Brim's May 2025 Feature Release: Clarity from Complexity

At Brim Analytics, we’re building tools that make chart abstraction cheaper, faster, and more accurate—especially for teams using AI in healthcare. Our May release is another step forward in that mission, bringing greater clarity, control, and cost-efficiency to the abstraction process through smarter coordination between variables, enhanced logic, and more intuitive interfaces.

We’re constantly learning from our users and our own internal projects. One recent example: while abstracting `cancer_progression`, we noticed we were wasting time and compute abstracting it for all patients—even those who had already been determined not to have cancer. This is the kind of mistake humans wouldn’t make—but LLMs don’t know what not to abstract unless you tell them. Thanks to our new inter-variable referencing feature, we can now explicitly reference other variables (like `has_cancer`) to gate when and how we run expensive abstractions. No cancer? No need to chase down progression.

It’s a simple example, but it shows what’s possible when abstraction is modular—something we explored in depth in our recent post.

‍

Inter-Variable Referencing

‍

This powerful new feature lets you define logic in one variable based on the values of others. For example:

- Only abstract `cancer_progression` if `has_cancer` is `True`
- Set `recent_treatment` to `None` unless `treatment_history` exists
- Use `gender_identity` to filter how pronouns or anatomical terms are interpreted

This reduces compute, helps avoid AI hallucination, and improves clarity. It also helps ensure that LLMs focus their attention where it matters.

‍

The Brim API

‍

Our new Brim API allows you to incorporate Brim directly into your ongoing workflows—so you can automate chart abstraction on an ongoing basis.

You can use the API to:

- Upload incremental patient data
- Update or revise existing records
- Trigger new generations of structured output programmatically

Whether you're building a clinical workflow, refreshing a research dataset, or pre-screening for a trial, the Brim API makes it easier to embed Brim into a live pipeline.

Read more about how to use the Brim API here.

‍

Why This Matters

‍

AI in healthcare only delivers value when it produces accurate, explainable, and relevant results. These updates make Brim even more precise by reducing unnecessary abstraction and by helping researchers tell AI when to abstract an answer—not just how.

It’s all part of our broader mission: to help researchers get from raw notes to meaningful insights—faster, cheaper, and more accurately using LLMs and AI in a way that’s grounded, not guesswork.

Thanks for being part of the journey—and stay tuned for what’s next.
— The Brim Analytics Team

‍