---
name: systematic-review-protocol
description: Develop or revise a systematic or scoping review protocol, including eligibility, screening, extraction, appraisal, and synthesis plans.
---

# Systematic review protocol

Use when a user wants a reproducible plan for a review of research. Agree the review question, purpose, intended audience, and review type from the request and available context. Make unresolved decisions visible; do not present an unregistered plan as registered or imply a search or screening has happened.

## Build the protocol

1. Specify the question and its relevant elements, such as population, intervention or exposure, comparator, outcomes, setting, and study designs. For a scoping review, specify the concepts and context where these fit better. Define inclusion and exclusion criteria that screeners can apply, including language, dates and publication types only when justified.
2. Identify information sources, search concepts, supplementary discovery methods, anticipated search dates, and a plan to record complete database-specific searches. Distinguish a proposed strategy from one actually executed. State how records will be imported, deduplicated, screened at title/abstract and full text, and how exclusions and disagreements will be recorded.
3. Set out extraction fields, study-level appraisal appropriate to the designs, and how reviewers will pilot and resolve discrepancies. Plan outcomes, measures, time points, and handling of multiple reports of the same study.
4. State when quantitative pooling is defensible, proposed effect measures and models, methods for heterogeneity and missing data, and alternatives when pooling is inappropriate. Describe any planned subgroup or sensitivity analyses and distinguish them from later exploratory analyses.
5. Document amendments, authorship, funding or conflicts, and a prospective registration destination if suitable and requested. Verify current registry and reporting requirements before claiming compliance.

## Deliverable and dependencies

Provide an editable protocol organized around decisions, with assumptions, open questions, and a practical search and review record. Use sources or reporting guidance appropriate to the review and cite versions consulted. Database access, full texts, reviewer capacity, and registration may be unavailable; identify their effect and do not invent any of them. A completed protocol does not establish that the review has been conducted.