Prepare a protocol-led systematic-review evidence package

Why use this

1. A systematic review can accumulate many records while losing track of why studies were excluded, which reports describe the same study or where reviewers disagreed. A polished narrative cannot repair an incomplete selection trail after the fact.

2. A protocol-linked evidence package could expose those decisions before synthesis proceeds. Researchers could inspect screening and extraction assumptions, retain unresolved disagreements and assess whether the intended review is reproducible without treating the package as validated scientific evidence.

Why an agent / simpler alternative

Useful judgment

Organize heterogeneous study descriptions and flag inconsistencies against predefined criteria while preserving researcher decisions.

Simpler alternative

Use a conventional review-management tool when human screeners already maintain a complete, consistent audit trail.

Unsuitable when

The process lacks an agreed protocol or would replace independent screening and domain assessment.

Proposed inputs

1. Name

   Review protocol

   Owner

   Research lead

   Minimum scope

   Question, inclusion criteria, outcomes and disagreement rules

   Quality check

   Identify amendments and decisions that require explicit researcher approval.

2. Name

   Study-record sample

   Owner

   Authorized researcher

   Minimum scope

   Lawfully supplied citations, abstracts and screening notes

   Quality check

   Separate duplicate reports from distinct studies.

Access and setup — proposed or unknown

Proposed minimum

1. Fictional records and a local screening-ledger design without literature access.

Compatibility unknowns

1. Unknown: current runtime, database access, full-text rights and extraction fidelity.

External actions requiring approval

1. Protocol changes, paid access, data processing beyond scope and manuscript sharing or submission.

Expected deliverable and review checks

Proposed artifact

Protocol-to-record audit ledger with screening reasons, disagreements and extraction questions.

Critical checks

1. Every inclusion or exclusion has a criterion-linked rationale.

2. Unresolved disagreements and missing evidence remain visible before synthesis.

Missing output evidence

No new workflow run or output inspection was performed for this overlay; use the canonical source-outcome references.

Failure signals and proposed responses

1. Signal

   Two publications from one study are counted as independent evidence.

   Why it matters

   The synthesis may overweight that study.

   Proposed response

   Create a tentative study family and ask researchers to confirm report relationships.

2. Signal

   A screening decision uses a criterion absent from the protocol.

   Why it matters

   Selection can drift without an auditable rationale.

   Proposed response

   Flag the deviation and request an explicit protocol decision before applying it.

Effort and cost — unknown

Setup work

1. Define protocol versions, record identities and disagreement handling.

Ongoing review

1. Audit screening, extraction and proposed bias assessments.

Unknowns

1. Unknown: access fees, screening volume and specialist review effort.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

Five invented study records include companion reports, a screening disagreement and an exclusion using an undeclared criterion.

Acceptance checks

1. Retain the unresolved study relationship.

2. Flag the disagreement and protocol deviation without resolving scientific merit.

Stop conditions

1. Live searches, full-text retrieval, research synthesis or manuscript submission is required.

Limits

1. This proposed ledger check cannot establish search completeness, unbiased selection or the correctness of scientific conclusions.