Faculty
Ul Balis, MD – course architect & pedagogy lead (informatics director/educator & data scientist)
Ron Jackups, MD, PhD – course co-architect & end-user perspective (clinical pathologist & LIS medical director)
Prerequisites
Laptop with admin rights, modern browser, ChatGPT/Bing/Copilot account (free tier acceptable), basic familiarity with spreadsheets and CLIA/CAP document structure.
High-Level Learning Objectives
-
Explain core LLM concepts (models, embeddings, RAG, prompt engineering, governance).
-
Generate lab-specific documents (policies, validation plans, letters, e-mail, CV extracts) with an audit trail suitable for CLIA/CAP inspection.
-
Build a procedure-manual lookup chatbot using Retrieval-Augmented Generation.
-
Integrate & automate LLM workflows with LIS/LIMS, scheduling, and EHR messaging.
-
Evaluate outputs for accuracy, bias, confidentiality, and regulatory compliance.
-
Leave with actionable templates, code snippets, and a personal “starter RAG” repo.
Day 1 – Deep Dive & Foundational Labs (8 hrs)
| Time |
Session |
Format |
| 08:30-08:45 |
Welcome & logistics |
Plenary |
| 08:45-10:15 |
Plenary: “LLM Essentials for Lab Medicine”
- Model families & tokens
- Prompt engineering patterns
- Regulatory & privacy guard-rails
- Where LLMs fail (hallucinations, bias)
|
Interactive lecture + live demos |
| 10:15-10:30 |
Break |
|
| 10:30-12:00 |
Hands-on 1: Prompt Engineering for Productivity
Attendees craft:
- SOP draft (template provided)
- Letter of recommendation
- CV “skills matrix”
- Inbox triage rules
|
Guided lab, Peer Share-out |
| 12:00-13:00 |
Hosted lunch & vendor fair (LLM tool sponsors) |
|
| 13:00-14:30 |
Hands-on 2: Hypothesis Generation & Literature Triaging
- ChatGPT + PubMed plug-ins
- Writing testable aims
- Summarizing 510(k) predicates
|
Small-group challenge |
| 14:30-14:45 |
Break |
|
| 14:45-16:15 |
Hands-on 3: Assay Validation Protocol Builder
- CLSI template ingestion
- Automating statistical section in Excel
- Producing a CAP-ready validation report
|
Lab + walk-through |
| 16:15-17:00 |
Panel & Wrap-Up
Q&A, day-one reflections, overnight mini-project assigned (“Bring a policy manual snippet to turn into a RAG”). |
Discussion |
| Time |
Session |
Format |
| 08:30-09:00 |
Coffee, overnight project check-in |
Informal |
| 09:00-10:30 |
Hands-on 4: Build Your Own Procedure-Manual Chatbot
- Chunking & embedding docs
- Vector store options (Chroma, FAISS)
- LangChain/OpenAI function calls
- Prompted citations for inspectors
|
Step-by-step lab |
| 10:30-10:45 |
Break |
|
| 10:45-12:15 |
Hands-on 5: Automation & Governance
- Connecting to LIS via FHIR/HL7
- Auto-drafting critical-value notifications
- Scheduling & throttling with Power Automate
- Risk register & audit logging
|
Demo + sandbox exercise |
| 12:15-13:00 |
Capstone Challenge & Debrief
Teams solve a real-world scenario (e.g., “Newborn bilirubin follow-up protocol”) using their chatbot + prompts. Peers vote; faculty give rapid feedback; course evaluation; CE certificates.
| Competition |
Day 2 – Applied RAG & Automation (4 hrs)
A full refund will be issued for cancellations received no later than 7 days prior to the start of the meeting. ASCP reserves the right to cancel a program due to circumstances beyond its control or for insufficient registration. Registrants will be notified immediately and allowed to obtain a full refund. The Society is not responsible for penalties incurred as a result of cancelled transportation. For additional assistance, contact ASCP Customer Relations at 800.267.2727 (International callers: 312.767.1553) or www.ascp.org/content/about-ascp/contact-ascp#