Low-cost future for sensor alarms in falls prevention
The University of Adelaide is developing a low-cost movement sensor alarm system to aid falls prevention in the acute care setting with plans to trial the technology in residential aged care.
Using radio-frequency identification (RFID) technology, the researchers from the university's Auto-ID Lab in association with the Adelaide Geriatrics Training and Research with Aged Care (G-TRAC) Centre are developing a system using very low-cost, wearable sensors allied to a customisable alarm system based on individual falls risk, along with a more intuitive, discreet alarm system that can be integrated with common nurse call systems.
Auto-ID Lab director Damith Ranasinghe told the Health Informatics Conference (HIC 2013) last week that there are a number of movement sensors currently being used to study and prevent falls, particularly in hospitals, but many of them are only able to detect when a person gets out of bed and many are too expensive for widespread use.
Dr Ranasinghe said there were two different methods commonly used: “instrument the person” or “instrument the environment”. The latter involves placing movement sensors near hospital beds, video cameras for monitoring or using pressure mats attached to the actual bed that are able to alert nurses that a person has got out. These methods often have drawbacks such as false alarms, privacy concerns, high cost and susceptibility to damage.
“Instrumenting the person” using wearable sensors is an area of real interest for researchers. The Irish company Shimmer, for example, manufactures and markets a range of wearable credit card-sized sensors both for research and for commercial use that are based on wireless technology and are particularly good for biofeedback and athletic training, but also for healthcare applications.
Shimmer sensors can be rather expensive, however, with the company's kinematic sensor platform costing upwards of 200 euros. What the Auto-ID Lab is working on is very low-cost technology using the Wireless Identification and Sensing Platform (WISP), developed by Intel Research and the University of Washington in the US.
The sensors are designed to be very low-cost – between $3 and $4, Dr Ranasinghe said – and are maintenance free as they don't use batteries. As they are so cheap, they are disposable, important in a hospital or residential aged care for hygiene and infection control.
Wireless sensors require an antenna, which in a hospital setting can be placed unobtrusively on the ceiling, but Dr Ranasinghe's team is also investigating the use of conductive fabrics made from silver thread and plastics, These can be sewn into clothing, can produce a barrier between the antenna and the body and can also be laundered. The team is also investigating snap-on tags that can be sterilised and reused.
Of even more practical application is the ability to use the sensors to produce a customised alarm system based on the individual, following a falls risk assessment that most elderly people are given when entering hospital.
“With most of these [alternative] devices, you can't really customise them based on personal needs,” Dr Ranasinghe said. “But we can individualise alarms so we can say if Mrs Jones isn't supposed to get off the bed but she can get off a chair, we can customise the alarm for that person.
“With our technology we can uniquely identify each patient with RFID. We now have the ability to customise alarms individually based on falls risk.”
Dr Ranasinghe said most alarm systems concentrated on monitoring bed exits, which is where most falls in hospitals occur. However, the technology his team is developing also aims to monitor people when they are sitting in or rising from a chair, walking around the room, leaving the room or moving without the use of their walking aids.
Dr Ranasinghe is currently working on a project with the university's Associate Professor Michael Sheng, a specialist in developing context-aware computer algorithms for inference engines, which can interpret what a person is doing automatically.
“Traditionally, if you look at algorithms that are in practice in terms of deciding whether a person is on or off the bed, they are almost all pressure mat-based,” Dr Ranasinghe said. “If the pressure is below [a certain level] or above it, you are in or out of bed. We have a different approach … we train the computer to recognise patterns and the computer can learn what the person is doing based on these patterns.”What the team is now developing is a monitoring application that will allow hospitals to customise an alarm system for that individual, which can send a discreet alert telling a nurse that a patient is active, and when and where they are doing it.
“What we are proposing is to send an alert signal to a pager that is carried by nurses – pager systems that are already in use in hospitals, so we don't have to add anything on and we can use the existing infrastructure – and now it will be who, where, what and when,” he said.
“No longer will it simply be a buzz. We can send a discreet message that is more directed to clinicians. Who is the patient? Where? Room 2. What is she trying to do? Get off the bed. And at what time.”
By using RFID sensors attached to the nurse's name badge, alarms can also be automatically turned off. Pressure mat systems often require an alarm to be turned off manually, but if a nurse wearing an RFID tag is in the immediate vicinity of the patient, the alarm will automatically turn off, he said.
“That will reduce alarm frustrations and false alarms. It will also reduce disturbance and agitation because the alarm is not a buzz but a discreet message.
“An alternative approach is, if privacy can be worked out, we could install a video camera that can only be used when there is an alarm going off. So if a nurse knew that someone was trying to get off the bed, they could turn on the video link for just that moment to check if the person is all right. These are all alternative things that we can do.”
The team is also developing a hand-held falls risk management tool based on the commonly used traffic light system, in which high-risk patients are identified by using falls risk alert cards above their beds. By automating this process, the patient's falls risk information is then fed back into the system.
The Auto-ID Lab has been working with Adelaide's Queen Elizabeth Hospital and the Hospital Research Foundation on a number of trials of the new technology, and hopes to begin trialling it with residential aged care providers as well. Dr Ranasinghe said there was also great potential for the technology in keeping older people in their own homes for longer.
This research is only part of a wider effort into researching what has been dubbed the “Internet of Things”, or the ability to connect objects in the physical world to digital information on the internet. The Adelaide Auto-ID Lab is one of several labs worldwide leading research efforts this research. It also works closely with GS1 to help it develop global standards for RFID traceability networks.
Posted in Aged Care