> 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/review-workflow/meta-analytical-extraction/configure-meta-analytical-extraction.md).

# Configure Meta-Analytical Extraction

Learn how to configure meta-analytical extraction.

Before extracting meta-analytical data, configure which parts of your tag hierarchy represent **interventions** and which represent **data elements** — this reuses the hierarchy you've already built in Tagging rather than asking you to rebuild it.

{% 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#format-tags) is a lighter-weight and recommended option than using MA Extraction.
{% endhint %}

Interventions are the therapies, treatments, or medications compared across your studies. They appear purple in the tag hierarchy. Data elements are every other value you want analyzed in Quantitative Synthesis — outcomes, baseline characteristics, anything measured. They appear gold.

### Toggle On And Configure MA Extraction Mode

Once you have decided MA Extraction fits into your nest, open Nest Settings and toggle it on.

Then decide on reviewer structure. Like in Screening, reviewer structure determines how many reviewer seats assess each record. Single MA Extraction uses one reviewer seat. Dual MA Extraction uses two reviewer seats and human adjudication for disagreements. Toggle this in Nest Settings.

### Configure Interventions

{% stepper %}
{% step %}

#### Open Configure MA Extraction

Select **Configure MA Extraction** <i class="fa-gear">:gear:</i> in the navigation menu. The configuration panel opens automatically.

{% hint style="info" %}
Only one branch of your hierarchy can be designated as interventions, but any tag anywhere in the hierarchy can be configured as a data element. Not every tag needs to be one.
{% endhint %}
{% endstep %}

{% step %}

#### Define Your Intervention Hierarchy

Select the root tag that encompasses the interventions you want to assign, from the dropdown. Selecting a tag turns it and every child tag purple.

<figure><img src="https://305208404-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fxb7uNGATUgwQAjDVQRXG%2Fuploads%2FqvO9Bh3xXLOZOc9eBovE%2Fma%20extraction%20hierarchy.png?alt=media&amp;token=51d3267e-d34a-4e11-87c1-097b28e9c815" alt="The Configure MA Extraction panel shows the intervention hierarchy dropdown, with a selected root tag and its child tags highlighted purple."><figcaption><p>Select the root intervention tag to mark its branch purple.</p></figcaption></figure>
{% endstep %}
{% endstepper %}

### Configure Data Elements

{% stepper %}
{% step %}

#### Select A Tag To Configure

Select any tag in the hierarchy to turn it into a data element. Its settings open in the right-hand panel.

<figure><img src="https://305208404-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fxb7uNGATUgwQAjDVQRXG%2Fuploads%2FJNbIDhcCayqsrNuppYnD%2Fdata%20element.png?alt=media&amp;token=238e6ecf-0041-4149-b203-289d135bffd3" alt="The MA Extraction configuration screen shows a selected tag in the hierarchy and its data element settings in the configuration panel."><figcaption><p>Select a tag to open its data element settings.</p></figcaption></figure>
{% endstep %}

{% step %}

#### Choose The Data Type

Choose **Continuous**, **Dichotomous**, or **Categorical** — required for every data element. Different inputs appear depending on which you choose; a categorical element, for instance, lets you specify its categories.
{% endstep %}

{% step %}

#### Choose The Expected Timepoint

Choose **Baseline** or **Outcome**, based on when the data was collected — required for every data element. If a value could reasonably be either, choose **Baseline**.
{% endstep %}

{% step %}

#### Choose The Direction, If Applicable

Choose **Higher Better** or **Lower Better** where direction matters — for a data element like Mortality, **Lower Better** is typically correct.
{% endstep %}

{% step %}

#### Add Units And Central Tendency, For Continuous Data

For continuous elements, add a unit (minutes, mL, kg) and choose **Mean** or **Median** — required for continuous variables. Choosing one prompts a dispersion measure: **SD** for a mean, **IQR** or **Range** for a median.
{% endstep %}

{% step %}

#### Add Categories, For Categorical Data

Add the custom categories a categorical element should offer.
{% endstep %}

{% step %}

#### Save

Select **Save**. The tag turns gold to confirm the configuration, and you can move on to the next data element.
{% endstep %}
{% endstepper %}

### How Many Data Elements Should You Configure?

Every data element you configure is collected across every included study, adding to reviewer time, and displayed on Synthesis, adding to what a reader has to take in. Unlike primary studies, meta-analytical research is also limited to aggregated data, so multivariate comparisons aren't available the way they would be on individual patient data. Therefore every data element should be of direct interest for comparing interventions.

There's no single correct number, but the minimum number of metrics collected is the number of study arms multiplied by the number of data elements configured, and multiplies further with additional timepoints. As a rough guide:

| Data elements | Typical time per study | Result                                                                                                                                        |
| ------------- | ---------------------- | --------------------------------------------------------------------------------------------------------------------------------------------- |
| 1–3           | Under 5 minutes        | Highly focused synthesis; fast, but may under-contextualize key outcomes.                                                                     |
| 4–7           | 5–10 minutes           | Focused synthesis covering primary and secondary outcomes plus key background data.                                                           |
| 8–10          | 10–20 minutes          | Detailed synthesis covering most outcomes of interest. Data sparsity becomes a risk — we recommend staying at or below this range.            |
| 11–20         | Up to an hour          | Intensively detailed; usually appropriate only for extensive reviews with undefined endpoints.                                                |
| 20+           | Several hours          | Risks data sparsity and collection-quality problems, and often signals an unfocused research question. Consider splitting the review instead. |

### FAQs

<details>

<summary>Can a tag be both an intervention and a data element?</summary>

No — a tag is either part of the single intervention branch or configured as a data element, not both. Keep your interventions and the values you're measuring about them as separate parts of the hierarchy.

</details>

<details>

<summary>Can I assign interventions, create data elements then run Smart MA Extraction?</summary>

No — Smart MA Extraction generates and extracts MA data based on your research question, it doesn't pair with your created tags.

</details>

<details>

<summary>Do I need to configure every tag in my hierarchy as a data element?</summary>

No. Configure only the tags you actually want analyzed and presented in Quantitative Synthesis.

</details>

<details>

<summary>Can I switch between Single and Dual MA Extraction partway through a review?</summary>

Yes, but switching from Dual back to Single keeps only final adjudicated decisions — any individual reviewer extractions still awaiting adjudication are lost.

</details>

### Next Steps

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><i class="fa-calculator">:calculator:</i><br><strong>Extract Meta-Analytical Data</strong></td><td>Extract quantitative data manually, study by study.</td><td><a href="https://app.gitbook.com/s/xb7uNGATUgwQAjDVQRXG/review-workflow/extraction-for-meta-analysis/extract-meta-analytical-data">https://app.gitbook.com/s/xb7uNGATUgwQAjDVQRXG/review-workflow/extraction-for-meta-analysis/extract-meta-analytical-data</a></td></tr><tr><td><i class="fa-sparkles">:sparkles:</i><br><strong>Smart MA Extraction</strong></td><td>Extract quantitative data from full texts automatically.</td><td><a href="https://app.gitbook.com/s/xb7uNGATUgwQAjDVQRXG/review-workflow/extraction-for-meta-analysis/smart-ma-extraction">https://app.gitbook.com/s/xb7uNGATUgwQAjDVQRXG/review-workflow/extraction-for-meta-analysis/smart-ma-extraction</a></td></tr><tr><td><i class="fa-chart-line">:chart-line:</i><br><strong>Quantitative Synthesis</strong></td><td>View pooled estimates and Network Meta-Analysis outputs.</td><td><a href="/review-workflow/synthesis/navigate-synthesis-outputs/quantitative-synthesis.md">Quantitative Synthesis</a></td></tr></tbody></table>


---

# 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/review-workflow/meta-analytical-extraction/configure-meta-analytical-extraction.md?ask=<question>&goal=<user_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.