Clinicians Turn to Unobtrusive Sensors for Patient Monitoring

Researchers at the University of South Australia developed a multimode optical fiber (MMF) sensor array to unobtrusively monitor the movement and vital signs of patients in hospitals and other health care settings. The sensors, which are designed to be embedded in the upper portion of a patient’s mattress, could help prevent falls, reduce bedsore injuries, monitor respiration rates, and otherwise improve health outcomes for patients while reducing the need for manual monitoring.

An array of seven MMF cables comprises the smart bed sensor, bundled at the sensor’s input and output locations. The bundled fiber ends are mounted for the positioning of a laser beam and camera; a single laser light source and a single camera are all that are required to interrogate the sensor. Light from a coherent, visible wavelength (633 nm) laser source is coupled with a single-mode fiber and then collimated to yield a beam approximately 3.4 mm in diameter.

The collimated beam is directed onto the end face of the MMF bundle, which exhibits relatively uniform excitation across all seven fibers. The far end of the fiber bundle is then imaged onto a CCD camera.

The propagation of coherent light through the MMF array produces a speckle output. Any physical perturbations that occur on the optical fiber, such as strain and bending, cause a change in the optical path length of the MMF modes. This change, in turn, induces changes in the speckle pattern that is output. An algorithm is used to quantify how much the speckle pattern changes when a physical perturbation occurs.

Tiny optical fiber sensors are attached to the upper surface of a hospital mattress to detect movement. Courtesy of the University of South Australia.

In demonstration, the researchers fixed the seven MMFs to a spring mattress with a foam top. Three fibers were laid vertically along the bed while four were laid horizontally, leading to 12 crossing points. The researchers used an algorithm to multiply the signals from the vertical and horizontal points and obtain basic spatial information on the movement that occurs at the crossing points.

The team performed three procedures to demonstrate the MMF sensor array for smart bed monitoring applications — getting on and off the mattress, rolling across the mattress, and breathing while on the mattress. The sensor array was shown to be sufficiently sensitive to detect breathing, and it identified the location of the participant. Additionally, it detected the presence of a person on the mattress in the absence of movement.

In another procedure, the sensor array responded strongly to the physiological effects of rigorous exercise — primarily increased heart rate and rate of breathing.

Hospitals use weight-based sensors or cameras installed in the room to monitor patients. “Existing weight-based hospital sensors cannot predict when a patient leaves the bed until their feet touch the floor, leaving little time for nursing staff to respond in the event of a fall,” said researcher Stephen Warren-Smith. “Also, there are privacy issues with camera-based technology.”

Warren-Smith said that the smart sensor technology could also relieve the burden on hospital staff who must constantly monitor patients for pressure sores.

The smart bed sensor requires no additional optical sensor elements to operate. The researchers anticipate that the sensor’s sensitivity and spatial resolution could be expanded without increasing interrogation costs through further multiplexing and the use of emerging machine learning algorithms.

“Respiration rates are often the first sign that a patient is deteriorating. This normally requires devices to be attached to the patient, either on the chest, as a mask on the face, or ventilator. These can be restrictive and sometimes inappropriate in an aged care setting,” Warren-Smith said. “Monitoring vital signs continuously, unobtrusively, and cheaply via the mattress-embedded sensors is a far better solution for both patient and nurse.”

The research was published in the Journal of Biomedical Optics (www.doi.org/10.1117/1.JBO.27.6.067002).