Scaling medical content review at Flo Health using Amazon Bedrock (Part 1)

This article describes how Flo Health partnered with AWS to develop MACROS, an AI-powered solution using Amazon Bedrock to automate medical content review and ensure accuracy at scale.

• MACROS processes large volumes of medical articles to identify inaccuracies and propose updates based on latest research
• Solution uses Amazon Bedrock foundation models for content review, revision, and rule extraction from medical guidelines
• Architecture leverages ECS, Lambda, Step Functions, S3, Textract, and CloudWatch for end-to-end workflow orchestration
• Content review pipeline includes filtering, chunking, review, revision, and post-processing stages
• Rule Optimizer extracts actionable guidelines from unstructured medical documents in style/tonality or medical modes
• PoC achieved 80% accuracy, 90%+ recall, and 10x faster processing than manual review
• Human-AI collaboration model maintains medical expert oversight while significantly reducing review workload

The solution demonstrates how generative AI can scale medical content verification while maintaining clinical accuracy through hybrid human-AI workflows.