US 12,237,089 B2
Online monitoring of clinical data drifts
Sagi Schein, Kiryat Tivon (IL); Ole Jakob Utkilen, Tel Aviv (IL); Ruth Bergman, Ceasarea (IL); Eyal Hayun, Adi (IL); Edwar Bkheet, Haifa (IL); Antonio Zaitoun, Haifa (IL); and Vitaly Burshtein, Kiryat Ata (IL)
Assigned to GE Precision Healthcare LLC, Waukesha, WI (US)
Filed by GE Precision Healthcare LLC, Milwaukee, WI (US)
Filed on Dec. 23, 2021, as Appl. No. 17/645,892.
Prior Publication US 2023/0207138 A1, Jun. 29, 2023
Int. Cl. G16H 50/70 (2018.01); G06F 40/20 (2020.01); G08B 21/18 (2006.01)
CPC G16H 50/70 (2018.01) [G06F 40/20 (2020.01); G08B 21/182 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system, comprising:
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
a reception component that:
intercepts a live data comprising medical information identifying clinical events for one or more patients prior to processing of the medical information by a clinical application, wherein the live data stream aggregates the medical information as sent to the system from a plurality of different clinical data information systems;
records reception times of the clinical events as received by the system within the live data stream; and
identifies different event types of the clinical events; and
a detection component that, for each event type of the different event types:
determines time differences between the reception times of respective ones of the clinical events corresponding to the event type;
estimates respective probabilities that the time differences are expected using a time-to-event model tailored to the event type included amongst a plurality of different time-to-event models tailored to the different event types;
detects whether a data failure event associated with the event type has occurred based on the respective probabilities relative to a threshold probability; and
controls the processing of the medical information by the clinical application based on a detection that the data failure event associated with the event type has occurred.