A Model for Safe Transport of Critical Patients in Unmanned Drones with a ‘Watch’ Style Continuous Anesthesia Sensor

Abstract

We envision unmanned aerial vehicles (UAV) for rapid evacuation of critically-ill patients from hazardous locations to health care facilities in safe zones. For safety, medical teams accompany patients to monitor vital signs and titrate anesthesia dose during transport. UAV transports would require continuous automated remote monitoring of both vital signs and of sedative dose to be feasible and safe. Volatile anesthetics (isoflurane) are the only anesthetic agents that can be monitored continuously with infrared spectroscopy (IR) devices; but unsuitable for transport. Our objective is to devise a safe UAV transport protocol incorporating novel technology for gas monitoring. Our group has developed and tested a miniaturized wearable fuel cell sensor that can detect isoflurane gas vapors as low as 40 ppm (within therapeutic range) with a sensitivity of 0.0112 nA ppm −1 cm −2 . Ambient signal interference was resolved by principal component analysis (PCA). Data variance of 1 st and 2 nd principal components was 88.68% and 11.31%, respectively. The PCA regression model reported here can determine accurate Isoflurane concentrations. Electronic IoT platform has been built constituting micro-fuel cell, miniaturized electronic components with Bluetooth. This wearable sensor can be incorporated into a comprehensive life support system for casualty evacuation in conjunction with autonomous UAV emergency medical operations.

The Application of Drones in Healthcare and Health-Related Services in North America: A Scoping Review

“…Authors of each of the 29 articles included in this review examined how drones were utilized for a specific health or health-related issue. Using drones to deliver medical supplies and treatments (e.g., gauze, testing kits, and medications) was the most common (N = 11) health application [3,15,20,[22][23][24][25][26][27][28][29][30][31]. Environmental monitoring (e.g., wildfire, landslide, and air quality monitoring) was the second most (N = 8) examined intervention [16,18,19,21,25,26,[31][32][33], and using drones to deliver automated external defibrillators (AEDs) for cardiac emergencies was the third most (N = 8) examined intervention [4,17,22,30,[33][34][35].…”

“…Using drones to deliver medical supplies and treatments (e.g., gauze, testing kits, and medications) was the most common (N = 11) health application [3,15,20,[22][23][24][25][26][27][28][29][30][31]. Environmental monitoring (e.g., wildfire, landslide, and air quality monitoring) was the second most (N = 8) examined intervention [16,18,19,21,25,26,[31][32][33], and using drones to deliver automated external defibrillators (AEDs) for cardiac emergencies was the third most (N = 8) examined intervention [4,17,22,30,[33][34][35]. Less frequently reported in the research literature was the use of drones to transport biological samples (e.g., blood, plasma, organs, and other tissues) (N = 6) [20,22,27,[36][37][38], to facilitate search and rescue operations (N = 4) [29,32,39,40], for emergency service delivery (N = 3) [24,25,28], to support first responder safety (N = 3) [24,25,…”

Early diagnosis with alternative approaches: innovation in lung cancer care

“…The available medical metadata of LC patients and treatments can be accurately visualize using different machine learning approaches (107). Furthermore, artificial intelligence can also be utilized to analyze patients' behavior, lifestyle, and food habits to detect possible LC patients at an early stage (107)(108)(109). Additionally, medical data from different patients can be analyzed using machine learning and artificial intelligence to propose a better treatment plan.…”

Prospects and Challenges of Volatile Organic Compound Sensors in Human Healthcare

“…Different machine learning tools, such as principle component analysis (PCA), canonic discriminant analysis (CDA), independent component analysis (ICA), discriminant factorial analysis (DFA), partial least-squares analysis (PLS), artificial neural networks (ANNs), support vector machine (SVM), and hierarchical cluster analysis (HCA) have been reported to be adopted for specific VOC detection. ,,, However, once a method of prediction has been established, the predictive accuracy is obtained through cross-validation of all the data sets. Prediction model improves the correlation of the data sets and infers a pattern whereas regression technique provides specific concentration of unknown value of future data streams in real-time through fitting with this pattern. ,, Mazzone’s group (Cleveland Clinic) demonstrated chemiresistive sensor array for the detection of lung cancer from exhaled breath . They applied PCA and CDA techniques for the classification of malignancy and benignity of lung cells and SVM was employed to generate a prediction model from the data.…”

“…22,30,144 Mazzone's group (Cleveland Clinic) demonstrated chemiresistive sensor array for the detection of lung cancer from exhaled breath 124. They applied PCA and CDA techniques for the classification of malignancy and benignity of lung cells and SVM was employed to generate a prediction model from the data.…”