Systematic review — PRISMA 2020

// PRISMA 2020 systematic-review flow. Same arithmetic validation
// as CONSORT — every exclusion must account for the delta.
//
// The `study` block is what makes this figure a PRISMA flow. The
// renderer used to stamp that banner on any cohort chain with no study
// wrapper, which put a clinical reporting standard over twelve figures
// that had never heard of one. A standard is now something a document
// declares, not something a renderer assumes.

study screening_2023 {
  design: "systematic review · PRISMA 2020"
}

cohort records_identified {
  n: 4521 source "Database search 2023-06"
}

cohort after_dedup {
  n: 3210
  from: records_identified
  excluded: { duplicates: 1311 }
}

cohort title_abstract_screened {
  n: 3210
  from: after_dedup
}

cohort full_text_assessed {
  n: 412
  from: title_abstract_screened
  excluded: { off_topic: 2798 }
}

cohort included {
  n: 47
  from: full_text_assessed
  excluded: {
    wrong_population: 142
    wrong_intervention: 89
    wrong_outcome: 76
    insufficient_data: 58
  }
}

// The argument is the STUDY id, not the last cohort. The renderer looks
// up a `study` block by it and takes the banner from that block's
// `design:`; handed a cohort id it matches no study, falls back to the
// headerless layout, and draws a systematic-review flow naming no
// standard at all. That is the defect in the comment at the top of this
// document run the other way round — the one document that does declare
// the standard getting no banner — and it is invisible in the issues
// panel, because a view argument that resolves to nothing is not a
// compile error.
view flow: prisma(screening_2023)