An algorithm for processing vital sign monitoring data to remotely identify operating room occupancy in real-time.

We developed an algorithm for processing networked vital signs (VS) to remotely identify in real-time when a patient enters and leaves a given operating room (OR). The algorithm addresses two types of mismatches between OR occupancy and VS: a patient is in the OR but no VS are available (e.g., patient is being hooked up), and no patient is in the OR but artifactual VS are present (e.g., because of staff handling of sensors). The algorithm was developed with data from 7 consecutive days (122 cases) in a 6 OR trauma center. The algorithm was then tested on data from another 7 consecutive days (98 cases), against patient in- and out-times captured by OR surveillance videos. When pulse oximetry, electrocardiogram, and temperature readings were used, OR occupancy was correctly identified 96% (95% confidence interval [CI] 95%-97%) and OR vacancy >99% of the time. Identified patient in- and out-times were accurate within 4.9 min (CI 4.2-5.7) and 2.8 min (CI 2.3-3.5), respectively, and were not different in accuracy from times reported by staff on OR records. The algorithm's usefulness was demonstrated partly by its continued operational use. We conclude that VS can be processed to accurately report OR occupancy in real-time.