Ultimo is an Intensive Care patient monitoring system that changes the monitor behavior based on who is in the patient room. Ultimo is developed with a vision of silent ICU, which aims to reduce stress and anxiety in intensive care. It helps to create a calm place for patients to recover, allows visitors to feel less anxious about their loved ones and improves the clinicians work environment.

ICU (Intensive Care Unit) is a highly critical environment in which patients are monitored continuously. The alarms are crucial, but they also create a noisy and stressful environment for all users (patients, visitors, and clinicians).undefined

Detection, Identification, and Trigger are the three key elements of the Ultimo. The detection system can detect if a person is entering or exiting the patient room, the identification system can differentiate between visitors and clinicians and trigger system switches the monitor between different modes. Ultimo monitor has three modes, patient mode, clinician mode, and visitor mode. When the patient is alone, the system triggers the monitor into the patient mode, in which the monitor mutes all the alarms and hides the complicated graphics with a simple standby screen. When a person is entering the room, Ultimo checks with the detection system, and if it is detected as a clinician, the monitor switches to clinician mode by unmuting all the alarms and showing all the graphical data of patient vitals. If the detection system doesn't identify the person as a clinician, Ultimo understands it as a visitor and switches to visitor mode, where all the alarms are muted again but little information of vitals is displayed on the screen.

Problem

The Intensive Care is a hospital department for patients in the critical state are put on continuous observation. Each patient has a dedicated nurse that observes them and each nurse works in shifts of 8 hours. The monitoring system notifies the nurses by sounding an alarm, blinking the light on the monitor and sending alarms to the nurses' pagers. Moreover, the patient monitor can be watched from the nurses' ward. Monitors generate very frequent alarms (39 to 352 alarms per patient, per day) and that a high proportion is false, defined as not being actionable (over 90% of pediatric intensive care unit and over 70% of adult intensive care alarms).

During the night the current monitors can be set on 'private mode'. The private mode dims the screen and hides the values and alarms(with a simple pop up window showing private mode). However, turning on this mode has to be done manually by the nurse. When nurses enter the patient room, they have to turn it off again. During the day the monitors are not on private mode and show all values and alarms. The patient still needs to recover and often sleep during the day. The monitor and its alarms disturb the patient during this time.

Visitors often get anxious by seeing the various values on the screen which they can't understand. Depending on the nurse, they can get an explanation but more often they look online for an answer. Talking to nurses shows that the monitor demands attention from the visitors. So much that often, the visitors are not focused on the patient anymore, but on the monitor and every little change it makes. They compare values from other days to current values and make conclusions. They might adjust patients positions so that the values seem to be optimal.

Solution

Silent ICU is the right solution for this problem where the alarms are muted and visuals are hidden when not necessary. To make this happen, the system needs to be intelligent enough to know who is present in the room and how is walking in and out. Various ways to detect people were thought through and tried out. The final decision, which is using Bluetooth beacons combined with people counter camera, was primarily based on two things; fit within the medical context and complexity of implementation.

Ultimo works as a system involving five different devices; camera, beacon, the sensor unit, PC and monitor screen. The sensor unit collects the patient data from sensors. The camera and Bluetooth beacons to receive the data of people going in and out. This data is processed in a PC and displayed over a monitor screen. The camera counts the number of people entering and leaving the room. The PC identifies clinicians by the beacons placed on their identity card. This data enables the Ultimo to see when a patient is alone in the room so that the monitor can be set to the silent patient mode. Furthermore, due to the beacons, the system is able to differentiate between visitors and clinicians. The PC being the heart of the system, the other devices work as the peripherals for Ultimo to function. The patient monitor screen has also been redesigned along with other peripherals to represents the Ultimo well. The new design has a clean, friendly and modern look that conveys the intelligence of the system.

The decision on which screen to show is based on input data from camera and beacons. All logic used in the code can be divided into 3 parts; detection, identification and triggering the display.

Detection: The Ultimo uses a ceiling-mounted camera to count the people entering and leaving the room. The camera provides the sequential frames which can be analyzed to detect shapes (human shapes). First, several frames are used to establish a background image. This background image is used to differentiate the foreground from the background. When an object stands underneath the camera for a longer period, the shape will be masked and the background reference updates using the entire frame. This is necessary to keep the background 'up-to-date' since room lighting might change and new stationary objects can be placed in the frame. Both need to become part of the background.

Identification: Bluetooth beacons are placed attached to clinician's identity cards. The beacons are scanned as often as possible inside a continuous loop. Each loop updates the scanned beacon's data and posts this to a local server. Thereby keeping a database of all beacons with their distance. Since there are other Bluetooth devices nearby, the scanner will check if the detected Bluetooth device is of the correct type, prior to updating the data.

Triggering display: The screen logic interprets this data and decides what to show using various if statements. During this process, the beacons have priority over people counting, this ensures that the clinician's screen will always be shown when there is a beacon nearby even when the counting did not detect due to a disconnection.

Conclusion

All in all this project proved that the vision of a silent ICU is not just wishful thinking, but actually feasible with plenty of stepping stones to complete integration. The foundation of a smart monitor has been made with the development of Ultimo. The system was presented in the Dutch Design Week 2018 in Eindhoven in the Netherlands. The intention was to show the industry experts and manufacturers that silent ICU is not a far future and quite possibly any sooner by improving the accuracy of the current system.