individuals (encompassing autism, ADHD, dyslexia, etc.) present with high variability.    - There is a lack of a comprehensive, centralized database of neurodivergent features, which acts as a barrier to identifying specific needs and accommodations.    - Practical application: The findings could inform the creation of an app or questionnaire to help individuals model their own brain presentation and identify necessary supports (e.g., movement breaks for children, standing desks or walking meetings for adults).    - The rationale must state why the topic matters and be supported by references from existing literature.

Strategic Database Selection and Search Construction

• Database selection depends on the breadth of the research question and where relevant literature is likely to be published.

• Example Databases for Neurodivergence:    - Psychology: PsychARTICLES or similar psychiatric databases.    - Medical: MEDLINE.    - Allied Health: CINAHL (referred to colloquially as "Sonal").    - Education: ERIC (relevant for school-based accommodations).

• Search String Layers:    - A standard search strategy often employs $3$ to $4$ layers connected by the "AND" operator.    - Adding more layers increases specificity but reduces the number of results returned.    - Fewer layers results in a broader, more comprehensive list of papers.

• Boolean Operators and Syntax Tips:    - Use an asterisk ($*$) as a wildcard to capture different word endings (e.g., feature* for "feature" and "features"; autis* for "autism" and "autistic").    - Use inverted commas for exact phrases (e.g., "attention deficit hyperactivity disorder") to prevent the search engine from looking for the words individually.    - Synonyms should be grouped in the search string (e.g., features, characteristics, symptoms).

Inclusion and Exclusion Criteria

• Establishing clear criteria is essential to justify why certain papers were included or omitted from the final review.

• Common Inclusion Criteria:    - English language publications (pragmatic limit based on researcher resources).    - Access to full-text articles (the library may assist in bypassing paywalls, but availability is a factor).    - Specific age ranges (e.g., all ages vs. under $18$).    - International perspectives.

• Common Exclusion Criteria:    - Publication date limits: For the neurodivergence mockup, the cutoff is post-$2013$ because that year marked the release of the DSM-$5$, which significantly changed diagnostic categories for autism.    - Grey literature: Excluding personal anecdotes and non-peer-reviewed sources to maintain focus on formal research data.    - Study types: While systematic reviews found during searching provide useful reference lists for citation searching (primary source mining), they are typically excluded from the final list of included papers to avoid secondary data extraction with existing slants.

Practical Demonstration: Library Database Searching (MEDLINE)

• Keyword searches scan the entire text, while title searches are more restrictive and often yield more relevant results.

• Initial searches are iterative; they serve to test if the search string is effective or if it generates thousands of irrelevant hits.

• Managing High Result Counts:    - Searching only the "Title" field instead of "All Fields" can reduce results from tens of thousands down to a manageable number.    - Use the "NOT" operator with caution to exclude distracting topics (e.g., NOT drugs or NOT pharmaceuticals) if a search on behavioral features is returning too much medical/pharmacological data.    - Check for terminology variations in initial results, such as slashes (e.g., ADHD/Autism), and incorporate them back into the refined search string.

Software for Review Management: Covidence

• Covidence is a specialized tool for managing systematic and scoping reviews. It offers a $7$-day free trial, which should be timed to coincide with high-intensity screening periods.

• The Review Workflow in Covidence:    1. Import References: Upload the final search results from selected databases.    2. Remove Duplicates: The software automatically identifies and removes identical articles found across multiple databases.    3. Title and Abstract Screening: Reviewers quickly tag papers as "Yes," "No," or "Maybe."    4. Full Text Review: Detailed reading of remaining articles to ensure they strictly meet eligibility criteria.    5. Data Extraction: Pulling specific numbers, findings, and methodologies from the final set of papers.

• The PRISMA Flow Diagram: Covidence automatically generates this flowchart, documenting exact numbers (e.g., $1,336$ records identified, $261$ duplicates removed, $917$ excluded at screening, etc.) until the final study count (e.g., $22$ studies) is reached.

Quality Appraisal and Data Characteristics Table

• Critical Appraisal: Assessing the methodological quality of the research (e.g., using the CASP tool). This involves evaluating sample sizes, follow-up testing, and potential biases.

• Deciding on Inclusion Based on Quality:    - If many high-quality papers exist (e.g., Randomized Controlled Trials), lower-quality papers may be excluded.    - If the topic is niche with few publications, low-quality studies may be included with a caveat in the results regarding their limitations.

• Study Characteristics Table: This table appears at the start of the Results section to provide a snapshot of the final papers. Columns typically include:    - Author and year of publication.    - Country of origin.    - Participant characteristics (sample size, ages).    - Data collection methods.    - Study methodology.

Questions & Discussion

• Question Regarding Systematic Reviews: If a systematic review is found that is relevant but not the exact same question, should it be included?    - Response: No, typically you exclude the review itself. Instead, perform "citation searching"—inspect the review's reference list to find the original primary studies and include those in your own review.

• Question Regarding Search Limits: Is it okay to have different strategies for different databases (e.g., filtering for only RCTs in Cochrane but using broader searches in ERIC)?    - Response: Yes, as long as it is documented clearly in the "Methods" section. For scientific transparency, anyone should be able to replicate the exact search string in that specific database to find the same results.

• Question Regarding Documentation: Do students need to document every single refinement step from $30,000$ papers down to $10$?    - Response: No. Only the final search strategy and final numbers need to be documented in the PRISMA flow diagram.

• Question Regarding Quality Appraisal: When does quality assessment happen?    - Response: Usually after you have identified your final set of included papers (e.g., the last $20$ or so articles). This appraisal informs the analysis and helps determine how much weight to give to the study's findings.

• Question Regarding Appraisal Representation: Must all included articles appear in the appraisal table?    - Response: Yes. If you have $18$ final articles, the PowerPoint should have a table summarizing the strengths, limitations, and overall quality for all $18$. However, only one full, detailed appraisal for a single article is required in the appendix.