> For the complete documentation index, see [llms.txt](https://docs.nested-knowledge.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.nested-knowledge.com/guides/guidance-by-review-type/meta-analysis.md).
# Meta-Analysis
Learn how to set up a meta-analysis in Nested Knowledge.
A meta-analysis statistically pools results from multiple comparable studies into a single summary estimate. Choose a meta-analysis when you have several studies reporting the same outcome and want a combined effect size. It is typically built on top of a systematic literature review (SLR); see the [SLR guide](/guides/guidance-by-review-type/systematic-literature-review.md).
For general methodology guidance independent of the software, see [Meta-analysis best practices](/best-practices/best-practices-by-review-type/meta-analyses.md).
### Run A Meta-Analysis In Nested Knowledge
**Inputs:** an already built SLR to import and studies reporting comparable, poolable outcomes.
**Outputs:** a pooled estimate for each intervention and outcome, network meta-analysis comparisons with forest plots and heterogeneity statistics and a PRISMA flow diagram.
{% hint style="info" %}
If you're extracting quantitative data but don't need pooled statistical outputs, a [Table tag in Tagging](/review-workflow/tagging/configure-tagging.md#table) is a lighter-weight option than turning on MA Extraction. Smart MA Extraction is in beta, so manual MA Extraction is the recommended approach. Plan for it to take longer than tagging.
{% endhint %}
| Stage | Recommended tools | Why |
| ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Search | [**Automatic Search**](/review-workflow/literature-search/add-search/automatic-search.md) across every database your protocol calls for; [**File Import**](/review-workflow/literature-search/add-search/import-search-files.md) for external databases; [**Smart Search**](/review-workflow/literature-search/ai-tools-in-search/smart-search.md) to build the starting query; [**Other Sources**](/review-workflow/literature-search/other-sources.md) for gray literature or hand-searched records | A pooled estimate is only as complete as the study set behind it, so search as comprehensively as an SLR. If your meta-analysis builds on a completed SLR, use its search as it stands. |
| Screening | [**Dual Two Pass Screening**](/review-workflow/screening/configure-screening.md#choose-your-screening-mode) (two reviewers, one adjudicator, Title/Abstract then Full Text); [**Exclusion-based Screening**](/review-workflow/screening/configure-screening.md#choose-your-screening-paradigm) | Two independent reviewers and a documented reason for every exclusion maintain SLR-level rigor. At Full Text, an exclusion reason can record when a study doesn't report the outcome in a form you can extract. |
| | Optional: [**Robot Screener**](/review-workflow/screening/exclusion-based-screening/robot-screener.md) | Once your team has adjudicated enough decisions, Robot Screener can fill the second reviewer seat under Exclusion Reasons. It doesn't replace the human adjudicator. |
| Tagging and extraction | [**Core Smart Tags**](/review-workflow/tagging/ai-tools-in-tagging/core-smart-tags.md) for a preliminary PICO read; [**Smart Config: Tagging**](/review-workflow/tagging/ai-tools-in-tagging/smart-config-tagging.md) or [**Import Tag Hierarchy Templates**](/review-workflow/tagging/configure-tagging/import-tag-hierarchy-templates.md) to build your hierarchy; [**Adaptive Smart Tags**](/review-workflow/tagging/ai-tools-in-tagging/adaptive-smart-tags.md) to extract it | Tags capture the qualitative content: study design, population, and intervention details. Numeric outcome data goes into Meta-Analytical Extraction instead. |
| Critical appraisal | [**Dual Critical Appraisal**](/review-workflow/critical-appraisal/appraise-studies.md#dual-critical-appraisal); optional: [**Smart Critical Appraisal**](/review-workflow/critical-appraisal/smart-critical-appraisal.md) | Risk of bias in the included studies affects how confidently you can interpret a pooled estimate. Two reviewers and one adjudicator appraise every included study. |
| **Meta-Analytical Extraction** | Turn on the optional module: [**Meta-Analytical Extraction**](/review-workflow/meta-analytical-extraction.md) | MA Extraction captures dichotomous, categorical, and continuous data, including arm sizes, event counts, means, and timepoints. Quantitative Synthesis is built from this data. |
| Synthesis | [**Quantitative Synthesis**](/review-workflow/synthesis/navigate-synthesis-outputs/quantitative-synthesis.md): **Summary** for a pooled estimate per intervention; **Distribution** for study-level spread without pooling; [**NMA**](/review-workflow/synthesis/navigate-synthesis-outputs/quantitative-synthesis/explore-nma.md) to compare interventions using direct and indirect evidence, with forest plots. [**PRISMA**](/review-workflow/synthesis/navigate-synthesis-outputs/prisma-chart.md) flow diagram export; [**Synthesis Chat**](/review-workflow/synthesis/navigate-synthesis-outputs/synthesis-chat.md) | Estimates default to Random Effects; switch to Fixed Effects only as a deliberate methodological choice. See [Meta-Analytical Methods](/review-workflow/synthesis/navigate-synthesis-outputs/quantitative-synthesis/meta-analytical-methods.md) and [Meta-analysis best practices](/best-practices/best-practices-by-review-type/meta-analyses.md) for the method decisions. |
### Next Steps
:chart-line: Meta-Analysis Best Practices
Apply methodology guidance alongside this workflow.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.
Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:
```
GET https://docs.nested-knowledge.com/guides/guidance-by-review-type/meta-analysis.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.