University of Texas Medical Branch screens 10x more patients with AI-powered Retool apps
Table of contents
Impact
- 66% reduction in patient testing time (from 15 minutes to 5 minutes per case)
- 10x more patients screened (from 50 to 500 cases per week)
- 5 week dev cycle for a custom AI app
“We built an entire app in five weeks without needing other resources. That’s unheard of.” Dr. Peter McCaffrey, Director of Pathology Informatics and Chief AI Officer, UTMB
Executive Summary
The University of Texas Medical Branch (UTMB), a major academic health science center, needed to scale its diagnostic testing capabilities to meet growing patient demand. Using Retool, Dr. Peter McCaffrey and his team built AutoTox, an AI-powered application↗ that dramatically improved the efficiency and accuracy of patient toxicology tests. These tests analyze a patient's biological samples to detect the presence and levels of drugs, alcohol, or other potentially harmful substances. This solution allowed UTMB to review ten times more patient cases each week, significantly expanding its ability to provide quality care to a diverse patient population.
Company Overview
- Size: Enterprise large medical institution
- Industry: Healthcare
- Region: Texas, USA
- Retool products: Retool Apps, Retool AI, Retool Workflows
- Use case: Automated medical testing
- Impact: AI innovation, business efficiency
- Data sources: Medical records system, Postgres database, OpenAI API
Challenge
A manual and time-consuming patient testing process meant UTMB’s pathology team could only review 50 toxicology cases each week. Every case required a pathologist to spend 15 minutes reviewing more than 40 metrics alongside patient data and medications. This bottleneck prevented UTMB from meeting the increasing demand for diagnostic services and limited its ability to effectively serve its diverse patient population.
Solution
Dr. McCaffrey and his team used Retool to build AutoTox—a custom app that integrates with UTMB's existing database, medical records system, and OpenAI’s API↗. It uses GPT models to generate simplified reports, and presents case data for pathologist review. Retool's flexibility and built-in AI integrations allowed for rapid development and iteration, resulting in a HIPAA-compliant solution they could deploy in just five weeks.
Results
AutoTox reduced patient testing time by 66% and enabled the pathology team to review 500 cases per week instead of 50—a dramatic transformation of its diagnostic capabilities. The AI-powered app made the team more efficient and improved diagnostic accuracy by catching errors they might’ve missed in manual reviews. This success has paved the way for further AI-powered applications at UTMB, demonstrating the potential for rapid, impactful innovation in healthcare using Retool.
Check out the full story on our blog↗
Watch a demo↗ to see UTMB’s app.