Performance, Scalability and Maintenance: An Experimental Study of RPA vs. Autonomous Agents in the Healthcare Domain
Summary
- Researchers compared a deterministic Robotic Process Automation (RPA) solution using Playwright with an autonomous Agent Process Automation (APA) solution using browser-use, LangGraph, and Gemini 2.5 Flash-lite in the healthcare domain.
- The evaluation focused on retrieving mammography and breast ultrasound reports from a legacy laboratory portal across two rounds to assess performance, scalability, and maintenance.
- RPA achieved 98.90% accuracy and a 98.27% F1-score, while the APA solution reached 96.49% accuracy and a 94.48% F1-score.
- RPA was approximately 13 times faster (16.07 seconds vs. 207.21 seconds per patient) and cost about six times less per report (USD 0.030 vs. 0.187) compared to the autonomous agent solution.
- Both approaches achieved 100% step success rates, but the APA solution showed more variance over time, indicating that RPA is superior for stable web interfaces.
Why it matters
This study provides empirical evidence comparing deterministic RPA and LLM-based autonomous agents for web data retrieval in healthcare, highlighting significant cost, speed, and accuracy trade-offs.