AI systematic review PRISMA flow diagram narrative

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-research-ai-systematic-review-prisma-flow-narrative-creators

1. A researcher wants to use AI to help document a systematic review. What is the primary task AI performs when creating a PRISMA narrative?

   1. AI automatically searches databases and retrieves new studies
   2. AI decides which studies meet inclusion criteria based on title and abstract
   3. AI converts screening log data into a written explanation of how many records remained at each stage
   4. AI resolves disagreements between two human reviewers about study eligibility

2. According to the systematic review workflow, at which stage should reason codes be attached to excluded records?

   1. During the protocol design phase
   2. At the time of each exclusion decision
   3. After the narrative is written
   4. Only when the final report is drafted

3. A screening log contains 15 records marked as excluded but with no reason code listed. How should the PRISMA narrative handle these?

   1. Mark them as [missing reason] as specified for undocumented exclusions
   2. Exclude them from the narrative entirely since they lack documentation
   3. Assume the most common reason code from other exclusions
   4. List them under a generic 'other reasons' category

4. Two reviewers independently screened the same 200 abstracts. They disagreed on whether 12 studies should be included. What role can AI play in resolving this?

   1. AI can make the final decision since it analyzed both sets of abstracts
   2. AI can suggest a compromise but a human must confirm the final decision
   3. AI can act as a third reviewer and break the tie automatically
   4. AI can override the disagreement and include both sets of results

5. What does it mean for exclusion reasons to be 'traceable to the protocol'?

   1. Each exclusion reason corresponds to a pre-specified criterion in the review protocol
   2. The reasons are written in the final published paper
   3. Reviewers can trace which AI generated each reason code
   4. The protocol is stored in the same folder as the screening log

6. Which statement describes a limitation of using AI in systematic review documentation?

   1. AI automatically generates new exclusion criteria not in the protocol
   2. AI might miscategorize a borderline exclusion and classify it incorrectly
   3. AI cannot process more than 100 records at once
   4. AI requires internet access to function during screening

7. In a PRISMA flow diagram, what do the stage-by-stage counts represent?

   1. The number of records identified, screened, assessed for eligibility, and included in the review
   2. The number of databases searched
   3. The number of reviewers working at each stage
   4. The number of articles cited in each section of the paper

8. A researcher asks AI to generate the PRISMA narrative while the screening is still in progress. What is the concern with this approach?

   1. It violates copyright rules for systematic reviews
   2. The final counts may change, requiring narrative revisions
   3. AI will include studies that were later excluded
   4. AI cannot generate text until all screening is complete

9. What distinguishes the AI's role from a second reviewer in systematic review screening?

   1. AI screens records that the first reviewer already approved
   2. AI and the second reviewer perform identical functions
   3. AI assists with drafting the narrative while the second reviewer verifies inclusion decisions
   4. AI can screen faster so it replaces the second reviewer entirely

10. Why is it important to document records that are 'lost' or 'unrecoverable' in a systematic review?

    1. It demonstrates that the researchers worked hard to find all studies
    2. It allows the AI to learn from its mistakes
    3. It satisfies journal submission requirements
    4. It provides transparency about potential selection bias from missing data

11. When AI generates a PRISMA narrative, what input does it require from the researchers?

    1. A completed protocol with pre-specified exclusion criteria
    2. The full text of all potentially relevant articles
    3. A list of conclusions from similar previous reviews
    4. The names of all team members involved

12. What information should be included in a PRISMA narrative about the identification stage?

    1. A summary of the study designs found in the search
    2. The names of databases searched and the total number of records identified
    3. The full citation for every article retrieved
    4. The number of reviewers who conducted the search

13. If an AI-generated PRISMA narrative shows that 80% of exclusions at the eligibility stage were marked with [missing reason], what should researchers do?

    1. Publish the narrative as-is since AI generated it
    2. Replace missing reasons with the most common reason code
    3. Investigate why reason codes were not consistently applied during screening
    4. Delete all exclusions without reason codes from the dataset

14. What is the purpose of having a prospective protocol in a systematic review?

    1. To satisfy journal requirements for publication
    2. To allow researchers to change exclusion criteria as needed
    3. To pre-specify criteria so decisions are not made after seeing results
    4. To give AI something to reference when generating narratives

15. During the eligibility stage (full-text assessment), a study is borderline for an exclusion criterion. How should this be handled in the PRISMA narrative?

    1. It should be included to avoid conflict with the AI narrative
    2. AI will automatically categorize it correctly based on its training
    3. It should be documented with the specific criterion it was assessed against and the resolution
    4. It should be excluded by default to be conservative