Measuring Patient Mobility in the ICU using a Novel Non-Invasive Sensor

Unmet Need
Patient mobilization is a critical variable that can predict future physical impairments and the ultimate length of stay for intensive care unit (ICU) patients. Measuring patient mobilization requires a clinical staff member to monitor a video feed of multiple patients in an ICU ward and make subjective distinctions on their level of mobility. This is a costly and time consuming process that distracts key clinical staff from other areas of patient care.

Technical Description
This disclosure describes the Non-Invasive Mobility Sensor (NIMS) technology that provides automated identification and classification of patient mobility using color/depth imaging and machine learning analytics.
Presently this invention uses multiple motion sensing systems to extract 3D scene information from an ICU ward. These image streams are then processed by a proprietary machine learning algorithm that contains: 1) Person Localization, 2) Patient Identification, 3) Patient Pose Classification, 4) Motion Analysis, and 5) Mobility Classification. The algorithm classifies the time-course of patient activity into mobility clusters for diagnostic review.

Status of Development
A prospective study of the NIMS system was conducted in an ICU at the Johns Hopkins Hospital. Three motion sensors were used to monitor 5 patients for 362 hours and classify their activity. The output of the NIMS system was validated by independent physicians. The NIMS system agreed with the expert opinion for 87%  (72/83) segments of data with efforts to continue to increase this percent agreement ongoing.

Publications
Ma et al., (2014) Critical Care Medicine ,42(12):A1389 [Poster Abstract]

 

Category(s):