Investigating a novel technological solution to prevent falls in older people in hospital

AmbIGeM intervention involves patients wearing a Bluetooth Low Energy (BLE) device with integrated sensors. The device positioned in a singlet pocket transmits movement signals, to a base station. Signals are interpreted by software that identifies risk circumstances and responds ‘intelligently’ (tailored response) to patient movements that lead to situations of increased falls risk. When a risk circumstance is identified by the system, clinical staff will be alerted via a hand held mobile device (vibration and/or alarm). Staff may then intervene and supervise the patient. 1. Wearable device and singlet. The device used in the trial has a number of inertial sensors (e.g. tri-axial accelerometer) and sensors for measuring elevation above ground. It is low cost and powered by a replaceable coin sized cell battery (typical battery life is 30 days). The device will sit in a plastic encasing to protect the device, improve comfort, and ensure the sensor is correctly orientated. The encasing has a black side and white side allowing nurses to easily determine the correct placement. A nurse will place the device within an ‘envelope type’ pocket on the inner side of a singlet worn by the patient under their hospital/own clothing, at the sternum position. Singlets will be changed as needed. The wearable device will be used one per patient, then disposed of. Lost or misplaced sensors will be replaced. A further device in an encasing will be attached to patient walking aids, for patients who are classified as not requiring supervision for transfers out of bed and chair but being risk takers who are unsafe in mobilizing without their walking aid during the admission. Wearable devices will capture patient physical orientation and movement information and wirelessly transmit this data. Different movements between the wearable device and the device attached to walking aids will allow detection that the walking aid is being used (or not) by the patient. 2. The modified hospital room environments and monitoring systems. To capture the data from wearable devices worn by patients and devices on walking aids, small form factor BLE enabled single board computer based listening devices (base stations) are installed on ceiling locations above each bed and door exit. This will provide adequate coverage for various configurations of one, two and four bed rooms. Base stations will collect and pre-process data from wearable devices worn by patients and forward this data for analysis by the backend AmbIGeM software to identify risky movements. Each base station will communicate with backend AmbIGeM systems over a Local Area Network. Activity recognition algorithms in the backend systems will process and analyse the data from all the devices forwarded from the base stations and in the context of the personalized information entered on the system by the staff, determine whether or not an alarm should be triggered, in real-time. The backend AmbIGeM systems will alert clinical staff when an assessed patient is undertaking high falls risk activities. 3. The AmbIGeM system. An electronic interface will run on a mobile device. The AmbIGeM Mobile App will be executed on an Android smart phone. Using this App, nurses, at the beginning of every shift, will select patients allocated to their care, determine the movement circumstances of individual patients where there is a risk of falling and the App will record this information onto the AmbIGeM system where the information collected will be used: i) to generate and print an individualized falls risk poster for display by the bedside, which will also act as a visual aid for falls prevention; and ii) to activate an alarm when a movement pattern indicating the identified risk movement occurs. Clinical staff, such as nurses and physiotherapists, will carry the mobile device and be alerted if a patient is undertaking falls risk related movements. Alert notifications from the AmbIGeM server monitoring patient activities will be sent over the existing Wi-Fi network and be received by the Mobile App. Staff carrying the Mobile App will then attend the patient with the aim to mitigate the risk of falling through timely supervision. The alert notification received will include i) identity of the patient (who); ii) physical location of the patient (where); iii) type of high-risk activity (what); and iv) timestamp of when the high-risk activity was detected (when). Risky movements that may be set to activate the alarm include: Sitting up from lying on the bed; Standing up from bed or chair; Walking (can be limited to a area); and Walking without a required gait aid (ie if a patient is identified as safe to walk with their frame, but is considered not safe without their frame, starts walking without the frame the system can alarm, through the gait aid being tagged). Staff will be able to deactivate the alarm from the same Mobile App. Staff will record if an incident was a false or true alarm and what they noted when they attended the patient. During the intervention period, the AmbIGeM system will replace other sensor alarms for all participants. A dedicated desktop falls management application (AmbIGeM Desktop App) at the nurses station will provide the capability to: i) enrol patients in the trial by assigning a wearable device to a patient; ii) visualize real-time updates of patient activity and current alarm information as well as a log of past alarms for individual patients; ii) discharge patients or un-enrol patients as required; iii) provide a facility for nurse managers to alter alert settings including sensitivity of alerts; iv) visually observe when falls risk related movements need updating based on user defined expiry times; and v) provide warnings to replace the battery when a patient stay lasts longer than typical battery life. Strategies to maximise fidelity include staff in-service (fall definition and staff reporting of falls), and staff training/protocols for specific key project activities (eg singlet and Mobile App use). Research staff will check the AmbIGeM Desktop App to ensure eligible patients are wearing the sensor. The system will detect when there is no sensor movement for patients enrolled in the study, potentially indicating failure in the system (eg sensor, battery, reader, computer) so that research staff can check and intervene. If the Mobile App risk movements are not updated within 24 hours, nurses will receive a reminder. Ward 1 intervention commences after block 1 (control phase) for 3 x 25 week blocks (75 week intervention); Ward 2 intervention commences after block 2 for 2 x 25 week blocks (50 week intervention); and Ward 3 intervention commences after block 3 for 1 x 25 week blocks (25 week intervention).