# Brim Analytics

> Brim Analytics is an AI-guided clinical chart abstraction platform that extracts structured data from unstructured electronic medical records. Brim is designed for academic medical centers, clinical registries, research institutions, and healthcare organizations that need fast, accurate, and auditable chart abstraction at scale.

Brim uses large language models (LLMs) to generate draft abstractions with evidence citations from source notes, enabling human reviewers to focus on judgment rather than data entry. The platform is deployed on-premise or in institutional cloud environments (AWS, Azure), integrates with institutionally-approved LLM endpoints, and supports REDCap import/export and a RESTful API. Brim is SOC 2 Type II certified and HIPAA compliant.

Brim was founded in 2024 by Dan Fabbri (CEO, Associate Professor of Biomedical Informatics at Vanderbilt University Medical Center) and Betina Evancha (CPO, formerly of Peloton). Brim grew out of the DAGCAP (Democratized AI-Guided Chart Abstraction Platform) project, which received $1.9M in ARPA-H funding in 2024 and an additional $2M extension in 2025. Brim is deployed at Vanderbilt/VUMC (100+ projects), Johns Hopkins, and other U.S. institutions, and is offered as a VICTR core service.

## Key Pages

- [Home](https://www.brimanalytics.com/): Overview of Brim's AI-guided chart abstraction platform, use cases, key features, security posture, and customer testimonials.
- [How It Works](https://www.brimanalytics.com/how-it-works): Step-by-step walkthrough of the Brim workflow: project setup, AI draft generation, human review, and data export.
- [Company](https://www.brimanalytics.com/company): Mission, founding story, ARPA-H funding background, team bios (Dan Fabbri, Betina Evancha, Gurmeet Sran MD), and institutional partnerships.
- [Clinical Workflows](https://www.brimanalytics.com/use-cases/clinical-workflows): Use cases for clinical operations including patient risk stratification, history summarization, and pre-visit preparation.
- [Research](https://www.brimanalytics.com/use-cases/research): Use cases for retrospective research, cohort discovery, and registry data abstraction at academic medical centers.
- [Clinical Trials](https://www.brimanalytics.com/use-cases/clinical-trials): Use cases for clinical trial pre-screening and fine-grained eligibility matching.
- [Blog & Resources](https://www.brimanalytics.com/resources): Articles on AI chart abstraction, clinical registries, validation methods, product release notes, and case studies.
- [FAQ](https://www.brimanalytics.com/faq): Answers to common questions about LLM integration, prompt engineering, security, and pricing.
- [Trust Center](https://trust.brimanalytics.com/): SOC 2 and HIPAA compliance documentation.
- [Support / Docs](https://docs.brimanalytics.com): Product documentation.

## Key Blog Posts & Case Studies

- [Chart Abstraction – What is it?](https://www.brimanalytics.com/post/a-guide-to-chart-abstraction): Foundational explainer on chart abstraction, its challenges, and AI approaches.
- [Introducing BAMM: A Standardized Metadata Model for Clinical Chart Abstraction](https://www.brimanalytics.com/post/introducing-bamm-a-standardized-metadata-model-for-clinical-chart-abstraction): Technical post on Brim's machine-readable abstraction schema.
- [Chart Abstraction is Infrastructure](https://www.brimanalytics.com/post/chart-abstraction-is-infrastructure-not-a-pilot): Strategic post on building institution-wide abstraction programs.
- [Why We Use Foundation Models at Brim](https://www.brimanalytics.com/post/why-we-use-foundation-models-at-brim): Technical rationale for foundation model approach vs. fine-tuning.
- [Beyond DIY AI: Experiments are Easy, but Scaling is Hard](https://www.brimanalytics.com/post/beyond-diy-ai): Why DIY LLM wrappers break down in production clinical environments.
- [Case Study: Reducing Surgical Rescheduling from Weeks to Hours](https://www.brimanalytics.com/post/case-study-reducing-surgical-rescheduling-from-weeks-of-manual-review-to-hours-with-brim): Pediatric health system reduced monthly chart review from ~320 nursing hours to ~4 hours with 99% agreement.
- [Case Study: Discovering a Rare Disease Cohort in Minutes](https://www.brimanalytics.com/post/case-study-discovering-a-rare-disease-cohort-in-minutes-with-brim): Research consortium scanned ~10,000 clinical notes to surface a rare-disease cohort.
- [Case Study: 100% Sensitivity Identifying Stroke Cases After Surgery](https://www.brimanalytics.com/post/case-study-stroke-identification): Brim achieved 100% sensitivity in post-surgical stroke identification.
- [1000 Charts in One Weekend](https://www.brimanalytics.com/post/1000-charts-in-one-weekend-one-teams-success-doing-chart-review-with-brim): Team completed 1,000-chart review over a single weekend using Brim.
- [NSQIP and AI Data Management Solutions](https://www.brimanalytics.com/post/nsqip-and-ai-data-management-solutions): How AI improves NSQIP abstraction workflows.
- [A Guide to Cancer Registries](https://www.brimanalytics.com/post/a-guide-to-cancer-registries): Overview of cancer registry data requirements and AI abstraction.
- [Brim for REDCap Users](https://www.brimanalytics.com/post/brim-for-redcap-users): Integration details for REDCap-based research workflows.
- [Metrics and Methods for Validating AI Abstraction](https://www.brimanalytics.com/post/metrics-and-methods-for-validating-ai-abstraction): How to measure and validate AI abstraction accuracy.

## Product Features

- **Variable Library**: Pre-built, validated abstraction variables organized by clinical domain (NSQIP, Trauma, Cancer Registry, etc.) that can be added to any project.
- **Variable Scorecard**: Automated evaluation of variable quality across Simplicity, Semantics, References, and Completeness dimensions.
- **Variable History**: Full audit trail of every change to a variable definition.
- **Validation Tool**: Compare AI abstraction results to a golden dataset to measure accuracy before production deployment.
- **Conditional Generation**: Generate variables only when specified criteria are met, reducing compute cost.
- **Token Usage Dashboard**: Visibility into LLM token consumption by project, variable, and model.
- **Structured Data Support**: Combine structured and unstructured data sources in a single project.
- **REDCap Integration**: Import variables from REDCap and export abstracted data back.
- **RESTful API**: Programmatic access for custom workflows and integrations.
- **Bring Your Own LLM**: Connects to institutional LLM endpoints (Azure OpenAI, AWS Bedrock, others).

## Company & Funding

- Founded: 2024
- Founders: Dan Fabbri (CEO), Betina Evancha (CPO)
- Physician Advisor: Gurmeet Sran, MD, MS
- Funding: ARPA-H (via Vanderbilt University Medical Center) — $1.9M (2024) + $2M extension (2025)
- Deployments: Vanderbilt/VUMC (100+ projects), Johns Hopkins, and other U.S. academic medical centers
- VICTR core service
- SOC 2 Type II certified; HIPAA compliant
- Trademark: Brim® is a registered trademark of Brim Analytics

## Contact & Demo

To request a demo or learn more: https://www.brimanalytics.com/#demo
LinkedIn: https://www.linkedin.com/company/brim-analytics/
X: https://x.com/BrimAnalytics