Component-based app design for platform-oriented devices in a medical device coordination framework

Abstract

Frameworks that promote the intelligent coordination of medical devices are required in ubiquitous health care environments. Such environments are essential to achieve closed-loop behavior, and an emphasis on interoperability standards and reconfigurable hardware/software components will allow these systems to adapt to changing patient/provider needs and technologies. Intelligent medical system design often focuses on smart, complex on-device algorithms. This paper addresses the use of a Medical Device Coordination Framework from the complementary perspective, where devices stream data to an information infrastructure populated with rearrangeable components that can be cascaded to form complex processing apps, or "transformers." This approach enables a new paradigm for the design of medical devices that we refer to as "medical platform-oriented devices" (MPODs): devices constructed from sensors/actuators integrated with a medical acquisition and analysis platform that hosts sophisticated software component collections while providing a variety of safety and real-time features relevant to the medical device domain. These components provide capabilities that may have been unsuitable for direct microprocessor implementation, promote the use of parallel analyses (e.g., for parameter verification), and enable the construction of reconfigurable apps. To demonstrate this approach, a set of reconfigurable, component-based software apps was developed to address medical devices that utilize photoplethysmograms (PPGs). The first app builds upon the behavior of an existing wireless reflectance pulse oximeter that (a) processes raw PPGs to ascertain heart rate, blood oxygen saturation, and respiration rate and (b) makes decisions regarding PPG viability. Other apps address the calculation of pulse wave velocity using multiple PPGs and closed-loop control of PPG quality when determining blood pressure and stroke volume. These scenarios demonstrate the immense potential that a component pool and its related app library provide for biomedical signal analyses and the creation of reconfigurable virtual medical devices.