Measuring Radiologist Workload After AI Triage in Breast Cancer Screening

Artificial intelligence triage can sharply reduce the number of human readings in breast cancer screening. Reading count, however, does not show how much radiologist time is saved or whether work shifts to arbitration, consensus, or other tasks. We propose a session-level reporting standard based on total active human interpretation minutes per 1000 women screened, work shifted elsewhere in the pathway, and diagnostic outcomes, together with a matched-session design to measure these effects directly.