Claude-Powered Lab Report Processing for Personalized Nutrition
A HIPAA-compliant platform for personalized nutrition insights and guidance for food and dietary supplements for targeted wellness.
Client/Industry
A nutritional genetics company leveraged DNA insights to automate personalized dietary recommendations and wellness management.
Enhancements
- Claude-based Lab Report Processing
- Blood Biomarker Extraction
- PHI Detection and Redaction
- Personalized Supplement Formulation
- Support for Multiple Lab Report Formats
Services Delivered
- Claude API Integration
- Amazon Bedrock Integration
- Event-Driven AWS Pipeline Development
- PHI De-Identification Workflow
- Lab Report Data Extraction
- Biomarker Mapping
Tech Stack
- AWS Amazon Bedrock Claude API
- Claude Haiku 4.5
- Node.js
- React
Project Overview:
Challenges Faced by Our Client
- Inconsistent Lab Report Formats Members submitted bloodwork from many different laboratories, with different layouts, labels, units, and document structures.
- Limited Report Compatibility The existing system worked with predictable report formats, but new ones required new rules to read and extract the data correctly.
- Sensitive Health Data Protection Lab reports contained PHI, which had to be detected and removed to meet HIPAA requirements before the document could move through the AI workflow.
- Inconsistent Biomarker Data Usage Processing lab reports inconsistently made it difficult to use their biomarker data reliably in GenoPalate’s personalization model.
- Lack of Third-Party Integration Billing systems & external health tools (e.g., wearables, EHRs, or fitness apps) remained disconnected from the main system.
Solutions We Gave
Claude-Powered Biomarker Extraction
- Claude Haiku 4.5 runs through the Claude API on Amazon Bedrock and reads the lab report content.
- It extracts the clinical biomarker values GenoPalate needs from reports with different layouts, labels, and units, reducing reliance on format-specific parsers.
Structured Biomarker Mapping
- The extracted values are converted into structured biomarker data and mapped to GenoPalate’s nutrient model.
- This allows blood test results to become part of the information used to create personalized nutritional supplement formulations.
Secure Lab Report Processing
- We built a HIPAA-compliant platform with an event-driven AWS pipeline that validates uploaded lab PDFs, detects and redacts PHI, and prepares Markdown for further processing.
- This gives GenoPalate a consistent way to prepare reports from different laboratories before sending the required information to the AI workflow.
HIPAA-Compliant Subscription Flexibility
- Our HIPAA-compliant platform's try-before-you-subscribe option allows users to experience personalized nutrition insights before committing to supplement subscriptions.
- For continued guidance, a six-month subscription includes optional consultations with a resident dietitian.
Impact Produced
- 6x Increase in Lab Report AdoptionGenoPalate processed 368 lab reports in the 3 months after Claude integration, compared with 64 in the 3 months before. That is 304 more reports than the previous quarter.
- Deeper Biomarker-Based Nutrition PersonalizationSupplement formulations can use blood biomarker data alongside DNA and dietary information, giving GenoPalate a more complete picture of the member’s biology.
- Scalable Lab-to-Formulation ProcessingMembers can upload their lab report, have the required biomarkers extracted, and use those results as part of their personalized supplement formulation.