Research has revealed that telecare, monitoring a patient’s health in their own home, can be used safely with active patients in place of the standard clinic visit
[Whiteman, 2012]. The emergence of telecare adds a new paradigm in health-
care, where the patient is monitored between physician office visits. This has been shown to significantly reduce hospitalizations, while improving the patient’s
quality of life [Breen, 2011]. Telecare can also benefit patients where traditional
delivery of health services is affected by distance and lack of local specialist clin- icians to deliver services. Generally, there are two different telecare approaches:
• Real-time model: a telecommunications link allows instantaneous interac- tion between patients and caregivers. Video-conferencing equipment is one of the most widespread forms of real-time telecare. With the availability of better and cheaper communication channels, direct two-way audio and video streaming between clients is leading to lower costs.
• Store-and-review model: digital images, video, audio, observations of daily living and healthcare data are captured and stored on the server; then at a convenient time they are retrieved by caregivers at another location where they are studied and reviewed.
The CARA system provides remote at-home monitoring services that sup- port both of the telecare models. It allows the incremental use of the system and thereby encourages the adoption of the technology. Two use case scenarios con- cerning different telecare models for remote monitoring are presented and their implementations are discussed in this section.
7.3.1
Real-time Interactive Remote Monitoring
The first scenario involves real-time at-home monitoring under remote supervision by a caregiver. Real-time sensor data are collected and transferred to a remote caregiver, who might be any suitable healthcare worker, specialist or medical consultant. The monitoring session also involves the use of a two-way video link whereby the patient and remote caregiver can communicate with each other. This
is an important aspect of making this scenario low-disturbance and non-stressful, and thereby gaining acceptance for the technology.
Specifically, the client application running on a home gateway gathers infor- mation from smart home sensors and medical equipment through wireless con- nections. The sensor data are then transmitted to the remote server in real-time
through the Action Message Format (AMF) protocol [Adobe, 2011]. The care-
giver can login to the system and select the patient who is eligible to be monitored. A video-conferencing component is integrated with real-time remote monitoring,
Figure 7.5: Real-time interactive remote monitoring
which can aid in an interactive examination and allow reassurance for patients who are new to the system. In this case, continuous monitoring of the patient is carried out by the remote caregiver in real-time with a live video communication
channel established as shown in Figure 7.5. Real-time sensor data are published
along with the video stream through the Flash Media Server (FMS) to the re- mote client of a caregiver. (The FMS is a proprietary data and media server from Adobe Systems). The live video is captured and encoded by Flash Media Live Encoder, and streamed to FMS. The video streams are then broadcast to the remote client through the internet. Moreover, a social utility is also implemented in the application which allows the user and caregiver to communicate with each
other by sending messages. It can be used during or outside of the monitoring session.
7.3.2
Automated Remote At-home Monitoring
The second scenario, following easily from the first, involves the most innovative use of the CARA system that is a later stage of the incremental incorporation of pervasive healthcare technology into medical practice. This scenario is where the system is analyzing the real-time sensor data, reasoning with all the con- texts to identify critical patient conditions and forwarding healthcare data to the healthcare server for inspection by caregivers. This makes use of the reasoning technology described in the previous chapter.
The fully automated intelligence of the system is achieved by incorporating real-time remote monitoring with the activity recognition application and the reasoning framework. In this scenario, the patient is monitored in a smart home environment without any supervision while wearing BAN continuously. Informa- tion around the patient is gathered and analyzed to detect any possible anomalies
(as demonstrated in Figure 7.6).
The contextualization involves the processing of raw data coming from smart home sensors and wearable wireless sensors, producing higher level information. While the activity recognition application keeps tracking the movement of the patient and providing activity contexts to the system. Note that, dynamic activ- ities (e.g. walking stairs, washing hands, sweeping) can be identified directly by the trained classification models, while other static activities (e.g. watching TV, sleeping, toileting) can be inferred from the posture of the body and relative am- bient context. The contexts can be used by the reasoning system to identify the current state of the patient using the domain knowledge (in terms of fuzzy sets and rules) and previous experience (in terms of case bases). The reasoning engine executes in real-time and can offer immediate notification of critical conditions. The vital signs of the patient and all the available contexts are transmitted along with the reasoning output to the healthcare server, where they are stored and can be examined or reviewed by a remote caregiver at the proper time.
Figure 7.6: Automated remote monitoring of the patient without supervision