Systematic Review Authors Editing and Proofreading Services

Twelve items have to fit inside a word allowance somebody else sets. The PRISMA 2020 statement carries a checklist for abstracts as well as one for the paper, and Table 2 lists twelve numbered items for the abstract alone.¹ The main checklist "includes seven sections with 27 items, some of which include sub-items."¹ The BMJ asks for a structured abstract of 250 to 300 words, and allows up to 400 for a PRISMA style abstract.² JAMA asks systematic reviews for a "structured abstract of no more than 350 words."³

Two of the twelve carry the same title

Items 6 and 8 of the abstract checklist are both called Synthesis of results.¹ One sits under Methods and asks for the methods used to present and synthesize results, and the other sits under Results and asks what the synthesis found.¹ A checklist mapped by title rather than by number will record one of them twice and the other not at all.

The checklist asks where each item is reported

The PRISMA 2020 checklist has four columns, and the last is headed "Location where item is reported."⁴ The BMJ asks for a "PRISMA checklist and flowchart" with a systematic review.² Its requirement that a checklist show "on which page of your manuscript each checklist item appears" is stated for CONSORT, filed with a randomized trial, and not for PRISMA.² JAMA asks that "a completed PRISMA checklist should be submitted for the items completed that apply to systematic reviews," and adds that the checklist "will be used during review but will not be published."³ The BMJ states that it does "not use reporting guidelines as critical appraisal tools to evaluate study quality or filter out articles."² Filling that last column is work done on a finished manuscript, and every later revision changes the pagination it was built on.

The search is reported by transcription

PRISMA-S is "a 16-item checklist that covers multiple aspects of the search process for systematic reviews."⁵ It asks that a strategy be reported "exactly as run, typically by copying and pasting the search strategy directly as entered into the search platform."⁵ Among the eight noteworthy changes the 2020 statement lists against PRISMA 2009 is a modification of the search item.¹ That item now asks for full search strategies for all databases, registers, and websites searched, rather than for at least one database.¹ A pasted strategy is not prose, and the usual instinct to tidy it is the wrong instinct.

Conduct and reporting are governed separately

The Cochrane Handbook, version 6.5, distinguishes the two. Cochrane's Methodological Expectations of Cochrane Intervention Reviews "detail standards for the conduct of Cochrane reviews of interventions," while PRISMA covers reporting.⁶ The MECIR manual of August 2023 carries 75 standards for conducting new reviews and 11 for updates.⁷ A review can be conducted to one document and reported against another, and the two are not versions of each other.

The abstract may be all a reader sees

The explanation and elaboration paper puts it this way: "For some readers, the abstract may be all that they have access to."⁸ The same paper notes that "Journals and publishers might impose word and section limits, and limits on the number of tables and figures allowed in the main report."⁸ Twelve required items, a limit set by the journal, and a readership that may see nothing else all bear on the same few hundred words.

The ICMJE recommendations give two reasons in one sentence. "Because abstracts are the only substantive portion of the article indexed in many electronic databases, and the only portion many readers read, authors need to ensure that they accurately reflect the content of the article."⁹ The recommendations then observe that "Unfortunately, information in abstracts often differs from that in the text."⁹

What the review team owns

The review team owns the judgments. Eligibility, risk of bias, and what a synthesis supports are theirs alone. We do not tell a team whether a manuscript satisfies a reporting guideline.

We read the same count in three places and see whether it is the same count. One appears in the abstract, one in the results text, and one in the flow diagram. An eligibility criterion is written twice as well, once in the methods and once in the abstract, and comparing the two wordings takes one pass. A checklist entry naming a page is settled by opening that page. Whatever arrives here stays confidential.

Each box of the 2020 flow diagram carries a count, and one row under records removed before screening reads "Records marked as ineligible by automation tools."¹ The explanation and elaboration paper directs authors to report records eliminated by a classifier in that row.⁸ A count appears in the diagram, in the results text, and in the abstract, and those three places are usually written weeks apart. Changing the count in one place does not change it in the other two.