Lab-on-Mask for Remote Respiratory Monitoring.

A smart mask integrated with a remote, noncontact multiplexed sensor system, or "Lab-on-Mask" (LOM) is designed for monitoring respiratory diseases, such as the COVID-19. This LOM can monitor the heart rate, blood oxygen saturation, blood pressure, and body temperature associated with symptoms of pneumonia caused by coronaviruses in real time. Because of this remote monitoring system, frontline healthcare staff can minimize the exposure they face from close contact with the patients and reduce the risks of being infected.

The COVID-19 outbreak and the containment measures to stem its spread have brought many nations to an economic halt. At the time of writing >16 million people have been tested positive, leaving 645 000 dead. Researchers and physicians desperately need more data to better understand and treat this extremely contagious and deadly disease. SARS-CoV-2, the coronavirus that causes the respiratory disease behind COVID-19, like other coronaviruses, is believed to be transmitted mainly through respiratory droplets released when an infected patient exhales, speaks, sings, or coughs.[1,2] Face masks are a physical barrier to the dispersal of such respiratory droplets, which are potentially infectious as some contain viable virus.[3] More than 50 countries have since made masks mandatory in public spaces in the current COVID-19 pandemic. However, frontline healthcare staff face heightened risks of being infected, due to the high viral concentration in the air and on the surfaces of COVID-19 wards.[4,5] In many hospitals in Singapore, once community transmission of COVID-19 is evident, patients whose conditions would permit have also been given masks.[6] A system that remotely monitors patients’ vital parameters can help reduce face-to-face contact between healthcare workers and patients. Such a remote system also benefits recovering patients by helping them track their progress, relieving the stress on overwhelmed healthcare systems during a pandemic.

In this Materials Express, we present a smart mask, monolithically integrated with a remote, noncontact multiplexed sensor system, or “Lab-on-Mask” (LOM), streamlining the concept of the “body to device to cloud” platform.[7] Details of the LOM are shown in Figures a and S1a. It can in real-time simultaneously monitor several physical parameters associated with pneumonia,[2,8,9] including heart rate (HR), blood pressure (BP), and blood oxygen saturation (SpO2). In this paper, for convenience and compliance, we use the terms of HR, BP, and SpO2 to refer those measured in the area of LOM covered. In addition, it measures the interior temperature (T) of the LOM. Since commercial printed circuit boards (PCB) and human skin are not completely conformable, polydimethylsiloxane (PDMS), which is the substrate for flexible electronics,[10,11] has been used to embed the entire monitoring system (Figure b). According to the unidirectional stress tensile test, the Young’s Modulus of PDMS (1.086 ± 0.043 MPa)[7,12] is 3–4 orders of magnitude lower than that of the PCB[13] and 2 orders of magnitude lower than that of the innermost non-woven fabric of the mask, indicating better compatibility to human facial skin (∼200 kPa)[14,15] (Figure ). The embedded health monitoring system attached to the inside of the mask can effectively enhance conformability with the human facial skin, making it more comfortable to wear and conducive to obtaining stable signal output.

Figure 1

Schematic of the LOM. (a) Different sensors embedded in the PDMS. (b) Comparison of Young’s modulus based on PCB, skin, and PDMS. (c) Scheme of the different parts of the LOM on skin. Embedded in PDMS, the Young’s modulus of the system is more similar to that of our skin. (d) Strain–stress curve of and Young’s modulus of PDMS and the nonwoven fabric of the mask.

The noncontact LOM integrates three components: signal-receiving sensors, data processing modules, and Bluetooth data output, schematically shown in Figure a. Different sensors first collect signals from the face surface and, subsequently, convert and send electrical signals to the respective data processing modules. For example, a photoplethysmography (PPG) sensor containing an infrared light-emitting diode (LED), a photodetector and preamplifier is used for SpO2, which receives light signals reflected from the blood vessels under the epidermis and converts them into electrical signals to identify vasoconstriction and vasodilation. Similarly, a green LED is used for detecting HR and BP. A thermistor converts the skin temperature inside the mask into electrical signals. A microprogrammed control unit (MCU) then collects the physiological information through separate preamplifiers and controls data output synchronously. Finally, through wireless Bluetooth, the real-time data can be presented remotely to the receiving terminal, such as an app in a mobile phone, as shown in Figure S1b and Movie S1.

Figure 2

Recorded data from the LOM. (a) Scheme of the different parts of the system on the mask. (b–e) HR, SpO2, T, and BP, compared with data collected from commercial products. (f) Remote real-time monitoring of a person using the mask for HR, SpO2, T, and BP.

In medical diagnosis, HR, BP and SpO2 should be read out from heartbeats, dedicated arteries and fingertips. In our LOM, we recorded the vital signs from the superficial temporal arteries and facial arteries as references. Therefore, to validate the accuracy of the LOM, we have compared those signs with commercial devices on the same individual simultaneously. Figure b–e show the HR, SpO2, T, and BP of a person under resting state. It is evident the HR of this person ranges from 62 to 72 when wearing the LOM, which are typical values for a healthy individual. Meanwhile, the recorded HR reflection of the same person from the pulse oximeter (PO, from Andon Health Co., Ltd.) at the same time ranges from 62 to 72 as well. Here, we define “c” (c = m/σ, m is the mean; σ is the standard deviation) as the coefficient of data dispersion to confirm the accuracy of our system. From Figure S2a, the c of HR based on LOM is 4.34 %, comparable to the 4.45 % from the commercial PO. SpO2 is another important parameter in patients with respiratory infection such as the COVID-19. For a healthy person, the values are above 95 %. The LOM shows the same SpO2 measurements as the PO for the same person, at ∼98% (Figure c). The c of SpO2 based on LOM is 0.77 %, while that of commercial PO is 0.84 % (Figure S2b).

Body temperature is a core parameter in the evaluation of a person with a suspected or confirmed infection. The LOM records the interior temperature of the mask, reflecting one’s cheek surface temperature. As a strong correlation exists between that and one’s forehead temperature, which commercial temporal artery thermometers commonly read from, our LOM provides a frame of reference for one’s body temperature and fevers. While there is a compensation value of 2.9 °C between the LOM and infrared temperature (IR-T) (Figure S3), Figure d shows the calibrated temperatures based on the LOM and IR-T. In addition, their statistical c values are similar (LOM 0.25 %, IR-T 0.22 %, shown in Figure S2c), suggesting that the LOM is as accurate as the commercial IR-T.

Finally, we compare the BP based on the LOM and a commercial monitor (Electronic Blood Pressure Monitor from Omron) (Figure e). From Figure S2d, the c of LOM (1.94 % and 1.12 % for SYS and DIA, respectively) is lower than that of Omron (3.06 % and 4.06 % for SYS and DIA), suggesting our device is more stable. It is noteworthy that we have measured ten other healthy individuals using the LOM, with the data shown in Figures S4–S8. Long-time and real-time monitoring of vital signs are also crucial elements. Here, we have in real-time monitored the HR, SpO2, T, and BP of a person remotely over 7 consecutive hours. In Figure f, we observe all vital signs are in the normal range with occasional fluctuations.

In the future, we plan to further optimize the LOM and collaborate with hospitals for remote, noncontact, and real-time monitoring of patients who have been infected with SARS-CoV-2. In addition, the LOM may provide an additional function of directly monitoring and deactivating viruses when integrating antiviral chemical sensors and surface coating. Macroscopically, cloud-based data collection from these point-of-care devices can improve surveillance of pandemic evolution across communities geographically, providing information for quick mass screening of problematic areas (for healthcare personnel to zone in), regional transmission pattern, policy-setting for quarantine and response measures, and post-analysis in better understanding the disease. Similarly, in hospitals, the device could be useful in tracking the vital signs of patients who are infectious but whose condition allows them to wear a mask. It has also been suggested that when patients who are able to do wear a mask, they are less likely to transmit the infection to their fellow patients. This is important in hospitals where multi-bedded wards are common. Meanwhile, the “body to device to cloud” concept aligns with the move towards digital healthcare, where sensors interface between the body and the digital, data science and machine learning analyze what is normal and not, and a cloud-based information collection system arches over various services in healthcare.