Covid-19: Getting to Grips with Respiration

As healthcare systems are struggling to keep up with the number of Covid-19 patients requiring hospitalization, a common dilemma being faced by healthcare practitioners is how best to triage patients. Upon admission to a hospital, the first step for a patient is to have their condition assessed and for priority levels to be assigned for their care, this is called triage.

Battling a disease which can cause rapid deterioration in patient health, staff shortages and overwhelming patient numbers; hospitals have turned to technology to better facilitate patient care.

Some hospitals are turning to AI, many for the first time, to speed up the triage process. The Royal Bolton Hospital, located in the UK, has adopted an AI-based chest x-ray system called qXR. Developed by Qure.ai, qXR can detect Covid-19 induced pneumonia in the lungs and support radiologists in their diagnosis. OSF HealthCare, a medical group based in Illinois, USA, is rolling out a new mobile app which they believe could limit hospital visits by at least 20%. E-Triage, developed by StoCastic, leverages machine learning to make triage level recommendations by assessing a patient’s vulnerability to COVID-19 complications.

Lessons learned from the Lombardy region

Adapting procedure to meet the demands of Covid-19 is not without risk but one of the biggest challenges faced by hospitals is “we still need lots of data to be able to understand which patients will deteriorate and which will not”, says Dr. Thomas Sull, an emergency physician at Mackenzie Health.

In Italy, once the epicenter of the Covid-19 pandemic, an algorithm has been devised to help doctors determine the clinical severity of a patient’s infection. Applying four testing criteria, the “BCRSS algorithm determines how sick a patient is and if they require critical care”, says Edward Shim, Co-founder and Managing Director at Studio 1 Labs. “One of the four testing criteria which is assessed is the respiratory rate of the patient”.

Edward Shim, Studio 1 Labs

Studio 1 Labs, a Canadian startup, have been developing a fabric sensor using e-textiles which can be deployed under the guise of a bedsheet. More than simply monitoring respiratory rate, Studio 1 Labs can identify irregular and abnormal breathing patterns.

Devised in Brescia, part of the Lombardy region, the Brescia-COVID Respiratory Severity Scale (BCRSS)/Algorithm was developed at a time of crisis for Italy. While the BCRSS algorithm is yet to be validated outside of Italy – a Swiss study published by BMC prior to the global pandemic – confirmed the importance of monitoring respiration volume to track declining lung function.

To find a consistent baseline to measure health indicators the placement of the sensor is crucial. “During clinical studies, we determined that attaining a baseline for health indicators is easiest when a person is resting. As such, the most natural environment for our sensor is in bedsheet form”, said Shim.

Currently, healthcare practitioners are conducting manual counts to assess irregularities in breathing but at a time of crisis, accuracy of such tests can prove unreliable. Dr Sull explains: “If there is a significant discrepancy between the true reading and the inaccurate data, it could lead to significant disability or death”.

Turning fabric into a circuit

Despite being first marketed as a solution for fall prevention and pressure ulcer monitoring, collecting respiration data was always a fundamental part of the technology.

“Covid-19 has created the need for us to reach the market quickly”, says Shims. “This has resulted in us focusing on respiratory indicators, putting aside fall prevention and pressure ulcer monitoring. We’ve shrunk the sensor down to a large strip which focuses on the torso area.”

Still an emerging field, e-textiles is rapidly moving beyond imbedding LEDs into clothing. Today, the garment itself is the sensor and companies are creating a range of pressure, strain, and temperature solutions.

E-textiles are commonly created using one of two methods, coating the fabric with an electrically conductive chemical layer or conductive material which has been interwoven with the fabric. Weaving electrically conductive material into the fabric creates a textile that effectively becomes a circuit. Shim says, “much like PCBs have different components, we’re creating different components on the fabric itself”.

This is a 3-Dimensional outline of a person lying on the Studio 1 Labs fabric sensor. The sensor can map the body in real-time and display the expansion and contraction of the chest. (Source: Studio 1 Labs)

Getting a consistent signal over distance

Shim views pressure sensors and other binary systems as inadequate tools to measure real-world trends. “Our goal is to measure the dynamic range. If your focus is on whether a handshake took place or not, conventional systems alone wouldn’t give you a good indication to the success of that handshake. By measuring a dynamic signal range, we’re able to capture the strength of the squeeze, the level of movement and use this as an indicator to better understand the outcome”.

The core strength of this technology is combining dynamic vibration and pressure into a single sensor. This is not too dissimilar to how human touch allows us to feel with different levels of sensitivity. The ability to collect a greater number of correlational indicators from a single source can also minimize error rates.

To minimize noise in the signal, Shim says “we’re currently working on sensor level filtering so that we can output clean signals at the edge”. This includes a dispersion technique which entangles the signal and energy across the sensor then measures the output based on the interaction between the sensor and corresponding applied force.

E-textiles is still a developing field and one of the largest obstacles impacting adoption of the technology is getting a consistent signal over distance. To put an electrical charge or voltage across longer distances doesn’t work the same as it would with other metal wires. This has forced companies in the e-textiles space to reduce the size of their fabric sensors to increase the accuracy of their results.

This 4x8ft sheet is in fact a fabric sensor which was used to monitor the heart rate and respiration of a horse during surgery. (Source: Studio 1 Labs)

Shim believes this is a constraint they’ve overcome: “We’ve managed to find a way to work around this limitation but we’re still researching different material combinations and structural designs so that we can get a stronger signal. We’ve trialed our device in particularly challenging environments, which included creating a 4x8ft sensor that was big enough for a horse”.

Several pilot projects are currently in the pipeline. One project in Taiwan is testing the viability of using fabric sensors to remotely monitor non-critical patients in long-term care facilities.

“Further preparation is required before the system can be implemented in hospitals”, says Markus Van Kempen, Venture Capitalist in Residence, IBM Corporate Technical Strategy. Referring to a collaborative project between IBM, Studio 1 Labs and Marist College, Van Kempen said, “these projects support controlled studies by testing environments and making iterations of speed of data to the cloud and high accuracy output for edge computing”.

Joining the PPE supply chain

In amongst the economic doom and gloom of today, Studio 1 Labs has found a new opportunity in an unlikely place.

“We’re constantly asked, ‘is the bedsheet washable?’. During early stage testing we developed an outer layer which could withstand disinfectants. We have since become remarkably familiar with the types of materials suitable for medical environments”, says Shim.

The need for washable, flame retardant materials made for a simple transition to personal protective equipment (PPE) – of which there is a widely documented global shortage.

With the support “of our partners and both the Canadian and Taiwanese government, we’ve been able to transition from learning about PPE standards to shipping materials to Canada. These materials can now be used to start producing medical gowns which have since been cleared by Health Canada, all in the space of just five weeks”.

Shim feels that it’s the unique nature of startups that has provided Studio 1 Labs with the agility to pivot during the global pandemic.

“Startups are often working very lean, with limited resources and budget. To pivot into something that’s unfamiliar is high risk. For a startup, the goal should always be to fulfil the needs of the customer, but you must also weigh the opportunity cost. During such unprecedented times, we now find ourselves selling fabrics as opposed to fabric sensors”.

The post Covid-19: Getting to Grips with Respiration appeared first on EE Times Asia.

from EE Times Asia https://ift.tt/2yhyY2q