Open-Source Synthetic Hospital Dataset Fooled Doctors

SparkCPark released Synthetic Hospital, an open-source, fully synthetic longitudinal EHR benchmark with 1,268 patients and 5,602 visits containing zero PHI. The validated data is realistic enough that doctors cannot distinguish synthetic from real patients, making it ideal for RL training.

2026-09-25 ~ 2026-09-26 · 2 related posts

• Synthetic Hospital: Open Fully Synthetic EHR Benchmark With 1,268 Patients and Verifiable Ground Truth — Tim_Dettmers · 2026-09-25
• All-synthetic hospital simulation: doctors can't tell real from fake patients, built for RL — AccBalanced · 2026-09-26