# biologit MLM-AI

AI-powered Medical Literature Screening for Safety Surveillance

![](/files/-MWAzqocKgUxDZaZLOXU)

[**biologit MLM-AI**](https://www.biologit.com/factsheet) is a complete scientific literature monitoring platform built for safety teams of all sizes, covering human and veterinary pharmacovigilance, medical devices, cosmetics and nutraceuticals. Its flexible workflow, unified scientific database, and unique AI productivity features deliver fast, inexpensive, and fully traceable results for any safety screening needs.

<figure><img src="/files/JOXjLclnzMIskPdXIWZx" alt=""><figcaption></figcaption></figure>

## New to Biologit?

Get started learning the Biologit platform screening capabilities and how it can help your team with the resources below:

{% content-ref url="/pages/-MWAkDq9MJQmWV7X\_PGG" %}
[MLM-AI Concepts](/topics/mlm-ai-concepts)
{% endcontent-ref %}

{% content-ref url="/pages/-MX-3EDG1eoZmQNK7Z4b" %}
[AI-Enabled Screening Workflows](/topics/ai-enabled-screening-workflows)
{% endcontent-ref %}

{% content-ref url="/pages/-MWAeK\_Q8GSm3ionGUHA" %}
[Sign on and Navigation](/application/navigation)
{% endcontent-ref %}

## Reviews and Abstract Screening

{% content-ref url="/pages/-MWAeHFuZ40-84xqWpb8" %}
[Reviews](/application/reviews)
{% endcontent-ref %}

{% content-ref url="/pages/-MWAk3VfBbpM\_QmgB5yR" %}
[Review Details](/application/review-details)
{% endcontent-ref %}

{% content-ref url="/pages/-MWAvaz89ng-LgJLfMLS" %}
[Article Screening](/application/article-screening)
{% endcontent-ref %}

{% content-ref url="/pages/-MWAvdvQmp9c-NkhrHwD" %}
[Batch Screening](/application/batch-screening)
{% endcontent-ref %}

## Searching the Biologit Database

{% content-ref url="/pages/V80yPmo2pVLYXYKyc4vG" %}
[Article Search](/application/article-search)
{% endcontent-ref %}

## Configuration

{% content-ref url="/pages/-MWB0izVuvIBWCZdkqkj" %}
[Monitors](/configuration/monitor-configuration)
{% endcontent-ref %}

{% content-ref url="/pages/-MWB0\_d5P78BUQXV3poq" %}
[Users, Teams and Permissions](/configuration/team-and-user-management)
{% endcontent-ref %}

{% content-ref url="/pages/hqFQ8i4LEsxJ6y9c7APr" %}
[Settings](/configuration/settings)
{% endcontent-ref %}

{% content-ref url="/pages/DlNJCfeO8M7qjQKWVPbP" %}
[Custom Workflows](/configuration/custom-workflows)
{% endcontent-ref %}

{% content-ref url="/pages/bz01OXRSgU0z0ZC1nFuR" %}
[E2B Configuration Guide](/configuration/e2b-configuration-guide)
{% endcontent-ref %}

{% content-ref url="/pages/M7q2gaaSh42xQyc4ye7E" %}
[Notifications](/configuration/notifications)
{% endcontent-ref %}

## More Help

Tutorials and deep dives:

{% content-ref url="/pages/YXPa48kFiuNngSDVTELN" %}
[TOPICS](/topics/mlm-ai-concepts)
{% endcontent-ref %}

{% content-ref url="/pages/-MWAgIsHJIRUZdEYQIi3" %}
[FAQ](/more-help/faq)
{% endcontent-ref %}

{% content-ref url="/pages/23DYtshbsrIFcxHdxGhY" %}
[Support and Platform Details](/more-help/support-and-platform-details)
{% endcontent-ref %}


# Sign on and Navigation

To login to MLM-AI, enter the username and password provided to you by your administrator.

<figure><img src="/files/1TTgRESumid8LsZwoLuw" alt=""><figcaption></figcaption></figure>

Once logged in, you can navigate via the launch page to relevant tabs via a left-hand menu. You can also access the biologit website, help pages and product tours from the landing page:

{% embed url="<https://youtu.be/gga2PtfXcGQ>" %}

### Forgot password?

Registered users in the platform can ask for a temporary password. This password can be obtained by clicking Forgot password and then writing the user's login email. If the email is registered in the platform the user will shortly receive an email with the temporary password he can use to login and create a new password. Note that to proceed with this request the user will need to complete a CAPTCHA verification step.<br>

<figure><img src="/files/yR4EFSQ1tqOQueDOEj7e" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/r41sMAA2Vq0f8rFziobg" alt=""><figcaption><p>Example temporary password email</p></figcaption></figure>

### Getting Help

From any screen, click the help (?) icon to get access to in-app help content, linking to the product documentation.

Where available, you can also take a **Tour** for the respective page. Click on the help icon and then the "Take the Tour" button:&#x20;

{% embed url="<https://youtu.be/2DXB6LaXAAc>" %}

### User Profile

The user profile menu appears when clicking the User icon (top right). From here users can:

* Log out from the application (**LOGOUT** link)
* View assigned [permissions ](/configuration/team-and-user-management#permissions)and [monitors](/configuration/monitor-configuration).
* Update your password.
* View MLM-AI release status and get access to the support desk.

<figure><img src="/files/hRv3X7Cdsvs9MDTDw4uF" alt=""><figcaption></figcaption></figure>

### Dark Mode

If you prefer, you can use the MLM-AI interface in dark mode:

{% embed url="<https://youtu.be/mx76MlL6Js4>" %}


# Reviews

Check reviews in progress, completed or planned to run; Submit new Reviews or Download review results for further processing

In MLM-AI, product searches are defined in a [**Monitor**](/configuration/monitor-configuration), and **Reviews** are their results for a date period. Reviews contain the search hits for users to analyse and screen.&#x20;

Click on the **Reviews menu** to see a summary list of reviews in progress and recently completed. This screen allows users to rank and select the next review to work on.

Clicking on the link of any Review will lead to the [Details page](/application/review-details) where users can select articles for screening.

Reviews are categorized in tabs:

1\.  **In Progress** - Reviews where articles are pending screening

2\.  **Completed** - Reviews where 100% of articles have been screened, or that have been set to Completed by the review owner

3\.  **Processing** - Reviews scheduled by the system, or actively being processing

<figure><img src="/files/e34Up6OYee2K966a6NrW" alt=""><figcaption></figcaption></figure>

Within a tab, it is possible to filter reviews using keywords and sort by a number of criteria.&#x20;

It is also possible to select other columns. The example below filters Reviews for "methotrexate":&#x20;

<figure><img src="/files/NDlb9Wmqs0C6mxm5O02y" alt=""><figcaption></figcaption></figure>

Adding terms to the search box matches the Monitor title field. The search box also accepts more specific queries using a mini-language. Examples:

| Example                         | Description                                                                       |
| ------------------------------- | --------------------------------------------------------------------------------- |
| `monitor:paracetamol start:jun` | Matches all results with "paracetamol" in monitor title and start date of June    |
| `rundate:july 2021`             | Matches all results with run date of July 2021                                    |
| `start:jun 2021 end:sep 2021`   | Matches all results with results start date between June 2021 and September 2021. |

## The Review Lifecycle

The tabbed view in the Review summary page groups reviews by completion. Reviews can have one of the following status:

* The **In Progress** tab are all reviews where screening decisions need to be finalized. A review stays in progress until 100% of articles receive a screening decision (unless the status is manually changed by the user)
* The **Completed** tab contains reviews where 100% of screening has been done
* Reviews may be **Archived** by users with [Review supervisor privilege](/configuration/team-and-user-management#permissions). Archiving a review no longer lists it in any tab removing clutter from already completed work, but the review may still be retrieved via Find Reviews (see below)
* Finally, when reviews are uploaded in error a user with [Review Supervisor privileges](/configuration/team-and-user-management#permissions) can set its status to **Retired**. Once set to Retired, reviews are no longer visible, and are not accounted for in reports or any other processing such as duplicate detection&#x20;

{% hint style="info" %}
If a screening decision sets any citation to a "Pending" status in Completed review, the Review moves back to "In Progress" as it now awaits review completion. This can happen on workflows where, for example, an article requires additional work after QC.&#x20;
{% endhint %}

### Changing Review Status

Users with Review Supervisor privileges can modify the status of reviews from the summary page.

* Select one or more reviews using the checkbox
* Select the "**...**" menu on top right
* Choose the target status for the selected reviews (see above discussion on each status)

<figure><img src="/files/TZIWEbS0LxBnADUVOOkz" alt=""><figcaption><p>Supervisors can change the status of a Review</p></figcaption></figure>

## Finding Reviews

For more comprehensive Review searches you can use the **Find Reviews** functionality from the Finder Icon:

![](/files/mX8g1qdIJVgMAnYWdOI9)

&#x20;Use Find Reviews to search by:

* A specific monitor
* By status: In Progress, Completed, Archived, etc.
* By assigned teams
* By target start and end dates

<figure><img src="/files/GksBL7vKQINh6Ynlg7Xb" alt=""><figcaption></figcaption></figure>

### The Processing Tab

The Processing tab shows the reviews processing or scheduled, for example:

* manually submitted reviews still being processed (see On Demand reviews below)&#x20;
* or *scheduled* by the system.&#x20;

If, for example, you have active scheduled monitors that run weekly, you will see results for the following week in progress and their respective ETA.

<figure><img src="/files/o8SB9USM6peIND42YIkg" alt=""><figcaption></figcaption></figure>

## Submitting Reviews On Demand

Privileged users can request on-demand Reviews for any active monitor:&#x20;

* Click the **Add Review** button to configure your submission
* Select **On Demand**
* Select desired monitor, start and end dates for the review

<figure><img src="/files/S6IgwjIgFeVfxvHgq7P3" alt=""><figcaption></figcaption></figure>

* Click the "Verify" button to check the submission options. This will perform a validation check on the dates selected, and overall size of the review (See [System Limits](/more-help/faq#system-limits))
* If verification passes, the "Submit" button will become enabled, and the review is ready for submission:

<figure><img src="/files/GlJeaboOeNRETPoN7TgP" alt=""><figcaption></figcaption></figure>

* Once submitted, the review will be queued for processing and progress can be monitored in the [Processing tab](#the-review-lifecycle).

{% hint style="info" %}
Reviews can also be created from an external data source (ex: Embase, Dialog, ADIS). See [**Uploading Review Data**](/application/upload-data).
{% endhint %}

## Batch Submissions

It is also possible to submit on-demand reviews in batch simplifying the process when many monitors require reviews.

* Use the "**Batch"** tab from "Add Reviews"
* Select monitor and date range
* Add or Copy the request to create a second submission
* Repeat the process for as many monitors as required

When ready users can verify and submit all monitors in a single request

<figure><img src="/files/4YAnrhVFXQ1eGJCLHkzF" alt=""><figcaption></figcaption></figure>

## Advanced Options

On-demand submissions allow advanced options that optionally provide more control on review processing behavior:

<figure><img src="/files/qwx1yXaq9dbVHJNlIgWO" alt=""><figcaption></figcaption></figure>

Users can use the **Publication Date Lookback** to override the publication date lookback period configured on the monitor. This is useful where hits by publication date for the on-demand review must comply with a specific period of time.

Users can optionally configure an on-demand submission to **Deduplicate against reviews of any period**, for the same monitor. By default the system will not take into account reviews from a date range ahead of the requested review date range (ie. reviews "in the future"), but it may be useful to override this behavior in certain scenarios (ex: a retrospective search for a date range in the past)

## Exporting Reviews

Reviews can be downloaded to your local computer as spreadsheets or HTML. From the Summary page:

* Select reviews you wish to download by checking the checkbox beside the review title
* Go to **Actions** and move the mouse to the download icon <img src="/files/-MWAug3MuvQi9XMTXF21" alt="" data-size="line"> &#x20;
* Select the file format to download from the options available

{% hint style="info" %}
Reviews can also be exported from the [Review Detail](/application/review-details) page; individual articles can also be exported from the [Article Detail](/application/article-screening) page.
{% endhint %}

{% hint style="info" %}
In addition to exporting reviews, users can run more comprehensive reports across multiple monitors and date periods from the [**Reports**](/application/reports) module.
{% endhint %}


# Find Articles from Reviews

Flexible search of Reviews by decision, tags, reviewers and any other criteria

Many workflows require searching for results across different monitors, larger time periods or using other specific criteria. For example:

* Search for articles to perform quality checks across a portfolio of many products
* Review results initially flagged for signal processing
* Find all articles assigned to my user and not yet assessed, regardless of monitors.

It is possible to achieve this from the **Find Articles** page, either from the Reviews page or from the main menu on the left:

<figure><img src="/files/5eKLyr0VXfP8EK6IOW11" alt=""><figcaption></figcaption></figure>

The Find Articles page works in two stages:

<figure><img src="/files/dzjFnqjzbe5WorwN26vY" alt=""><figcaption></figcaption></figure>

#### Choose what Reviews to Retrieve&#x20;

First, select which monitors, review status and date range to select reviews:

<figure><img src="/files/ANlKR5SE9f9LVVXmXXvz" alt=""><figcaption></figcaption></figure>

* Multiple monitors can be selected, type into the Monitors drop down for a shortlist

The date range for selecting monitors can be based on:

* Review Start Date
* Date of Creation (Day Zero)

To choose which criteria to apply, use the toggle as shown below:

<figure><img src="/files/2rFOKWWB5uPOs9mSTG0K" alt=""><figcaption></figcaption></figure>

#### Criteria for Selecting Articles

Once the criteria for reviews is chosen, you can determine what **articles** to select:&#x20;

* by any criteria available in the [Review Details](/application/review-details) page including assignment, sampling, decision, or open queries from the [search box](/application/review-details/searching-abstracts-inside-a-review)

#### Example

The example below selected from all reviews between a date range, with decision status of "Confirmed ICSR":

<figure><img src="/files/6Raj0B6jtOrPKXBaCdjA" alt=""><figcaption></figcaption></figure>

Once results are ready, users can:

* Navigate to the retrieved articles by clicking the article ID and complete screening
* Alternatively, [review results in batch mode](/application/batch-screening) by clicking **Batch Review**
* Select multiple items and perform an [assignment action](/application/review-details/work-assignment#assignment-process)
* From the Download Icon, it is possible to **export** your query results directly from this page:

<figure><img src="/files/21OexaB6UcWrzJUVE9OU" alt=""><figcaption></figcaption></figure>


# Uploading Review Data

Use MLM-AI to screen results of search results from external databases like PubMed, ProQuest or Embase

It is possible to upload literature search results exported from an external database:

* Use this option to perform literature screening using data from proprietary third party data sources
* All of MLM-AI's screening, auditing and AI-based automation features can be used on uploaded results

Currently, it is possible to upload results exported from these systems:

<table><thead><tr><th width="201.33333333333331">Source System</th><th>Expected Format from Source System</th></tr></thead><tbody><tr><td><strong>PubMed</strong></td><td>PubMed's <a href="https://pubmed.ncbi.nlm.nih.gov/help/#citation-management-software">nbib reference format</a>. </td></tr><tr><td><strong>PubMed</strong></td><td>PubMed email alerts in nbib reference format as <code>eml</code> email file</td></tr><tr><td><strong>Embase</strong></td><td>RIS citation management export, or <code>csv</code></td></tr><tr><td><strong>EBSCO</strong></td><td>RIS citation management export</td></tr><tr><td><strong>Adis</strong></td><td><code>csv</code> export</td></tr><tr><td><strong>ProQuest/Dialog</strong></td><td>Excel export</td></tr><tr><td><strong>EMA (Excel)</strong></td><td>Eudravigilance EMA MLM files for consolidation</td></tr></tbody></table>

{% hint style="info" %}
Depending on the source of exported results, uploads could be automated. Refer to [Integrations](/application/integrations) to find out how.
{% endhint %}

## Uploading Multiple Files

For convenience, uploading files can be done in bulk saving time in routine weekly uploads of search results.

* Select **Upload** as the review type
* You can drag files directly from your desktop anywhere into the area marked as "Drag and Drop files here"

<figure><img src="/files/S7xitS39RZzrOBSKBk3c" alt=""><figcaption></figcaption></figure>

* Once files are uploaded, configure them by mapping to the Monitor, file type and start-end date of search results

## Applying Default Values

Configuring uploads can be done one file at a time, or you can select multiple files and apply the same setting to all.

* Example: if the source data for your files is always the same you can select one default value and apply to all
* Select what files to apply defaults, then choose "Edit Selections"

<figure><img src="/files/COzo3tmi1cdBUqUemIOC" alt=""><figcaption></figcaption></figure>

* In the Dialog, select default values for File format, Dates and Monitor (leave empty the fields you do not need default values for)
* Click "Apply"
* Complete filling any remaining fields as needed

<figure><img src="/files/bKRudPVIZrGlQi9qYowG" alt=""><figcaption></figcaption></figure>

## Verification

Once entering parameters is complete, files can be Verified by clicking the "**Verify**" button. This will pre-process files to ensure the file format is correct and now rows are corrupted.

* If successful, file status will change from "**To Verify**" to a green indicator of the number of records to be processed
* Once verification is complete, the "**Upload**" button will become enabled and files can be uploaded

<figure><img src="/files/kvVnYBDAvCfmopnfy9O6" alt=""><figcaption></figcaption></figure>

## What Happens after the Upload

Uploaded files will generate new [Reviews](/application/reviews). They will be visible in the "Processing" tab in the [Review Summary](/application/reviews) page, and will become available (ie. move to the "In-Progress" tab) once processing finishes.

From this point, results can then be normally screened, like any other Review produced by MLM-AI.

### Unique Records During Upload

* Uploaded records are identified by their **unique ID from the source** when uploaded
* This ensures only a single record exists for the corresponding unique record from the source and facilitates duplicate detection
* Therefore, subsequent uploads of citations with an existing unique ID do not create new records in the MLM-AI database

During the verification step, users can verify if repeated IDs are present in their upload file and if they will be discarded. Example:

<div align="left"><figure><img src="/files/5wdj5DMnCf9gmkEegZ8w" alt=""><figcaption></figcaption></figure></div>

Clicking the badge shows further details of specific IDs detected as already existing.

{% hint style="info" %}
Article de-duplication is also available for uploaded data. Learn more in the [Duplicate Detection](/topics/duplicate-detection) page.
{% endhint %}

## Supporting Guides for Exporting from External Databases

* For PubMed, follow the steps for [exporting into a citation management format](https://pubmed.ncbi.nlm.nih.gov/help/#citation-management-software)
* For Embase, refer to [this guide](http://supportcontent.elsevier.com/Support%20Hub/Embase/Files%20&%20Attachements/5540-Embase%20E105%20-%20Saving%20Searches%20and%20Exporting%20Results%20-%20June%202015.pdf) for steps on how to search results


# Review Details

Review contents and select articles for screening

The **Review Details** page lists articles retrieved according to the review's [monitor definitions](/configuration/monitor-configuration) for the date range specified. In this page users can:

* Check screening status for this review: articles in queue, pending author follow up, etc
* Rank and filter articles to prioritize screening
* Select a sample of results for quality control
* Select and assign articles to a user for review or QC

A status pane indicating the review ID, date created, start-end dates, percentage of suspected adverse events, and status progress. Additional details on status of each articles of the review are also provided.

![Status pane](/files/Zv1tFhPxTmAA7VFExMTv)

Below the status pane, the full list of articles retrieved for the review is shown. Additional fields can be added with the **Show** selector and articles can be ranked via **Sort By** or filtered using the [**Search box**](/application/review-details/searching-abstracts-inside-a-review).

{% embed url="<https://youtu.be/Bhiyv2sNNrA>" %}

{% hint style="info" %}
The search box accepts both keywords and more complex queries. Learn more in [**Searching abstracts inside a Review**](/application/review-details/searching-abstracts-inside-a-review).
{% endhint %}

### Sampling

It is also possible to specify a sampling rate to select a random % of results. Sampling applies once other selection criteria is applied, such as Decision, or search criteria.

See also: <https://docs.biologit.com/more-help/faq#how-does-sampling-of-articles-work>

### Tabs

To facilitate screening, articles are grouped by status in tabs, so you can quickly focus on the ones that need attention:

![](/files/1EdRZ6Cub2zsq2CNAixs)

**Tabs by screening status**

* **In Queue:** articles not yet screened
* **Reviewed:** articles already screened (including articles [automatically screened](/topics/ai-enabled-screening-workflows#automatic-filtering-by-tag) by MLM-AI)
* **Duplicates** for articles pre-screened as duplicates of other articles&#x20;

Optional tabs may appear depending on the contents of the review.&#x20;

* **Non-English**: articles whose language is not English and may require special follow-up
* **No Abstract**: results where the citation contains only the article title

Additional tabs may appear if Work assignment feature is in use.

* Learn more about [Work Assignment feature](/application/review-details/work-assignment).

### Screening Articles&#x20;

From the Review Details page, articles can be selected for screening. The next step is to:

* Click on the title of any articles to [begin screening](/application/article-screening)
* Alternatively, select [Batch Review](/application/batch-screening) to review all articles in batch mode

### Tab, Filter, Sampling and Sorting Choices "Follow" User During Screening&#x20;

{% hint style="warning" %}
**Important**: The selected tab, sort order and filters and sampling will be reflected at the[ article screening](/application/article-screening) and [batch screening](/application/batch-screening) pages.
{% endhint %}

To streamline the screening workflow, all user selections on this page will "follow" user on subsequent screening.

**Example**: User wishes to screen articles tagged as "Identifiable Patient":

* Select "In Queue" Tab
* Use [the search box](/application/review-details/searching-abstracts-inside-a-review) to filter by `tag:patient`

The info bar will indicate selections:

<figure><img src="/files/p9A5FKbY0snPPxYGNpZc" alt=""><figcaption></figcaption></figure>

Clicking into any article will lead user onto the screening page, where the selections will be indicated to the user:

<figure><img src="/files/65MEh5MGGxxjIt9vp292" alt=""><figcaption></figcaption></figure>

The same behavior occurs when in [Batch Review](/application/batch-screening) screen.

## Learn More

* See the [Concepts guide](/topics/mlm-ai-concepts#duplicate-detection) for more details on MLM-AI's screening workflow
* Learn more about [Duplicate Detection](/topics/duplicate-detection)
* See [Searching abstracts inside a Review](/application/review-details/searching-abstracts-inside-a-review) for details on querying articles within a review


# Searching Inside a Review

Use the search box to select results inside a review for faster screening

In the [Review Details](/application/review-details) and in [Article Search](/application/article-search) the search box can be used to filter results using keywords, by tags and many other attributes.

## Keyword Matching

The simplest method is just to type the key terms to be matched. These will be searched in the title, abstract and publication name:

<figure><img src="/files/WGiW6tg6PQ69MQnTS1CT" alt=""><figcaption></figcaption></figure>

Users can also search according to AI tags, screening decisions, comments and source data. Lets look at how this is possible with the results query language:

## Query by Attribute

Results can be filtered by AI tag, contents of decision, author country of origin and many other attributes. They can also be combined in a single query.

Lets look at some examples:

### Example Queries

| Task                                                                                                                                    | Query                              |
| --------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------- |
| Retrieve all articles tagged as "Suspected AE"                                                                                          | `tag:ae` (or `tag:adverse)`        |
| Retrieve all articles containing the word "acetamoniphen" AND containing a patient mention (patient tag)                                | `match:acetaminophen tag:patient`  |
| Retrieve articles NOT flagged as a special situation AND flagged as aggregate report                                                    | `notag:special decision:aggregate` |
| Only articles from PubMed                                                                                                               | `source:pubmed`                    |
| Articles where the [main author affiliation](/application/review-details/author-country-of-origin) is from Portugal OR Ireland OR Japan | `mainauthor:PT,IE,JP`              |

### Querying with Multiple Criteria

When adding different attributes in the same query, they are considered as an AND query

Example:

* Match results with Suspect Adverse tag **AND** the keyword "acetaminophen" in contents
* `tag:adverse match:acetamoniphen`

If different criteria are combined by a comma (`,`) they are treated like an **OR** search:

Example:

* Results where main author is from US or Ireland or Portugal:
* `mainauthor:US,IE,PT`

Finally, both methods can be combined for more sophisticated searches:

* `tag:adverse mainauthor:US,IE,PT`
* (results tagged as Suspect Adverse AND where the main author is from US or Ireland or Portugal)

### Pre-defined Filters

Monitors can be configured with pre-defined filters that can later be used as shortcuts from the search box. This facilitates reusing complex queries.

👉 Check out the steps on [**how to configure pre-defined Monitor filters**](/configuration/monitor-configuration#filters)&#x20;

In the example below there are two filters configured for the sample Monitor:

<figure><img src="/files/203lXF8b4kKxtf5d0YoU" alt=""><figcaption></figcaption></figure>

In the Review search box, the filters appear as shortcuts in place of the actual query. Once selected the configured query runs:

<figure><img src="/files/TG5gQVUz2fPZIoyykg4s" alt=""><figcaption></figcaption></figure>

### Reference: All Query Options

<table><thead><tr><th width="169.90183822506123">Parameter</th><th>What it does</th></tr></thead><tbody><tr><td><a href="#match-less-than-value-greater-than"><code>match</code></a><code>:</code></td><td><ul><li>Match specified keyterms in article title and publication name</li><li>Use <code>nomatch:</code> for the opposite query.</li></ul></td></tr><tr><td><a href="#tag-less-than-value-greater-than"><code>tag</code></a><code>:</code></td><td><ul><li>Match only articles with a specific <a href="/pages/-MWAkDq9MJQmWV7X_PGG#ai-tags">AI Tag</a>. (for the opposite effect use <a href="#notag-less-than-value-greater-than"><code>notag</code></a><code>:</code></li></ul></td></tr><tr><td><a href="#by-less-than-user-greater-than"><code>by</code></a><code>:</code></td><td><ul><li>Matches articles last updated by user matching <code>&#x3C;user></code> (use <code>nouser:</code> for the opposite query)</li></ul></td></tr><tr><td><a href="#source-less-than-source-greater-than"><code>source</code></a><code>:</code></td><td><ul><li>Matches articles from a given source (pubmed, doaj, crossref, etc.)</li><li>For the opposite query use <code>nosource:</code></li></ul></td></tr><tr><td><a href="#comment-less-than-comment-greater-than"><code>comment</code></a><code>:</code></td><td><ul><li>Matches the contents in article's comment section of the most recent screening decision (if any)</li><li>For the opposite query use <code>nocomment:</code></li></ul></td></tr><tr><td><a href="#decision-less-than-decision-greater-than"><code>decision</code></a><code>:</code></td><td><ul><li>Select specific review or QC decisions. Ex: <code>decision:serious</code></li><li>For the opposite result use <code>nodecision:</code></li></ul></td></tr><tr><td><code>mainauthor:</code></td><td><ul><li>Select according to main author's country of origin.</li><li>See: <a data-mention href="/pages/wJMRZdZPacre8dDinH3l">/pages/wJMRZdZPacre8dDinH3l</a></li></ul></td></tr><tr><td><code>anyauthor:</code></td><td><ul><li>Select according to any of the author's country of origin.</li><li>See: <a data-mention href="/pages/wJMRZdZPacre8dDinH3l">/pages/wJMRZdZPacre8dDinH3l</a></li></ul></td></tr><tr><td><a href="https://app.gitbook.com/o/jj4cAbkJRlFV7NUEATGJ/s/6pIe8UPxEvsbrxTVXkGi/~/changes/111/application/review-details/searching-abstracts-inside-a-review#issn-less-than-issn-greater-than"><code>issn</code></a><code>:</code></td><td><ul><li>Matches based on source ISSN of an article</li></ul></td></tr><tr><td><a href="https://app.gitbook.com/o/jj4cAbkJRlFV7NUEATGJ/s/6pIe8UPxEvsbrxTVXkGi/~/changes/111/application/review-details/searching-abstracts-inside-a-review#dueby-less-than-days-greater-than"><code>dueby</code></a><code>:</code></td><td><ul><li>Select based on the number of days left before review submission deadline</li></ul></td></tr></tbody></table>

### Query Options

#### match: \<value>

Match specified keyterms in article title and publication name. This is also the default behavior when not using the `param: <value` syntax.

* **Usage**: `match:key terms`

Examples:

* `match:case`

#### tag: \<value>

Match only articles with a specific [AI Tag](/topics/mlm-ai-concepts#ai-tags). Tags can be specified as follows

* Usage: `tag:value`

Examples:

* `tag:ae`  (articles tagged "Suspected AE", can also use `tag: adverse`)
* `tag:patient` (articles with a confirmed patient mention)
* `tag:special` (articles tagged with any special situation)
* `tag:case` (suspected case tag)
* `tag:keyword` (articles matching [user defined keywords](/configuration/monitor-configuration#user-defined-keywords) configured in the monitor)

#### **notag: \<value>**

Performs the inverse of tag: matches articles that *do not* contain stated tag.

* `notag:ae` (articles not tagged "Suspected AE" )

#### by: \<user>

Matches articles last updated by user matching `<user>`

Example:

* `by:mlmai`

#### source: \<source>

Matches articles from a given source (pubmed, doaj, crossref, etc.)

Example:

* `source:doaj`

#### **comment:\<comment>**

Matches the contents in article's comment section of the most recent screening decision (if any)

Example:

* `comment:duplicate`

#### decision:\<decision>

Matches articles with a specific review decision:

Examples:

* Serious events: `decision:serious`
* Aggregate reporting: `decision:agg`
* Safety signal: `decision:signal`

Decision labels from [custom workflows](/configuration/custom-workflows) are also accepted:

* `decision:Pending Translation`

Note: use `nodecision:` to retrieve opposite results.

#### issn:\<issn>

Returns all articles that match a provided ISSN, or do not have an associated ISSN.

#### dueby:\<days>

Matches articles based on the number of days left before the deadline of review submission. For example, `dueby:5`  will fetch all articles a review for which:

* should be submitted within 5 days or less
* is already overdue (deadline for review submission has already passed)

See [Sources and Search Schedule](/configuration/monitor-configuration#sources-and-search-schedule) part of the monitor configuration guide for more details on how the deadline date is calculated.

{% hint style="info" %}
The search box is also available from the [**Find Articles**](/application/reviews/find-articles-from-reviews) screen, where users can search into multiple reviews based on a date range, or monitors
{% endhint %}


# Author Country of Origin

On cases where literature screening is conducted for a specific territory, it is desirable to filter articles according to the country of origin. The country of origin is typically determined by verifying the **country of affiliation** of the article's authors, or just the primary (first) author.

Biologit MLM-AI supports this workflow by detecting author's country of affiliation where this information is available.

#### How Much Can be Filtered by Country of Origin?

Results will vary depending on databases being used in the search. In general, PubMed has better affiliation coverage and you can expect more articles to be filtered from this source.

Based on our experimental results:

* When searching only PubMed as source, country of affiliation can be identified (and therefore filtered) in over **80%** of results
* When searching all sources, country of affiliation can be detected in **18%** of results

We are constantly making improvements to affiliation data and our detection algorithms to improve coverage.

### Filtering by Country Of Origin

Author country of affiliation can be filtered with the `mainauthor` and `anyauthor` filters.

For example:

* `mainauthor:us`
  * **Includes** articles where the main author's affiliation is the US
  * **Includes** any articles unable to detect country of affiliation
  * **Excludes** articles where author affiliation is not from the US

* `anyauthor:us`
  * **Includes** articles where the affiliation of any author is from the US
  * **Includes** articles unable to match affiliation country
  * **Excludes** articles that contain affiliation and where no author is from the US&#x20;

{% hint style="info" %}
The article **main author** is always the first author listed&#x20;
{% endhint %}

&#x20;In the Review Details screen, select the "Countries in Authors" column to see detected countries of affiliation:

<figure><img src="/files/oBkqmiY0RkwqOLPljDTw" alt=""><figcaption></figcaption></figure>

Each author's affiliation is also visible in the details tab:

<figure><img src="/files/jGEurQHrQQpNGHoEIETy" alt=""><figcaption></figcaption></figure>

### Frequently Asked Questions

#### How does affiliation detection work?

MLM-AI will determine the country according to mentions of any countries, nationalities, cities or regions from the author affiliation metadata.

#### Who is the primary author

The primary author is always the first author listed.

#### What happens when affiliation is not known

In some cases, author affiliation data is not present. MLM-AI will **not** filter such articles and leave up to users to make the determination on the country of origin.


# Work Assignment

Supervisors can assign, or configure MLM-AI to automatically assign abstracts for review to team members

## About Work Assignment

When it is necessary to directly assign screening or QC activities to users, MLM-AI can operate in *Work Assignment* mode.&#x20;

When Work Assignment is enabled:

* Supervisors can assign abstracts for screening or QC to other team members, and&#x20;
* Users are only allowed to create decisions on abstracts directly assigned to them, either by a supervisor, or automatically by the system

### Enabling Work Assignment

* Work assignment feature is enabled via Settings. See [Work assignment configuration page](/configuration/settings#work-assignment) for mode details
* To make assignments, a supervisor user needs to be part of a team with the ["Work assignment" permission](/configuration/team-and-user-management#role-based-permissions) enabled

## Work Assignment Mode

Once work assignment is enabled, users will be able to see new tabs on the Review Details screen.&#x20;

The amount of tabs visible to a user differs based on the their role in the system:

* Supervisors or users given the "Work Assignment" system privilege have access to all tabs, including **Assigned**, **Unassigned** and **My Items**
* Other users have access to the **My Items** tab only

<figure><img src="/files/7BZvXzyfhpTL8hwtnXfd" alt=""><figcaption><p>Review Details view for a supervisor</p></figcaption></figure>

Assignment tabs appear in green. They are organized as follows:

* **Assigned** (visible to team supervisors only) - it is a view of all assigned articles across the review. From this tab user is free to filter articles by Assignee (person assigned to an article), Assignment type and Work Status
* There are two types of work status: "Assigned Work To Do" means decision is not made by assignee yet, otherwise article has status "Assigned Work Done". Filtering by status is useful to track completed work or when looking for articles that could be re-assigned

<figure><img src="/files/JQCoiPTxxS3phFiG56T6" alt=""><figcaption><p>Showing articles assigned for review from Assigned tab</p></figcaption></figure>

* **Unassigned** (visible to team supervisors only) - all articles that don't have an assignee yet will appear in this tab
* **My Items** - everything that is assigned to current users is in this tab. The "Work Status" drop-down box helps identify work that waits to be done: when "Assigned Work To Do" is selected, it leaves articles without a decision or where decision is made by another user

<figure><img src="/files/px3vGGamXVZjGcg9yOFm" alt=""><figcaption><p>Using "Work Status" filter from My Items tab</p></figcaption></figure>

#### Toggle Tab Groups

Use the **Toggle Groups** widget to reduce the number of visible tabs and customize the screen to your needs. can help achieving this. Once clicked, it will show a pop-up window providing a possibility to hide any group of tabs (Assignment tabs - green, Status tabs - dark blue, Miscellaneous tabs - brown).

<figure><img src="/files/gR7lBxq4ZnGok3KQOhfn" alt=""><figcaption><p>Example on Toggling Tab Groups</p></figcaption></figure>

### Assignment Process

To make an assignment, a supervisor needs to select one or more articles using check boxes on the left-hand side and click on the **Assignments** button:

<figure><img src="/files/EQKMhZNrpzZr9jEmqCYO" alt=""><figcaption><p>Assignment from Review Detail</p></figcaption></figure>

In the appeared pop-up window users can assign or unassign selected articles.&#x20;

While assigning, a desired assignee and type of assignment have to be selected. On un-assignment, just a confirmation from a user is required to proceed:

<figure><img src="/files/NqZ2uSvr0Ogny8KsnvWf" alt=""><figcaption><p>Assignment/un-assignment process</p></figcaption></figure>

The following rules apply on assignment conflicts:

* If user wants to assign an article that is already assigned, assignment will be overwritten
* If user wants to unassign an article without an assignment, nothing happens

A user with [assignment privileges](/configuration/team-and-user-management#role-based-permissions) can make assignments from any tab of the review, including Assignment, Status and Miscellaneous groups.

Assignment status is visible from Article Detail view:

* Green icon - article is assigned to a current user
* Red icon - article is assigned to another user
* Grey icon - article is not assigned to anyone

<figure><img src="/files/2u2tenluc9noIodRiYYF" alt=""><figcaption><p>Assignment states</p></figcaption></figure>

### Auto-Assignment

It is possible for MLM-AI to assign work evenly and automatically, avoiding manual assignment of huge amounts of articles.

Auto-assignment is a monitor-level setting, it could be enabled or disabled from Create Monitor or Edit Monitor view. Abstracts are assigned to users configured with the "Abstract Review" role in round-robin fashion.

<figure><img src="/files/hYNGLQeD6rXto3bt5iIN" alt=""><figcaption><p>Auto-assignment scope</p></figcaption></figure>

To enable auto-assignment for monitor, check "Auto-Assign Articles to Reviewer Teams":

<figure><img src="/files/7ZkbL98dQx5eRD2TvosL" alt=""><figcaption><p>Enabling auto-assignment on monitor</p></figcaption></figure>

Once the monitor is saved, results for this monitor's review are going to be assigned to team members. To avoid favoring articles to certain users, an algorithm randomizes the list of article first, and assigns them to users in a round-robin way.&#x20;

### Self-Assignment

When the **Self-Assign Citations** option is enabled in [Settings](/configuration/settings#advanced-screening-settings), users can optionally assign articles to themselves from the Related tab

Use this option to optimize the assessment of related articles by allowing the reassignment of articles to the same user.

When this option is enabled users see the "**Reassign**" option from the Related Articles tab:

<figure><img src="/files/ufXtt97Ng3pOZsSFfN9j" alt=""><figcaption></figcaption></figure>

At this point users can choose to reassign the related article and complete the assessment for both articles together. **Users can only reassign articles to themselves**.


# Article Screening

AI powered screening, save assessments according to your workflow in fully auditable workflow

In this page, all details of the citation are presented in order for users to complete their assessment. Here, users can:

* Record screening decisions according to the desired workflow: Individual Case Safety Reports, Signal detection, etc
* Upload supporting documentation as attachments: full text articles, translations, etc. in `pdf` format
  * Learn more in: [Attachments and Full text Screening](/application/article-screening/attachments-and-full-text-screening)
* Enter adverse event details for case processing and export as an E2B file (R2 or R3)
* Review all decisions recorded by users in chronological order in the Audit Log

The Main tab in the displays citation details: title, abstract and keywords, and a side bar with navigation, highlighting and input fields for recording decisions:

![Article Screening Page](/files/rRxRQkQZVPcRXeu6moG5)

{% hint style="info" %}
**Tip**: To make more room for abstracts at the screening page try collapsing the left menus. See the [navigation help page](/application/navigation#minimizing-the-navigation-menus) for tips on how to do that.
{% endhint %}

## Navigation Pane

Use the arrows to navigate to previous/next articles according to the criteria selected in the [Review Details](/application/review-details) screen.

Click the filter icon to see a summary of chosen sort and filter options in use. (recall that selected tab, filters and sort options "follow" the user to the detail screen)

<figure><img src="/files/pXJB681ygf7LR13fpz6b" alt=""><figcaption></figcaption></figure>

## Highlighted Annotations

MLM-AI automatically annotates the abstract for patient information, special situations, medications, etc. You can toggle highlights on/off for these annotations to help screen abstracts quicker.

<figure><img src="/files/3QyPlsBkdRJlcqxWEIPi" alt="" width="450"><figcaption><p>The highlighter toolbox</p></figcaption></figure>

* Any identifiable patients immediately found in the abstract can be found with the **patient** highlighter:

![](/files/-MWfSHVEBCyr7TluNdOm)

{% hint style="info" %}
Identifiable patients may also be found in the full text of the article (but not the abstract) depending on the type of study. The **suspected case** AI tag is used to predict such occurrences. See more details in [MLM-AI Concepts](/topics/mlm-ai-concepts#ai-tags)
{% endhint %}

* **Sections of interest** relate to passages in the abstract where adverse events descriptions are likely to be discussed

![](/files/-MWfQ1L6-MHg6pt3I8w8)

* The **Target** highlighter indicates the target product, while the **Medications** highlighter indicates all other medications mentioned in the abstract.
* The **special situation** highlighter will indicate any occurrence of terms related to pregnancy, elderly or pediatric uses, occupational exposure, drug misuse, etc:

![](/files/-MWfSimlKFZk3QXs5lQO)

## Saving Decisions

When screening is done, add results and click **Submit**. The decision pane provides a number of options to suit your workflow:

* The *Serious Event*, *Aggregate Reporting* and *Safety Signal* checkboxes to mark articles of interest for later reporting
* Select one or more special situations relevant to the abstract
* Select one or more categories (exclusions)
  * When the review decision is of type "Not ICSR - Exclusion Required", selecting at least one category is mandatory
* You can optionally add free text commenting on your decision

<figure><img src="/files/dVzRivxilReIvZuY44nf" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
The decisions list and values for categories and special situations are configurable. Learn more in the [Monitor Configuration](/configuration/monitor-configuration) section.
{% endhint %}

### Expand Decision View

An expanded view of the decision window is available by clicking the widget next to the Submit button:

<figure><img src="/files/sDu7rh233ZJxFRXvu3QP" alt="" width="325"><figcaption></figcaption></figure>

Clicking on it opens an expanded view which also shows the article audit log for reference. This view is useful when the comments box requires more space for editing:

<figure><img src="/files/VOfab0tHnt7ropUl4dli" alt=""><figcaption></figcaption></figure>

### Related Articles

When the same article is returned on different monitors, they become visible in the "Related" tab. Users can directly inspect related articles and also save decisions to all articles from a single operation. Learn more on:

{% content-ref url="/spaces/6pIe8UPxEvsbrxTVXkGi/pages/Mu0LriBlnPFnsziJLWNu" %}
[Reviewing Articles in Related Reviews](/application/article-screening/reviewing-articles-in-related-reviews)
{% endcontent-ref %}

### Tips for Faster Screening

When saving screening decisions, users can type into the drop down to narrow the available options:

<img src="/files/HIMAlxLQ5uz3gD2EQmyF" alt="" data-size="original">

Selecting **Next Article on Submit** using the settings toolbox will automatically move to the next article to screen after saving a decision.

![](/files/pkTEOQnqQJ1gzE0uXDAR)

## Page Locking

When Article [page locking](/configuration/settings#article-locking) is enabled, users will see a lock icon indicating lock status:

* A **green** lock means the article has been locked by the user and editing is allowed
* A **red** lock means the article is locked by another user, editing the article is not allowed.

<figure><img src="/files/bWeJN6Rf0i9xXSgJdykk" alt=""><figcaption></figcaption></figure>

After a configured number of seconds, users will see a prompt warning the lock on the page is about to expire. This ensures articles don't remain locked indefinitely:

<figure><img src="/files/QXfeEZzHs5FFJzKzJ2WC" alt=""><figcaption></figcaption></figure>

* See also: [Article Locking in System Settings](/configuration/settings#article-locking)

## Editing Annotations

Users can modify the annotations presented by MLM-AI and save their changes. This information can be used in later stages of case management.

The highlighter toolbox on the right-side menu allows users to add and remove annotations as needed:

<figure><img src="/files/g9ZM0D8SjQKlFMvDdUmq" alt=""><figcaption></figcaption></figure>

* Select the "editor" tool and select text with the mouse anywhere in the abstract to display the annotations pop-up
* Select the "eraser" tool and select any highlighter to remove it

#### Reviewing Changes

The Data tab summarizes captured terms by category.&#x20;

* Removing a term from this page will remove all the annotations for that term
* Clicking the **Reset** button restores the original annotations captured by MLM-AI

<figure><img src="/files/ojM2Qi9blYWPfeeSRtXs" alt=""><figcaption></figcaption></figure>

Any changes will be saved together with the article decision when user clicks the **Submit** button.&#x20;

## Article Details

Additional information about the article such as authors, journal and publication date are available from the "Details" tab. This tab will also provide external links to the article and full text (if freely available).

<figure><img src="/files/oBvUg0zwdbZ3Aw6Ym4Vo" alt=""><figcaption></figcaption></figure>

## Email Article

From this screen articles can be shared out to an approved (to add to this list please contact your system administrator) list of recipients using the button highlighted below:

<div align="left"><figure><img src="/files/Sdmu0760D5IagDOnUgJj" alt=""><figcaption></figcaption></figure></div>

## The Audit Log

All decisions saved by users or automatically generated by MLM-AI will be timestamped and recorded in the article's audit log ("Audit Log" tab):

<figure><img src="/files/50j7TztSPHz9bSIUt8Mz" alt=""><figcaption></figcaption></figure>

* The article audit log is also visible on data exports (see below)
* The E2B icon details whether any changes were performed on the E2B record associated with this citation, if applicable

## Exporting Data

It is possible to export article and assessment data in various formats. From the Article screen the download icon presents possible options:

<figure><img src="/files/U3eyH4TPVmd5wGqzhzPI" alt=""><figcaption></figcaption></figure>

Formats available:

* Excel containing article and assessment decision data. The "QC" version includes a column view of all recorded decisions for the article
* HTML containing the article abstract and latest assessment decision (if any)
* E2B export in XML format (R2 or R3, see E2B section below)
* ZIP: A single `zip` file containing:
  * HTML version of the abstract
  * E2B XML (if enabled)
  * All associated attachments

## Attachments

It is possible to upload reference data related to the assessment as Attachments, keeping all relevant information in a single place for the entire team.

Attachments are seamlessly displayed from the application and support automatic machine translation from a source language into English.

Learn more in the [**Attachments page**](#attachments).

## E2B

When screening for cases, it may be useful to also record detailed adverse reaction information from the article. Reaction data can be directly exported as an **E2B XML** file in either R2 or R3 formats, for processing in a safety database.

* In the E2B tab you can create one or more patients corresponding to cases from the article by clicking "Add Patient"
* Each patient can be populated with verbatim information describing the event
* MLM-AI will suggest terms for patient, medication and events, based on highlights from the abstract text
* MLM-AI will auto-populate reporter as the article's main author

<figure><img src="/files/Zo9W6JaSl2aUZKn6uUBr" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Reactions are exported into the "E2B" export format only for articles with a **confirmed ICSR decision** (See also [Custom Workflows](/configuration/custom-workflows)).
{% endhint %}

Once data is saved, an E2B R2 or R3 XML file can be generated directly from the download icon in the article page:

<figure><img src="/files/OGd1Mjr0cnoBzVNHKj64" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
See also [**E2B Configuration section**](/configuration/settings#e2b-settings) in the Settings page
{% endhint %}

{% hint style="info" %}
Depending on the final destination of E2B documentation, exports could be automated. Find out how by referring to [Integrations](/application/integrations).
{% endhint %}


# Attachments and Full Text Screening

Upload full text PDFs, translations and other relevant content to complete article assessments

From the Attachments tab it is possible to upload related files from your desktop. Attachments help consolidate and make visible all relevant assessment information in a single place. Files can be uploaded by clicking "Add Attachment":

<figure><img src="/files/ClWDX4ypvbaWEGTlJVvp" alt=""><figcaption></figcaption></figure>

* Users can attach relevant information for an article as PDF, MS-Word, Excel of Text files.
* Each attachment can not exceed `50Mb` in size

## PDF Attachments

Users can directly review full text articles and translations from the platform by uploading the corresponding article `pdf`

{% hint style="info" %}
Only valid `pdf` files that under `50Mb` in size are supported
{% endhint %}

After uploading the system will process and render the file. Progress can be monitored from the Attachments tab. Once complete the article will be viewable by clicking "Rendered" green badge:

<figure><img src="/files/rVHKUJ1YSlWWONLmhq7X" alt=""><figcaption></figcaption></figure>

Or by navigating to the Main tab and switching subtab to the full text that you recently uploaded.

<figure><img src="https://docs.biologit.com/~gitbook/image?url=https%3A%2F%2F1171269993-files.gitbook.io%2F%7E%2Ffiles%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252F6pIe8UPxEvsbrxTVXkGi%252Fuploads%252FJWnysRaIucG3b9SDrRS4%252Fimage.png%3Falt%3Dmedia%26token%3D2ba52a8c-cbbf-492c-bfa8-897bf3814264&#x26;width=768&#x26;dpr=4&#x26;quality=100&#x26;sign=2b591926&#x26;sv=1" alt=""><figcaption></figcaption></figure>

From this view some relevant annotations and highlights are provided to assist users to review the article.

### Switching PDF Viewers

You can also switch to the browser default PDF viewer by clicking the `PDF` icon. This is useful for rotating PDFs if required (the default viewer does not perform highlighting).

<figure><img src="/files/cFJWDgAvp94ngsHsKcnV" alt=""><figcaption></figcaption></figure>

### Search inside the PDF

Keywords captured from the article abstract, or any free text can be searched in the processed PDF. Use the arrows to quickly navigate to matches.

<figure><img src="/files/VkGApXtoeNzwrigcw1jx" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
**Only** keywords found in abstract will be highlighted in the PDF using the highlighter menu
{% endhint %}

## Requesting Translations

If the Translation setting is enabled on your account (via [Settings](/configuration/settings)), users can optionally request an automated machine translation into English for any uploaded `pdf` file:

<figure><img src="/files/szzJUgxpWCvDZyU87UIq" alt=""><figcaption></figcaption></figure>

* Upload file in source language
* Once processed, the translation icon will appear. Click to request a translation:

<figure><img src="/files/zAHfqXhn5G7hGcIqzuTg" alt=""><figcaption></figcaption></figure>

* Both original and translated  `pdf` files can be downloaded from the Actions menu

<figure><img src="/files/UBYeQ5LqFA45k1pWTdsP" alt=""><figcaption></figcaption></figure>

* An additional translated file will become available from the Article detail screen's "Main" tab

### Partial Translations

In certain documents it is preferrable to request a translation for a range of pages instead of the entire document (ex: if the original PDF is for an entire journal issue but only one article is relevant)

Users can customize the page range for a translation request:

<figure><img src="/files/lpDWcWSzuc6xCcGXwYSY" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Translation requests may incur additional charges to your account. Please [contact Biologit support](/more-help/support-and-platform-details) for further details.
{% endhint %}


# Reviewing Articles in Related Reviews

Complete assessment only once when the same article appears on different monitor results

## Related Articles

If your configuration contains multiple active monitors, you may see the same article present on different monitor results.&#x20;

This can happen for example when an article discussing concomitant products is returned, and these products have been configured as separate monitors.

In biologit MLM-AI, a **related article** is an article that appears in the results of more than one monitor.

Users can streamline the assessment of related articles, preventing duplicate work and increasing quality of assessments.

## Navigate to a Related Article

Any time a related article is found in another monitor result, it is presented to the user with an additional drop down:

<figure><img src="/files/OEgQMN4AW9AUSCA11Vb2" alt="" width="375"><figcaption></figcaption></figure>

Select the monitor name from the drop down to directly navigate to the related article in the corresponding review.

Alternatively the "Related" tab presents details of the related article, including the review ID and recorded decision (if any), with links to navigate to the article.

<figure><img src="/files/DsCi4Wra2tospujEc3KG" alt=""><figcaption></figcaption></figure>

## Saving Assessments for Multiple Articles

To streamline assessments, users can enter their decision once and save to all selected related articles in a single operation.

This is useful when the assessment decision *is common to all articles*. For example: no safety event is found and same exclusions apply to all related articles.

* From the Related Articles drop down, select all monitors where the decision will apply. The "Submit" button will change to "Review and Submit":

<figure><img src="/files/nfztTzz3BfwU2F9PDV9L" alt=""><figcaption></figcaption></figure>

* Click on "Review and Submit", where the common decision can be entered:

<figure><img src="/files/c9HnFi2xgkP5IS5eKZFk" alt=""><figcaption></figcaption></figure>

* Click "Submit" to save the same decision on all articles.

**Pre Requisites**

* User must have permissions to save decisions on all monitors where the related article appeared
* The monitors for each related article must have same workflow decisions


# Article-Level Day Zero

Track and update regulatory "day zero" per article for more granular workflows

Certain workflows require granular tracking of regulatory "day zero" (day where information became available for assessment).

* As standard, Biologit MLM-AI uses the date the search has run, or search results were uploaded into the system as day zero
* It is also possible to track and update Day zero information per article for more granular workflows
* This is possible by enabling "Allow Day Zero Resetting" by the Administrator in the [Settings](/configuration/settings) page

### Enable Day Zero per Article

It is required that the administrator user enables "Allow Day Zero Resetting" from the [Settings ](/configuration/settings)page:

<figure><img src="/files/Gt5BtlzV4zgtpPk330kq" alt=""><figcaption></figcaption></figure>

### Events that trigger a Day Zero Update

Once enabled users can optionally update Day Zero for an article:

* Upon uploading relevant information for the assessment (i.e. upload an attachment from the [Attachments](/application/article-screening/attachments-and-full-text-screening) tab)
* Monitor Administrators can also update day zero as an override

### Updating Day Zero on Attachment Upload

This option requires the article to be saved with a **Pending** decision (see [Workflow ](/configuration/custom-workflows#configuring-workflows)configuration)

* When uploading a file, an optional flag to update day zerobecomes available
* This will reset day zero only for this article to today's date (date of file upload)

<figure><img src="/files/N8mZ973ODGJzweyCKzJU" alt=""><figcaption></figcaption></figure>

### Updating Day Zero as Administrative Override

If necessary, users with "Administrator" or "Work Assignment" [permissions](/configuration/team-and-user-management#team-permissions) can override the day zero for a particular Article.&#x20;

For these users the day zero widget becomes editable:

<figure><img src="/files/su26EMWNtNc7O81fmP0A" alt=""><figcaption></figcaption></figure>

And the day zero date can be manually updated:

<figure><img src="/files/pQKUetkUI05Vg3ul8JhJ" alt=""><figcaption></figcaption></figure>

### Exports

In the Excel exports, the following additional columns are available to track by-article day zero

* `day_zero_owner`: User who performed the day zero update&#x20;
* `day_zero_set_date`: New day zero for this article&#x20;
* `day_zero_timestamp`: Timestamp when action was performed


# Batch Screening

Fast assessment and of abstracts meeting a given criteria

MLM-AI provides a second method for screening results "in batch". Use it for fast assessment of many articles at once according to a given criteria.

Batch Screening lets you **add a single screening decision to a collection of selected articles at once**. It is useful to quickly add decisions on results that are not likely to be relevant to your specific workflow, and to conduct faster QC tasks.

{% embed url="<https://youtu.be/B4pxywx7Bh0>" %}

To screen abstracts in batch, select results first according to desired criteria:

* From the [Review Details](/application/review-details) page: select the relevant status Tab, and desired filters. You can also include specific criteria [using the search box](/application/review-details/searching-abstracts-inside-a-review)
* Or, from the [Find Articles](/application/reviews/find-articles-from-reviews) page, select results according to desired criteria

The example below shows abstracts "In Queue" (pending screening), sorted by Animal/In Vitro tag:

* Once abstracts are of the desired status and sort order, click "**Batch Review**"
* The batch review screen will display abstracts following the same selection from the review details. In this screen, the title and abstract of selected articles are displayed
* The checkbox beside each abstract indicates the article is part of a single batch review. A single screening decision can be saved for all selected abstracts

<figure><img src="/files/SVH1hP28pJjjj4dp1fk5" alt=""><figcaption></figcaption></figure>


# Article Search

Search the MLM-AI database for hits on your custom search

The Article Search page gives you full access to the [biologit MLM-AI database](https://www.biologit.com/database) with a familiar interface and advanced features to design, preview and test your searches. You can also seamlessly [create Monitors](#create-monitors-from-searches) from your search string.

{% embed url="<https://www.youtube.com/watch?v=B0WYiHC30Hk>" %}

## Creating Searches

Search strings can be composed using logical operators like `AND`,`OR`, `NOT`.

* Use the `(` and `)` brackets to group terms together
* To search for matches of a multi-word term use quotes. Ex: `"liver damage"`
* You can end a term with `*` to match similar words.
  * Ex: `pregnan*` will match pregnant, pregnancy, pregnancies...

Examples:

| Goal                                                         | Example Query                                                                                             |
| ------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------- |
| Search for any query terms with an `OR` query                | `acetaminophen OR paracetamol OR panadol`                                                                 |
| Search for multi-word terms with quotes                      | `"liver damage" OR "liver toxicity"`                                                                      |
| Search for keywords `AND` the presence of one or more events | (`acetaminophen OR paracetamol OR panadol) AND (`"`liver damage" OR "liver toxicity" OR "liver failure")` |

### Auto-Expansion of Medication Names

Selecting a medication keyword with the mouse will display the "Expand" icon. Clicking on it will bring up a list of synonyms for selection. These will be included back in the query as an `OR` section:

<figure><img src="/files/jaxu87uPfZK7GJQRpxm8" alt=""><figcaption><p>(1) Click the expand icon to see known synonyms</p></figcaption></figure>

<figure><img src="/files/i5Nw4vVr5CBKl1LI8YJ6" alt=""><figcaption><p>(2) Select desired synonyms</p></figcaption></figure>

<figure><img src="/files/4oU74Y6nQA7fncJqfYfp" alt=""><figcaption><p>(3) The search string now expanded with the selected synonyms</p></figcaption></figure>

### Select Dates and Source Databases

<figure><img src="/files/5Lw177DRJdk6ulISg3z8" alt=""><figcaption></figcaption></figure>

Use the Start/End dates to specify the date range to search. This will retrieve articles by **publication or entry date** within the date range.

Alternatively use the range buttons to quickly select last week, month or year.

Each data source available can be toggled using the selectors.

### Advanced Query Editor

More complex searches can be edited in a larger view by clicking the advanced button

<figure><img src="/files/8KMJXeEp8xYkBMA7gYfy" alt=""><figcaption></figcaption></figure>

This view also performs syntax checking as you type.

<figure><img src="/files/iro6f8aJ2rETJVYSXEC7" alt=""><figcaption></figcaption></figure>

### Adverse Terms and Query Strings

In MLM-AI, adverse events are captured by [AI tags indicative of suspected events, animal studies and special situations](/topics/ai-enabled-screening-workflows). The AI tags are designed to replace long lists of adverse "keywords" such as `adverse event`, `adverse reaction`, etc.&#x20;

Where possible it is recommended to use the AI tags in place of explicit lists of adverse terms.

### Query String Limits

Queries are limited to 1,000 results. When this limit is reached you will be informed with the warning below. Try reducing results with a shorter date range or a less broad query.

<figure><img src="/files/PMDZCH3CuftaFPrn7lP7" alt=""><figcaption></figcaption></figure>

## Comparison Mode

When designing a query string it is important to understand the impact of adding and removing query terms. Use the **Compare tab** to perform analysis of results between two searches.

The example below evaluates differences between an open query and a more restrictive one (with `AND` clause). Click on the image to expand:

<figure><img src="/files/6yVRE4iv5q91V9u6aAlb" alt=""><figcaption></figcaption></figure>

## Analyze Mode

Finally, it is also important to estimate the expected volume of hits from the query string to help users understand and plan for screening effort, in particular for periodic searches.&#x20;

The **Analyze tab** automates this analysis and presents estimated hits obtained from historical periods.

The example below illustrates weekly volumes for a sample search over a 2 month period.

<figure><img src="/files/y8897tbwffYTJfC6wOPY" alt=""><figcaption></figcaption></figure>

## Create Monitors from Searches

When your search is ready a [Monitor](/configuration/monitor-configuration) can be created directly from this page using the *same search string and databases selected*.&#x20;

Clicking the "pill" button will direct to the Monitor configuration page, where additional monitor details can be entered.

<figure><img src="/files/a7V4zmQUnwZMZ184iqpJ" alt=""><figcaption></figcaption></figure>

The monitor configuration is similar to the default [monitor configuration steps](https://docs.biologit.com/configuration/monitor-configuration) only the search string and database information will be pre-loaded into the monitor creation form.


# Reports

Generate audit and screening activity reports for all your needs

Biologit MLM-AI supports extensive reporting of screening actions and system activity from its audit log. The reporting module is accessible from the **Reports** link from the main menu.

<figure><img src="/files/RmcqLyd4DLkq5v859KVE" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
**Reports access**: To have access to reports users must be Administrators or be part of a team with the **Reports management** system privileges. Learn more on the [Permissions section](/configuration/team-and-user-management).
{% endhint %}

To submit a new report:

* Click on **Submit Report** button on top right.
* Reports are grouped into categories. Select category and report to run (see [table below](#available-reports))
* Enter any report parameters as appropriate.
* Select delivery settings:
  * Reports can be automatically sent by email to the requesting user or to a team.
  * Alternatively reports will be available for download from the user interface.

<figure><img src="/files/jlmHbu96aJErGqLGTYNE" alt=""><figcaption><p>Reports Submission Form</p></figcaption></figure>

### Available Reports

<table><thead><tr><th width="156.33333333333331">Category</th><th>Report</th><th>Description</th></tr></thead><tbody><tr><td>Analytics</td><td>AI Tag Statistics</td><td>Detailed reconciliation report of results, decisions recorded and AI tags present</td></tr><tr><td>Analytics</td><td>Decision Statistics</td><td>Summary report of decision breakdown per monitor, for a given time period. Useful for at-a-glance view of assessment outcomes per monitor.</td></tr><tr><td>Auditing</td><td>System Activity Log</td><td>Reports all system actions (configuration changes, report submissions) from the audit trail, for a given time period</td></tr><tr><td>Auditing</td><td>Failed Logins</td><td>Reports on failed login attempts for a given time period</td></tr><tr><td>Auditing</td><td>Integration Runs</td><td>For clients with enabled integrations this audit report lists all integration actions during a time period</td></tr><tr><td>Auditing</td><td>Review Activity</td><td>Reports all review-related administrative actions (review submit, upload, status changes) for a given time period</td></tr><tr><td>Auditing</td><td>Translation Requests</td><td>Detailed reconciliation report of requests for translations performed by users</td></tr><tr><td>Auditing</td><td>Monitor Query Strings</td><td>For each monitor provide a timeline of changes that specifically altered the query string </td></tr><tr><td>Screening Activity</td><td>Reviews Activity</td><td>Complete log of screening activities for a given monitor (or all monitors) and time period</td></tr><tr><td>Screening Activity</td><td>Reviews Summary</td><td>Summary status of all reviews for a given monitor (or all monitors) and time period</td></tr><tr><td>Screening Activity</td><td><p>Decision Summary Report </p><p>(MS Word)</p></td><td>MS Word report of selected citations by date range, monitor and decision. The report includes citation and assessor details</td></tr><tr><td>Screening Activity</td><td>Review Summary by Journal (Local Literature)</td><td>Summary of results retrieved grouped by journals</td></tr><tr><td>System Configuration</td><td>Monitors</td><td>Export of all Monitor definitions</td></tr><tr><td>System Configuration</td><td>Users and Teams</td><td>Export of all configured users and corresponding teams</td></tr><tr><td>System Configuration</td><td>Monitors to Team Permissions</td><td>Mapping of monitors to teams and their roles (review or QC)</td></tr><tr><td>Data Archiving</td><td>Archive completed reviews</td><td>Moves all reviews in Completed status to Archived, for a given monitor and minimum review age</td></tr></tbody></table>

### Report Results

After submission, reports are queued for processing and will become available in the **Recent** tab once done. As reports complete, older runs can be found from the **Find Reports** tab.

Reports can be downloaded via the links indicated below. Report execution logs can also be downloaded.

<figure><img src="/files/CUbWIZwh5kqu3MjA6sQK" alt=""><figcaption></figcaption></figure>

### Day Zero Option

In Review-based reports you can optionally select reviews based on **day zero** (date of creation) instead of review date from-to range:

<figure><img src="/files/lHrSwSf6y5dxkDKT1EwI" alt=""><figcaption></figcaption></figure>


# Dashboards

Analyze data screened in MLM-AI with powerful dashboards

## Summary View

The summary review provides visualization of review statistics, by progress, assignment status, AI tags and due dates.

<figure><img src="/files/sxCXV5j0Y3yxZ4NyDKcE" alt=""><figcaption></figcaption></figure>

**Days Open Chart**

* The Days Open chart organizes selected Reviews in descending order of days open
* Chart will be colored green/red according to days open vs. ETA, [as defined in the Monitor](/configuration/monitor-configuration#sources-and-search-schedule)
* Clicking on the corresponding review in the Days Open chart will lead directly to the Review screen

## Trends

Trends dashboards can visualize literature screening process metrics over time:

<figure><img src="/files/Ruf5i9TYsXhIiVQVjWZz" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
**Access:** Dashboards view is available to Administrators and members of teams with Monitor management privileges. Depending on a dashboard, a user might also need Reviews and Reports management access. More details could be found under the [Permissions section](/configuration/team-and-user-management#users-and-teams-management).
{% endhint %}

### Usage

* Choose a target dashboard and click on a blue filter button
* Fill in the parameters, like start and end date, monitor, etc.

<div align="center"><figure><img src="/files/jvl9cSHZ8D59jQat4Hfb" alt="" width="380"><figcaption></figcaption></figure></div>

* Click **Find**. Once data is retrieved, it will show up on the dashboard

<div align="center"><figure><img src="/files/NpSX1Aq55NrLpm2yQORm" alt="" width="563"><figcaption></figcaption></figure></div>

### Table View and Downloads

* You can enable Chart and Table views by clicking the respective chart/table icons near the filter.
* Data can also be exported using the Download icon. The dashboard exports (Charts view) or Excel (Table view)

<figure><img src="/files/glrvl6ofIoO4nOBK0skE" alt=""><figcaption></figcaption></figure>

## Available Dashboards

<table><thead><tr><th width="182">Category</th><th width="393">Dashboard</th><th>Parameters</th></tr></thead><tbody><tr><td>Result Trends</td><td><strong>Screening Volume</strong>: Returns a number of articles retrieved in MLM-AI searches on a given day for a given monitor</td><td><ul><li>Start date</li><li>End date</li><li>Monitors</li></ul></td></tr><tr><td>Result Trends</td><td>Screening Volume AI Tags: For a single monitor, shows a number of articles retrieved on a given day by associated AI tags</td><td><ul><li>Start date</li><li>End date</li><li>Monitor</li></ul></td></tr><tr><td>Assessment Insights</td><td>Decision Breakdown: Compare decisions recorded for a group of monitors and date range</td><td><ul><li>Start date</li><li>End date</li><li>Monitors</li></ul></td></tr></tbody></table>


# Integrations

Automate recurring data operations: prior or following work performed in MLM-AI

Integrations enable automated, event-driven data flows by allowing data to move seamlessly into and out of the platform as part of predefined workflows. This removes the need for manual file handling and ensures that data exchanges happen reliably and at the right time.

{% hint style="info" %}
Integrations are currently available on request. Automated workflows are configured and managed by our team in close collaboration with customers. Contact the [support desk](/more-help/support-and-platform-details) for further details.
{% endhint %}

Once integrations are enabled, configured, and the required conditions are fulfilled, execution details are available under the Integrations icon:

<figure><img src="/files/ciZLffc5xnzVAjgFHKkb" alt=""><figcaption></figcaption></figure>

Automated workflows can be initiated by specific events, including:

* Scheduled or time-based triggers
* Completion of a process within the platform (submitted decision, processed review, etc.)

Integrations currently support a limited but growing set of automation scenarios, including:

* Ingestion of results received by email, enabling files to be uploaded into the platform without manual intervention. Refer to [Uploading Review Data](/application/upload-data) for more details about the manual process.
* Export of documentation to external destinations, allowing outputs generated in the platform to be delivered to third-party systems (safety databases) or storage locations (SFTP). See [Article Screening](/application/article-screening) guide for a full list of documentation for export.


# Monitors

A Monitor configures a product search, its schedule, automation settings, and what teams get to screen and QC results.

In MLM-AI, a **Monitor** defines the product search and how results will be screened. In a monitor you define:

* The search terms to include in the query (ex: product name and relevant synonyms)
* What literature databases to search
* How often should results be produced (weekly, daily, or on demand)
* Permissions: what teams will perform reviews
* Notification and automation to be carried out by MLM-AI

Watch the walk through below to learn more about configuring search strategies in MLM-AI using Monitors:

{% embed url="<https://youtu.be/KmV9S-yvYhA>" %}

{% hint style="info" %}
To configure monitors, users must be part of a team that grants one of the following [permissions](/configuration/team-and-user-management#role-based-permissions):

* System Administrator
* Monitor Management
  {% endhint %}

## Creating Monitors

To create a new monitor, click "**+ Add Monitor**" from the Monitors menu.

![Click - Add Monitor button](/files/n3pGE5KQKfBAKlRelqZC)

Alternatively chose to **Copy** the definitions from an existing monitor if you wish to replicate existing configuration settings.

Give your new monitor a name: **e.g. Oxytocin**

![A new setup page open once you click on Add Monitor](/files/qhp7aphh1e7gjc1OA7iG)

### **Search Terms**

To add search terms just enter the product name or compound in the keyword or compound field.&#x20;

This will auto-populate a list of synonyms to choose from in the box below titled Available synonyms.&#x20;

You can **select the synonyms** from the list or click on **select all**. After that click on **"Add Selected"**.&#x20;

<figure><img src="/files/bPSXb4GdwByySqcQ7GkY" alt=""><figcaption></figcaption></figure>

Your selected terms will appear in the "**Chosen Search Terms**" box. Items in this list will be searched as an OR-based search strategy.

{% hint style="info" %}
Click "🔍**Preview Results**" to run the search in the MLM-AI database. This will let you browse results and estimate the expected volume of hits.
{% endhint %}

### User Defined Keyterm Highlights

Additional relevant terms can be added to a monitor in the "User Defined Keyterms" section. Use this feature to create additional highlights that facilitate screening. For example: terms of interest (symptoms, concomitant medications) relevant to particular products

Select the "Apply user defined keyterms" checkbox to enable the keywords list. Terms can be added one at a time or by copy/pasting a list using "Multi-Line"

<figure><img src="/files/f14kVKqi6TWO5dz7dtuJ" alt=""><figcaption></figcaption></figure>

### **Sources and Search Schedule**

The MLM-AI Database ingests data from various literature sources daily, including repositories of global reach (PubMed, Crossref, DOAJ) complemented by various regional repositories.

Select which ones to use when searching with this monitor as shown below.

{% hint style="info" %}
Learn more about the repositories available in the MLM-AI Database in our [FAQ on data sources](/more-help/faq#data-sources-and-searching-abstracts).
{% endhint %}

You can also customize search schedule the Monitors to run periodically (weekly or daily) or on demand. When a monitor runs on a schedule, "**Day Zero**" is the weekday when results will be produced (if monitor runs weekly). "**ETA for Completion**" specifies a number of days expected for screening completion.

<figure><img src="/files/6gSTrk4cSm87uOAnFhbB" alt=""><figcaption><p>Monitor repositories and schedule configuration</p></figcaption></figure>

### **Team Assignment, Special Situations and Exclusions**

This section assigns teams to screening stages, giving users permissions to update decisions on incoming abstracts.

**Team Assignment**

Teams can be assigned the following roles in a monitor:

* Abstract Review - members of teams in this role can add screening decisions
* Quality Review - members in this role can add or update other user's decisions.

The Quality Review role is assigned to teams in charge of "QC" activities. Multiple teams can be assigned to either Abstract Review or Quality Review roles.&#x20;

![Team Assignment and Workflow - Review Settings](/files/CUtp4vUVdfEuJlPaDWkk)

{% hint style="info" %}
Learn more about team and user setup on: [Team and User Management](/configuration/team-and-user-management)
{% endhint %}

**Special Situations and Exclusions**

When users save decisions, they can select categories for exclusions and special situations. The selections are multi-choice and come pre-populated with values typical for ICSR screening.

The values can be configured per your workflow: existing entries can be removed and customized as needed.

![Workflow - Review Settings](/files/CWqvLfZY4W3gzjSBTRRZ)

## **Monitor Options**

Additional settings control monitor behavior on automatic screening and date searches

### **Tag-based Screening**

For high volume screening workloads, users can configure [automated screening based on AI tags](/topics/ai-enabled-screening-workflows#options-for-screening-workflows-with-ai-tags). Pre-screened results receive a corresponding exclusion and are always available in MLM-AI Review details for inspection.

<figure><img src="/files/6iItQEmB8K9qwlSKaKhM" alt=""><figcaption></figcaption></figure>

| Setting                                                            | Description                                                                                                                      |
| ------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------- |
| Automatic EXCLUSION of tag: Animal/In Vitro                        | Results with the tag "Animal/In-Vitro" are automatically screened with Animal/In-Vitro study as the exclusion criteria.          |
| Automatic INCLUSION of tag: Suspect AE                             | All results except the ones with the tag "Suspect AE" are automatically screened with exclusion "Missing adverse event"          |
| Automatic INCLUSION of tag: Suspected Case or Identifiable Patient | All results except the ones with tags "Patient" and "Suspected Case" are automatically screened with exclusion "Missing patient" |

#### **Combining Settings**

* When more than one INCLUSION setting is used together, MLM-AI will accept any of the chosen criteria for inclusion.

### Duplicate Handling

Selecting "Automatic Exclusion of duplicate articles" will automatically [screen duplicates](#duplicate-handling) with the corresponding exclusion. This option is on by default.

<figure><img src="/files/qHyDw1aoqMz9jIi6Ajjl" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Learn more about how the techniques used in automated duplicate detection on: [**Feature  spotlight: Duplicate Detection in Biologit MLM-AI**](https://www.biologit.com/post/feature-spotlight-duplicate-detection-in-biologit-mlm-ai)
{% endhint %}

### **Date Search Criteria**

The following settings control how dates are applied on search results:

<figure><img src="/files/7EA9KaKmZpKiEUv5Sjad" alt=""><figcaption></figcaption></figure>

* **Include abstracts found by publication date** - retrieve abstracts where the publication date matches the selected date range of the review. This option is on by default and helps maintain correspondence to existing manual search strategies
* **Discard old publication dates** - if the publication date is older than a number of days from the search start date, results are discarded. It is possible to select the "cut off" point for old publication dates (e.g. 60, 90 or 120 days). This option is useful to avoid spurious results from re-publication of journals going back many years that can cause duplication in assessments
* **Postpone weekend searches** - when selected, schedule for daily searches will adjust accordingly:
  * The following searches will not be created: Friday (runs on Saturday), Saturday (runs on Sunday), Sunday (runs on Monday)
  * One consolidated search will be created: Friday-Sunday (runs on Monday)

{% hint style="info" %}
See also [**Handling Article Dates**](/topics/handling-article-dates) to learn more about how MLM-AI uses dates effectively.&#x20;
{% endhint %}

### Notifications

Monitors support a number of notification options that let users:

* Be notified when a review is ready for screening, or when its status changes (ex: from *In Progress* to *Completed*)
* Be notified of any assessment updates, when they are part of an assessment (per the article audit log)
* Send the details of an abstract and its assessments to a pre-configured list of recipients

<figure><img src="/files/9bqIyvZJ2egEgHnAqApd" alt=""><figcaption></figcaption></figure>

For further details, visit the [**guide to configuring Notifications**](/configuration/notifications) in MLM-AI.&#x20;

### Filters

#### Query Filters

Monitors can be configured with pre-defined query filters. Pre-defined filters use the same options available on the [results search box](/application/review-details/searching-abstracts-inside-a-review) and appear as shortcuts to speed up common queries.

In the example below, the monitor is configured with one pre-defined filters matching conditions of interest and one for author country affiliation.

<figure><img src="/files/0LWWzukfGzyWpMzaPqNP" alt=""><figcaption></figcaption></figure>

👉 Learn more about querying review results in : [**Searching Inside a Review**](/application/review-details/searching-abstracts-inside-a-review)

#### Filter by Publication

When assessment workflows require the monitoring of only a subset of publications (ex: for a given coutry or region) the Monitor can be configured to automatically filter based on a list. Biologit will automatically filters articles based on the list.&#x20;

Source filters are available inside the Filter tab of monitor configuration and allow users to create a pre-defined list of relevant sources:

<figure><img src="/files/yWt5d6iSR1uOJd8HoFpk" alt=""><figcaption></figcaption></figure>

Relevance of an article is defined based on the ISSN of the publication, and an article is considered relevant if ISSN is present in the configuration list. Otherwise, article is marked as non-relevant and is automatically excluded.&#x20;

If source filters are configured for a monitor used inside a review, relevant articles would will stay in "In Queue" tab, while non-relevant ones will be moved to "Reviewed" tab and marked with a respective comment:

<figure><img src="/files/eNMbPIGbNW0HgSmQrl35" alt=""><figcaption></figcaption></figure>

Learn more about tabs used in MLM-AI reviews in [**Review Details - Tabs**](/application/review-details#tabs).

### Annotations

If additional data entry fields are required to complete the assessment, they can be configured as Annotations:

* Annotations are shown to users under the "Annotations" tab in Article Details. Saving a decision will also save any entered annotation data
* Annotations are visible as additional columns in the Excel export and the Article summary in HTML
* Annotations allow for free text or drop down (single or multiple choice)

Example:

<figure><img src="/files/uyL0CAai3yP2L1Es3E3E" alt=""><figcaption></figcaption></figure>

### Custom Workflows

Custom screening workflows can be configured at monitor level. See the [**Custom Workflows**](/configuration/custom-workflows) page for more details.

### 1-Click Decision

For workflows where many assessments result in a straightforward action (ex: a rejection), it is possible to configure a visual shortcut to allow users to complete the action with one click.

In the workflow section users can add up to three 1-click decisions that will appear as shortcuts in the Article Screening page:

<figure><img src="/files/ZrGVH8556QkOUni4QhXZ" alt=""><figcaption></figcaption></figure>

For each shortcut, it is possible to configure the label, decision and any additional annotations, exclusions or special situatios to be pre-filled:

<figure><img src="/files/bCt4UEJTRowm10pVFp1a" alt=""><figcaption></figcaption></figure>

It is also possible to select the color of the shortcut button.

Once configured, this option will become visible from the Article Screening page. Clicking the shortcut will:

* Populate decision and annotations according to the 1-click configuration
* Save the decision and automatically navigate to the next article

<figure><img src="/files/idMTNc80dvoI42d7xT0k" alt=""><figcaption></figcaption></figure>

## Updating Monitors in Batch

When managing a large collection of monitors, it is sometimes useful to consistently update the same settings to all monitors at once. Monitors can be selected and updated in batch in a single operation.

* From the Monitors screen, select relevant monitors
* Then from the ". . ." menu (top right) choose "Batch Update"

<figure><img src="/files/qtZwHt8S2pISUz7UN46B" alt=""><figcaption></figcaption></figure>

You will be asked to select a reference monitor from the chosen ones. The details of the selected reference monitor will be populated for editing and can be used as the starting point for batch update.

### How Batch Updates Work

* When in batch mode, only certain aspects of monitors can be updated. Individual data items such as monitor title and query string details can not be updated in batch
* Additional checkboxes will appear in batch mode, representing areas that will be updated. Only selected areas will be applied when saving the batch operation

<figure><img src="/files/q96OPNMAn06F2KHnPd9H" alt=""><figcaption></figcaption></figure>

* Selecting a specific area enables it for editing. This means this area is now marked for batch update.

<figure><img src="/files/ffOSVmIjAjtBRIHPy046" alt=""><figcaption><p>Once selected, the "Database Sources" area became editable and will be updated when batch-saving</p></figcaption></figure>

{% hint style="info" %}
Any configuration area "checked" during batch review will save its content as they appear on the screen to all selected monitors
{% endhint %}

#### Confirm Save

A confirmation dialog will inform users what sections of the monitor configuration will update.

Ex:

<figure><img src="/files/EkMlVUQ7qBX4QhgaPICi" alt=""><figcaption></figcaption></figure>


# Users, Teams and Permissions

Configure users and permissions in MLM-AI

## Users and Teams Management

Administrators or users with **User Management** privileges can manipulate teams and users. From the User Management menu, use the Teams and Users tabs to create, modify or deactivate users and teams:

<figure><img src="/files/IMADg1BTLnCF9yesv5pB" alt=""><figcaption></figcaption></figure>

## Adding Users

Selecting the Users tab, then "Add User" opens the user creation page.

* The user email address is also its login ID
* Administrators can toggle the "Require password reset" when creating a user for the first time, in order to force password change
* Select teams to assign to the user. Multiple teams can be chosen

<figure><img src="/files/3NrBm2ye85KQjk4rM0Rj" alt=""><figcaption></figcaption></figure>

### Overriding User ID

By default a user's ID is the user email address. This can be overridden in cases where these need to be different (for example with test accounts). Use the **Custom User ID** toggle to override the user ID.

Resetting Passwords

Supervisors can quickly reset a user password to a random password using the **🔑action** from the user summary view:

<figure><img src="/files/eVYKQoAaBx9hB2Iy066W" alt=""><figcaption></figcaption></figure>

The system will present the administrator with the random password and login instructions that can be shared with the user.

## Team Permissions

In MLM-AI, teams are containers for permissions. Users are assigned teams who in turn gives access to the specific product functionality.

There are two ways users are granted permissions to MLM-AI functionality:

* By being added to a team configured with one or more application-level permissions
* By being added to a team who is then assigned to a monitor

Lets look at each of these in turn:

### Application-level Permissions

Users are granted role-based permissions according to the teams they are added to. These permissions can be selected from the Teams configuration page:

<figure><img src="/files/sBYtXnMeJKi41S5vWS2S" alt=""><figcaption></figcaption></figure>

The following permissions can be configured to a team:

| Permission/Role                             | Permissions granted                                                                                                              |
| ------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| System Administrator                        | Full access to all of MLM-AI functionality.                                                                                      |
| User and team management                    | Create and modify teams and users.                                                                                               |
| Monitor management                          | Create and modify [Monitors](/configuration/monitor-configuration).                                                              |
| Reviews Supervisor (ALL monitors)           | Can manipulate any reviews for any monitors: change status, submit, upload, add screening decisions.                             |
| Review upload (ANY assigned monitor)        | Can create [upload reviews](/application/upload-data) for any assigned monitor.                                                  |
| Reports Management                          | Submit and schedule reports.                                                                                                     |
| Review submit ad-hoc (ANY assigned monitor) | Can submit [on demand](/application/reviews#submitting-reviews-on-demand) reviews for any assigned monitor.                      |
| Data export (ANY assigned monitor)          | Can [export ](/application/reviews#downloading-reviews)contents of reviews (Excel, etc) for external processing.                 |
| Work assignment                             | Ability to assign/unassign abstracts when the [Work Assignment](/application/review-details/work-assignment) feature is enabled. |

{% hint style="info" %}
Making a team inactive will disable any permissions granted by this team.
{% endhint %}

### Read-only Permissions

Users who need to be restricted from doing changes in the platform (independently from application-level permissions) can be configured as read-only users by enabling this user flag on creation/edition.

The following restrictions will be applied to read-only users:

| **Configuration**                          |                                       |
| ------------------------------------------ | ------------------------------------- |
| Create or update monitors                  | `forbidden`                           |
| Export monitors                            | `allowed` (if core permissions allow) |
| Create or update teams and users           | `forbidden`                           |
| Export teams and users                     | `allowed` (if core permissions allow) |
| Create or update system settings           | `forbidden`                           |
| Export system settings                     | `allowed` (if core permissions allow) |
| Change my password                         | `allowed`                             |
| **Reports/Exports**                        |                                       |
| Run Reports                                | `allowed` (if core permissions allow) |
| Export reviews                             | `allowed` (if core permissions allow) |
| **Reviews**                                |                                       |
| Navigate to review results and assessments | `allowed` (if core permissions allow) |
| Submit or upload reviews                   | `forbidden`                           |
| Save decisions                             | `forbidden`                           |
| Transfer articles                          | `forbidden`                           |
| Add attachments                            | `forbidden`                           |
| Request translations                       | `forbidden`                           |
| Article search                             | `allowed`                             |
| Email articles                             | `forbidden`                           |
| Access Dashboards                          | `allowed`                             |
| AI actions                                 | `forbidden`                           |
| **Assignment**                             |                                       |
| Assign citations                           | `forbidden`                           |
| Be assigned citations                      | `forbidden`                           |

### Assign Teams to a Monitor

When a team is assigned to a Monitor, users in that team are granted access to review results from this Monitor, in order to complete screening activities.

This step is done in the Monitor screen, during [Monitor configuration](/configuration/monitor-configuration)

<figure><img src="/files/7Kn53ICKHPn9uiWGQuU6" alt=""><figcaption></figcaption></figure>

Members of a team assigned to "Abstract Review" can:

* Access to search and inspect contents of the review results
* Save screening decisions when no decision yet exists (initial assessment), or update their own decision
* Users in this role can not update screening decisions from another user (*except* when Work Assignment is enabled - see below)

Members of a team assigned "Quality Review" role can:

* Access to search and inspect contents of the review results
* Save screening decisions to any article, whether or not they have been updated previously by another user

#### Work Assignment Mode

When the [Work Assignment](/application/review-details/work-assignment) feature is enabled, the permissions of a user may be *extended* so that the user can complete a review:

* Ordinarily, a user in a team given the "Abstract Review" role is unable to update another user's decision.
* However, if the user is *directly assigned* the abstract, the user is granted permission to update that article. This permission is only valid for as long the user is assigned the article.&#x20;

## Password Resets

Users have the ability to reset their passwords at any time, however in case of credentials lossor reactivation, Administrators can quickly reset a user password from the "key" action icon:

<figure><img src="/files/j7ISEWS1UhvgdTVZ1A5l" alt=""><figcaption></figcaption></figure>

Doing do will auto-generate a random password and details that can be shared with the user by copy/pasting into an email:

<figure><img src="/files/Y5pkh2Ax0UG93kpwHMAP" alt=""><figcaption></figcaption></figure>


# Settings

Global settings affecting application behavior can be configured here.

Administrators can customize application behavior by configuring global settings in the Settings page.

<figure><img src="/files/XZ5pJbadsdhBeUvKMcpL" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
**Permissions**: Only Administrators (ie. members of a team that grants the Administrator permission) have access to the Settings page.
{% endhint %}

### Saving Settings

* Settings will become effective once user logs off and back into the application.

## Settings Sections

### Additional Products

This section configures what additional products are configured on your environment. Note some settings are read-only and enabled by the Biologit support team.

* **Local Literature**: This setting indicates whether Biologit Local Literature product is enabled
* **Translation Engine**: Selecting a translation engine will enable citation and full text machine translation capabilities to all results.

<figure><img src="/files/6FxsYuCmqiRfb8bzUNIW" alt=""><figcaption></figcaption></figure>

### AI

The MLM-AI Model Release setting determines the version of AI Models used for new reviews. For more information on models used see [MLM-AI factsheets](https://github.com/biologit-engineering/factsheets/raw/main/mlm-ai/PROD5/biologit_factsheet_1.0.5.1.pdf)

The **Latest** setting is recommended: this will always use the most recently available model version.

Changing model versions only affects results produced since the change - historical results are unaffected.&#x20;

<figure><img src="/files/8pe1D2fKvVEzj5SLzsvV" alt=""><figcaption></figcaption></figure>

### Article Locking

When Article Locking is enabled, only one user is allowed on any given article at a time. This ensures that no two team users will update the same article without each other's knowledge.

There are two settings that determine locking behavior:

* **Article Page Lock**: (Yes/No) This enables/disables the locking feature for all users.
* **Article Page Lock Timeout:** The number of seconds a user is allowed to stay in an article page before the lock expires. The user will see a prompt to continue on the page.

The below example configures the locking feature with a timeout of 10 minutes:

&#x20;

<figure><img src="/files/nwNoID4z7Ij5Y5DAoDUU" alt="" width="563"><figcaption></figcaption></figure>

### E2B Settings

MLM-AI users can produce XML exports in [E2B R2 and R3 formats](https://www.ich.org/page/e2br2-icsr-specification-and-related-files) for screening results marked as ICSR.&#x20;

{% hint style="info" %}
See the [**E2B Configuration Guide**](/configuration/e2b-configuration-guide) for further details on setting up this functionality.
{% endhint %}

### Security

The system can be configured with timeout (in minutes) that forces users to log back in due to inactivity:

<figure><img src="/files/8xXqsMy442tKKUwKh16g" alt="" width="347"><figcaption></figcaption></figure>

To protect against password break-in attempts administrators can also configure maximum login fail attempts and account lock period (in minutes)

<figure><img src="/files/2YLysw0DZWr1wI8IeUq9" alt=""><figcaption></figcaption></figure>

### Advanced Screening Settings

<figure><img src="/files/kjMoTPRPdgQqjoZe4Ilz" alt=""><figcaption></figcaption></figure>

<table><thead><tr><th width="288">Setting</th><th>Description</th></tr></thead><tbody><tr><td><strong>Use Screening Checkboxes</strong></td><td><p>When set to <strong>No</strong> will cause the checkboxes for signal, serious event and aggregate reporting to be disabled</p><p>(this is the recommended setting) </p></td></tr><tr><td><strong>Work Assignment Mode</strong></td><td>Enables the <a href="#work-assignment"><strong>Work assignment</strong></a> feature to all Monitors.</td></tr><tr><td><strong>Allow Day Zero Resetting</strong></td><td>When uploading additional information as attachment, allow users to optionally reset article "day zero" to reflect the data new information became available.</td></tr><tr><td><strong>Assessment labels</strong></td><td>Not in use, please contact support</td></tr><tr><td><strong>Review splitting</strong></td><td>Split each review created in the system <code>300</code> hits at a time (recommended)</td></tr><tr><td><strong>Self Assign Citations</strong></td><td>Allow users self-assign articles from the Related Articles tab. (<a href="/pages/djYljwCKpBefD8WcgjVB#self-assignment">more on self-assignment</a>)</td></tr></tbody></table>

{% hint style="info" %}
[Contact support](/more-help/support-and-platform-details#customer-support) to further discuss how to best apply advanced screening settings to your requirements
{% endhint %}

## The Workflow Tab

Within the Settings page, the **Workflow** tab shows the system's default screening workflow.&#x20;

The default workflow is applied to all reviews where a custom workflow has not been defined in the monitor configuration. It can be configured from this tab, and is pre-populated with a standard ICSR (Individual Case Safety Report) screening workflow.

{% hint style="info" %}
To configure workflows, see [**Custom Workflows**](/configuration/custom-workflows) for more details
{% endhint %}


# Custom Workflows

Customize screening decisions to suit your workflow

## Introduction

The Workflow is how assessment options can be configured in Biologit, allowing decisions to map your assessment and QC workflow steps.

* A default Workflow can be configured in the [Settings](/configuration/settings) page, which is available when creating new Monitors
* For greater flexibility Workflows can also be configured for each specific monitor&#x20;

<figure><img src="/files/SOuE5CF4PByzraTL008G" alt=""><figcaption></figcaption></figure>

## Configuring Workflows

* Add as many **decisions** as needed
* Each decision is made of:
  * A **label**, indicating the choice presented to users and shown on reports
  * A **type**, indicating how the system should behave for a particular screening decision
  * Optional rules that determine "next stage" decisions can be selected, what annotations and categories are visible (see Decision Rules below)

### Decision Types

Decision types affect how MLM-AI handles the specific screening decision. The table below outlines how decision types work:

<table><thead><tr><th>Type</th><th width="241.33333333333331">Description</th><th>Behavior</th></tr></thead><tbody><tr><td>Reviewed - ICSR</td><td>Article considered screened and flagged as an ICSR</td><td><ul><li>Appears in "Reviewed" tab</li><li>Exported in E2B files</li></ul></td></tr><tr><td>Reviewed - Exclusion Required</td><td>Article considered screened; not flagged as ICSR</td><td><ul><li>Appears in "Reviewed" tab</li><li>Requires one or more exclusion categories when saving</li><li>Not exported in E2B files</li></ul></td></tr><tr><td>Reviewed - Other</td><td>Article considered screened; not flagged as ICSR</td><td><ul><li>Appears in "Reviewed" tab</li><li>Not exported in E2B files</li></ul></td></tr><tr><td>Pending</td><td>Article pending further action in order to complete screening</td><td><ul><li>Appears in "Pending" tab</li></ul></td></tr></tbody></table>

### Example: Workflow for ICSR

In the below example we configure a simple ICSR workflow with 3 x decisions:

* "Confirmed ICSR" - the result of the assessment identified an ICSR
* "Pending further information" - the assessment is not yet complete and needs further details (full text, translation, author follow up, etc)
* "Invalid ICSR" - the result of the assessment resulted in not an ICSR
  * In this case we selected the decision type "Exclusion Required": users selecting this option must enter at least one exclusion

<figure><img src="/files/yoE0dRai6dRgPEmB5r0P" alt=""><figcaption></figcaption></figure>

#### Enhancement: Multiple Pending Decisions

Workflows can be as specific as needed. Lets expand the options of "Pending" decisions to fully describe what additional information is required:

<figure><img src="/files/WkwSItyDZHbv4UmEw2dP" alt=""><figcaption></figcaption></figure>

## Adding Rules to Decisions

In multi-stage workflows it is helpful to configure the possible decisions allowed once a certain decision is recorded.

Consider for example an assessment workflow with distinct "Initial" and "Confirmed" stages. With **decision rules** the workflow can be configured such that:

* When no decision is yet recorded, only decisions for the "Initial" stage can be selected
* When the assessment is recorded, only decision for the "Confirmed" stage can be selected

Configuring rules reduces the risk of user error when saving decisions, improving overall process quality.

To configure decision rules:

* populate the "show decisions" drop down with the allowed decisions
* select "default decision" if the decision is to appear by default, when no assessment has yet been recorded

The below example configures "Confirmed" decisions as allowed once "Initial Assessment" decision has been recorded:

<figure><img src="/files/Su8WR3CeYLMzC7fjvxH9" alt=""><figcaption></figcaption></figure>

### Mandatory and Disabled Annotations

Decisions can control the behavior of [Annotations](/configuration/monitor-configuration#annotations) on data entry: when user selects a decision corresponding annotations can be configured as mandatory, or disabled (field greyed out).

In the below example "Product" is mandatory for the "Valid ICSR" decision, and "Custom" is disabled:

<figure><img src="/files/P4LJ9WOQatLfAFbl7Dr1" alt=""><figcaption></figcaption></figure>

### Categories to Display

Decisions can also control what categories to display. This can help simplify data entry on complex workflows with granular categories to select from.

You can select which of the categories configured in the monitor are valid for a specific decision, per the example below:

<figure><img src="/files/adrzzYS55fnOFfwHAW02" alt=""><figcaption></figcaption></figure>

### Testing Rules behavior&#x20;

The Workflow Settings tab also include a testing tool to verify rules behavior is configured as expected. From here users can simulate saving a decision and verify values allowed for the next decision on the workflow.

<figure><img src="/files/2do8X1QWlrGiU6C3Zund" alt=""><figcaption></figcaption></figure>

### Overriding Decision Rules

If an assessment has been recorded by mistake, users have the option to override decision rules in order to see the original selection of allowed decisions.

From the Article Detail screen, select the expanded decision view -> ![](/files/lIBJ0TOUVFeduTeu0WIh)

Then, select "Override Decision Rules" to enable all possible decisions.

Alternatively click the "clock" icon to "Revert to Previous Decision": this will display drop down values that were available prior to the current decision being recorded.

<figure><img src="/files/B5xhHTMjefTV4BHAw524" alt=""><figcaption></figcaption></figure>

&#x20;&#x20;


# E2B Configuration Guide

Export results screened as ICSR direct into E2B XML files

MLM-AI can generate E2B XML files in R2 and R3 format as an export option, containing verbatim information on patient, drug and event data. E2B is a widely used format for exchanging safety information, for example for case creation in a safety database.

Once enabled, a new "E2B" tab is shown in the [Article Details](/application/article-screening#e2b) screen allowing users to add case details to fields expected by the E2B standard. The E2B tab is itself a tabbed view where the fields correspond to E2B elements in the exported XML file (see below for a mapping of each field in the R2 and R3 versions).

<figure><img src="/files/GnPdsCJKO63XLnhBpaYs" alt=""><figcaption></figcaption></figure>

This guide walks you through how to setup MLM-AI for E2B exports according to your company requirements.

## E2B Settings

In the [Settings](/configuration/settings) page, the following E2B settings are configured *once*, and will be used at every XML export.

* To enable E2B tab, set **E2B Enabled** to either `Yes - R2 only` or `Yes - R2 & R3`(Recommended)
* To enable append-only behavior when adding details to E2B, Set **Allow Patient Deletion** to `No` (Recommended)&#x20;
* To configure E2B exports for your environment, see the reference field mapping below.

Below is an example of a test configuration sending from an "other" organization (ex: service provider) to a pharmaceutical company:

<figure><img src="/files/Eg8A72JVdOFHLHJofRO8" alt=""><figcaption></figcaption></figure>

## Reference E2B Field Mapping

In the below tables we provide a mapping between the fields in MLM-AI and the International Council for Harmonisation's Individual Case Safety Report (ICSR) Document Type Definition.&#x20;

* For E2B R2 we reference version 2.1 which is available [here](https://www.ich.org/page/e2br2-icsr-specification-and-related-files)&#x20;
* For E2B R3 we reference the Implementation Guide for ICSR's which can be downloaded [here](https://admin.ich.org/sites/default/files/inline-files/E2B%28R3%29_IG_Complete_Package_v1_09.zip).

### Patient sub-tab

| UI Field Name | Corresponding E2B R2 Field | Element Code (R2) | Element Code (R3) |
| ------------- | -------------------------- | ----------------- | ----------------- |
| Patient       | patientinitial             | B.1.1             | D.1               |
| Gender        | patientsex                 | B.1.5             | D.5               |
| Age           | patientonsetage            | B.1.2.2a          | D.2.2a            |
| Unit          | patientonsetageunit        | B.1.2.2b          | D.2.2b            |
| Group         | patientagegroup            | B.1.2.3           | D.2.3             |

### Drugs sub-tab

| UI Field Name | Corresponding E2B R2 Field                                                                         | Element Code (R2) | Element Code (R3) |
| ------------- | -------------------------------------------------------------------------------------------------- | ----------------- | ----------------- |
| Target        | <p>medicinalproduct<br>{<em>with drugcharacterization (B.4.k.1) set to 1 for target drug}</em></p> | B.4.k.2.1         | G.k.2.2           |
| Concomitant   | <p>medicinalproduct<br>{<em>with drugcharacterization (B.4.k.1) set to 2 for target drug}</em></p> | B.4.k.2.1         | G.k.2.2           |

### Events sub-tab

| UI Field Name | Corresponding E2B R2 Field | Element Code (R2) | Element Code (R3) |
| ------------- | -------------------------- | ----------------- | ----------------- |
| Events        | primarysourcereaction      | B.2.i.0           | E.i.1.1a          |

### Other sub-tab

| UI Field Name            | Corresponding E2B R2 Field                          | Element Code (R2)     | Element Code (R3)                                             |
| ------------------------ | --------------------------------------------------- | --------------------- | ------------------------------------------------------------- |
| Serious                  | serious                                             | A.1.5.1               |                                                               |
| Reporter                 | reporterfamilyname                                  | A.2.1.1d              | C.2.r.1.4                                                     |
| Reporter Country         | reportercountry                                     | A.2.1.3               | C.2.r.3                                                       |
| Primary Source Country   | <p>primarysourcecountry<br>occurcountry</p>         | <p>A.1.1<br>A.1.2</p> | C.2.r.3                                                       |
| Report Type              | reporttype                                          | A.1.4                 | C.1.3                                                         |
| Report Version           | safetyreportversion                                 | B.5                   | <p>N/A <br>Creation of Creation is used for versioning R3</p> |
| Date First Received      | receivedate                                         | A.1.6b                | C.1.4                                                         |
| Date of Last Information | receiptdate                                         | A.1.7b                | C.1.5                                                         |
| Case ID                  | <p>N/A<br>Added for internal sender record ONLY</p> | N/A                   | N/A                                                           |

### Comment sub-tab

| UI Field Name | Corresponding E2B R2 Field | Element Code (R2) | Element Code (R3) |
| ------------- | -------------------------- | ----------------- | ----------------- |
| Comment Box   | reportercomment            | B.5.2             | H.2               |

### System Settings

The field below are configured at [system settings](/configuration/settings) level:

<table><thead><tr><th width="150.33333333333331">Setting</th><th width="175">Description</th><th width="275">Corresponding E2B R2 Field</th><th>Element Code (R2)</th><th>Element Code (R3)</th></tr></thead><tbody><tr><td>E2B Enabled</td><td><strong>Yes</strong>: E2B Feature is enabled</td><td>N/A</td><td>N/A</td><td>N/A</td></tr><tr><td>Receiver Country Code</td><td>2-Character country code (Ex: FR)</td><td>recevercountrycode</td><td>A.3.2.3e</td><td>N/A</td></tr><tr><td>Receiver E-mail</td><td>Email address for receiver contact</td><td>receiveremailaddress</td><td>A.3.2.3l</td><td>N/A</td></tr><tr><td>Receiver ID</td><td>Unique ID of receiver</td><td>messagereceiveridentifier</td><td>M.1.6</td><td>N.1.4<br>N.2.r.3</td></tr><tr><td>Receiver Organization</td><td>Name or receiver organization</td><td>receiverorganization</td><td>A.3.2.2a</td><td>N/A</td></tr><tr><td>Receiver Type</td><td>Organization type of the receiver</td><td>receivertype</td><td>A.3.2.1</td><td>N/A</td></tr><tr><td>Sender Country Code</td><td>2-char country code for the sender organization</td><td>sendercountrycode</td><td>A.3.1.4e</td><td>C.2.r.3</td></tr><tr><td>Sender E-Mail</td><td>Email address for sender contact</td><td>senderemailaddress</td><td>A.3.1.4l</td><td>C.3.4.8</td></tr><tr><td>Sender ID</td><td>Unique ID of sender</td><td>messagesenderidentifier</td><td>M.1.5</td><td>N.1.3<br>N.2.r.2</td></tr><tr><td>Sender Organization</td><td>Name of sender organization</td><td>senderorganization</td><td>A.3.1.2</td><td>C.3.2</td></tr><tr><td>Sender Type</td><td>Organization type of the sender</td><td>sendertype</td><td>A.3.1.1</td><td>C.3.1</td></tr></tbody></table>

### Automatically Generated Fields

MLM-AI will automatically generate these E2B fields during the export:

<table><thead><tr><th width="199.33333333333331">Data Element</th><th>MLM-AI Uses</th><th width="159">XML Field</th><th>Element Code (R2)</th><th>Element Code (R3)</th></tr></thead><tbody><tr><td>Message Number</td><td>A unique, randomly generated identifier derived from the Review ID</td><td>messagenumb</td><td>M.1.4</td><td>N.2.r.1</td></tr><tr><td>Safety Report ID</td><td>Article/patient unique identifier in MLM-AI</td><td>safetyreportid</td><td>A.1.0.1</td><td>C.1.1</td></tr><tr><td>Transmission Date</td><td>Date user exported the XML file</td><td>transmissiondate</td><td>A.1.3b</td><td>N.1.5<br>N.2.r.4<br>C.1.2</td></tr><tr><td>Literature Reference</td><td>Citation information including title</td><td>bibliographicDesignationText</td><td>N/A</td><td>C.4.r.1</td></tr></tbody></table>


# Notifications

MLM-AI notify users by email on key events

## Notifications in MLM-AI

MLM-AI supports email notifications of application events. Users can:

* Be notified when a review is ready for screening, or when its status changes (ex: from *In Progress* to *Completed*)
* Be notified of any assessment updates, when they are part of an assessment (per the article audit log)
* Send the details of an abstract and its assessments to a pre-configured list of recipients

## Notify on Assessment Updates

When an assessment is added to an abstract, users involved in the assessment - ie. the **interested parties** - can be notified of the update.

### How Automatic Notifications Work?

If a user is taking part in the assessment of an article (by saving a decision to the audit log) they can receive automatic notification of any further updates.&#x20;

See example scenarios below. Note that the user that creates the new decision does not get notified.

| Scenario                    | Users in Audit Log | New Decision | Who is Notified? |
| --------------------------- | ------------------ | ------------ | ---------------- |
| Assessment                  | -                  | User A       | -                |
| User Corrects Assessment    | User A             | User A       | -                |
| QC                          | User A             | User B       | User A           |
| User A corrections after QC | User A, User B     | User A       | User B           |
| 2nd level QC                | User A, User B     | User C       | User A, User B   |

### Configuration

When configuring a Monitor, select Notify on decision updates in the Notifications section of the Screening tab:

<figure><img src="/files/SeG2iQrHw4hoUWs4CGii" alt=""><figcaption></figcaption></figure>

## Email Article Details

Users can optionally send a summary of a citation and screening decisions to a pre-configured list of email recipients.

From the [Abstract Detail](/application/review-details#screening-abstracts) screen, the email widget allows users to select an email recipient to send to:

<figure><img src="/files/4kHRABCjz2sBwf7OdKna" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
The **Interested Parties** option will send an email to all users present in the article audit log - ie. all users that previously saved a screening decision
{% endhint %}

### Configuration

In the [Settings](/configuration/settings) page, add the email recipients users are allowed to select:

<figure><img src="/files/FqEQ9bcFMDDx4IGW5uRy" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Email recipients can be individual inboxes or aliases/groups configured to receive emails from external sources.&#x20;
{% endhint %}

In the [Monitor configuration](/configuration/monitor-configuration#monitor-options), check the Enable Email articles feature to make the email widget visible:

<figure><img src="/files/EvEqApkeFLjbsTny8e00" alt=""><figcaption></figcaption></figure>

## Review Status Changes

Users can be notified when:

* A new Review submitted for processing is ready for screening
* A Review changes status (ex: from In Progress to Completed)

### Who is Notified

All users that are part of a [Reviewer or QC team in the Monitor](/configuration/monitor-configuration#team-assignment-special-situations-and-exclusions) for the review are notified of updates.

### Configuration

In the [Monitor configuration](/configuration/monitor-configuration#monitor-options) select the Notify on review status changes option:

<figure><img src="/files/f4LP0DG7qWfkpSMqucMQ" alt=""><figcaption></figcaption></figure>

&#x20;


# MLM-AI Concepts

Learn more about how MLM-AI works and key concepts that deliver more efficient workflows.

## Literature Monitoring with MLM-AI

[Biologit MLM-AI](https://www.biologit.com/factsheet) addresses all productivity challenges on literature monitoring workflows for safety surveillance. The diagram below outlines key stages in a typical workflow, and their productivity challenges:

<figure><img src="/files/4CckH9vrVFX6fpDrXibx" alt=""><figcaption></figcaption></figure>

* One or more literature databases (PubMed, Embase, etc.) are **searched**, each with their specific query language
* Results (title and abstract) are **extracted** and collated together
* Articles are **de-duplicated**, removing repeated and articles previously seen on past runs
* Abstracts are then **screened** for relevant safety events
* If screen is positive, the article may then proceed for a **full text review**
* Results are **quality controlled (QC)** and must be **exported in a suitable format** for further processing downstream

The Biologit platform addresses the above challenges with an integrated system that is easy to adopt and to scale to your needs. In Biologit the following components were built from the ground up with the needs of literature monitoring of safety events, working seamlessly together:

* **Biologit Database**: comprehensive database of scientific and medical literature covering global and regional sources
* Robust and validated **AI models** tailored for pharmacovigilance and safety surveillance
* Easy to use web application delivering integrated workflow, collaboration, reporting and integration capabilities

&#x20;

<figure><img src="/files/kHwYj1WsOhFXiJM8r2MB" alt=""><figcaption></figcaption></figure>

## The Biologit Database

The [Biologit Database](https://www.biologit.com/database) is a comprehensive and continuously updated repository of scientific literature, ready for compliant regulatory searches, integrating databases of global and regional reach into an easy to use interface. It provides:

* High quality results for literature monitoring of adverse events and possible new risks for human and veterinary products, medical devices, nutraceuticals and cosmetics.
* Full integration into [biologit MLM-AI ](https://www.biologit.com/factsheet)for fast and audit-ready literature screening.

## Application Concepts

### Monitors and Reviews

The literature screening process can be setup in the platform with these simple steps:

* Users [configure a monitor](/configuration/monitor-configuration) defining the search criteria for a product, literature sources to search, and workflow details. This is a one-time setup.
* Once configured, Monitors produce results periodically, or on request.
  * Search results [can optionally be uploaded](/application/upload-data) from an external source.
* Results are de-duplicated and presented to users as [Reviews](/application/reviews). They are also tagged by AI models for faster screening. (see AI Tags below)
* Users can then [complete assessments ](/application/article-screening)according to their workflow. during article assessment:
  * Full text articles can be uploaded to the workflow as [attachments](/application/article-screening/attachments-and-full-text-screening), they are permanently stored in the platform for future reference.
  * [Article translations](/application/article-screening/attachments-and-full-text-screening) can be directly requested from the system.
* On completion assessment results can be [exported ](/application/article-screening#e2b)in various formats for further processing (addition to the safety database) or reporting.

<figure><img src="/files/EzG3jA6NpPo5QeIXYWti" alt=""><figcaption></figcaption></figure>

### AI Tags

Every article presented to users receives a number of tags based on AI predictions. These tags can be used to filter and rank results, [facilitating more efficient workflows.](/topics/ai-enabled-screening-workflows)

![Example tags associated to an abstract](/files/-MWLPaoAml4TTmi5dzoe)

| Tag                      | Description                                                                                                                                                      |
| ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Suspected AE**         | <p></p><p>The abstract or full text may contain the description of an adverse event. Thus, MLM-AI flagged the abstract as a candidate for further screening.</p> |
| **Suspected Case**       | The abstract or full text may contain the description of a case (identifiable patient).                                                                          |
| **Animal/In-Vitro**      | The abstract describes animal or in-vitro studies.                                                                                                               |
| **Special Situations**   | Any detected special situation (elderly, pediatric, pregnancy) is also flagged as a tag.                                                                         |
| **Identifiable Patient** | A confirmed patient mention (ex: "a 72-year old woman" etc)                                                                                                      |

**Why "suspected" adverse events?**

A "suspected" event indicates an event may have been described in the full text and indirectly mentioned in the abstract and was considered by MLM-AI as worth further inspection.&#x20;

Suspected events are drug-agnostic: they can refer to any drug or treatment mentioned in the abstract.

{% hint style="success" %}
**Learn more:**

* See our page on [**AI-Enabled Screening Workflows**](/topics/ai-enabled-screening-workflows) for approaches that integrate AI tags into your screening workflow.&#x20;
* See the [**FAQ on AI models**](/more-help/faq#ai-models) for more details about how MLM-AI makes predictions.
  {% endhint %}

## Next Steps

Follow the links below to:

* [Configure your first Monitor](/configuration/monitor-configuration#creating-monitors)
* [Submit an review to retrieve your first results](/application/reviews#submitting-reviews-on-demand)
* [Learn how to use AI Tags for faster screening](/topics/ai-enabled-screening-workflows)


# Handling Article Dates

## Summary

* MLM-AI always searches articles by the date they were added to the source database, ie. the "entry date" or "uploaded date"
  * This approach avoids missed articles in periodic searches and returns more comprehensive results
* MLM-AI can optionally search by publication dates in addition to the entry date. This option is **On** by default and can be configured in each Monitor

## How MLM-AI Handles Article Dates

The start and end dates of search results are important parameters when performing periodic MLM. When searching for articles there are two important dates to consider:

* The **publication date** of an article - the date the article was published, according to the source journal
* The date an article first becomes *visible* in a database - ie. the article's indexation date, or **entry** **date**

For example, an article may be published by a journal today, but [it may be a few days](https://support.jmir.org/hc/en-us/articles/115001813331-How-long-after-publication-does-it-take-for-an-article-to-show-up-in-PubMed-) or weeks before it enters the PubMed database and is visible on search results. Note however the *publication date remains unchanged*.

The diagram below illustrates the effects of this publication delay: a weekly search strategy that uses only the publication date could miss this article entirely.

![](/files/-MjjAW3piCev6k5eo93d)

To address this issue, MLM-AI *always* searches articles using the "entry date". This ensures that the article is retrieved if it became available at the source within the date range specified, even in cases where a publication is delayed.&#x20;

In addition, MLM-AI provides monitoring configuration options to fine-tune the strategy:

* **Include abstracts found by publication date** - searches the same date range matching the publication date. This option is useful to maintain correspondence to existing manual search strategies.
* **Discard publication dates older than 60 days** - if the difference of an article first seen date and publication date is older than 60 days, results are discarded. This option is useful to avoid spurious results from "blanket" publication of journals going back many years.&#x20;

### Configuration Options

The above options are configurable for individual monitors under [Monitor configuration](/configuration/monitor-configuration):

![](/files/-Mjj9YbnpwrEg6xDYBaH)

## PubMed and Publication Dates

Another reason to use article "first seen" date in searches is: *publication dates can change over time in PubMed*.&#x20;

This behavior is documented in PubMed [help manual](https://pubmed.ncbi.nlm.nih.gov/help/) (excerpt shown below):

*<mark style="color:purple;">Most journals now publish articles online on a continuous basis, as soon as they are ready for publication (after peer review and editing, etc.) instead of, or in addition to, publishing collections of articles as an "issue" on a periodic basis.</mark>*&#x20;

*<mark style="color:purple;">When a journal deposits a citation for an "online first" article in PubMed, NLM appends the note "\[Online ahead of print]" to the online publication date.</mark> <mark style="color:purple;"></mark><mark style="color:purple;">**The citation is updated, and the ahead of print notation removed, when the article is included in a journal issue**</mark><mark style="color:purple;">. The lag between the "online first" and "issue" publication dates may be days, weeks, months, or more than a year.</mark>*

This behavior is also reported by PubMed and searching by "entry date" brings more consistent results:

*<mark style="color:purple;">The time between an article’s publication and the citation’s availability in PubMed varies depending on when the publisher deposits the citation to PubMed.</mark> <mark style="color:purple;"></mark><mark style="color:purple;">**Because of this, searching with Create Date (the date a citation was created in PubMed) is often more comprehensive than Publication Date**</mark> <mark style="color:purple;"></mark><mark style="color:purple;">when checking PubMed on a regular basis for new citations.</mark>*

'&#x20;


# AI-Enabled Screening Workflows

Learn how to make the most of AI tags in MLM-AI

Biologit MLM-AI supports various strategies to improve quality and speed up screening workflows using AI features. this can be achieved by focusing efforts on the results *more likely to contain relevant safety information*.&#x20;

In this document we explore common AI-based screening strategies with examples.&#x20;

## AI Tags

Every abstract retrieved by Biologit MLM-AI can receive *AI tags* representing predictions that correspond to relevant safety surveillance information:

* Potential safety event,&#x20;
* Special situations,&#x20;
* Identifiable patients,
* Animal or In-vitro study&#x20;

The diagram below illustrates how AI tags helps users focus on the most relevant results. Consider for example an ICSR pharmacovigilance workflow focusing on adverse events on humans. From all results retrieved, users can prioritize the ones with an identifiable patient (patient tag), progressively moving to the abstracts tagged with a suspect adverse event (suspect AE). &#x20;

![](/files/p5RAZQUiSdSx1jAt2q3s)

The following AI Tags are available in MLM-AI:&#x20;

| **Suspected AE**         | <p></p><p>The abstract or full text may contain the description of an adverse event. Thus, MLM-AI flagged the abstract as a candidate for further screening.</p> |
| ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Suspected Case**       | The abstract or full text may contain the description of a case (identifiable patient).                                                                          |
| **Animal/In-Vitro**      | The abstract describes animal or in-vitro studies.                                                                                                               |
| **Special Situations**   | Any detected special situation (elderly, pediatric, pregnancy) is also flagged as a tag.                                                                         |
| **Identifiable Patient** | A confirmed patient mention (ex: "a 72-year old woman" etc)                                                                                                      |

With AI tags, users can:

* Prioritize abstracts classified as Suspected AE, by **ranking or filtering results**.
* Use **less resource intensive screening methods** where safety events are unlikely. For example:
  * [Batch Screening](/application/batch-screening) of Non Suspected AE abstracts.
* Finally, users can [Automatically filter](#automatic-filtering-by-tag) articles by Tag, skipping screening and perform only QC.

{% hint style="info" %}
Ensure automated screening using AI Tags is implemented with adequate risk-based quality controls in place for your scenario. [Batch Review](/application/batch-screening) and [Sampled QC](/application/review-details#tab-filter-sampling-and-sorting-choices-follow-user-during-screening) can help with your requirements.
{% endhint %}

## Screening with AI Tags

This section contains examples of how to employ AI tags for faster screening.

### Prioritize Abstracts by AI Tag

Abstracts in a review can be ranked according to their AI Tags. In the [Reviews page](/application/review-details) simply sort the abstracts by a tag such as "Suspected AE" or "Identifiable Patient":

<figure><img src="/files/WLFANnYFm2273rDVCanP" alt=""><figcaption></figcaption></figure>

The ranking is preserved once users start reviewing articles in the detail page: users will screen the prioritized articles first.

### Screening Less Relevant Articles

Abstracts *not* tagged can in principle receive less scrutiny during reviews. Examples of how this can be done:

* Update your screening process so that the "Not Suspected AE" abstracts receive fewer QC steps (ie. "sampled" QC)
  * These can be retrieved from any review by searching `notag:adverse`(see more details on [search options inside a review](/application/review-details/searching-abstracts-inside-a-review))
* Screen quicker with [Batch Review](/application/batch-screening)

### Example with Batch Review

[Batch review](/application/batch-screening) enables faster screening of many abstracts at once.&#x20;

For example: In an ICSR workflow, Batch Review can be used to refute all animal/in vitro studies at once (typically not a valid ICSR):

* From the search box, select articles tagged as Animal/In Vitro with `tag:animal`
* Then click "Batch Review" to quickly review all abstracts from the same screen, according to the search criteria specified
* Use the checkbox to select irrelevant articles and save your decision for all selections at once

<figure><img src="/files/KvnINDqKDYeZq0KT1CvU" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
You can also use the `notag:` option to search for the absence of a given tag.&#x20;

Example: `notag:adverse` returns all articles *without* the "Suspected AE" tag.

See also [Batch Screening](/application/batch-screening) and [Search options in the Review Details](/application/review-details/searching-abstracts-inside-a-review) page.
{% endhint %}

### Automated Filtering by Tag

Users can further reduce effort by configuring monitors to *automatically pre-screen* abstracts based on AI tags.

* All pre-screened articles remain visible in the results. They are presented in the "Reviewed" tab and can still be reviewed as part of the quality control process.

In the [Monitor configuration screen](/configuration/monitor-configuration#tag-based-screening) users can select the level of pre-screening for a given Monitor according to which AI tags are present in an article:&#x20;

<figure><img src="/files/1Q45ALJ59qBuUhPh0E03" alt=""><figcaption><p>Screening automation options during Monitor configuration</p></figcaption></figure>

{% hint style="info" %}
See also [Monitor Configuration ](/configuration/monitor-configuration)for more details on this feature.
{% endhint %}


# Duplicate Detection

Duplicate articles are often found when screening the scientific literature. Correctly identifying them is time consuming if performed manually, and becomes even more burdensome as more sources of data are added.

MLM-AI natively performs duplicate detection using different strategies, simplifying screening workflows and removing manual effort.&#x20;

This page describes how MLM-AI performs duplicate detection, and how users can enable automated screening of duplicates in their results.

### Design Principles

The MLM-AI duplicate engine is designed for **high precision:** to minimize the risk of an erroneous duplicate match, the engine only marks articles as duplicates with high confidence, employing multiple verification approaches.&#x20;

Because the engine emphasizes safe duplicate detections, certain duplicate articles may not be flagged in all cases, although such instances are expected to be infrequent.

The methods and verification approaches used in MLM-AI are described in the following sections.

### How it Works? Duplicate Detection in the MLM-AI Database

MLM-AI identifies duplicate articles according to these criteria:

* Duplicate articles are only detected for results from the [same Monitor](/configuration/monitor-configuration).&#x20;
* To be marked as a duplicate the article must:
  * Have same ID and same source database, or
  * Have same Document Object Identifier (DOI), or
  * Have similar abstract content (with high confidence)&#x20;
* Candidate duplicate articles must also pass an *article title* similarity check

#### Same ID and Source Database

In MLM-AI each article receives a unique identifier generated at the source database (in PubMed, for example, this is the article PMID). Articles with same ID are marked as duplicates.

This method is straightforward and useful when Reviews return results for overlapping date ranges.

#### Same Document Object Identifier

This approach relies on the [Digital Object Identifier](https://www.doi.org/) (DOI) of the article. The DOI is a unique ID assigned for each publication. Hence, if the same article appears on different databases it will still preserve the same DOI.&#x20;

When an article DOI is available, MLM-AI will present it in the Details tab, linking the article directly to its authoritative source:

<figure><img src="/files/PL1RGkeNaapatofmSXWw" alt=""><figcaption><p>Article DOI visible from the Details tab</p></figcaption></figure>

If the source journal has published valid DOI information, MLM-AI can use it to detect duplicates.

#### Similar Content

Finally, MLM-AI can also detect duplicates using content similarity. This happens in two stages:

* Select articles with similar content from the MLM-AI database, based on title and abstract
  * Only articles with a valid abstract are eligible for comparison
  * Duplicate abstracts do not need to match exactly, but must be highly similar

This option is only available for results obtained from the MLM-AI database. This excludes results from [uploads from external sources](/application/upload-data).

**Title Match**

* For all types of duplicate, perform a second verification step on the article titles, to prevent detection of false positives

### Enabling Automated Duplicate Detection

MLM-AI can perform automated screening on duplicate articles. This can be enabled when [configuring Monitors](/configuration/monitor-configuration#duplicate-handling):&#x20;

<figure><img src="/files/3YcVPbYVpZ5JuAnppHOY" alt=""><figcaption><p>Enabling Duplicate Screening</p></figcaption></figure>

This option is enabled by default. When enabled, duplicate articles are pre-screened and appear in the "Duplicates" tab of the Review results:

<figure><img src="/files/k5k0VvqKsfk5pJBLwu8t" alt=""><figcaption><p>Duplicates tab</p></figcaption></figure>

### Inspecting Duplicate Articles

Any article where a duplicate has been detected will also contain a "Duplicates" tab, displaying the duplicate reason and linking to the duplicate article. This tab is always available for convenient inspection of duplicates, irrespective of whether pre-screening of duplicates is enabled for your Monitor.

<figure><img src="/files/BEpreko2TK7C1LSMhwGx" alt=""><figcaption></figcaption></figure>

### Limitations

* Duplicate detection works only on results from the same Monitor. This is by design, to prevent results from different monitors from interfering with one another.
* Duplicate detection by content similarity is only available to results generated from the MLM-AI database. However duplicates by ID and DOI are available on uploaded results.&#x20;

### Learn More

* [Configuring Monitors](/configuration/monitor-configuration)
* Learn more about how the techniques used in automated duplicate detection on: [**Feature  spotlight: Duplicate Detection in Biologit MLM-AI**](https://www.biologit.com/post/feature-spotlight-duplicate-detection-in-biologit-mlm-ai)

&#x20;&#x20;

&#x20;


# Support and Platform Details

### Service Availability

biologit MLM-AI service availability can be tracked in real time from our public status page:

* [**https://status.prodeu.biologit-mlmai.com/**](https://status.prodeu.biologit-mlmai.com/)

### **Customer Support**

Users with an active subscription can access the support helpdesk:

* Go to [**biologit helpdesk**](https://biologit.atlassian.net/servicedesk/customer/portal/1)

### **AI Technical Specification**

Technical fact sheet on models in use by MLM-AI are periodically issued by our engineering team. They can be found in the team public Github project:

* [**https://github.com/biologit-engineering/factsheets**](https://github.com/biologit-engineering/factsheets)

### **CFR-11**

Details of MLM-AI capabilities that contribute to CFR-11 compliance are detailed on:

* [CFR-11](/more-help/notices/cfr-11)


# Release Notes

Biologit MLM-AI Release Notes

## MLM-AI 1.1.56.2 (19-AUG-2026)

**🎉 New Report: Audit - Integrations Runs**

* For clients with enabled integrations, this report will list all integration runs during a time period

**🎉 Enhancements**

* For consistency, all review-based reports now report date fields in the DD-MON-YYYY format

:wrench: **Fixes**

* Short user keywords configured can show an entire highlighted word if they partially match that word on PDF's rendered by the system (ex: "aba" is the keyword, where "abacus" ends up being fully highlighted)

## MLM-AI 1.1.55.11 (13-AUG-2026)

**🎉 Enhancements**

* Added support for [keyword highlighters derived from abstracts](/application/article-screening/attachments-and-full-text-screening#search-inside-the-pdf), including a new lookup feature to improve visibility and navigation of relevant terms within full-text documents
* Introduced a confirmation step when modifying user-defined keywords within a monitor ensuring keyword selections are explicitly confirmed

## MLM-AI 1.1.55.8 (30-JUL-2026)

**🎉 Enhancements**

[**System Limits**](/more-help/faq#system-limits)

* Article Search
  * Article search query string limit increased to `30000` characters
* Monitor
  * Search terms limit in monitor configuration increased to `1800` keyterms

## MLM-AI 1.1.55.3 (21-JUL-2026)

**🎉 Enhancements**

* Email-ed articles now include day zero in body
* Column `inferred_language` is now included in workflow export report. The inferred language indicates the language detected by the platform based upon article content.
* **Reports:**
  * Various Reporting Engine efficiency improvements
  * AI Tag Analysis report now reports metrics for the special situations tag
  * Screening activity reports can now optionally be filtered by (1) the last recorded decision and (2) the user who made the last decision&#x20;

<figure><img src="/files/cSaiD4aTSbY8Ian5yuMG" alt=""><figcaption></figcaption></figure>

:wrench: **Fixes**

* Navigating to related articles can sometimes display incorrect integration run details from the integrations widget
* Exports (via download icon) sometimes display "UNK" when publisher article language is not known

## MLM-AI 1.1.54.9 (01-JUL-2026)

**🎉 Enhancements**

* We introduced performance enhancements that speed up PDF loading and rendering
* Uploads: results uploaded from a 3rd party source will report additional details in case duplicate records are found
* Transfer reviews: it is now possible to see details of a review of type "Transfer" from the Reviews Summary screen:

<figure><img src="/files/RrQhs1XGZwwVfB33ihdJ" alt=""><figcaption></figcaption></figure>

:wrench: **Fixes**

* The web application token expiration timeout was adjusted to prevent premature user logouts

## MLM-AI 1.1.54.7 (22-JUN-2026)

**🎉 Enhancements**

* Reports: Added decision comments column to QC Review reports
* Reports: Added filter by monitor to Query String Update report

:wrench: **Fixes**

* Reports: The column order in all review-based reports is now fixed
* Reports: The Query String Update report sometimes shows redundant entries unrelated to query string changes

## MLM-AI 1.1.53.12 (04-JUN-2026)

**🎉 Enhancements**

* As a safety feature, 1-click buttons no longer visible after a decision save, where the new decisions do not have a configured button
* Improved performance of monitor list drop down for clients with large number of monitors

## MLM-AI 1.1.53.2 (13-MAY-2026)

**🎉 New Feature: 1-Click Decisions**

For workflows where many assessments result in a straightforward action (ex: a rejection), it is possible to configure a visual shortcut to allow users to complete the action with **one click**.

Once configured, this option will become visible from the Article Screening page. Clicking the shortcut will:

* Populate decision and annotations according to the 1-click configuration
* Save the decision and automatically navigate to the next article

<figure><img src="/files/eCtqnKPTh2YP3Z80Xac9" alt=""><figcaption></figcaption></figure>

Learn more 👉 [Monitors](/configuration/monitor-configuration#id-1-click-decision)

**🎉 Enhancements**

* Review-based reports now include a new column for article AI tags
* Better de-duplication support for retrospective reviews
  * Retrospective review requests can now be configured to de-duplicate against reviews of any time period.&#x20;
  * Learn more about Advanced Options in: [Reviews](/application/reviews)
* The duplicate detection engine ships more robust logic to capture scenarios where the same article may appear more than twice, across distinct reviews and data sources

:wrench: **Fixes**

* Excel exports may fail due to rare unsupported characters in source data
* Find Articles search could get stuck on rare cases that include transfer reviews

## MLM-AI 1.1.52.3 (23-APR-2026)

:wrench: **Fixes**

* Deleted attachments are no longer transferred during a Transfer Article operation
* UI would sometimes display stale ISSN lists in monitor configuration
* Improved message for failed decision submission because of article assignment permissions during related decisions
* Column in Review and Find Articles Page  “Due Date“ was relabeled to correct meaning “Day Zero“
* Display clearer error message on query string parsing errors in Article Search
* Related Articles decisions are refreshed to display more accurate decision information
* In a Transfer Article operation, UI correctly displays only monitors user has access to&#x20;

## MLM-AI 1.1.52.2 (20-APR-2026)

:wrench: **Fixes**

* In rare occasions, not all articles present in an review with auto-assignment enabled have been assigned to respective team members

## MLM-AI 1.1.51.8 (31-MAR-2026)

**🎉 Enhancements**

**New Reports**

The Reporting module now offers new reports to facilitate auditing and analysis:

<table><thead><tr><th width="165.7999267578125">Section</th><th width="247">Report</th><th>Description</th></tr></thead><tbody><tr><td>Auditing</td><td>Monitor Query String Changes</td><td>For each monitor provide a timeline of changes that specifically altered the query string </td></tr><tr><td>Analytics</td><td>Decision Statistics</td><td>Summary report of decision breakdown per monitor, for a given time period. Useful for at-a-glance view of assessment outcomes per monitor.</td></tr></tbody></table>

**Configurable Cut Off Days for Old Publication Dates**

Monitors allow discarding articles with old publication dates. It is now possible to select the "cut off" point for old publication dates (e.g. 60, 90 or 120 days).

* See more on: [Monitors](/configuration/monitor-configuration#filter-by-publication)

:wrench: **Fixes**

* After user picks a decision from the drop down (without saving), correctly reflect it across all decision widgets (ie. in the expanded decision dialog)

## MLM-AI 1.1.51.5 (19-MAR-2026)

:wrench: **Fixes**

* Stability enhancements: updated caching and retry logic to improve overall response rates in the Biologit API
* E2B XML: Correctly override author information when these fields are set to blank in the E2B tab
* Article un-assignment: in some cases application did not correctly indicate an unsuccessful un-assignment operation&#x20;

## MLM-AI 1.1.50.5 (06-MAR-2026)

:wrench: **Fixes**

* Admin-level reset of an article day zero incorrectly displays action timestamp
* Use shorter time periods to validate a query string update in Monitor configuration (thus preventing timeouts when saving the Monitor)
* More robust logic to detect valid article ISSNs when calling an AI action (iBot)

**Find Articles**

* It is now possible to select reviews in [Find Articles](/application/reviews/find-articles-from-reviews) using both **review start date** *or* **date of creation (day zero)** as the date criteria. Day zero can now be selected using the toggle shown below
* It is now also possible to sort results in Find Articles by date - sorting options shown below

<figure><img src="/files/pvlQ0OI080OZ7Xkl6VQh" alt=""><figcaption></figcaption></figure>

**Other Enhancements**

* Additional attachment extension support: users can now upload other file formats commonly found as reference data in literature workflows: `.pdf, .docx, .doc, .txt, .xlsx, .xls, .csv, .ris, .nbib, .eml, .xml`

## MLM-AI 1.1.50.4 (25-FEB-2026)

:wrench: **Fixes**

* In some scenarios where monitor workflows use annotation validation, saving multiple decisions of related articles is blocked due to enforcement of incorrect validation rules. Users can still save articles one at a time as a workaround.
* In some cases annotation validation may not be enforced (affects related articles)

## MLM-AI 1.1.50.2 (20-FEB-2026)

:wrench: **Fixes**

* [E2B R3 XML export](/configuration/e2b-configuration-guide) better supports associating multiple patients to the same article
* Some local literature articles with noisy content may not display citation abstract (fixed)&#x20;
* While transferring an article to a new review, UI would not indicate the target review ID (fixed)

## MLM-AI 1.1.49.2 (09-FEB-2026)

**🎉 Enhancements**

**Reports**

* Users reports now indicate read-only status of users
* Add support for all review-based reports to optionally select reviews based on **day zero** instead of review date from-to range

<figure><img src="/files/6wjEgEB6jEQe3BLWGbuB" alt=""><figcaption></figcaption></figure>

**Assignments**

* Assign many/all related articles to current user at once.

**Article Transfer**

* Transfer articles to other monitors in batch
* Article attachments can be transferred with their translations to related articles

**Find Articles**

* Filter citations by Latest Decision Date from-to range

:wrench: **Fixes**

**Local literature ingestion:** a small amount of documents ingested through Biologit's local literature manual process were identified as not available on search results. This release implements the fix. &#x20;

## MLM-AI 1.1.49.1 (22-JAN-2026)

**🎉 New Features**

* **Integrations**: Communication to external systems and services could now be facilitated by MLM-AI. This feature is best suited for teams looking to automate recurring data processes and reduce manual operations across systems. Find more details and example use cases under [Integrations](/application/integrations).

:wrench: **Fixes**

* Translation audit report could fail under certain conditions.

## MLM-AI 1.1.48.11 (16-JAN-2026)

🎉 **Enhancements**

* Users can request a [temporary password](/more-help/faq#i-forgot-my-password) by email to login and update their previous password.
* Read-only users feature introduced. Please refer to [Read-only Permissions](/configuration/team-and-user-management#read-only-permissions) for more details

:wrench: **Fixes**

* Login button on the login screen needs to be clicked twice sometimes to successfully log in.

## MLM-AI 1.1.48.10 (05-JAN-2026)

:wrench: **Fixes**

* Creating multiple users in sequence from user management would leave incorrect user details when copy/pasting onboarding message on [password resets](/configuration/team-and-user-management#password-resets)&#x20;
* Single sign on logout flow: User explicit logout or logout due to inactivity timeout also logs out user from single sign on provider
* [Article lock timeout setting](/configuration/settings#security) supports 3 digits (seconds)&#x20;

## MLM-AI 1.1.48.3 (24-NOV-2025)

🎉 **Enhancements**

* All related articles can be selected in single click
* Duplicated articles can now be togged in/out of view when viewing related articles

## MLM-AI 1.1.48.1 (16-NOV-2025)

🎉 **Enhancements**

* Daily monitors now support functionality to consolidate jobs scheduled for weekend into a single review on Monday. Find more details under [Date Search Criteria](/configuration/monitor-configuration#date-search-criteria) section of monitor configuration guide

:wrench: **Fixes**

* Improved error handling for report requests

## MLM-AI 1.1.47.8 (28-OCT-2025)

🎉 **Enhancements**

* New filter by review submission due date is now available in both [Review Details](/application/review-details) and [Find Articles](/application/reviews/find-articles-from-reviews). Refer to [**Query Options**](/application/review-details/searching-abstracts-inside-a-review#query-options) for more details.
* Search selections in [Find Articles](/application/reviews/find-articles-from-reviews) screen are persisted between different sessions, allowing for faster reviews:

<figure><img src="/files/I2sjKgpBTmedpFfZ2R7q" alt=""><figcaption></figcaption></figure>

* Layout of the [Find Articles](/application/reviews/find-articles-from-reviews) screen is simplified with all available filters consolidated into a single :heavy\_plus\_sign: icon next to the search bar:

<div align="center"><figure><img src="/files/GBdw53NUlTulsieZmbMv" alt="" width="471"><figcaption></figcaption></figure></div>

## MLM-AI 1.1.47.6 (24-OCT-2025)

🎉 **Enhancements**

* **Find Articles improvements**
  * Annotation can now be toggled on and off per monitor in search to better use screen space
  * Due date can now be included in searches
* **Audit log redesign:** Audit log now clearly incorporates user annotations

<figure><img src="https://ci3.googleusercontent.com/meips/ADKq_NZkwBGKyyrD3ZZ9qaKjyZzCrzuVbVlmaj4g07SbZzouEqThvIsZZKlaHxt_dvTdGO9bQ2hVi4G9241ip_kQrDYHp3ETURsQ5inSVZ9eY0ezKHo9qXvQZaIdr_ZVbRKy42OHctvN4AyjY1vE7y0Bs121t9FW7tnYwY3zPrKVEkWg4VdYr5ftYlYtXYqEuCSlZmAc4gYmdwHDndCX5QLA3rHtByumBd1J9bDpALKrdMUMu4sW0x5jKrN4wyYMKlYiezJi3AQmsU_GiE6HELnyL0iZVtQYzC0BTvEPuM226yNKmnnC5wTaHKTYw6hTKF0iiCwMEighSArqLSM9bsG7SBpP5BQMIFEHRfdBUwoSORqz4vJVrYJWzzcouMwoXP8CbTCd2CikLMS5VGRlSRSeY7kc_iL5u3osASlnlGlv0sqeEiP9rcYK0rrMneeqZS2e4r-N0RptWQ=s0-d-e1-ft#https://images.gitbook.com/__img/dpr=2,width=760,onerror=redirect,format=auto,signature=-1980287826/https%3A%2F%2Ffiles.gitbook.com%2Fv0%2Fb%2Fgitbook-x-prod.appspot.com%2Fo%2Fspaces%252F6pIe8UPxEvsbrxTVXkGi%252Fuploads%252FPJ4O0RAqPrbHQ2VeD05X%252Fimage.png%3Falt%3Dmedia%26token%3D1b899430-32d2-4a5c-bfb0-8bcd5330cd30" alt=""><figcaption></figcaption></figure>

* **Automation on Decision Save:** Monitors can now be configured to prompt users for next steps in their workflow on certain decisions. Both user email notifications and suggested reassignments have been developed

## MLM-AI 1.1.47.5 (20-OCT-2025)

:wrench: **Fixes**

* Email Notification can now be generated without a saved decision

## MLM-AI 1.1.45.6 (9-OCT-2025)

🎉**Enhancements**

* **Rearranging annotation fields**: The order of annotation fields can be rearranged from the Monitor Configuration screen, which will be reflected when annotations a presented to users
* **Filter and select drop downs**: it is now easier to select all values based on keyword match in drop downs: users can apply a filter and use "Select all" to select all matches in one action.

<figure><img src="/files/Ru7JC98ZfSYRYjuElZ8L" alt=""><figcaption></figcaption></figure>

* **Easier batch-submit reviews**: It is now easier to complete a batch submission of reviews for many monitors reusing the same date range parameters.

<figure><img src="/files/krHvpDX6YqBdihalyqcW" alt=""><figcaption></figcaption></figure>

**Fixes**

* **Decision overrides notification**: users are now prompted when overriding a decision that may erase previously completed annotations or E2B data
* **Consistent use of save buttons when saving decisions**: all save buttons (main tab, annotations, E2B) obey the same behaviors and become enabled upon user edits in any field

## MLM-AI 1.1.44.5 (5-AUG-2025)

🎉**Enhancements**

* **Streamlined editing of workflow**: we updated the user interface for defining workflows in Monitors with a better layout and options to copy decisions and easier drag and drop

<figure><img src="/files/jHLbeYL07QQBpCyleFsh" alt=""><figcaption></figcaption></figure>

* **Get PDFs from system configuration**: all system configuration elements (System settings, Users, Teams, Monitors) can now export a timestamped PDF for documentation reference.
* **Export full text PDFs with highlights**: A new download option "PDF + Tags" allows download of uploaded PDFs with system highlight

&#x20;

<figure><img src="/files/YYUCDzTwEIpffs2UKegQ" alt=""><figcaption></figcaption></figure>

## MLM-AI 1.1.44.1 (11-JUL-2025)

🎉**Enhancements**

[**User Management**](/configuration/team-and-user-management#adding-users)

* Quick password reset: administrators can now reset user passwords quickly from User Management, with the system providing a random password and password reset  + login instructions
* User management: user Ids that default to user email address can be overridden when creating users

[**Reports**](/application/reports)

* New reports added to the reports module: translation requests, journal-wise summary report (local literature clients) and Results vs. AI tag analysis

**Fixes**

* Find Articles progress counter not correctly reflecting progress of results retrieved

## MLM-AI 1.1.43.5 (23-JUN-2025)

**🎉New Features**

* [**Batch review submission**](/application/reviews#batch-submissions): Users can now conveniently submit on-demand reviews for multiple monitors simultaneously, saving time on routine process for large portfolios.
* [**Batch monitor update**](/configuration/monitor-configuration#updating-monitors-in-batch): This useful feature will allow users to update monitor definitions "in batch", making it easier and less error prone to maintain large collections of monitors &#x20;

## MLM-AI 1.1.43.4 (17-JUN-2025)

**Fixes**

* Full text translations: Articles from the Biologit local literature data source would not allow requests for PDF translations
* Decision override option: unable to save a decision override when custom annotations are configured in Monitor
* Find Articles: added validation of review status selections (at least one required)

## MLM-AI 1.1.42.4 (05-JUN-2025)

**🎉New Features**

**Partial translation requests** - for large documents it is now possible to request a page range for translation, facilitating assessments and managing translation costs.&#x20;

* Learn more in:[Attachments and Full Text Screening](/application/article-screening/attachments-and-full-text-screening#partial-translations)

**Self-Assignment of Articles** - When work assignment is enabled, users can optionally self-assign related citations to themselves, completing the assessment faster. &#x20;

* Learn more in: [Work Assignment](/application/review-details/work-assignment#self-assignment) &#x20;

**Enhancements**

* Database selection widget updated in Article Search including database categories and "select all" options

**Fixes**

* In some cases the E2B "Excel" file generated in the Zip export option would not correctly fill in target and concomitant products&#x20;

## MLM-AI 1.1.42.2 (23-MAY-2025)

**Enhancements**

* **New source: Scielo Spain** has been added to available sources in the Biologit Database
  * Data start date is `01-JAN-2025`

**Fixes**

* **Duplicate detection by content**: on some cases the duplicate engine would ignore certain content-based duplicates, affecting a small fraction of candidate duplicates. This fix enhances detection of duplicates in this scenario. ([more about duplicate detection](/topics/duplicate-detection))

## MLM-AI 1.1.42.1 (16-MAY-2025)

**Enhancements**

* **Performance improvements**: this release includes general performance enhancements when navigating Review Details, Find Articles and Monitors
* **E2B (Excel) download**: Date of Latest Information now included in E2B Excel file contained in zip folder&#x20;

**Fixes**

* Zip download extraction error when including very large file names

## MLM-AI 1.1.41.6 (25-APR-2025)

**Enhancements**

* **Incremental loading of** [**attachments**](/application/article-screening/attachments-and-full-text-screening): Large attachment files are now rendered incrementally, and users can start reviewing the article as soon as initial pages are retrieved
* **Customizable sampling rates**: Users can now specify sampling rates directly in the sampling drop down when selecting articles for review or QC:

<div align="left"><figure><img src="/files/5EiXMBDDs5WbVeiJj4zh" alt=""><figcaption></figcaption></figure></div>

* **Search by annotations, exclusions and special situations**: from Reviews or Find Articles pages it is now possible to search results matching specific [annotations](/configuration/monitor-configuration#annotations), [exclusions](/configuration/monitor-configuration#team-assignment-special-situations-and-exclusions) or special situations. The **+** icon on the search bar expands to additional search criteria.

<figure><img src="/files/hV27HOos7OECQFueZwTj" alt=""><figcaption></figcaption></figure>

**Fixes**

* Dashboards: clearer color scheme in Summary dashboard, include monitor name in downloaded chart for Result Trends
* In rare cases user may lose a locked article while on screen, lock re-acquisition timing has been adjusted&#x20;
* Day zero resetting option limits date selection to up to today's date

## MLM-AI 1.1.41.1 (08-APR-2025)

**Enhanced Review Dashboard**

We have updated the [Summary View Dashboard](/application/dashboards#summary-view) with table view, downloadable option and a new "Days Open" chart to help prioritize in-progress reviews according to number of days open.&#x20;

<figure><img src="/files/RvStSVfOHSZQi9MJUzYy" alt=""><figcaption></figcaption></figure>

**Fixes**

* Users unable to remove certain attachments in the Attachments tab

## MLM-AI 1.1.40.5 (27-MAR-2025)

**Improved usability of Attachments**

* [Attachments ](/application/article-screening/attachments-and-full-text-screening)tab now consolidate each upload, rendered status and translation into a single view for easy reference

<figure><img src="/files/JhHPJubsXHBYZSVQE5mL" alt=""><figcaption></figcaption></figure>

* Each attached file or translation will only download on click, thus preventing high network usage in case of multiple attached files.

## MLM-AI 1.1.40.4 (22-MAR-2025)

:tada: **New Features**

* [**Annotation Validations**](/configuration/custom-workflows#mandatory-and-disabled-annotations): This is a new exciting feature where user annotations can be configured as mandatory or disabled depending on assessor decision, enhancing data quality and reducing mistakes.
* [**Categories Filtering**](/configuration/custom-workflows#categories-to-display): Similarly, decisions can determine what categories user can select, reducing the chance of data entry mistakes

**Enhancements**

* The Article Detail info panel has been redesigned to better fit screens of all sizes:

<figure><img src="/files/RiX0RXPdWInaCDHYu0bj" alt=""><figcaption></figcaption></figure>

## MLM-AI 1.1.40.1 (14-MAR-2025)

:tada: **Enhancements**

* **Performance improvements** on Login, Article navigation and Find Articles :rocket:

**Fixes**

* **Translations**: Several enhancements improving Abstract and Full Text translations; support for translation of PDF's of up to 200 pages&#x20;
* **Password auto-lock**: admins can immediately unlock accounts with a password reset (skipping lock wait time)

## MLM-AI 1.1.39.2 (04-MAR-2025)

:tada: **Enhancements**

* **Performance improvements** on Article navigation and Find Reviews :rocket:
* **Reports**: Report submission now allows the selection of multiple monitors
* **Reports**: Can now send report output to designated emails, per [Settings](/configuration/settings)
* **QC Sampling rate**: new 5% sampling rate option now available

**Fixes**

* [Article Search](/application/article-search) now supports running large query strings (previously failing with "Bad Request")

## MLM-AI 1.1.39.1 (15-FEB-2025)

:tada:**Enhancement: Alternative PDF view**

* Files requiring rotation can be visualized using the default browser viewer, by clicking the `PDF`icon:

<figure><img src="/files/HgsGlRDOjIWa3DkXk7u8" alt=""><figcaption></figcaption></figure>

:tada:**Enhancement: Assign articles from the detail screen**

* Users with work assignment permissions now can assign articles directly from the Article Detail screen
* Clicking the assignment icon (top right) will open the asignment dialog

## MLM-AI 1.1.38.2 (03-FEB-2025)

:tada: **New Feature -** [**Extra Annotations**](/configuration/monitor-configuration#monitor-options)**:** Monitors can be configured with additional fields to help users capture customized data during assessment.

:tada: **New Feature - Password auto-lock:** this functionality prevents account break-in attempts by temporarily locking the account after too many failed logins. This is a system configuration that administrators can setup via the Settings page. Learn more more on [Security Settings](/configuration/settings#security).

## MLM-AI 1.1.38.1 (22-JAN-2025)

This is a maintenance release introducing the following enhancements

* **Hide auxiliary tabs**: when in Work Assignment mode the number of tabs visible to users by default can be reduced to minimize clutter. This is a system setting under "Advanced Screening"

<figure><img src="/files/MqJAt9TS4JFxRwUTXOg8" alt=""><figcaption></figcaption></figure>

* **Fix**: on certain cases the a "red" lock could be displayed in Article Detail screen (visual issue only, does not affect locking behavior)

## MLM-AI 1.1.37.3 (09-JAN-2025)

:tada: **New Feature:** [**Track Day Zero per Article**](/application/article-screening/article-level-day-zero)

* It is now possible to track regulatory "day zero" (day information became available for assessment) per article. Learn more in: [Article-Level Day Zero](/application/article-screening/article-level-day-zero#enable-day-zero-per-article)

**Fixes**

* Monitor export/import functionality not importing across environments
* Article lock icon sometimes not showing user details (locking still works as expected)

## MLM-AI 1.1.35.3 (22-OCT-2024)

This is a maintenance release introducing minor fixes and improvements:

* Fix: When creating a copy of a monitor based on a query string, user defined keywords need to be re-defined manually
* Fix: Creating a copy of a patient record in the E2B tab may sometimes ignore edited fields&#x20;

## MLM-AI 1.1.34.2 (04-OCT-2024)

This is a maintenance release introducing minor fixes and enhancements:

* Fix: on certain cases a review that retrieved no results would not automatically move from the "In Progress" to "Completed" tab
* Fix: Downloading E2B/Zip exports from an article now produces unique file names
* Fix: When inspecting a monitor details from the Review tab, it will correctly display ISSN filter information
* Fix: Uploads of EMA Excel files

:tada: **Enhancement**: See Additional Citation Columns in Batch Review

From the Batch Review screen users can choose additional fields corresponding to other citation details like journal, publication dates, etc:

&#x20;&#x20;

<figure><img src="/files/femBsYg7Ugseh0xexaFk" alt=""><figcaption></figcaption></figure>

## MLM-AI 1.1.33.2 (19-SEP-2024)

:tada: **New Feature:** [**Source Filters**](/configuration/monitor-configuration#source-filters)

Source filters is a monitor configuration that allows users to exclude "non-relevant" articles from review and focus on screening of "relevant" articles only. Learn more about the feature in [Source Filters](/configuration/monitor-configuration#source-filters).

## MLM-AI 1.1.32.3 (29-JUL-2024)

:tada: [**New feature: Decision Rules**](/configuration/custom-workflows#adding-rules-to-decisions) - Multi-stage workflows can be configured with decision rules to control options available to users at each stage, such as allowing only "Preliminary" decisions initially and "Final" decisions only after preliminary ones are recorded.

:tada: [**New feature: Related Articles**](/application/article-screening/reviewing-articles-in-related-reviews) - Biologit MLM-AI helps streamline assessments by identifying related articles that appear in multiple monitor results, preventing duplicate work and improving assessment quality.

**Other Enhancements**

* Fix: faster rendering of uploaded PDF articles

## MLM-AI 1.1.32.1 (11-JUL-2024)

* Fix: Performance issue when clicking the "Processing" tab in the Reviews summary page.

## MLM-AI 1.1.31.1 (13-JUN-2024)

:tada: **Automated machine translation of abstracts (in Beta)**

We are rolling out support for automatic machine translations as a Beta feature. Users interesting in evaluating machine translation for their workflows can enable the feature from the [Settings](/configuration/settings) page.&#x20;

Select the `biologit` machine translation vendor:

![](/files/n7G2lM8QLlpuqDANMwSE)&#x20;

Once enabled, articles with title and abstract not in English will display an option to request a translation:

<figure><img src="/files/YDjPh0Qk2eQlQb5RCpd3" alt=""><figcaption></figcaption></figure>

The Machine Translation feature is **free to use during Beta**. Please contact support for more details on future translation charges.

## MLM-AI 1.1.30.2 (21-MAY-2024)

**🎉Enhancement: Article Search Advanced Features**

The [Article Search](/application/article-search) page now includes a number of enhancements:

* Query syntax checking as you type
* Suggested medication synonyms for expansion
* Export results as `csv`
* Compare mode to investigate the effects of adding/removing query terms
* Analyze mode to get insights on query hit volume over time

## MLM-AI 1.1.29.3 (16-APR-2024)

This is a maintenance release covering various small fixes and enhancements:

* Fix: monitors are consistently listed in alphabetical order in all drop down selectors (submission, reports, etc)
* Fix: Do not show DOI badge in Article Search when citation has no DOI data
* Enhancement: validation of email addresses when creating new users
* Enhancement: the [Reports module](/application/reports) now contains new tabs to prevent clutter of old report runs.
  * On submission, users can quickly search their results in the "Recent" tab
  * Users can also search any previously run report in the "Find Reports" tab

<figure><img src="/files/Qe7ikGoqfdJlxCz6MKD7" alt=""><figcaption></figcaption></figure>

## MLM-AI 1.1.29.2 (05-APR-2024)

**🎉E2B R3 Now Available**

Biologit MLM-AI has extended support of E2B XML exports to include the [R3 version](broken://spaces/o9FZ6OwG5pRp4pCglbxy/pages/bz01OXRSgU0z0ZC1nFuR) of the specification, ensuring smoother integration with systems that require this version of the E2B spec.

Learn more in the [E2B Configuration Guide](/configuration/e2b-configuration-guide).

**🎉More space for entering your decision**

A new, optional decision window is available for cases where more space is needed to enter screening decisions. This view also shows the complete audit trail of the article as a reference. Users can access it from the pop-up icon from the Article Detail screen:

<figure><img src="/files/VOfab0tHnt7ropUl4dli" alt=""><figcaption></figcaption></figure>

**🎉Making Aggregate Report Submissions Easier!**

It is now possible to obtain a MS Word report of selected articles to be used for submissions such as aggregate reporting, CERs and others. From the reports module select the Decisions Summary Report:

<figure><img src="/files/AcToGwQdEURzvsHt1P6C" alt="" width="563"><figcaption></figcaption></figure>

The report output is in editable MS-Word containing a citation summary along with assessor review decisions:

<figure><img src="/files/0M6NQbKQMVurrZ80wX8U" alt="" width="554"><figcaption></figcaption></figure>

## MLM-AI 1.1.27.3 (01-MAR-2024)

**🎉New Database: MEDES**

MEDES has become a new literature source that is being ingested into MLM-AI database. In order to start retrieving articles from a new database in your searches, [create a new monitor](/configuration/monitor-configuration#creating-monitors) or update an already existing one to include MEDES into [a list of sources](/configuration/monitor-configuration#sources-and-search-schedule):

<figure><img src="/files/MsrPwNugoeEPIeMhedju" alt=""><figcaption></figcaption></figure>

While using [Article Search](/application/article-search), just make sure the checkbox located near MEDES label is selected:

<figure><img src="/files/8FduvBmwTVepOIHLE7qi" alt=""><figcaption></figcaption></figure>

## MLM-AI 1.1.27.1 (15-FEB-2024)

**Find Articles ➡️ Batch Review**

With the [Find Articles](https://docs.biologit.com/application/reviews/find-articles-from-reviews) search, it is possible to search results by monitors, date, assessment decisions or any criteria allowed in the review search box. It is now possible to review results from Find Articles in [batch mode](https://docs.biologit.com/application/batch-screening). This is useful for consolidating review or QC activities in a single operation.

**Export Articles as ZIP**&#x20;

A new `zip` format is available to download from the Article details screen including Article `HTML` export, all attachments and E2B data (if available).

&#x20;

## MLM-AI 1.1.25.3 (11-JAN-2024)

**🎉New Feature: Dashboards (beta)**

Accessible from the menu on the left, [dashboards](/application/dashboards) provide a new way to analyze data retrieved and screened inside MLM-AI. They show total daily literature volumes, volumes by tags including case reports, suspected AE special situations, animal/in-vitro and many else.

Additional visualizations coming soon, let us know your feedback!

<figure><img src="/files/LlESB6GcyauhET1xV7LX" alt=""><figcaption></figcaption></figure>

## Archived Release Notes

{% content-ref url="/pages/UfPNrAab7mUzEB6zy3lW" %}
[Release Notes - 2023](/more-help/release-notes/release-notes-2023)
{% endcontent-ref %}

{% content-ref url="/pages/U2ceT0j61QJC1ySCn8hY" %}
[Release Notes - 2022](/more-help/release-notes/release-notes-2022)
{% endcontent-ref %}


# Release Notes - 2023

## MLM-AI 1.1.24.2 (24-NOV-2023)

**🎉New Feature: Configurable Filters**

Monitors now support [creation of pre-defined filters](/configuration/monitor-configuration#filters). Filters can be used directly in the search box in the Review Details screen to simplify complex searches.

<figure><img src="/files/hjGHuyTe0XhIQY21kUP8" alt=""><figcaption></figcaption></figure>

**Fixes**

* Duplicates by DOI: a small fraction of relevant articles may not be flagged as a duplicate [by DOI matching](/topics/duplicate-detection#same-document-object-identifier), which is one of the methods used in MLM-AI's duplicate engine. This fix improves the accuracy of this detection method
* Fixed regression in Monitor Export to file

## MLM-AI 1.1.22.3 (12-OCT-2023)

**Enhancement - Syntax highlight support**

The [Article Search](/application/article-search) page now supports syntax highlights, line numbering and easier to read font style for writing queries:

<figure><img src="/files/ik5FAkRZWFX9QwAUOtmk" alt=""><figcaption></figcaption></figure>

**Uploading Review Data**

When [uploading review data](/application/upload-data) from a 3rd party source, users may see additional informational message indicating when article IDs have already been previously uploaded and will be discarded by the upload process:

&#x20;![](/files/HY2QiDbnFWwzpVYPYeVi)

Further information is visible by clicking on the verification badge.

**Help pane content has been fully refreshed, and is now easier to find next to page titles**

<div align="left"><figure><img src="/files/GV0KNlJSVd1DyOZhogHu" alt=""><figcaption></figcaption></figure></div>

## MLM-AI 1.1.21.1 (08-AUG-2023)

**Enhancement - User Inactivity Timeout**

* For additional security, the system logs off users due to inactivity. This feature is enabled by default and can be configured [from the Settings view](/configuration/settings). The default inactivity timeout setting is 30 minutes

**Fixes**

* Highlights: the user defined highlighter takes precedence over other system highlighters when they apply to the same word
* Attachments: full text PDF attachments would not display correctly after an attachment was deleted and re-uploaded

## MLM-AI 1.1.20.6 (31-JUL-2023)

* **Feature Improvements**
  * [**Find Articles from Reviews**](/application/reviews/find-articles-from-reviews) **-** Users can now select multiple monitors in a single query the multi-choice drop down. &#x20;
  * **Attachment Size Limits -** Attachment maximum size has now been increased to 50mb from 10mb

## MLM-AI 1.1.20.5 (25-JUL-2023)

**Enhancement - Truncated Abstract Notification**

* In some rare cases, publishers might send article abstracts too large for display in MLM-AI.&#x20;
* The UI now shows a banner on articles that have been truncated by the MLM-AI database, notifying that users should inspect the source article for the complete text.

<figure><img src="/files/e8tLW7OrNyHEI3hzmbqu" alt=""><figcaption></figcaption></figure>

## MLM-AI 1.1.20.1 (13-JUL-2023)

:tada: **New Feature -** [**Full Text Review**](/application/article-screening#full-text-screening)

* Users can now upload pdf versions of an articles full text and review its content inside MLM-AI
* This feature supports key term and target term highlighting&#x20;
* Full decision making capabilities are accessible from the full text view

## MLM-AI 1.1.19.4 (3-JUL-2023)

:tada: **New Feature -** [**Automatic Notifications**](/configuration/notifications)

* You can now configure monitors to send automatic notifications on assessment updates on articles.
* Whenever a new assessment is added, all users involved in the article assessment (per audit log)
* Users can optionally send to interested parties directly, using the email article feature&#x20;

**Enhancements**

* Input text limits: multi-line input fields were increased to `1500` characters to allow greater level of detail when recording screening decisions

## MLM-AI 1.1.18.5 (29-MAY-2023)

**Enhancements**

* **Improved readability of filter descriptors**: Clicking the filter icon in the Article Detail page shows user-friendly summary of selected filters.

![](/files/fzMKD1kXc2yi35mZUBz7)

* **Upload PubMed Email Alerts**: MLM-AI now supports screening from PubMed email alerts in MEDLINE format by [uploading from an email file](/application/upload-data) (in `eml` format)

**Fixes**

* Reports: the Review Activity Report may fail on citations where main author affiliation has ambiguous country
* Uploads: Embase uploads capture and present keywords and drug taxonomy from source file
* Uploads: Dialog uploads now disambiguate main author affiliations
* Exports: attachment details for fill text free/paid uploads now correctly displayed

## MLM-AI 1.1.18.3 (17-MAY-2023)

**Enhancements**

* Text input on any field now clearly displays maximum allowed sizes when entering data:

![](/files/twVIrKLyI3RxHuTJgOWj)

**Fixes**

* In some cases the top-right counter in Review Details can double-count reviewed articles (does not affect tab behavior or data in reports)

## MLM-AI 1.1.17.5 (27-APR-2023)

#### Enhancements

* **Validation and Feedback on Article Search**: The [Article Search](/application/article-search) page has a new layout and more detailed validation feedback on user queries:

<figure><img src="/files/akF1CyB5k5A8i21FR9xg" alt=""><figcaption></figcaption></figure>

* **PubMed Mesh terms updates**: PubMed will sometimes update the MeSH taxonomy associated with an article only after it is already published. To increase coverage, MLM-AI now makes this information available by re-scanning and updating article MeSH keywords periodically

#### Fixes

* The Article Search page was previously available only to users with "Administrator" privilege. It is now available to all users (creating Monitors from a query is still limited to administrators or users with "Monitor management" privilege)&#x20;
* Monitor exports to file not correctly exporting Query-based monitors
* Searching an article ID sometimes retrieve multiple results (by string matching) instead of a single ID as expected

## MLM-AI 1.1.17.4 (14-APR-2023)

#### Enhancements

* **Review Validation**: Better user experience when submitting [On-Demand reviews](/application/reviews#submitting-reviews-on-demand). Users now get instant feedback on selected date range and how many hits the review is expected to return.
* **Exports from the Global View**: It is now possible to export results from the [Global View (Find Articles)](/application/reviews/find-articles-from-reviews) view, using similar to export options available in Review and Article Details.

#### Fixes

* Copy/pasting a link from the browser tab works even if user is not authenticated: on accessing the link users are redirected to the login page, and immediately redirected to the target link if login is correct

## MLM-AI 1.1.16.4 (03-APR-2023)

**New Features**

* With the new [**Find Articles**](/application/reviews/find-articles-from-reviews) view, users can select articles from different Review results (by monitor, or by date range across all monitors), and filter results by status, assignment, user and any other available criteria&#x20;

**Fixes and Enhancements**

* If a screening decision sets any citation to a "Pending" status in Completed review, the Review moves back to "In Progress" as it now awaits review completion. This can happen on workflows where, for example, an article requires additional work after QC. Learn more about the [Review Lifecycle](/application/reviews#the-review-lifecycle).

## MLM-AI 1.1.15.3 (13-MAR-2023)

**New Features** :tada:

* [**Work Assignment**](/application/review-details/work-assignment) - With new feature enabled, a team supervisor gets an opportunity to maintain an evenly balanced workload by splitting the screening activity between team members.

**Fixes and Enhancements** :gear:

* Supported sources for Upload Review now support ingesting of authors' affiliations.
* **Fix**: duplicate detection support for reviews using a mixture of upload and MLMAI results.&#x20;

## MLM-AI 1.1.15.1 (13-FEB-2023)

**New Features** :tada:

* **E2B** - Comprehensive support for capturing verbatim event data during screening and produce E2B files in R2 format. See the [E2B configuration guide](/configuration/e2b-configuration-guide) for more details on how to use this feature.

**Enhancements**

* [**Retiring Reviews**](/application/reviews#the-review-lifecycle) - in cases when a review is produced in error, an administrator can sideline it from  any screening activity and reporting by setting its status to **Retired**.

## MLM-AI 1.1.14.4 (6-FEB-2023)

**New Features** :tada:

* [**AI Model Selection**](/configuration/settings#mlm-ai-model-release)**:** New setting that allows users to alter the version of AI Models that reviews use to infer.
* **AI Model Version JAN2023A\_051:** Newly trained and tuned models for MLM. This model uses more data that has been selected, labelled and annotated by biologit.&#x20;

**Enhancements** :gear:

* Upload Review process:
  * Reporter country is ingested from Adis to be used for territory-based searches.
  * Semicolon-separated data coming from Dialog is presented in a more user friendly format.

## MLM-AI 1.1.14.3 (27-JAN-2023)

**New Features** :tada:

* [**Article Page Locking**](/configuration/settings#article-locking): Delivers better support for team-based screening activities where an article page remains locked for a user ensuring no two users update the same article at the same time.

**Enhancements**

* Per-article download option: Data for individual articles (in Excel, E2B XML formats) can now be downloaded directly from the Article Detail page.
* Better support for QC permissions: Only users assigned the "QC" role in a monitor will be able to update decisions from a different user.
  * A user can still update its own screening decisions as needed.

## MLM-AI 1.1.14.2 (23-JAN-2023)

This is a maintenance release delivering updates to our internal infrastructure and the following improvements:

* Review Details Search box now supports [searching decisions](/application/review-details/searching-abstracts-inside-a-review#decision-less-than-decision-greater-than) for a custom workflow with the `decision:` syntax.
* Title check on DOI-based duplicates. As an extra verification measure, DOI-based duplicate articles are confirmed with an additional title similarity check.
* &#x20;**Fix**: Pending tab would not render correctly when saving a "Pending" decision using a custom workflow.

## MLM-AI 1.1.13.2 (16-JAN-2023)

**Enhancements**

* [**Multi-Uploads Screen**](/application/upload-data) **-** Search results produced by external databases can now be imported in bulk saving time. Uploaded data is now validated up-front, clearly indicating the number of processed rows, and reporting validation errors as appropriate.&#x20;

<figure><img src="/files/s1Skf77OBoecb6XtMqfq" alt=""><figcaption></figcaption></figure>

* **Duplicate Detection by DOI -** Automated duplicate detection now supports [DOI-based detection](/topics/duplicate-detection#same-document-object-identifier) in addition to existing methods such as detection by content similarity and unique identifier. This enhancement helps detect duplicate articles published on different databases.
  * This feature is automatically enabled for all Monitors where duplicate detection is enabled.&#x20;
  * Learn more at the [Duplicate Detection](/topics/duplicate-detection) page.
* **Duplicate Tab in Article Details -** It is now easier to inspect the details of duplicate articles. A new "Duplicates" tab will appear on every article where a duplicate is found (according to the criteria outlined in the [Duplicate Detection](/topics/duplicate-detection) page)&#x20;

<figure><img src="/files/BEpreko2TK7C1LSMhwGx" alt=""><figcaption></figcaption></figure>

## Archived Release Notes

{% content-ref url="/pages/U2ceT0j61QJC1ySCn8hY" %}
[Release Notes - 2022](/more-help/release-notes/release-notes-2022)
{% endcontent-ref %}


# Release Notes - 2022

Biologit MLM-AI Release Notes Archive - 2022

## MLM-AI 1.1.12.3 (08-DEC-2022)

**Enhancements**

* **Search by article topics and keywords**: The biologit MLM-AI database is now fully searchable by key terms and topics associated to abstract by the source publishers. This is now available for the PubMed and DOAJ data sources.
  * Article topics are also visible in Article Search results, and in the Article Review page.
* **Duplicate detector**: Improved accuracy on duplicate detection by content: this feature now includes an additional check on title similarity to improve accuracy.

**Fixes**

* **Fix**: Tab navigation in article detail on certain cases was retrieving incorrect counters
* **Fix**: Article detail asking users to save/discard changes when no changes were made

## MLM-AI 1.1.11.3 (24-NOV-2022)

**New Features**

* **NEW:** [**Custom Workflows**](/configuration/custom-workflows) - users can now configure choice of screening decisions and how they are interpreted by MLM-AI via Custom workflows. This can be setup at system level or for individual monitors.&#x20;

**Enhancements**

* Session preferences: user selections of columns to display, highlighters and other screening settings are preserved
* Search history: All search boxes now display recent searches in a drop down
* Reloadable tabs: Clicking on a tag on Review Summary, Review Details or Article Details will refresh data.
* Extracted country data in text or author affiliation now available in review exports (excel, csv).
* Users are warned before leaving the review page when changes are made to decisions, reactions of highlights.

**Bug Fixes**

* Attachment downloads on some browser versions not rendering PDF files correctly.
* Reactions screen - unable to add reactions after editing highlights.
* List of synonyms in [Monitor configuration](#monitors) page is now sorted.

## MLM-AI 1.1.11.1 (24-OCT-2022)

**New Features**

* **NEW** :tada:: **Create your own queries** using the MLM-AI database with the new [Article Search](/application/article-search#creating-searches) feature, then use your queries to [Create Monitors](/application/article-search#create-monitors-from-searches) and begin screening your results with AI!

{% embed url="<https://youtu.be/tNC7lZbu1mk>" %}

* **NEW**: **Dark Mode**! MLM-AI now supports dark mode interface.

{% embed url="<https://youtu.be/mx76MlL6Js4>" %}

## MLM-AI 1.1.10.3 (11-OCT-2022)

**New Features**

* **NEW 🎉** : MLM-AI now has the ability to filter author's country of origin according to author affiliation. Learn more on [**Author Country of Origin**](https://docs.biologit.com/application/review-details/author-country-of-origin).

**Fixes and Improvements**

* **Fix**: Audit log tab now refreshes with latest data on click.
* **Fix**: navigation issues when "Next Article on Submit" enabled
* **Enhancement**: All product tours can be played together from the main login page.
* **Enhancement**: Article detail tab now displays author affiliation and source database.

## MLM-AI 1.1.10.1 (16-SEP-2022)

* **NEW** :tada: : Product tours are now available! Click on the help icon and then "Take the tour" on any page where tours are available.

{% embed url="<https://youtu.be/2DXB6LaXAAc>" %}

* **Fix**: Disable clicking into a review when it is still in processing stage & results are not yet ready.
* **Fix**: Improve color scheme of highlighters

## MLM-AI 1.1.9.2 (08-SEP-2022)

* **Enhancement**: When saving screening decisions, users can now type into the drop down to narrow the available options:

![](/files/HIMAlxLQ5uz3gD2EQmyF)

* **Enhancement**: Selecting Next Article on Submit using the settings toolbox will automatically move to the next article to screen after saving a decision.

![](/files/pkTEOQnqQJ1gzE0uXDAR)

## MLM-AI 1.1.9.1 (02-SEP-2022)

* **Enhancement**: MLM-AI now informs users when a search period starts before the [system's data start date](/more-help/faq#system-limits).
* **Enhancement**: The search function In the Review details page no accepts [querying for QC status ](/application/review-details/searching-abstracts-inside-a-review#decision-less-than-decision-greater-than)with `decision:qc` or `nodecision:qc`
* **Enhancement**: the search function in the reports module now also searches for matches in report arguments.
* **Fix**: article search corrects certain cases of malformed quoting

## MLM-AI 1.1.8.4 (15-AUG-2022)

**MLM-AI 1.1 Launched!** :rocket:

A modern, faster and fully featured new interface to MLM-AI has been launched. Both 1.0 and 1.1 versions will remain available, with new features being added to the latest 1.1 going forward.

To access MLM-AI 1.1, users will now see a banner at the top menu bar:

![](/files/BLN7ySZNWjl1NOz0Ni2H)

By clicking the "Try it out" link, you will be redirected to the new login page, and can then update your bookmarks.

MLM-AI 1.1 has been **fully validated** and is now ready for service.

![](/files/77gY1wbYvXMvsM0tRSZe)

**\[Beta] Search the MLM-AI scientific database**

It is now possible to directly search articles in the MLM-AI database. This functionality is currently in beta, and will receive further updates in coming releases.

**System Settings**

Administrators can now configure certain account-level settings for all users. Settings will be available from the&#x20;

**Bug Fixes**

* [Monitor search](#monitors) from the UI now searches for keyword in monitor title, description or any defined product synonym.&#x20;

## MLM-AI 1.0.6.1 (05-JUL-2022)

**Reviews**

* **New**: "[Send to inbox](https://docs.biologit.com/application/monitor-configuration#monitor-options)" email feature now includes citation in Vancouver format.
* **FIX**: Incorrect treatment of attachment upload/download functionality on binary files.

**Monitors**

* **FIX**: Display correct monitor schedule on monitor summary screen.

## MLM-AI 1.0.5.2 (14-JUN-2022)

**Reviews**

* **FIX**: Reviews can retrieve results outside of desired start-end periods, when searching by Publication Date.
* **New**: [Keyword search](https://docs.biologit.com/application/review-details/searching-abstracts-inside-a-review#keyword-matching) in Reviews Summary page now searches title, abstract and publication name.
* **New**: [Review uploads](/application/upload-data) now supports EMBASE csv format.

**AI Tags**

* Model **1.0.5.2** released. Technical specification is available from [1.0.5.2 FactSheet](https://github.com/biologit-engineering/factsheets/blob/main/mlm-ai/PROD5/biologit_factsheet_1.0.5.1.pdf):&#x20;

## MLM-AI 1.0.4.3 (01-JUN-2022)

**Monitor settings**

* **Updated** [monitor configuration options](/configuration/monitor-configuration#monitor-options) for automated screening inclusion and exclusion criteria
* **New**: [Push notifications setting](/configuration/monitor-configuration#notifications): get notified by email when reviews are ready or change status.
* **New**: [User defined keywords](/configuration/monitor-configuration#user-defined-keywords), can be used as tags and for [automatic screening results](/configuration/monitor-configuration#keywords-screening).

**Reports**

* **New**: Additional reporting capabilities in the [reporting engine](/application/reports).


# Notices

{% content-ref url="/pages/UwV1b5IxNqLMh0VoT8a0" %}
[Data Privacy](/more-help/notices/data-privacy)
{% endcontent-ref %}

{% content-ref url="/pages/D6IpU5doBVvrjc0UfLT7" %}
[CFR-11](/more-help/notices/cfr-11)
{% endcontent-ref %}


# Data Privacy

This Privacy Policy explains how biologit uses personal data we collect when users access our website and the biologit MLM-AI service.

## Privacy Policy

This Privacy Policy explains how biologit uses personal data we collect when users access our website and the biologit MLM-AI service.

### Data Protection Principles

We take our responsibility for the personal information we collect about you seriously, and we aim to be transparent about how we manage it, and give you control over it. below is a summary of the data protection principles which guide how we use or manage your personal information. These principles provide that personal data should be:

| 1. Used lawfully, fairly and in a transparent way.                                                                                          | 2. Accurate and kept up to date.                                           |
| ------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------- |
| 3. Collected only for valid purposes that we have clearly explained to you and not used in any way that is incompatible with those purposes | 4. Kept only as long as necessary for the purposes we have told you about. |
| 5. Relevant to the purposes we have told you about and limited only to those purposes.                                                      | 6. Kept securely.                                                          |

### What data do we collect?

<table data-header-hidden><thead><tr><th width="235.77777777777777"></th><th></th></tr></thead><tbody><tr><td><strong>User Personal Information</strong></td><td>Any information about one of our Users which could, alone or together with other information, personally identify them or otherwise be linked or connected with them. such as username and password, email addresses, real names, Internet protocol (IP) address.</td></tr><tr><td><strong>Usage Information</strong></td><td>We collect information about how you use the biologit MLM-AI service and our websites, such as the pages accessed, operations performed, your IP address and the date and time of each request.</td></tr><tr><td><strong>Data not intentionally collected</strong></td><td>We do not intentionally collect sensitive user information that may have been stored in the system through free-form content inputs (for example comment boxes).</td></tr></tbody></table>

### How do we collect your data?

The law requires us to have a “legal basis” or “lawful ground” to collect and use your personal information. Some of the grounds for processing may overlap and there may be several grounds which justify our use of your personal information. Most commonly, we will use your personal information in the following circumstances:

* To comply with a legal obligations and regulatory requirements.
* Use your browser to access the biologit MLM-AI application, documentation, or company websites.
* Voluntarily complete customer surveys or provide us feedback via support channels or email.
* biologit may also receive your data indirectly through online marketing campaigns and traffic data collected by third party sources.
* Cookies are text files placed on your computer to collect standard Internet log information and visitor behavior information. When you visit our websites, we may collect information from you automatically through cookies or similar technology. For further information, visit: [https://allaboutcookies.org](https://allaboutcookies.org/) .

### How we use your data?

biologit collects your data to:

* Operate and improve the biologit MLM-AI service.
* Respond to support requests.
* Contact you for updates about the biologit MLM-AI service, or specific queries in relation to your account, subject to your consent.
* Invite you to take part in surveys, test programs or other research purposes subject, where necessary, to your consent.
* Investigate technical incidents, including security incidents.
* Comply with our legal obligations, protect our intellectual property, and comply with our Terms of Service.

### How we store your data?

Data for the biologit MLM-AI service is stored on secure cloud servers hosted in the EU region (Ireland). biologit retains your data on our servers for as long as your account is active, and as needed to comply with our legal obligations. If your account is deactivated your data may still be retained in company backups according to the company backup policies current at the time.

Separately, data that you provide via the biologit website, customer surveys and our support portal are stored on secure cloud servers by third party companies.

### Data Security

biologit have put in place best-practice security measures and training to assist with preventing your personal information from being accidentally lost, used, or accessed in an unauthorized way, altered, or disclosed. In addition, we limit access to your personal information to those employees, contractors and other third parties who have a business need-to-know. They will only process your personal information on our instructions, and they are subject to a duty of confidentiality.

### Data Retention

We keep personal data in accordance with our regulatory obligations and good industry practice. If you have any questions or concerns about how long we retain your personal data, please contact [our support channels](/more-help/support-and-platform-details#customer-support).

### Your Data Protection Rights

biologit wishes to ensure you are fully aware of all your data protection rights. Data protection requests are subject to technical viability and applicable laws. In some cases, a processing fee may be charged in relation to the request. You are entitled to the following:

| **The right to access**                            | You have the right to request biologit for copies of your personal data.                                                                                                                                                                                                                                                                                                                                                       |
| -------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **The right to rectification**                     | You have the right to request that biologit corrects any information you believe is inaccurate. You also have the right to request us to complete the information you believe is incomplete.                                                                                                                                                                                                                                   |
| **The right to erasure**                           | You have the right to request that we erase your personal data, under certain conditions subject to technical viability.                                                                                                                                                                                                                                                                                                       |
| **The right to restrict processing**               | You have the right to request that biologit restricts the processing of your personal data.                                                                                                                                                                                                                                                                                                                                    |
| **The right to object to processing**              | You have the right to object to biologit’s processing of your personal data.                                                                                                                                                                                                                                                                                                                                                   |
| **The right to data portability**                  | You have the right to request that biologit transfer the data that we have collected to another organization, or directly to you, subject to technical viability.                                                                                                                                                                                                                                                              |
| **Request erasure of your personal information\*** | This enables you to ask us to delete or remove personal information in limited circumstances, where: (i) it is no longer needed for the purposes for which it was collected; (ii) you have withdrawn your consent (where the data processing was based on consent); (iii) following a successful right to object (iv) it has been processed unlawfully; or (v) to comply with a legal obligation to which biologit is subject. |

**\****We are not required to comply with your request to erase personal information if the processing of your personal information is necessary for a number of reasons, including compliance with a legal obligation, or for the establishment, exercise, or defence of legal claims***.**

If you make a request, we have one month to respond to you. If you would like to exercise any of these rights, please [contact support](/more-help/support-and-platform-details#customer-support).

### How to Contact Us

If you have any questions about Our Company’s privacy notice, the data we hold on you, or you would like to exercise one of your data protection rights, please do not hesitate to [contact us](/more-help/support-and-platform-details#customer-support).

Privacy notice last updated: 08-Dec-2022

Document code: QA-POL-002


# CFR-11

This document describes how biologit MLM-AI core features help customers implement processes compliant with CFR Part 11 requirements.

Biologit MLM-AI was designed from the ground up with the requirements of pharmacovigilance in mind, including FDA's [CFR Part 11](https://www.ecfr.gov/current/title-21/chapter-I/subchapter-A/part-11) guidance on electronic records for computerized systems.&#x20;

### Key Features <a href="#key-features" id="key-features"></a>

The key features of MLM-AI supporting CFR-11 compliance are:

**Auditing**

* Full audit trail of user actions.
  * The audit trail has indefinite retention and can not be tampered with.
* Reporting capabilities to inspect the audit trail.
  * Limited to privileged users.
* Data export capability to facilitate customer’s records management policies.

**Permissions**

* [Granular permissions management](https://docs.biologit.com/application/team-and-user-management#permissions) that allows users access to specific resources.
* Flexible [assignment of users to teams](https://docs.biologit.com/application/team-and-user-management#users-and-teams-management) according to their function.

**Workflow**

* [Configurable workflow settings](https://docs.biologit.com/application/monitor-configuration#creating-monitors) that allow users only specific access to perform the necessary steps according to their roles.

**User authentication**

* Authentication based on unique user ID + password combination.
* Configurable password complexity policy.

## CFR Part 11 and Biologit MLM-AI <a href="#fda-title-21-cfr-part-11-and-biologit" id="fda-title-21-cfr-part-11-and-biologit"></a>

The table below maps relevant sections of the CFR Part-11 specification to biologit MLM-AI capabilities addressing the requirement.

### Closed Systems Controls (11.10) <a href="#closed-systems-controls-11.10" id="closed-systems-controls-11.10"></a>

<table data-header-hidden><thead><tr><th width="162.00000000000003"></th><th></th><th></th></tr></thead><tbody><tr><td><strong>Section</strong></td><td><strong>Requirement</strong></td><td><strong>biologit MLM-AI Compliance</strong></td></tr><tr><td><p><br><br></p><p><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (a)</a></p></td><td>Validation of systems to ensure accuracy, reliability, consistent intended performance, and the ability to discern invalid or altered records.</td><td><ul><li>biologit MLM-AI includes full audit trail of user actions within the system.</li><li>The audit trail can be accessed via the system’s reporting capabilities.</li><li>The above functionality can be leveraged as part of a customer’s validation plan.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (b)</a></td><td>The ability to generate accurate and complete copies of records in both human-readable and electronic form suitable for inspection, review, and copying by the agency. Persons should contact the agency if there are any questions regarding the ability of the agency to perform such review and copying of the electronic records.</td><td><ul><li>biologit MLM-AI can report all activity recorded on the audit trail from its reporting functionality.</li><li>Reports can be issued in human and machine readable formats.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (c)</a></td><td>Protection of records to enable their accurate and ready retrieval throughout the records retention period.</td><td><ul><li>biologit MLM-AI has <a href="/pages/-MWB0_d5P78BUQXV3poq#permissions">configurable, granular permissions</a> that determine which actions and data users have access to.</li><li><a href="/pages/-MWAeHFuZ40-84xqWpb8#downloading-reviews">Data export</a> capabilities facilitate the desirable level records retention by customers.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (d)</a></td><td>Limiting system access to authorized individuals.</td><td><ul><li>biologit MLM-AI implements a security model based on<a href="/pages/-MWB0_d5P78BUQXV3poq#users-and-teams-management"> users and user groups (teams)</a> allowing customers to limit access to users according to their function.</li><li>Users are uniquely identified by their ID and password combination.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (e)</a></td><td>Use of secure, computer-generated, time-stamped audit trails to independently record the date and time of operator entries and actions that create, modify, or delete electronic records. Record changes shall not obscure previously recorded information. Such audit trail documentation shall be retained for a period at least as long as that required for the subject electronic records and shall be available for agency review and copying.</td><td><ul><li>biologit MLM-AI includes full audit trail of user actions within the system.</li><li>The audit trail can be accessed via the system’s reporting capabilities.</li><li>The audit trail can not be tampered with from within biologit MLM-AI.</li><li>The audit trail is retained indefinitely and is available for export to meet customer’s specific records retention policies.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (f)</a></td><td>Use of operational system checks to enforce permitted sequencing of steps and events, as appropriate.</td><td><ul><li>biologit MLM-AI includes <a href="/pages/-MWB0izVuvIBWCZdkqkj#monitor-options">configurable workflow settings</a> to enforce the completion of screening activities in accordance to a pre-determined sequence of steps.</li><li><a href="/pages/-MWB0_d5P78BUQXV3poq#role-based-permissions">User access</a> can be configured such that steps are executed only by authorized individuals.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (g)</a></td><td>Use of authority checks to ensure that only authorized individuals can use the system, electronically sign a record, access the operation or computer system input or output device, alter a record, or perform the operation at hand.</td><td><ul><li>User access can be configured such that specific actions are executed only by authorized individuals.</li><li>Users are uniquely identified by their ID and password combination.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10"><em>11.10 (h)</em></a></td><td>Compliant Electronic Document Management Systems must use device (e.g., terminal) checks to determine, as appropriate, the validity of the source of data input or operational instruction.</td><td>N/A</td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (i)</a></td><td>Determination that persons who develop, maintain, or use electronic record/electronic signature systems have the education, training, and experience to perform their assigned tasks.</td><td><ul><li>N/A - Implementing a training plan and maintaining training records are responsibilities specific to organizations using biologit MLM-AI</li><li><p>To assist with this task biologit MLM-AI supplies:</p><ul><li>Extensive <a href="https://docs.biologit.com/">online product documentation</a> and tutorials.</li><li>The ability for users to access an isolated “sandbox” environment for training purposes.</li></ul></li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (j)</a></td><td>The establishment of, and adherence to, written policies that hold individuals accountable and responsible for actions initiated under their electronic signatures, in order to deter record and signature falsification.</td><td><ul><li>N/A - Preparing policies and ensuring their adherence are responsibilities specific to organizations using biologit MLM-AI.</li></ul></td></tr><tr><td><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.10">11.10 (k)(1)</a></td><td><p>Use of appropriate controls over systems documentation including:</p><ol><li>Adequate controls over the distribution of, access to, and use of documentation for system operation and maintenance.</li><li>Revision and change control procedures to maintain an audit trail that documents time-sequenced development and modification of systems documentation.</li></ol></td><td><ul><li>N/A - Documentation controls are processes specific to organizations.</li><li>biologit provides extensive, freely available versioned online product documentation. The same content can be made available in printed format.</li></ul></td></tr></tbody></table>

### Controls for Open Systems (11.30) <a href="#controls-for-open-systems-11.30" id="controls-for-open-systems-11.30"></a>

| **Section**                 | **Requirement**                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   | **biologit MLM-AI Compliance**                                                                                                                                                                                                                                                                                                                                                                  |
| --------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| <p><br><br></p><p>11.30</p> | <p>Persons who use open systems to create, modify, maintain, or transmit electronic records shall employ procedures and controls designed to ensure the authenticity, integrity, and, as appropriate, the confidentiality of electronic records from the point of their creation to the point of their receipt.</p><p>Such procedures and controls shall include those identified in Sec. 11.10, as appropriate, and additional measures such as document encryption and use of appropriate digital signature standards to ensure, as necessary under the circumstances, record authenticity, integrity, and confidentiality.</p> | <ul><li>biologit MLM-AI is a cloud-hosted platform accessible via web browsers, and thus is considered an open system.</li><li><p>In addition to the 11.10 compliance features discussed above:</p><ul><li>All customer data at-rest and in-transit is encrypted in biologit’s cloud platform.</li><li>biologit MLM-AI is only accessible via encrypted connections (HTTPS)</li></ul></li></ul> |

### Electronic Signatures (11.50/11.70/11.100/11.200/11.300) <a href="#electronic-signatures-11.50-11.70-11.100-11.200-11.300" id="electronic-signatures-11.50-11.70-11.100-11.200-11.300"></a>

| **Section**                                                                                                              | **Requirement**                                                                                                                                                                                                                                                                                                                                                                                                    | **biologit MLM-AI Compliance**                                                                                                                                                                                                                                                                                                                                  |
| ------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [11.50 (a)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.50)                              | <p>Compliant Electronic Document Management Systems ensure that signed electronic documents contain information associated with the signing, clearly indicating all of the following:</p><ol><li>The printed name of the signer.</li><li>The date and time when the signature was executed.</li><li>The meaning (such as review, approval, responsibility, or authorship) associated with the signature.</li></ol> | <ul><li>A unique user ID and password are mapped to individuals configured in the tool.</li><li>All activity in biologit MLM-AI is timestamped and recorded in the system audit trail.</li></ul>                                                                                                                                                                |
| [11.50 (b)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.50)                              | The items identified in paragraphs (a)(1), (a)(2), and (a)(3) of this section shall be subject to the same controls as for electronic records and shall be included as part of any human readable form of the electronic record (such as electronic display or printout).                                                                                                                                          | <ul><li>biologit MLM-AI provides extensive reporting of the system audit trail, including user names, timestamps and activity record.</li></ul><p> </p>                                                                                                                                                                                                         |
| <p><a href="http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.70">11.70</a></p><p><br><br></p> | Electronic signatures and handwritten signatures executed to electronic records shall be linked to their respective electronic records to ensure that the signatures cannot be excised, copied, or otherwise transferred to falsify an electronic record by ordinary means.                                                                                                                                        | <ul><li>The biologit MLM-AI audit trail records all activity carried out in the system. Each audit record is associated with its signer and is immutable: records cannot be tampered with from the application.</li></ul>                                                                                                                                       |
| [11.100 (a)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.100)                            | Each electronic signature shall be unique to one individual, and shall not be reused by, or reassigned to, anyone else.                                                                                                                                                                                                                                                                                            | <ul><li>Users are authenticated in biologit MLM-AI by their unique user ID and password. Once logged in records created by the user are permanently associated with their user ID.</li><li>It is the responsibility of organizations using biologit MLM-AI to ensure accounts are not shared or reassigned.</li></ul>                                           |
| [11.100 (b)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.100)                            | Before an organization establishes, assigns, certifies, or otherwise sanctions an individual's electronic signature, or any element of such electronic signature, the organization shall verify the identity of the individual.                                                                                                                                                                                    | <ul><li>Users with administrative privileges' in biologit MLM-AI are responsible for ensuring users are properly identified before a unique user ID and password is created in the system for that user.</li><li>biologit MLM-AI supports this process by associating a user ID with a unique identifier in the organization (typically the email ID)</li></ul> |
| [11.100 (c)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.100)                            | Persons using electronic signatures shall, prior to or at the time of such use, certify to the agency that the electronic signatures in their system, used on or after August 20, 1997, are intended to be the legally binding equivalent of traditional handwritten signatures.                                                                                                                                   | N/A                                                                                                                                                                                                                                                                                                                                                             |
| [11.200 (a)(1)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.200)                         | <p>Electronic signatures shall employ at least two distinct identification components, such as an identification code and password.</p><p><br><br></p>                                                                                                                                                                                                                                                             | <ul><li>biologit MLM-AI uses the unique user ID and password as the two distinct identification components.</li></ul>                                                                                                                                                                                                                                           |
| [11.200 (a)(1)(i)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.200)                      | When an individual executes a series of signings during a single, continuous period of controlled system access, the first signing shall be executed using all electronic signature components; subsequent signings shall be executed using at least one electronic signature component that is only executable by, and designed to be used only by, the individual.                                               | <ul><li>In biologit MLM-AI, no user access is granted before users authenticate with their user ID and password (all components).</li><li>After login, all user activity is associated with the user ID in the audit trail.</li></ul>                                                                                                                           |
| [11.200 (a)(1)(ii)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.200)                     | When an individual executes one or more signings not performed during a single, continuous period of controlled system access, each signing shall be executed using all of the electronic signature components.                                                                                                                                                                                                    | <ul><li>In biologit MLM-AI, no user access is granted before users authenticate with their user ID and password (all components).</li><li>Administrators can define a session inactivity period (defaults to 1 hour) corresponding to a contiguous period of controlled system access.</li></ul>                                                                |
| [11.200 (a)(2)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.200)                         | <p>(Electronic signatures) shall be used only by their genuine owners</p><p><br><br></p>                                                                                                                                                                                                                                                                                                                           | <ul><li>User IDs and passwords are uniquely associate with a user upon creation.</li><li>It is the responsibility of organizations using biologit MLM-AI to ensure user IDs are created to corresponding genuine users and passwords are distributed in accordance to company policy.</li></ul>                                                                 |
| [11.200 (a)(3)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.200)                         | <p>(Electronic signatures) shall be administered and executed to ensure that attempted use of an individual's electronic signature by anyone other than its genuine owner requires collaboration of two or more individuals.</p><p><br><br></p>                                                                                                                                                                    | <ul><li>This requirement should be implemented by the procedures in organizations implementing biologit MLM-AI.</li><li>biologit MLM-AI supports monitoring of unusual activity patterns (ex: login attempts) via its reporting capability.</li></ul>                                                                                                           |
| [11.300 (a)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.300)                            | <p>Persons who use electronic signatures based upon use of identification codes in combination with passwords shall employ controls to ensure their security and integrity. Such controls shall include:</p><p>(a) Maintaining the uniqueness of each combined identification code and password, such that no two individuals have the same combination of identification code and password.</p><p> </p>           | <ul><li>In biologit MLM-AI, users are assigned a unique user ID and password combination.</li></ul>                                                                                                                                                                                                                                                             |
| [11.300 (b)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.300)                            | <p>Ensuring that identification code and password issuance are periodically checked, recalled, or revised (e.g., to cover such events as password aging).</p><p> </p>                                                                                                                                                                                                                                              | <ul><li>MLMAI administrators can configure the system with password policies reflecting the organization requirements.</li></ul>                                                                                                                                                                                                                                |
| [11.300 (c)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.300)                            | <p>Following loss management procedures to electronically de-authorize lost, stolen, missing, or otherwise potentially compromised tokens, cards, and other devices that bear or generate identification code or password information, and to issue temporary or permanent replacements using suitable, rigorous controls.</p><p> </p>                                                                             | <ul><li>biologit MLM-AI does not use tokens or cards at this time.</li></ul>                                                                                                                                                                                                                                                                                    |
| [11.300 (d)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.300)                            | <p>Use of transaction safeguards to prevent unauthorized use of passwords and/or identification codes, and to detect and report in an immediate and urgent manner any attempts at their unauthorized use to the system security unit, and, as appropriate, to organizational management.</p><p> </p>                                                                                                               | <ul><li>To support this requirement, administrators can configure the system with minimum password complexity and expiration policies.</li><li>biologit MLM-AI supports monitoring of unusual activity patterns (ex: login attempts)</li></ul>                                                                                                                  |
| [11.300 (e)](http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=11.300)                            | Initial and periodic testing of devices, such as tokens or cards, that bear or generate identification code or password information to ensure that they function properly and have not been altered in an unauthorized manner.                                                                                                                                                                                     | <ul><li>biologit MLM-AI does not use tokens or cards at this time.</li></ul>                                                                                                                                                                                                                                                                                    |


# FAQ

Frequently asked questions about biologit MLM-AI

## Login

### I forgot my password

Users that are registered in the platform can request a temporary password to login and update the forgotten password. This is covered in [Sign on and Navigation](/application/navigation#forgot-password).

## Reviews and Monitors

### Why some abstract in my weekly review have already been screened?

Newly generated reviews may present pre-screened abstracts that were tagged at the [duplicate detection](/topics/duplicate-detection) stage. Duplicate abstracts relate to a corresponding abstract in the same review, or in a previous review for the same product.

### I updated a monitor, but the changes did not take effect on the monitor reviews

Reviews always use the monitor settings in effect at the time of submission. This means any monitor changes performed later will not be reflected. This is by design to ensure traceability: the system behaves and reflects monitor settings current at the time the review was created.

Note the same applies to *scheduled* reviews: once a review is scheduled (i.e. it appears in the "[Processing](/application/reviews#the-processing-tab)" tab), the settings are fixed.

### How does Duplicate Detection Work?

MLM-AI performs duplicate detection **by ID, DOI and content**. This is further detailed in the page:

* [Duplicate Detection](/topics/duplicate-detection)

Articles are flagged as duplicates **only when**:

* They appear in the same review results. One of the articles will be automatically reviewed as "duplicate".
* An article is a duplicate of another article seen in a past review for the *same Monitor*.

{% hint style="info" %}
Learn more about how the techniques used in automated duplicate detection on: [**Feature  spotlight: Duplicate Detection in Biologit MLM-AI**](https://www.biologit.com/post/feature-spotlight-duplicate-detection-in-biologit-mlm-ai)
{% endhint %}

### When will my review move from "In Progress" to "Completed"?

Reviews stay "In Progress" while there are still abstracts to be screened, and move to "Completed" once *all abstracts* have been screened ie. all abstracts were saved with a decision by a user or by MLM-AI via automated action.

* See also the [Review Lifecycle](/application/reviews#the-review-lifecycle) for more details

## Uploading Data

### What data can I upload to MLM-AI for screening?

Search results obtained from **PubMed, EMBASE, EBSCO, Adis, and ProQuest** can be directly uploaded into MLM-AI for screening.&#x20;

The supported formats and help with the export steps are detailed in: &#x20;

{% content-ref url="/pages/-MadTCZehlDKB1Uxd6in" %}
[Uploading Review Data](/application/upload-data)
{% endcontent-ref %}

## Article Search

### Why do I see duplicates on my query results?

Duplicate detection is a feature of [Monitors](#monitors) and is active only when Reviews are requested. Note that [you can create monitors based on the query you formulate in the Article Search](#article-search) screen, and benefit from duplicates detection this way.

### How do I perform term expansion?

The query syntax of the MLM-AI database allows for wildcards that can expand common suffixes, for example:

* `toxic*` -> toxic, toxicity, toxicities
* `pregna*` -> pregnant, pregnancy, pregnancies

### I am using "-" in my search terms, and got more results than expected

In certain cases the `-` is ignored to facilitate matching of commonly (non) hyphenated terms like "where by" and "where-by"

### Are there limits to results in Article Search?

Yes, a maximum of `1000` hits can be retrieved from a single query. Try reducing the date range or refining the search criteria.

## Screening Articles

### Does MLM-AI support E2B files?

Yes, once [E2B settings](/configuration/settings#e2b-settings) are configured for your system, users [can produce E2B XML files in R2 format](/application/article-screening#e2b) for articles that qualify for ICSR.

### Does MLM-AI support quality control (QC) workflows?

Quality control with [various levels of sampling](/application/review-details#tab-filter-sampling-and-sorting-choices-follow-user-during-screening) is supported. Users can also [configure their workflows](/configuration/custom-workflows) to clearly indicate QC status.&#x20;

Finally, if required users can directly assign articles for QC using the [assignment feature](/application/review-details/work-assignment).&#x20;

### How does sampling of articles work?

**Sampling Criteria**

* Sampling will return at least one article, irrespective of sampling rate
* Otherwise, sampling will round up or down to the nearest integer value, according to sampling rate

**Sampling and Reviews**

* In each review, articles are randomly shuffled into a list using a pre-defined method, and sampling draws articles from the shuffled list
  * This ensures sampling produces consistent results, allowing subsequent QC sessions to retrieve the same articles
* In Find Articles, sampling applies to each review separately (following the sampling approach above), and are then aggregated and retrieved to the user

## Data Sources and Searching Articles

### What literature sources are present in MLM-AI?

MLM-AI ingests data from established and open access databases to achieve a broad literature reach. Currently, we automatically scan and upload the following sources:

* [**PubMed** ](https://pubmed.ncbi.nlm.nih.gov/)- Comprehensive repository of biomedical and life sciences literature
* [**Crossref** ](https://www.crossref.org/services/)- Worldwide open access repository of scientific literature
* The [**Directory of Open Access Journals**](https://doaj.org/) (DOAJ) - Worldwide Open Access literature
* [**SciELO** ](https://www.scielo.br/)- Scientific life sciences literature focusing on Latin America and the Caribbean. The following Scielo repositories are ingested:
  * Scielo Brazil
  * Scielo Spain (from 01-JAN-2025)
* [**DialNet** ](https://dialnet.unirioja.es/)- open access repository focused on literature from Spain
* [**IBECS and LILACS**](https://red.bvsalud.org/en/vhl-search-portal-collection/) - open access repositories with regional coverage in Spain, Latin America and the Caribbean
  * **Note**: As of May/2024 the IBECS source is no longer being updated with new publications at the source
  * Data start date for LILACS is Nov/2023

### What is the search strategy employed by MLM-AI?

There are two options:

(1) When creating a monitor using product name and synonyms, MLM-AI will match any of the key terms specified in the [monitor configuration](/configuration/monitor-configuration).

In the example below, the key terms associated with "etanercept" are **OR**'ed together, and searched across all sources specified in the monitor.

<figure><img src="/files/4PnlJzYuqV4M7vtdJHIr" alt=""><figcaption></figcaption></figure>

(2) Monitors can also be created from a search string created by the user. From [Article Search](/application/article-search), create a query string to suit your needs and proceed to create a monitor using the "pill" button. The monitor query string is exactly the one created by the user.

{% embed url="<https://youtu.be/B0WYiHC30Hk>" %}

### Should I include additional sources available in the Biologit Database in my searches?

Including more repositories will broaden the reach of scientific literature available to you, and can contribute in identifying more relevant safety information.&#x20;

Additional databases can also help in meeting your regulatory requirements.

See also these resources for further information on this topic:

{% hint style="info" %}
[Case study using DOAJ and Crossref for safety surveillance](https://www.biologit.com/post/broader-searches-for-adverse-event-articles-a-case-study-with-doaj-and-crossref)
{% endhint %}

{% hint style="info" %}
[GVP-VI Guidelines and biologit MLM-AI](https://www.biologit.com/post/gvp-vi-guidelines-and-biologit-mlm-ai)
{% endhint %}

## AI Tags

### Will MLM-AI detect Suspected adverse events abstracts for my product?

MLM-AI models are product agnostic, having been trained with a cross-section of scientific literature covering a broad range of abstracts.

There are various strategies that can be employed for the safe implementation of AI-based screening. In [AI-Enabled Screening Workflows](/topics/ai-enabled-screening-workflows) we discuss techniques to leverage AI into your workflows.

### How accurate are MLM-AI models?

All MLM-AI models are configured for **high recall** on the target category of interest (Suspected AE, Suspected Case etc). This means AI tags produced by MLM-AI are "conservative" in that it tolerates some amount of false positives so that missing a relevant abstract is highly unlikely.&#x20;

There are many ways in which users can take advantage of AI Tags produced by MLM-AI. Please see the section on [AI-Enabled Screening Workflows](/topics/ai-enabled-screening-workflows) to learn more and use one that best suits your needs.

For performance results please refer to the latest fact sheet:

* <https://www.biologit.com/post/towards-ai-transparency-with-model-factsheets>

### Are AI-based tags safe to use? How can I take advantage of them?

The best resource on how to safely integrate AI tags is our [Guide to AI-Enabled Screening Workflows](/topics/ai-enabled-screening-workflows).

### What is the Behavior of Tags in Articles where only Title is Available?

Title-only citations may sometimes appear, depending on how journals decide to publish their data.&#x20;

* Without an abstract there is insufficient information to reliably make model predictions
* Hence, all "No Abstract" articles are tagged as "Suspected AE", indicating they should be screened&#x20;
* "No Abstract" citations are also presented in [a separate "No Abstarct" tab](/application/review-details#tabs) in Review results, to facilitate a separate screening workflow

### Why are MLM-AI Tags Called "Suspected" adverse event, and other "Suspected" categories?

MLM-AI models were designed so that relevant safety information is not missed. With this in mind, we have labeled and curated datasets that capture how likely safety information is present either in the abstract or in the full text of an article.

Consider the samples below extracted from abstracts. The most ambiguous cases may still be relevant, and will likely need review of the full text of the article:

![](/files/-MX1GKiMkmcIbdwB0luY)

By looking for "suspected" safety information, MLM-AI can still tag abstracts with incomplete information without missing important safety data.

### Where can I learn more about the AI models powering AI Tags?

See our article discussing the the use of AI for literature screening:

* <https://www.biologit.com/post/fast-mlm-with-ai>

See also the model technical specification and intended uses in the FactSheet:

* <https://www.biologit.com/post/towards-ai-transparency-with-model-factsheets>

Finally, our technical paper goes in more details on our AI development methodology and experimental results:

* [Validation and Transparency in AI systems for pharmacovigilance: a case study applied to the medical literature monitoring of adverse events](https://arxiv.org/ftp/arxiv/papers/2201/2201.00692.pdf)

## Monitor Configuration

### How does the synonyms expansion work?

Medication synonyms presented in the [Monitor Configuration](#monitor-configuration) pages are periodically (at least quarterly) collected from the following public sources: [MeSH](https://www.nlm.nih.gov/databases/download/mesh.html), [ChemID](https://www.nlm.nih.gov/databases/download/chemidplus.html), [OpenFDA](https://open.fda.gov/data/downloads/) and the [EMA list of marketed products](https://www.ema.europa.eu/en/human-regulatory/post-authorisation/data-medicines-iso-idmp-standards/public-data-article-57-database).

### I made changes to a monitor configuration. When will I see them in Review results?

Any modification to monitors apply *only* to Reviews created *after* the change.

Existing reviews always use the monitor configuration present *at the time of review creation*. This ensures consistent behavior with the then current settings. This also applies to "Scheduled" reviews that have been submitted and are already in the "Processing" tab.

## System Limits

Some system limits may be adjusted for your environment. Please [contact support](/more-help/support-and-platform-details#customer-support) with your requirements.

#### Monitors

* Maximum number of synonyms used in a monitor: `1800`
* Maximum configured monitors per account: `100`
* Maximum configured special situation options: `50`
* Maximum configured exclusion options: `50`

**Workflow**

* Maximum configurable workflow decisions (system settings or monitor): `50`

#### User Management

* Maximum number of teams assigned to a single user: `10`
* Maximum configured teams per account: `50`

#### Review Date Range

* Maximum number of days in a review: `95`
* Oldest start date of a review (number of days from today): `450`

#### Data Starting Point

* Date where data starts in the Biologit Database: `01-JAN-2020`&#x20;
  * Note: exceptions apply. [See also FAQ on available sources](/more-help/faq#data-sources-and-searching-abstracts)
* Note: this limit does not apply to [uploaded reviews](/application/upload-data).

**Attachments**

* Maximum size of uploaded attachment: `50Mb`

#### Article Search

* Maximum results in a request: `1000`&#x20;
* Query string limit (chars): `30000`

## Security

### What is the default password policy?

By default every user is subject to the following password rules:

* Minimum password length: `10` characters
* At least `2` numbers in password
* At least `1` lowercase letter
* At least `1` uppercase letter
* At least `1` special character (`$&?`, etc...)

### Can the default password policy change?

Yes, [contact support](/more-help/support-and-platform-details) to request different password rules for your account.

### How to configure password auto-lock after too many failed logins?

Password auto-lock prevents account break-in attempts by temporarily locking the account after too many failed logins. This is a system configuration that administrators can setup via the Settings page. Learn more more on [Security Settings](/configuration/settings#security).